repo_name
stringlengths
7
65
path
stringlengths
5
185
copies
stringlengths
1
4
size
stringlengths
4
6
content
stringlengths
977
990k
license
stringclasses
14 values
hash
stringlengths
32
32
line_mean
float64
7.18
99.4
line_max
int64
31
999
alpha_frac
float64
0.25
0.95
ratio
float64
1.5
7.84
autogenerated
bool
1 class
config_or_test
bool
2 classes
has_no_keywords
bool
2 classes
has_few_assignments
bool
1 class
yarikoptic/pystatsmodels
statsmodels/sandbox/distributions/try_pot.py
5
2283
# -*- coding: utf-8 -*- """ Created on Wed May 04 06:09:18 2011 @author: josef """ import numpy as np def mean_residual_life(x, frac=None, alpha=0.05): '''emprirical mean residual life or expected shortfall Parameters ---------- todo: check formula for std of mean doesn't include case for all observations last observations std is zero vectorize loop using cumsum frac doesn't work yet ''' axis = 0 #searchsorted is 1d only x = np.asarray(x) nobs = x.shape[axis] xsorted = np.sort(x, axis=axis) if frac is None: xthreshold = xsorted else: xthreshold = xsorted[np.floor(nobs * frac).astype(int)] #use searchsorted instead of simple index in case of ties xlargerindex = np.searchsorted(xsorted, xthreshold, side='right') #replace loop with cumsum ? result = [] for i in range(len(xthreshold)-1): k_ind = xlargerindex[i] rmean = x[k_ind:].mean() rstd = x[k_ind:].std() #this doesn't work for last observations, nans rmstd = rstd/np.sqrt(nobs-k_ind) #std error of mean, check formula result.append((k_ind, xthreshold[i], rmean, rmstd)) res = np.array(result) crit = 1.96 # todo: without loading stats, crit = -stats.t.ppf(0.05) confint = res[:,1:2] + crit * res[:,-1:] * np.array([[-1,1]]) return np.column_stack((res, confint)) expected_shortfall = mean_residual_life #alias if __name__ == "__main__": rvs = np.random.standard_t(5, size= 10) res = mean_residual_life(rvs) print res rmean = [rvs[i:].mean() for i in range(len(rvs))] print res[:,2] - rmean[1:] ''' >>> mean_residual_life(rvs, frac= 0.5) Traceback (most recent call last): File "<stdin>", line 1, in <module> File "E:\Josef\eclipsegworkspace\statsmodels-josef-experimental-030\scikits\statsmodels\sandbox\distributions\try_pot.py", line 35, in mean_residual_life for i in range(len(xthreshold)-1): TypeError: object of type 'numpy.float64' has no len() >>> mean_residual_life(rvs, frac= [0.5]) array([[ 1. , -1.16904459, 0.35165016, 0.41090978, -1.97442776, -0.36366142], [ 1. , -1.16904459, 0.35165016, 0.41090978, -1.97442776, -0.36366142], [ 1. , -1.1690445 '''
bsd-3-clause
7131197cd7ebe67e4dc1a93738974f68
30.273973
155
0.619799
3.023841
false
false
false
false
yarikoptic/pystatsmodels
statsmodels/examples/l1_demo/demo.py
3
13517
from optparse import OptionParser import statsmodels.api as sm import scipy as sp from scipy import linalg from scipy import stats import pdb # pdb.set_trace() docstr = """ Demonstrates l1 regularization for likelihood models. Use different models by setting mode = mnlogit, logit, or probit. Examples ------- $ python demo.py --get_l1_slsqp_results logit >>> import demo >>> demo.run_demo('logit') The Story --------- The maximum likelihood (ML) solution works well when the number of data points is large and the noise is small. When the ML solution starts "breaking", the regularized solution should do better. The l1 Solvers -------------- The solvers are slower than standard Newton, and sometimes have convergence issues Nonetheless, the final solution makes sense and is often better than the ML solution. The standard l1 solver is fmin_slsqp and is included with scipy. It sometimes has trouble verifying convergence when the data size is large. The l1_cvxopt_cp solver is part of CVXOPT and this package needs to be installed separately. It works well even for larger data sizes. """ def main(): """ Provides a CLI for the demo. """ usage = "usage: %prog [options] mode" usage += '\n'+docstr parser = OptionParser(usage=usage) # base_alpha parser.add_option("-a", "--base_alpha", help="Size of regularization param (the param actully used will "\ "automatically scale with data size in this demo) "\ "[default: %default]", dest='base_alpha', action='store', type='float', default=0.01) # num_samples parser.add_option("-N", "--num_samples", help="Number of data points to generate for fit "\ "[default: %default]", dest='N', action='store', type='int', default=500) # get_l1_slsqp_results parser.add_option("--get_l1_slsqp_results", help="Do an l1 fit using slsqp. [default: %default]", \ action="store_true",dest='get_l1_slsqp_results', default=False) # get_l1_cvxopt_results parser.add_option("--get_l1_cvxopt_results", help="Do an l1 fit using cvxopt. [default: %default]", \ action="store_true",dest='get_l1_cvxopt_results', default=False) # num_nonconst_covariates parser.add_option("--num_nonconst_covariates", help="Number of covariates that are not constant "\ "(a constant will be prepended) [default: %default]", dest='num_nonconst_covariates', action='store', type='int', default=10) # noise_level parser.add_option("--noise_level", help="Level of the noise relative to signal [default: %default]", dest='noise_level', action='store', type='float', default=0.2) # cor_length parser.add_option("--cor_length", help="Correlation length of the (Gaussian) independent variables"\ "[default: %default]", dest='cor_length', action='store', type='float', default=2) # num_zero_params parser.add_option("--num_zero_params", help="Number of parameters equal to zero for every target in "\ "logistic regression examples. [default: %default]", dest='num_zero_params', action='store', type='int', default=8) # num_targets parser.add_option("-J", "--num_targets", help="Number of choices for the endogenous response in "\ "multinomial logit example [default: %default]", dest='num_targets', action='store', type='int', default=3) # print_summaries parser.add_option("-s", "--print_summaries", help="Print the full fit summary. [default: %default]", \ action="store_true",dest='print_summaries', default=False) # save_arrays parser.add_option("--save_arrays", help="Save exog/endog/true_params to disk for future use. "\ "[default: %default]", action="store_true",dest='save_arrays', default=False) # load_old_arrays parser.add_option("--load_old_arrays", help="Load exog/endog/true_params arrays from disk. "\ "[default: %default]", action="store_true",dest='load_old_arrays', default=False) (options, args) = parser.parse_args() assert len(args) == 1 mode = args[0].lower() run_demo(mode, **options.__dict__) def run_demo(mode, base_alpha=0.01, N=500, get_l1_slsqp_results=False, get_l1_cvxopt_results=False, num_nonconst_covariates=10, noise_level=0.2, cor_length=2, num_zero_params=8, num_targets=3, print_summaries=False, save_arrays=False, load_old_arrays=False): """ Run the demo and print results. Parameters ---------- mode : String either 'logit', 'mnlogit', or 'probit' base_alpha : Float Size of regularization param (the param actually used will automatically scale with data size in this demo) N : Integer Number of data points to generate for fit get_l1_slsqp_results : boolean, Do an l1 fit using slsqp. get_l1_cvxopt_results : boolean Do an l1 fit using cvxopt num_nonconst_covariates : Integer Number of covariates that are not constant (a constant will be prepended) noise_level : Float (non-negative) Level of the noise relative to signal cor_length : Float (non-negative) Correlation length of the (Gaussian) independent variables num_zero_params : Integer Number of parameters equal to zero for every target in logistic regression examples. num_targets : Integer Number of choices for the endogenous response in multinomial logit example print_summaries : Boolean print the full fit summary. save_arrays : Boolean Save exog/endog/true_params to disk for future use. load_old_arrays Load exog/endog/true_params arrays from disk. """ if mode != 'mnlogit': print "Setting num_targets to 2 since mode != 'mnlogit'" num_targets = 2 models = { 'logit': sm.Logit, 'mnlogit': sm.MNLogit, 'probit': sm.Probit} endog_funcs = { 'logit': get_logit_endog, 'mnlogit': get_logit_endog, 'probit': get_probit_endog} # The regularization parameter # Here we scale it with N for simplicity. In practice, you should # use cross validation to pick alpha alpha = base_alpha * N * sp.ones((num_nonconst_covariates+1, num_targets-1)) alpha[0,:] = 0 # Don't regularize the intercept #### Make the data and model exog = get_exog(N, num_nonconst_covariates, cor_length) exog = sm.add_constant(exog) true_params = sp.rand(num_nonconst_covariates+1, num_targets-1) if num_zero_params: true_params[-num_zero_params:, :] = 0 endog = endog_funcs[mode](true_params, exog, noise_level) endog, exog, true_params = save_andor_load_arrays( endog, exog, true_params, save_arrays, load_old_arrays) model = models[mode](endog, exog) #### Get the results and print results = run_solvers(model, true_params, alpha, get_l1_slsqp_results, get_l1_cvxopt_results, print_summaries) summary_str = get_summary_str(results, true_params, get_l1_slsqp_results, get_l1_cvxopt_results, print_summaries) print summary_str def run_solvers(model, true_params, alpha, get_l1_slsqp_results, get_l1_cvxopt_results, print_summaries): """ Runs the solvers using the specified settings and returns a result string. Works the same for any l1 penalized likelihood model. """ results = {} #### Train the models # Get ML results results['results_ML'] = model.fit(method='newton') # Get l1 results start_params = results['results_ML'].params.ravel(order='F') if get_l1_slsqp_results: results['results_l1_slsqp'] = model.fit_regularized( method='l1', alpha=alpha, maxiter=1000, start_params=start_params, retall=True) if get_l1_cvxopt_results: results['results_l1_cvxopt_cp'] = model.fit_regularized( method='l1_cvxopt_cp', alpha=alpha, maxiter=50, start_params=start_params, retall=True, feastol=1e-5) return results def get_summary_str(results, true_params, get_l1_slsqp_results, get_l1_cvxopt_results, print_summaries): """ Gets a string summarizing the results. """ #### Extract specific results results_ML = results['results_ML'] RMSE_ML = get_RMSE(results_ML, true_params) if get_l1_slsqp_results: results_l1_slsqp = results['results_l1_slsqp'] if get_l1_cvxopt_results: results_l1_cvxopt_cp = results['results_l1_cvxopt_cp'] #### Format summaries # Short summary print_str = '\n\n=========== Short Error Summary ============' print_str += '\n\n The maximum likelihood fit RMS error = %.4f'%RMSE_ML if get_l1_slsqp_results: RMSE_l1_slsqp = get_RMSE(results_l1_slsqp, true_params) print_str += '\n The l1_slsqp fit RMS error = %.4f'%RMSE_l1_slsqp if get_l1_cvxopt_results: RMSE_l1_cvxopt_cp = get_RMSE(results_l1_cvxopt_cp, true_params) print_str += '\n The l1_cvxopt_cp fit RMS error = %.4f'%RMSE_l1_cvxopt_cp # Parameters print_str += '\n\n\n============== Parameters =================' print_str += "\n\nTrue parameters: \n%s"%true_params # Full summary if print_summaries: print_str += '\n' + results_ML.summary().as_text() if get_l1_slsqp_results: print_str += '\n' + results_l1_slsqp.summary().as_text() if get_l1_cvxopt_results: print_str += '\n' + results_l1_cvxopt_cp.summary().as_text() else: print_str += '\n\nThe maximum likelihood params are \n%s'%results_ML.params if get_l1_slsqp_results: print_str += '\n\nThe l1_slsqp params are \n%s'%results_l1_slsqp.params if get_l1_cvxopt_results: print_str += '\n\nThe l1_cvxopt_cp params are \n%s'%\ results_l1_cvxopt_cp.params # Return return print_str def save_andor_load_arrays( endog, exog, true_params, save_arrays, load_old_arrays): if save_arrays: sp.save('endog.npy', endog) sp.save('exog.npy', exog) sp.save('true_params.npy', true_params) if load_old_arrays: endog = sp.load('endog.npy') exog = sp.load('exog.npy') true_params = sp.load('true_params.npy') return endog, exog, true_params def get_RMSE(results, true_params): """ Gets the (normalized) root mean square error. """ diff = results.params.reshape(true_params.shape) - true_params raw_RMSE = sp.sqrt(((diff)**2).sum()) param_norm = sp.sqrt((true_params**2).sum()) return raw_RMSE / param_norm def get_logit_endog(true_params, exog, noise_level): """ Gets an endogenous response that is consistent with the true_params, perturbed by noise at noise_level. """ N = exog.shape[0] ### Create the probability of entering the different classes, ### given exog and true_params Xdotparams = sp.dot(exog, true_params) noise = noise_level * sp.randn(*Xdotparams.shape) eXB = sp.column_stack((sp.ones(len(Xdotparams)), sp.exp(Xdotparams))) class_probabilities = eXB / eXB.sum(1)[:, None] ### Create the endog cdf = class_probabilities.cumsum(axis=1) endog = sp.zeros(N) for i in xrange(N): endog[i] = sp.searchsorted(cdf[i, :], sp.rand()) return endog def get_probit_endog(true_params, exog, noise_level): """ Gets an endogenous response that is consistent with the true_params, perturbed by noise at noise_level. """ N = exog.shape[0] ### Create the probability of entering the different classes, ### given exog and true_params Xdotparams = sp.dot(exog, true_params) noise = noise_level * sp.randn(*Xdotparams.shape) ### Create the endog cdf = stats.norm._cdf(-Xdotparams) endog = sp.zeros(N) for i in xrange(N): endog[i] = sp.searchsorted(cdf[i, :], sp.rand()) return endog def get_exog(N, num_nonconst_covariates, cor_length): """ Returns an exog array with correlations determined by cor_length. The covariance matrix of exog will have (asymptotically, as :math:'N\\to\\inf') .. math:: Cov[i,j] = \\exp(-|i-j| / cor_length) Higher cor_length makes the problem more ill-posed, and easier to screw up with noise. BEWARE: With very long correlation lengths, you often get a singular KKT matrix (during the l1_cvxopt_cp fit) """ ## Create the noiseless exog uncorrelated_exog = sp.randn(N, num_nonconst_covariates) if cor_length == 0: exog = uncorrelated_exog else: cov_matrix = sp.zeros((num_nonconst_covariates, num_nonconst_covariates)) j = sp.arange(num_nonconst_covariates) for i in xrange(num_nonconst_covariates): cov_matrix[i,:] = sp.exp(-sp.fabs(i-j) / cor_length) chol = linalg.cholesky(cov_matrix) # cov_matrix = sp.dot(chol.T, chol) exog = sp.dot(uncorrelated_exog, chol) ## Return return exog if __name__ == '__main__': main()
bsd-3-clause
6ca53a0bae4d22236bd45be554894b30
37.076056
83
0.620256
3.466786
false
false
false
false
yarikoptic/pystatsmodels
statsmodels/datasets/grunfeld/data.py
3
2696
"""Grunfeld (1950) Investment Data""" __docformat__ = 'restructuredtext' COPYRIGHT = """This is public domain.""" TITLE = __doc__ SOURCE = """This is the Grunfeld (1950) Investment Data. The source for the data was the original 11-firm data set from Grunfeld's Ph.D. thesis recreated by Kleiber and Zeileis (2008) "The Grunfeld Data at 50". The data can be found here. http://statmath.wu-wien.ac.at/~zeileis/grunfeld/ For a note on the many versions of the Grunfeld data circulating see: http://www.stanford.edu/~clint/bench/grunfeld.htm """ DESCRSHORT = """Grunfeld (1950) Investment Data for 11 U.S. Firms.""" DESCRLONG = DESCRSHORT NOTE = """Number of observations - 220 (20 years for 11 firms) Number of variables - 5 Variables name definitions:: invest - Gross investment in 1947 dollars value - Market value as of Dec. 31 in 1947 dollars capital - Stock of plant and equipment in 1947 dollars firm - General Motors, US Steel, General Electric, Chrysler, Atlantic Refining, IBM, Union Oil, Westinghouse, Goodyear, Diamond Match, American Steel year - 1935 - 1954 Note that raw_data has firm expanded to dummy variables, since it is a string categorical variable. """ from numpy import recfromtxt, column_stack, array from statsmodels.tools import categorical from statsmodels.datasets import utils as du from os.path import dirname, abspath def load(): """ Loads the Grunfeld data and returns a Dataset class. Returns ------- Dataset instance: See DATASET_PROPOSAL.txt for more information. Notes ----- raw_data has the firm variable expanded to dummy variables for each firm (ie., there is no reference dummy) """ data = _get_data() raw_data = categorical(data, col='firm', drop=True) ds = du.process_recarray(data, endog_idx=0, stack=False) ds.raw_data = raw_data return ds def load_pandas(): """ Loads the Grunfeld data and returns a Dataset class. Returns ------- Dataset instance: See DATASET_PROPOSAL.txt for more information. Notes ----- raw_data has the firm variable expanded to dummy variables for each firm (ie., there is no reference dummy) """ from pandas import DataFrame data = _get_data() raw_data = categorical(data, col='firm', drop=True) ds = du.process_recarray_pandas(data, endog_idx=0) ds.raw_data = DataFrame(raw_data) return ds def _get_data(): filepath = dirname(abspath(__file__)) data = recfromtxt(open(filepath + '/grunfeld.csv','rb'), delimiter=",", names=True, dtype="f8,f8,f8,a17,f8") return data
bsd-3-clause
009ff21023652e3644e1d80681cc34e4
28.955556
79
0.671736
3.378446
false
false
false
false
yarikoptic/pystatsmodels
examples/example_rlm.py
2
3080
""" Robust Linear Models Notes ----- The syntax for the arguments will be shortened to accept string arguments in the future. """ import numpy as np import statsmodels.api as sm import matplotlib.pyplot as plt from statsmodels.sandbox.regression.predstd import wls_prediction_std #Estimating RLM #-------------- # Load data data = sm.datasets.stackloss.load() data.exog = sm.add_constant(data.exog) # Huber's T norm with the (default) median absolute deviation scaling huber_t = sm.RLM(data.endog, data.exog, M=sm.robust.norms.HuberT()) hub_results = huber_t.fit() print hub_results.params print hub_results.bse varnames = ['var_%d' % i for i in range(len(hub_results.params))] print hub_results.summary(yname='y', xname=varnames) # Huber's T norm with 'H2' covariance matrix hub_results2 = huber_t.fit(cov="H2") print hub_results2.params print hub_results2.bse # Andrew's Wave norm with Huber's Proposal 2 scaling and 'H3' covariance matrix andrew_mod = sm.RLM(data.endog, data.exog, M=sm.robust.norms.AndrewWave()) andrew_results = andrew_mod.fit(scale_est=sm.robust.scale.HuberScale(), cov='H3') print andrew_results.params # See ``help(sm.RLM.fit)`` for more options and ``module sm.robust.scale`` for # scale options #Comparing OLS and RLM #--------------------- #Artificial data #^^^^^^^^^^^^^^^ nsample = 50 x1 = np.linspace(0, 20, nsample) X = np.c_[x1, (x1 - 5)**2, np.ones(nsample)] sig = 0.3 # smaller error variance makes OLS<->RLM contrast bigger beta = [0.5, -0.0, 5.] y_true2 = np.dot(X, beta) y2 = y_true2 + sig * 1. * np.random.normal(size=nsample) y2[[39, 41, 43, 45, 48]] -= 5 # add some outliers (10% of nsample) #Example: quadratic function with linear truth #^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ # Note that the quadratic term in OLS regression will capture outlier effects. res = sm.OLS(y2, X).fit() print res.params print res.bse print res.predict # Estimate RLM resrlm = sm.RLM(y2, X).fit() print resrlm.params print resrlm.bse # Draw a plot to compare OLS estimates to the robust estimates plt.figure(); plt.plot(x1, y2, 'o', x1, y_true2, 'b-'); prstd, iv_l, iv_u = wls_prediction_std(res); plt.plot(x1, res.fittedvalues, 'r-'); plt.plot(x1, iv_u, 'r--'); plt.plot(x1, iv_l, 'r--'); plt.plot(x1, resrlm.fittedvalues, 'g.-'); #@savefig rlm_ols_0.png plt.title('blue: true, red: OLS, green: RLM'); #Example: linear function with linear truth #^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ # Fit a new OLS model using only the linear term and the constant X2 = X[:, [0, 2]] res2 = sm.OLS(y2, X2).fit() print res2.params print res2.bse # Estimate RLM resrlm2 = sm.RLM(y2, X2).fit() print resrlm2.params print resrlm2.bse # Draw a plot to compare OLS estimates to the robust estimates prstd, iv_l, iv_u = wls_prediction_std(res2) plt.figure(); plt.plot(x1, y2, 'o', x1, y_true2, 'b-'); plt.plot(x1, res2.fittedvalues, 'r-'); plt.plot(x1, iv_u, 'r--'); plt.plot(x1, iv_l, 'r--'); plt.plot(x1, resrlm2.fittedvalues, 'g.-'); #@savefig rlm_ols_1.png plt.title('blue: true, red: OLS, green: RLM');
bsd-3-clause
7bc1c78c2ed8830be8b1ccc8aca647db
28.615385
79
0.661688
2.652885
false
false
false
false
yarikoptic/pystatsmodels
statsmodels/sandbox/survival2.py
35
17924
#Kaplan-Meier Estimator import numpy as np import numpy.linalg as la import matplotlib.pyplot as plt from scipy import stats from statsmodels.iolib.table import SimpleTable class KaplanMeier(object): """ KaplanMeier(...) KaplanMeier(data, endog, exog=None, censoring=None) Create an object of class KaplanMeier for estimating Kaplan-Meier survival curves. Parameters ---------- data: array_like An array, with observations in each row, and variables in the columns endog: index (starting at zero) of the column containing the endogenous variable (time) exog: index of the column containing the exogenous variable (must be catagorical). If exog = None, this is equivalent to a single survival curve censoring: index of the column containing an indicator of whether an observation is an event, or a censored observation, with 0 for censored, and 1 for an event Attributes ----------- censorings: List of censorings associated with each unique time, at each value of exog events: List of the number of events at each unique time for each value of exog results: List of arrays containing estimates of the value value of the survival function and its standard error at each unique time, for each value of exog ts: List of unique times for each value of exog Methods ------- fit: Calcuate the Kaplan-Meier estimates of the survival function and its standard error at each time, for each value of exog plot: Plot the survival curves using matplotlib.plyplot summary: Display the results of fit in a table. Gives results for all (including censored) times test_diff: Test for difference between survival curves Examples -------- >>> import statsmodels.api as sm >>> import matplotlib.pyplot as plt >>> import numpy as np >>> from statsmodels.sandbox.survival2 import KaplanMeier >>> dta = sm.datasets.strikes.load() >>> dta = dta.values()[-1] >>> dta[range(5),:] array([[ 7.00000000e+00, 1.13800000e-02], [ 9.00000000e+00, 1.13800000e-02], [ 1.30000000e+01, 1.13800000e-02], [ 1.40000000e+01, 1.13800000e-02], [ 2.60000000e+01, 1.13800000e-02]]) >>> km = KaplanMeier(dta,0) >>> km.fit() >>> km.plot() Doing >>> km.summary() will display a table of the estimated survival and standard errors for each time. The first few lines are Kaplan-Meier Curve ===================================== Time Survival Std. Err ------------------------------------- 1.0 0.983870967742 0.0159984306572 2.0 0.91935483871 0.0345807888235 3.0 0.854838709677 0.0447374942184 4.0 0.838709677419 0.0467104592871 5.0 0.822580645161 0.0485169952543 Doing >>> plt.show() will plot the survival curve Mutliple survival curves: >>> km2 = KaplanMeier(dta,0,exog=1) >>> km2.fit() km2 will estimate a survival curve for each value of industrial production, the column of dta with index one (1). With censoring: >>> censoring = np.ones_like(dta[:,0]) >>> censoring[dta[:,0] > 80] = 0 >>> dta = np.c_[dta,censoring] >>> dta[range(5),:] array([[ 7.00000000e+00, 1.13800000e-02, 1.00000000e+00], [ 9.00000000e+00, 1.13800000e-02, 1.00000000e+00], [ 1.30000000e+01, 1.13800000e-02, 1.00000000e+00], [ 1.40000000e+01, 1.13800000e-02, 1.00000000e+00], [ 2.60000000e+01, 1.13800000e-02, 1.00000000e+00]]) >>> km3 = KaplanMeier(dta,0,exog=1,censoring=2) >>> km3.fit() Test for difference of survival curves >>> log_rank = km3.test_diff([0.0645,-0.03957]) The zeroth element of log_rank is the chi-square test statistic for the difference between the survival curves for exog = 0.0645 and exog = -0.03957, the index one element is the degrees of freedom for the test, and the index two element is the p-value for the test Groups with nan names >>> groups = np.ones_like(dta[:,1]) >>> groups = groups.astype('S4') >>> groups[dta[:,1] > 0] = 'high' >>> groups[dta[:,1] <= 0] = 'low' >>> dta = dta.astype('S4') >>> dta[:,1] = groups >>> dta[range(5),:] array([['7.0', 'high', '1.0'], ['9.0', 'high', '1.0'], ['13.0', 'high', '1.0'], ['14.0', 'high', '1.0'], ['26.0', 'high', '1.0']], dtype='|S4') >>> km4 = KaplanMeier(dta,0,exog=1,censoring=2) >>> km4.fit() """ def __init__(self, data, endog, exog=None, censoring=None): self.exog = exog self.censoring = censoring cols = [endog] self.endog = 0 if exog != None: cols.append(exog) self.exog = 1 if censoring != None: cols.append(censoring) if exog != None: self.censoring = 2 else: self.censoring = 1 data = data[:,cols] if data.dtype == float or data.dtype == int: self.data = data[~np.isnan(data).any(1)] else: t = (data[:,self.endog]).astype(float) if exog != None: evec = data[:,self.exog] evec = evec[~np.isnan(t)] if censoring != None: cvec = (data[:,self.censoring]).astype(float) cvec = cvec[~np.isnan(t)] t = t[~np.isnan(t)] if censoring != None: t = t[~np.isnan(cvec)] if exog != None: evec = evec[~np.isnan(cvec)] cvec = cvec[~np.isnan(cvec)] cols = [t] if exog != None: cols.append(evec) if censoring != None: cols.append(cvec) data = (np.array(cols)).transpose() self.data = data def fit(self): """ Calculate the Kaplan-Meier estimator of the survival function """ self.results = [] self.ts = [] self.censorings = [] self.event = [] if self.exog == None: self.fitting_proc(self.data) else: groups = np.unique(self.data[:,self.exog]) self.groups = groups for g in groups: group = self.data[self.data[:,self.exog] == g] self.fitting_proc(group) def plot(self): """ Plot the estimated survival curves. After using this method do plt.show() to display the plot """ plt.figure() if self.exog == None: self.plotting_proc(0) else: for g in range(len(self.groups)): self.plotting_proc(g) plt.ylim(ymax=1.05) plt.ylabel('Survival') plt.xlabel('Time') def summary(self): """ Print a set of tables containing the estimates of the survival function, and its standard errors """ if self.exog == None: self.summary_proc(0) else: for g in range(len(self.groups)): self.summary_proc(g) def fitting_proc(self, group): """ For internal use """ t = ((group[:,self.endog]).astype(float)).astype(int) if self.censoring == None: events = np.bincount(t) t = np.unique(t) events = events[:,list(t)] events = events.astype(float) eventsSum = np.cumsum(events) eventsSum = np.r_[0,eventsSum] n = len(group) - eventsSum[:-1] else: censoring = ((group[:,self.censoring]).astype(float)).astype(int) reverseCensoring = -1*(censoring - 1) events = np.bincount(t,censoring) censored = np.bincount(t,reverseCensoring) t = np.unique(t) censored = censored[:,list(t)] censored = censored.astype(float) censoredSum = np.cumsum(censored) censoredSum = np.r_[0,censoredSum] events = events[:,list(t)] events = events.astype(float) eventsSum = np.cumsum(events) eventsSum = np.r_[0,eventsSum] n = len(group) - eventsSum[:-1] - censoredSum[:-1] (self.censorings).append(censored) survival = np.cumprod(1-events/n) var = ((survival*survival) * np.cumsum(events/(n*(n-events)))) se = np.sqrt(var) (self.results).append(np.array([survival,se])) (self.ts).append(t) (self.event).append(events) def plotting_proc(self, g): """ For internal use """ survival = self.results[g][0] t = self.ts[g] e = (self.event)[g] if self.censoring != None: c = self.censorings[g] csurvival = survival[c != 0] ct = t[c != 0] if len(ct) != 0: plt.vlines(ct,csurvival+0.02,csurvival-0.02) x = np.repeat(t[e != 0], 2) y = np.repeat(survival[e != 0], 2) if self.ts[g][-1] in t[e != 0]: x = np.r_[0,x] y = np.r_[1,1,y[:-1]] else: x = np.r_[0,x,self.ts[g][-1]] y = np.r_[1,1,y] plt.plot(x,y) def summary_proc(self, g): """ For internal use """ if self.exog != None: myTitle = ('exog = ' + str(self.groups[g]) + '\n') else: myTitle = "Kaplan-Meier Curve" table = np.transpose(self.results[g]) table = np.c_[np.transpose(self.ts[g]),table] table = SimpleTable(table, headers=['Time','Survival','Std. Err'], title = myTitle) print(table) def test_diff(self, groups, rho=None, weight=None): """ test_diff(groups, rho=0) Test for difference between survival curves Parameters ---------- groups: A list of the values for exog to test for difference. tests the null hypothesis that the survival curves for all values of exog in groups are equal rho: compute the test statistic with weight S(t)^rho, where S(t) is the pooled estimate for the Kaplan-Meier survival function. If rho = 0, this is the logrank test, if rho = 0, this is the Peto and Peto modification to the Gehan-Wilcoxon test. weight: User specified function that accepts as its sole arguement an array of times, and returns an array of weights for each time to be used in the test Returns ------- An array whose zeroth element is the chi-square test statistic for the global null hypothesis, that all survival curves are equal, the index one element is degrees of freedom for the test, and the index two element is the p-value for the test. Examples -------- >>> import statsmodels.api as sm >>> import matplotlib.pyplot as plt >>> import numpy as np >>> from statsmodels.sandbox.survival2 import KaplanMeier >>> dta = sm.datasets.strikes.load() >>> dta = dta.values()[-1] >>> censoring = np.ones_like(dta[:,0]) >>> censoring[dta[:,0] > 80] = 0 >>> dta = np.c_[dta,censoring] >>> km = KaplanMeier(dta,0,exog=1,censoring=2) >>> km.fit() Test for difference of survival curves >>> log_rank = km3.test_diff([0.0645,-0.03957]) The zeroth element of log_rank is the chi-square test statistic for the difference between the survival curves using the log rank test for exog = 0.0645 and exog = -0.03957, the index one element is the degrees of freedom for the test, and the index two element is the p-value for the test >>> wilcoxon = km.test_diff([0.0645,-0.03957], rho=1) wilcoxon is the equivalent information as log_rank, but for the Peto and Peto modification to the Gehan-Wilcoxon test. User specified weight functions >>> log_rank = km3.test_diff([0.0645,-0.03957], weight=np.ones_like) This is equivalent to the log rank test More than two groups >>> log_rank = km.test_diff([0.0645,-0.03957,0.01138]) The test can be performed with arbitrarily many groups, so long as they are all in the column exog """ groups = np.asarray(groups) if self.exog == None: raise ValueError("Need an exogenous variable for logrank test") elif (np.in1d(groups,self.groups)).all(): data = self.data[np.in1d(self.data[:,self.exog],groups)] t = ((data[:,self.endog]).astype(float)).astype(int) tind = np.unique(t) NK = [] N = [] D = [] Z = [] if rho != None and weight != None: raise ValueError("Must use either rho or weights, not both") elif rho != None: s = KaplanMeier(data,self.endog,censoring=self.censoring) s.fit() s = (s.results[0][0]) ** (rho) s = np.r_[1,s[:-1]] elif weight != None: s = weight(tind) else: s = np.ones_like(tind) if self.censoring == None: for g in groups: dk = np.bincount((t[data[:,self.exog] == g])) d = np.bincount(t) if np.max(tind) != len(dk): dif = np.max(tind) - len(dk) + 1 dk = np.r_[dk,[0]*dif] dk = dk[:,list(tind)] d = d[:,list(tind)] dk = dk.astype(float) d = d.astype(float) dkSum = np.cumsum(dk) dSum = np.cumsum(d) dkSum = np.r_[0,dkSum] dSum = np.r_[0,dSum] nk = len(data[data[:,self.exog] == g]) - dkSum[:-1] n = len(data) - dSum[:-1] d = d[n>1] dk = dk[n>1] nk = nk[n>1] n = n[n>1] s = s[n>1] ek = (nk * d)/(n) Z.append(np.sum(s * (dk - ek))) NK.append(nk) N.append(n) D.append(d) else: for g in groups: censoring = ((data[:,self.censoring]).astype(float)).astype(int) reverseCensoring = -1*(censoring - 1) censored = np.bincount(t,reverseCensoring) ck = np.bincount((t[data[:,self.exog] == g]), reverseCensoring[data[:,self.exog] == g]) dk = np.bincount((t[data[:,self.exog] == g]), censoring[data[:,self.exog] == g]) d = np.bincount(t,censoring) if np.max(tind) != len(dk): dif = np.max(tind) - len(dk) + 1 dk = np.r_[dk,[0]*dif] ck = np.r_[ck,[0]*dif] dk = dk[:,list(tind)] ck = ck[:,list(tind)] d = d[:,list(tind)] dk = dk.astype(float) d = d.astype(float) ck = ck.astype(float) dkSum = np.cumsum(dk) dSum = np.cumsum(d) ck = np.cumsum(ck) ck = np.r_[0,ck] dkSum = np.r_[0,dkSum] dSum = np.r_[0,dSum] censored = censored[:,list(tind)] censored = censored.astype(float) censoredSum = np.cumsum(censored) censoredSum = np.r_[0,censoredSum] nk = (len(data[data[:,self.exog] == g]) - dkSum[:-1] - ck[:-1]) n = len(data) - dSum[:-1] - censoredSum[:-1] d = d[n>1] dk = dk[n>1] nk = nk[n>1] n = n[n>1] s = s[n>1] ek = (nk * d)/(n) Z.append(np.sum(s * (dk - ek))) NK.append(nk) N.append(n) D.append(d) Z = np.array(Z) N = np.array(N) D = np.array(D) NK = np.array(NK) sigma = -1 * np.dot((NK/N) * ((N - D)/(N - 1)) * D * np.array([(s ** 2)]*len(D)) ,np.transpose(NK/N)) np.fill_diagonal(sigma, np.diagonal(np.dot((NK/N) * ((N - D)/(N - 1)) * D * np.array([(s ** 2)]*len(D)) ,np.transpose(1 - (NK/N))))) chisq = np.dot(np.transpose(Z),np.dot(la.pinv(sigma), Z)) df = len(groups) - 1 return np.array([chisq, df, stats.chi2.sf(chisq,df)]) else: raise ValueError("groups must be in column exog")
bsd-3-clause
52ec49ecb1078ededdcf77e362f01971
34.91984
84
0.483765
3.637175
false
true
false
false
yarikoptic/pystatsmodels
statsmodels/sandbox/regression/penalized.py
4
12034
# -*- coding: utf-8 -*- """linear model with Theil prior probabilistic restrictions, generalized Ridge Created on Tue Dec 20 00:10:10 2011 Author: Josef Perktold License: BSD-3 open issues * selection of smoothing factor, strength of prior, cross validation * GLS, does this really work this way * None of inherited results have been checked yet, I'm not sure if any need to be adjusted or if only interpretation changes One question is which results are based on likelihood (residuals) and which are based on "posterior" as for example bse and cov_params * helper functions to construct priors? * increasing penalization for ordered regressors, e.g. polynomials * compare with random/mixed effects/coefficient, like estimated priors there is something fishy with the result instance, some things, e.g. normalized_cov_params, don't look like they update correctly as we search over lambda -> some stale state again ? I added df_model to result class using the hatmatrix, but df_model is defined in model instance not in result instance. -> not clear where refactoring should occur. df_resid doesn't get updated correctly. problem with definition of df_model, it has 1 subtracted for constant """ import numpy as np import statsmodels.base.model as base from statsmodels.regression.linear_model import OLS, GLS, RegressionResults def atleast_2dcols(x): x = np.asarray(x) if x.ndim == 1: x = x[:,None] return x class TheilGLS(GLS): '''GLS with probabilistic restrictions essentially Bayes with informative prior note: I'm making up the GLS part, might work only for OLS ''' def __init__(self, endog, exog, r_matrix, q_matrix=None, sigma_prior=None, sigma=None): self.r_matrix = np.asarray(r_matrix) self.q_matrix = atleast_2dcols(q_matrix) if np.size(sigma_prior) == 1: sigma_prior = sigma_prior * np.eye(self.r_matrix.shape[0]) #no numerical shortcuts self.sigma_prior = sigma_prior self.sigma_prior_inv = np.linalg.pinv(sigma_prior) #or inv super(self.__class__, self).__init__(endog, exog, sigma=sigma) def fit(self, lambd=1.): #this does duplicate transformation, but I need resid not wresid res_gls = GLS(self.endog, self.exog, sigma=self.sigma).fit() self.res_gls = res_gls sigma2_e = res_gls.mse_resid r_matrix = self.r_matrix q_matrix = self.q_matrix sigma_prior_inv = self.sigma_prior_inv x = self.wexog y = self.wendog[:,None] #why are sigma2_e * lambd multiplied, not ratio? #larger lambd -> stronger prior (it's not the variance) #print 'lambd inside fit', lambd xpx = np.dot(x.T, x) + \ sigma2_e * lambd * np.dot(r_matrix.T, np.dot(sigma_prior_inv, r_matrix)) xpy = np.dot(x.T, y) + \ sigma2_e * lambd * np.dot(r_matrix.T, np.dot(sigma_prior_inv, q_matrix)) #xpy = xpy[:,None] xpxi = np.linalg.pinv(xpx) params = np.dot(xpxi, xpy) #or solve params = np.squeeze(params) self.normalized_cov_params = xpxi #why attach it to self, i.e. model? lfit = TheilRegressionResults(self, params, normalized_cov_params=xpxi) lfit.penalization_factor = lambd return lfit def fit_minic(self): #this doesn't make sense, since number of parameters stays unchanged #need leave-one-out, gcv; or some penalization for weak priors #added extra penalization for lambd def get_bic(lambd): #return self.fit(lambd).bic #+lambd #+ 1./lambd #added 1/lambd for checking #return self.fit(lambd).gcv() #return self.fit(lambd).cv() return self.fit(lambd).aicc() from scipy import optimize lambd = optimize.fmin(get_bic, 1.) return lambd #TODO: #I need the hatmatrix in the model if I want to do iterative fitting, e.g. GCV #move to model or use it from a results instance inside the model, # each call to fit returns results instance class TheilRegressionResults(RegressionResults): #cache def hatmatrix_diag(self): ''' diag(X' xpxi X) where xpxi = (X'X + lambd * sigma_prior)^{-1} Notes ----- uses wexog, so this includes weights or sigma - check this case not clear whether I need to multiply by sigmahalf, i.e. (W^{-0.5} X) (X' W X)^{-1} (W^{-0.5} X)' or (W X) (X' W X)^{-1} (W X)' projection y_hat = H y or in terms of transformed variables (W^{-0.5} y) might be wrong for WLS and GLS case ''' xpxi = self.model.normalized_cov_params #something fishy with self.normalized_cov_params in result, doesn't update #print self.model.wexog.shape, np.dot(xpxi, self.model.wexog.T).shape return (self.model.wexog * np.dot(xpxi, self.model.wexog.T).T).sum(1) def hatmatrix_trace(self): return self.hatmatrix_diag().sum() #this doesn't update df_resid @property #needs to be property or attribute (no call) def df_model(self): return self.hatmatrix_trace() #Note: mse_resid uses df_resid not nobs-k_vars, which might differ if df_model, tr(H), is used #in paper for gcv ess/nobs is used instead of mse_resid def gcv(self): return self.mse_resid / (1. - self.hatmatrix_trace() / self.nobs)**2 def cv(self): return ((self.resid / (1. - self.hatmatrix_diag()))**2).sum() / self.nobs def aicc(self): aic = np.log(self.mse_resid) + 1 aic += 2 * (1. + self.hatmatrix_trace()) / (self.nobs - self.hatmatrix_trace() -2) return aic #contrast/restriction matrices, temporary location def coef_restriction_meandiff(n_coeffs, n_vars=None, position=0): reduced = np.eye(n_coeffs) - 1./n_coeffs if n_vars is None: return reduced else: full = np.zeros((n_coeffs, n_vars)) full[:, position:position+n_coeffs] = reduced return full def coef_restriction_diffbase(n_coeffs, n_vars=None, position=0, base_idx=0): reduced = -np.eye(n_coeffs) #make all rows, drop one row later reduced[:, base_idx] = 1 keep = range(n_coeffs) del keep[base_idx] reduced = np.take(reduced, keep, axis=0) if n_vars is None: return reduced else: full = np.zeros((n_coeffs-1, n_vars)) full[:, position:position+n_coeffs] = reduced return full def next_odd(d): return d + (1 - d % 2) def coef_restriction_diffseq(n_coeffs, degree=1, n_vars=None, position=0, base_idx=0): #check boundaries, returns "valid" ? if degree == 1: diff_coeffs = [-1, 1] n_points = 2 elif degree > 1: from scipy import misc n_points = next_odd(degree + 1) #next odd integer after degree+1 diff_coeffs = misc.central_diff_weights(n_points, ndiv=degree) dff = np.concatenate((diff_coeffs, np.zeros(n_coeffs - len(diff_coeffs)))) from scipy import linalg reduced = linalg.toeplitz(dff, np.zeros(n_coeffs - len(diff_coeffs) + 1)).T #reduced = np.kron(np.eye(n_coeffs-n_points), diff_coeffs) if n_vars is None: return reduced else: full = np.zeros((n_coeffs-1, n_vars)) full[:, position:position+n_coeffs] = reduced return full ## ## R = np.c_[np.zeros((n_groups, k_vars-1)), np.eye(n_groups)] ## r = np.zeros(n_groups) ## R = np.c_[np.zeros((n_groups-1, k_vars)), ## np.eye(n_groups-1)-1./n_groups * np.ones((n_groups-1, n_groups-1))] if __name__ == '__main__': import numpy as np import statsmodels.api as sm examples = [2] np.random.seed(765367) np.random.seed(97653679) nsample = 100 x = np.linspace(0,10, nsample) X = sm.add_constant(np.column_stack((x, x**2, (x/5.)**3)), prepend=True) beta = np.array([10, 1, 0.1, 0.5]) y = np.dot(X, beta) + np.random.normal(size=nsample) res_ols = sm.OLS(y, X).fit() R = [[0, 0, 0 , 1]] r = [0] #, 0, 0 , 0] lambd = 1 #1e-4 mod = TheilGLS(y, X, r_matrix=R, q_matrix=r, sigma_prior=lambd) res = mod.fit() print res_ols.params print res.params #example 2 #I need more flexible penalization in example, the penalization should #get stronger for higher order terms #np.random.seed(1) nobs = 200 k_vars = 10 k_true = 6 sig_e = 0.25 #0.5 x = np.linspace(-2,2, nobs) #X = sm.add_constant(np.column_stack((x, x**2, (x/5.)**3)), prepend=True) X = (x/x.max())[:,None]**np.arange(k_vars) beta = np.zeros(k_vars) beta[:k_true] = np.array([1, -2, 0.5, 1.5, -0.1, 0.1])[:k_true] y_true = np.dot(X, beta) y = y_true + sig_e * np.random.normal(size=nobs) res_ols = sm.OLS(y, X).fit() #R = np.c_[np.zeros((k_vars-4, 4)), np.eye(k_vars-4)] # has two large true coefficients penalized not_penalized = 4 R = np.c_[np.zeros((k_vars-not_penalized, not_penalized)), np.eye(k_vars-not_penalized)] #increasingly strong penalization R = np.c_[np.zeros((k_vars-not_penalized, not_penalized)), np.diag((1+2*np.arange(k_vars-not_penalized)))] r = np.zeros(k_vars-not_penalized) ## R = -coef_restriction_diffseq(6, 1, n_vars=10, position=4) #doesn't make sense for polynomial ## R = np.vstack((R, np.zeros(R.shape[1]))) ## R[-1,-1] = 1 r = np.zeros(R.shape[0]) lambd = 2 #1e-4 mod = TheilGLS(y, X, r_matrix=R, q_matrix=r, sigma_prior=lambd) res = mod.fit() print res_ols.params print res.params res_bic = mod.fit_minic() #this will just return zero res = mod.fit(res_bic) print res_bic for lambd in np.linspace(0, 80, 21): res_l = mod.fit(lambd) #print lambd, res_l.params[-2:], res_l.bic, res_l.bic + 1./lambd, res.df_model print lambd, res_l.params[-2:], res_l.bic, res.df_model, np.trace(res.normalized_cov_params) import matplotlib.pyplot as plt plt.figure() plt.plot(beta, 'k-o', label='true') plt.plot(res_ols.params, '-o', label='ols') plt.plot(res.params, '-o', label='theil') plt.legend() plt.title('Polynomial fitting: estimated coefficients') plt.figure() plt.plot(y, 'o') plt.plot(y_true, 'k-', label='true') plt.plot(res_ols.fittedvalues, '-', label='ols') plt.plot(res.fittedvalues, '-', label='theil') plt.legend() plt.title('Polynomial fitting: fitted values') #plt.show() if 3 in examples: #example 3 nobs = 600 nobs_i = 20 n_groups = nobs // nobs_i k_vars = 3 from statsmodels.sandbox.panel.random_panel import PanelSample dgp = PanelSample(nobs, k_vars, n_groups) dgp.group_means = 2 + np.random.randn(n_groups) #add random intercept print 'seed', dgp.seed y = dgp.generate_panel() X = np.column_stack((dgp.exog[:,1:], dgp.groups[:,None] == np.arange(n_groups))) res_ols = sm.OLS(y, X).fit() R = np.c_[np.zeros((n_groups, k_vars-1)), np.eye(n_groups)] r = np.zeros(n_groups) R = np.c_[np.zeros((n_groups-1, k_vars)), np.eye(n_groups-1)-1./n_groups * np.ones((n_groups-1, n_groups-1))] r = np.zeros(n_groups-1) R[:, k_vars-1] = -1 lambd = 1 #1e-4 mod = TheilGLS(y, X, r_matrix=R, q_matrix=r, sigma_prior=lambd) res = mod.fit() print res.params params_l = [] for lambd in np.linspace(0, 20, 21): params_l.append(mod.fit(5.*lambd).params) params_l = np.array(params_l) plt.figure() plt.plot(params_l.T) plt.title('Panel Data with random intercept: shrinkage to being equal') plt.xlabel('parameter index') plt.figure() plt.plot(params_l[:,k_vars:]) plt.title('Panel Data with random intercept: shrinkage to being equal') plt.xlabel('strength of prior') #plt.show()
bsd-3-clause
b8ccc613ab41c7ed13b0aa17e0ac537b
32.151515
110
0.609855
3.077749
false
false
false
false
yarikoptic/pystatsmodels
statsmodels/miscmodels/nonlinls.py
3
9316
'''Non-linear least squares Author: Josef Perktold based on scipy.optimize.curve_fit ''' import numpy as np from scipy import optimize from statsmodels.base.model import Model class Results(object): '''just a dummy placeholder for now most results from RegressionResults can be used here ''' pass ##def getjaccov(retval, n): ## '''calculate something and raw covariance matrix from return of optimize.leastsq ## ## I cannot figure out how to recover the Jacobian, or whether it is even ## possible ## ## this is a partial copy of scipy.optimize.leastsq ## ''' ## info = retval[-1] ## #n = len(x0) #nparams, where do I get this ## cov_x = None ## if info in [1,2,3,4]: ## from numpy.dual import inv ## from numpy.linalg import LinAlgError ## perm = np.take(np.eye(n), retval[1]['ipvt']-1,0) ## r = np.triu(np.transpose(retval[1]['fjac'])[:n,:]) ## R = np.dot(r, perm) ## try: ## cov_x = inv(np.dot(np.transpose(R),R)) ## except LinAlgError: ## print 'cov_x not available' ## pass ## return r, R, cov_x ## ##def _general_function(params, xdata, ydata, function): ## return function(xdata, *params) - ydata ## ##def _weighted_general_function(params, xdata, ydata, function, weights): ## return weights * (function(xdata, *params) - ydata) ## class NonlinearLS(Model): #or subclass a model '''Base class for estimation of a non-linear model with least squares This class is supposed to be subclassed, and the subclass has to provide a method `_predict` that defines the non-linear function `f(params) that is predicting the endogenous variable. The model is assumed to be :math: y = f(params) + error and the estimator minimizes the sum of squares of the estimated error. :math: min_parmas \sum (y - f(params))**2 f has to return the prediction for each observation. Exogenous or explanatory variables should be accessed as attributes of the class instance, and can be given as arguments when the instance is created. Warning: Weights are not correctly handled yet in the results statistics, but included when estimating the parameters. similar to scipy.optimize.curve_fit API difference: params are array_like not split up, need n_params information includes now weights similar to curve_fit no general sigma yet (OLS and WLS, but no GLS) This is currently holding on to intermediate results that are not necessary but useful for testing. Fit returns and instance of RegressionResult, in contrast to the linear model, results in this case are based on a local approximation, essentially y = f(X, params) is replaced by y = grad * params where grad is the Gradient or Jacobian with the shape (nobs, nparams). See for example Greene Examples -------- class Myfunc(NonlinearLS): def _predict(self, params): x = self.exog a, b, c = params return a*np.exp(-b*x) + c Ff we have data (y, x), we can create an instance and fit it with mymod = Myfunc(y, x) myres = mymod.fit(nparams=3) and use the non-linear regression results, for example myres.params myres.bse myres.tvalues ''' #NOTE: This needs to call super for data checking def __init__(self, endog=None, exog=None, weights=None, sigma=None, missing='none'): self.endog = endog self.exog = exog if not sigma is None: sigma = np.asarray(sigma) if sigma.ndim < 2: self.sigma = sigma self.weights = 1./sigma else: raise ValueError('correlated errors are not handled yet') else: self.weights = None def predict(self, exog, params=None): #copied from GLS, Model has different signature return self._predict(params) def _predict(self, params): pass def start_value(self): return None def geterrors(self, params, weights=None): if weights is None: if self.weights is None: return self.endog - self._predict(params) else: weights = self.weights return weights * (self.endog - self._predict(params)) def errorsumsquares(self, params): return (self.geterrors(params)**2).sum() def fit(self, start_value=None, nparams=None, **kw): #if hasattr(self, 'start_value'): #I added start_value even if it's empty, not sure about it #but it makes a visible placeholder if not start_value is None: p0 = start_value else: #nesting so that start_value is only calculated if it is needed p0 = self.start_value() if not p0 is None: pass elif not nparams is None: p0 = 0.1 * np.ones(nparams) else: raise ValueError('need information about start values for' + 'optimization') func = self.geterrors res = optimize.leastsq(func, p0, full_output=1, **kw) (popt, pcov, infodict, errmsg, ier) = res if ier not in [1,2,3,4]: msg = "Optimal parameters not found: " + errmsg raise RuntimeError(msg) err = infodict['fvec'] ydata = self.endog if (len(ydata) > len(p0)) and pcov is not None: #this can use the returned errors instead of recalculating s_sq = (err**2).sum()/(len(ydata)-len(p0)) pcov = pcov * s_sq else: pcov = None self.df_resid = len(ydata)-len(p0) self.df_model = len(p0) fitres = Results() fitres.params = popt fitres.pcov = pcov fitres.rawres = res self.wendog = self.endog #add weights self.wexog = self.jac_predict(popt) pinv_wexog = np.linalg.pinv(self.wexog) self.normalized_cov_params = np.dot(pinv_wexog, np.transpose(pinv_wexog)) #TODO: check effect of `weights` on result statistics #I think they are correctly included in cov_params #maybe not anymore, I'm not using pcov of leastsq #direct calculation with jac_predict misses the weights ## if not weights is None ## fitres.wexogw = self.weights * self.jacpredict(popt) from statsmodels.regression import RegressionResults results = RegressionResults beta = popt lfit = RegressionResults(self, beta, normalized_cov_params=self.normalized_cov_params) lfit.fitres = fitres #mainly for testing self._results = lfit return lfit def fit_minimal(self, start_value): '''minimal fitting with no extra calculations''' func = self.geterrors res = optimize.leastsq(func, start_value, full_output=0, **kw) return res def fit_random(self, ntries=10, rvs_generator=None, nparams=None): '''fit with random starting values this could be replaced with a global fitter ''' if nparams is None: nparams = self.nparams if rvs_generator is None: rvs = np.random.uniform(low=-10, high=10, size=(ntries, nparams)) else: rvs = rvs_generator(size=(ntries, nparams)) results = np.array([np.r_[self.fit_minimal(rv), rv] for rv in rvs]) #selct best results and check how many solutions are within 1e-6 of best #not sure what leastsq returns return results def jac_predict(self, params): '''jacobian of prediction function using complex step derivative This assumes that the predict function does not use complex variable but is designed to do so. ''' from statsmodels.tools.numdiff import approx_fprime_cs jaccs_err = approx_fprime_cs(params, self._predict) return jaccs_err class Myfunc(NonlinearLS): #predict model.Model has a different signature ## def predict(self, params, exog=None): ## if not exog is None: ## x = exog ## else: ## x = self.exog ## a, b, c = params ## return a*np.exp(-b*x) + c def _predict(self, params): x = self.exog a, b, c = params return a*np.exp(-b*x) + c if __name__ == '__main__': def func0(x, a, b, c): return a*np.exp(-b*x) + c def func(params, x): a, b, c = params return a*np.exp(-b*x) + c def error(params, x, y): return y - func(params, x) def error2(params, x, y): return (y - func(params, x))**2 x = np.linspace(0,4,50) params = np.array([2.5, 1.3, 0.5]) y0 = func(params, x) y = y0 + 0.2*np.random.normal(size=len(x)) res = optimize.leastsq(error, params, args=(x, y), full_output=True) ## r, R, c = getjaccov(res[1:], 3) mod = Myfunc(y, x) resmy = mod.fit(nparams=3) cf_params, cf_pcov = optimize.curve_fit(func0, x, y) cf_bse = np.sqrt(np.diag(cf_pcov)) print res[0] print cf_params print resmy.params print cf_bse print resmy.bse
bsd-3-clause
807804cbc44729be414a4b9a4d4f1a5c
29.148867
96
0.598755
3.686585
false
false
false
false
yarikoptic/pystatsmodels
statsmodels/robust/norms.py
3
19696
import numpy as np #TODO: add plots to weighting functions for online docs. class RobustNorm(object): """ The parent class for the norms used for robust regression. Lays out the methods expected of the robust norms to be used by statsmodels.RLM. Parameters ---------- None : Some subclasses have optional tuning constants. References ---------- PJ Huber. 'Robust Statistics' John Wiley and Sons, Inc., New York, 1981. DC Montgomery, EA Peck. 'Introduction to Linear Regression Analysis', John Wiley and Sons, Inc., New York, 2001. R Venables, B Ripley. 'Modern Applied Statistics in S' Springer, New York, 2002. See Also -------- statsmodels.rlm for more information on how the estimators are used and the inputs for the methods of RobustNorm and subclasses. Notes ----- Currently only M-estimators are available. """ def rho(self, z): """ The robust criterion estimator function. Abstract method: -2 loglike used in M-estimator """ raise NotImplementedError def psi(self, z): """ Derivative of rho. Sometimes referred to as the influence function. Abstract method: psi = rho' """ raise NotImplementedError def weights(self, z): """ Returns the value of psi(z) / z Abstract method: psi(z) / z """ raise NotImplementedError def psi_deriv(self, z): ''' Deriative of psi. Used to obtain robust covariance matrix. See statsmodels.rlm for more information. Abstract method: psi_derive = psi' ''' raise NotImplementedError def __call__(self, z): """ Returns the value of estimator rho applied to an input """ return self.rho(z) class LeastSquares(RobustNorm): """ Least squares rho for M-estimation and its derived functions. See also -------- statsmodels.robust.norms.RobustNorm for the methods. """ def rho(self, z): """ The least squares estimator rho function Parameters ----------- z : array 1d array Returns ------- rho : array rho(z) = (1/2.)*z**2 """ return z**2 * 0.5 def psi(self, z): """ The psi function for the least squares estimator The analytic derivative of rho Parameters ---------- z : array-like 1d array Returns ------- psi : array psi(z) = z """ return np.asarray(z) def weights(self, z): """ The least squares estimator weighting function for the IRLS algorithm. The psi function scaled by the input z Parameters ---------- z : array-like 1d array Returns ------- weights : array weights(z) = np.ones(z.shape) """ z = np.asarray(z) return np.ones(z.shape, np.float64) def psi_deriv(self, z): """ The derivative of the least squares psi function. Returns ------- psi_deriv : array ones(z.shape) Notes ----- Used to estimate the robust covariance matrix. """ return np.ones(z.shape, np.float64) class HuberT(RobustNorm): """ Huber's T for M estimation. Parameters ---------- t : float, optional The tuning constant for Huber's t function. The default value is 1.345. See also -------- statsmodels.robust.norms.RobustNorm """ def __init__(self, t=1.345): self.t = t def _subset(self, z): """ Huber's T is defined piecewise over the range for z """ z = np.asarray(z) return np.less_equal(np.fabs(z), self.t) def rho(self, z): """ The robust criterion function for Huber's t. Parameters ---------- z : array-like 1d array Returns ------- rho : array rho(z) = .5*z**2 for \|z\| <= t rho(z) = \|z\|*t - .5*t**2 for \|z\| > t """ z = np.asarray(z) test = self._subset(z) return (test * 0.5 * z**2 + (1 - test) * (np.fabs(z) * self.t - 0.5 * self.t**2)) def psi(self, z): """ The psi function for Huber's t estimator The analytic derivative of rho Parameters ---------- z : array-like 1d array Returns ------- psi : array psi(z) = z for \|z\| <= t psi(z) = sign(z)*t for \|z\| > t """ z = np.asarray(z) test = self._subset(z) return test * z + (1 - test) * self.t * np.sign(z) def weights(self, z): """ Huber's t weighting function for the IRLS algorithm The psi function scaled by z Parameters ---------- z : array-like 1d array Returns ------- weights : array weights(z) = 1 for \|z\| <= t weights(z) = t/\|z\| for \|z\| > t """ z = np.asarray(z) test = self._subset(z) absz = np.fabs(z) absz[test] = 1.0 return test + (1 - test) * self.t / absz def psi_deriv(self, z): """ The derivative of Huber's t psi function Notes ----- Used to estimate the robust covariance matrix. """ return np.less_equal(np.fabs(z), self.t) #TODO: untested, but looks right. RamsayE not available in R or SAS? class RamsayE(RobustNorm): """ Ramsay's Ea for M estimation. Parameters ---------- a : float, optional The tuning constant for Ramsay's Ea function. The default value is 0.3. See also -------- statsmodels.robust.norms.RobustNorm """ def __init__(self, a = .3): self.a = a def rho(self, z): """ The robust criterion function for Ramsay's Ea. Parameters ---------- z : array-like 1d array Returns ------- rho : array rho(z) = a**-2 * (1 - exp(-a*\|z\|)*(1 + a*\|z\|)) """ z = np.asarray(z) return (1 - np.exp(-self.a * np.fabs(z)) * (1 + self.a * np.fabs(z))) / self.a**2 def psi(self, z): """ The psi function for Ramsay's Ea estimator The analytic derivative of rho Parameters ---------- z : array-like 1d array Returns ------- psi : array psi(z) = z*exp(-a*\|z\|) """ z = np.asarray(z) return z * np.exp(-self.a * np.fabs(z)) def weights(self, z): """ Ramsay's Ea weighting function for the IRLS algorithm The psi function scaled by z Parameters ---------- z : array-like 1d array Returns ------- weights : array weights(z) = exp(-a*\|z\|) """ z = np.asarray(z) return np.exp(-self.a * np.fabs(z)) def psi_deriv(self, z): """ The derivative of Ramsay's Ea psi function. Notes ----- Used to estimate the robust covariance matrix. """ return np.exp(-self.a * np.fabs(z)) + z**2*\ np.exp(-self.a*np.fabs(z))*-self.a/np.fabs(z) class AndrewWave(RobustNorm): """ Andrew's wave for M estimation. Parameters ---------- a : float, optional The tuning constant for Andrew's Wave function. The default value is 1.339. See also -------- statsmodels.robust.norms.RobustNorm """ def __init__(self, a = 1.339): self.a = a def _subset(self, z): """ Andrew's wave is defined piecewise over the range of z. """ z = np.asarray(z) return np.less_equal(np.fabs(z), self.a * np.pi) def rho(self, z): """ The robust criterion function for Andrew's wave. Parameters ---------- z : array-like 1d array Returns ------- rho : array rho(z) = a*(1-cos(z/a)) for \|z\| <= a*pi rho(z) = 2*a for \|z\| > a*pi """ a = self.a z = np.asarray(z) test = self._subset(z) return (test * a * (1 - np.cos(z / a)) + (1 - test) * 2 * a) def psi(self, z): """ The psi function for Andrew's wave The analytic derivative of rho Parameters ---------- z : array-like 1d array Returns ------- psi : array psi(z) = sin(z/a) for \|z\| <= a*pi psi(z) = 0 for \|z\| > a*pi """ a = self.a z = np.asarray(z) test = self._subset(z) return test * np.sin(z / a) def weights(self, z): """ Andrew's wave weighting function for the IRLS algorithm The psi function scaled by z Parameters ---------- z : array-like 1d array Returns ------- weights : array weights(z) = sin(z/a)/(z/a) for \|z\| <= a*pi weights(z) = 0 for \|z\| > a*pi """ a = self.a z = np.asarray(z) test = self._subset(z) return test * np.sin(z / a) / (z / a) def psi_deriv(self, z): """ The derivative of Andrew's wave psi function Notes ----- Used to estimate the robust covariance matrix. """ test = self._subset(z) return test*np.cos(z / self.a)/self.a #TODO: this is untested class TrimmedMean(RobustNorm): """ Trimmed mean function for M-estimation. Parameters ---------- c : float, optional The tuning constant for Ramsay's Ea function. The default value is 2.0. See also -------- statsmodels.robust.norms.RobustNorm """ def __init__(self, c=2.): self.c = c def _subset(self, z): """ Least trimmed mean is defined piecewise over the range of z. """ z = np.asarray(z) return np.less_equal(np.fabs(z), self.c) def rho(self, z): """ The robust criterion function for least trimmed mean. Parameters ---------- z : array-like 1d array Returns ------- rho : array rho(z) = (1/2.)*z**2 for \|z\| <= c rho(z) = 0 for \|z\| > c """ z = np.asarray(z) test = self._subset(z) return test * z**2 * 0.5 def psi(self, z): """ The psi function for least trimmed mean The analytic derivative of rho Parameters ---------- z : array-like 1d array Returns ------- psi : array psi(z) = z for \|z\| <= c psi(z) = 0 for \|z\| > c """ z = np.asarray(z) test = self._subset(z) return test * z def weights(self, z): """ Least trimmed mean weighting function for the IRLS algorithm The psi function scaled by z Parameters ---------- z : array-like 1d array Returns ------- weights : array weights(z) = 1 for \|z\| <= c weights(z) = 0 for \|z\| > c """ z = np.asarray(z) test = self._subset(z) return test def psi_deriv(self, z): """ The derivative of least trimmed mean psi function Notes ----- Used to estimate the robust covariance matrix. """ test = self._subset(z) return test class Hampel(RobustNorm): """ Hampel function for M-estimation. Parameters ---------- a : float, optional b : float, optional c : float, optional The tuning constants for Hampel's function. The default values are a,b,c = 2, 4, 8. See also -------- statsmodels.robust.norms.RobustNorm """ def __init__(self, a = 2., b = 4., c = 8.): self.a = a self.b = b self.c = c def _subset(self, z): """ Hampel's function is defined piecewise over the range of z """ z = np.fabs(np.asarray(z)) t1 = np.less_equal(z, self.a) t2 = np.less_equal(z, self.b) * np.greater(z, self.a) t3 = np.less_equal(z, self.c) * np.greater(z, self.b) return t1, t2, t3 def rho(self, z): """ The robust criterion function for Hampel's estimator Parameters ---------- z : array-like 1d array Returns ------- rho : array rho(z) = (1/2.)*z**2 for \|z\| <= a rho(z) = a*\|z\| - 1/2.*a**2 for a < \|z\| <= b rho(z) = a*(c*\|z\|-(1/2.)*z**2)/(c-b) for b < \|z\| <= c rho(z) = a*(b + c - a) for \|z\| > c """ z = np.fabs(z) a = self.a; b = self.b; c = self.c t1, t2, t3 = self._subset(z) v = (t1 * z**2 * 0.5 + t2 * (a * z - a**2 * 0.5) + t3 * (a * (c * z - z**2 * 0.5) / (c - b) - 7 * a**2 / 6.) + (1 - t1 + t2 + t3) * a * (b + c - a)) return v def psi(self, z): """ The psi function for Hampel's estimator The analytic derivative of rho Parameters ---------- z : array-like 1d array Returns ------- psi : array psi(z) = z for \|z\| <= a psi(z) = a*sign(z) for a < \|z\| <= b psi(z) = a*sign(z)*(c - \|z\|)/(c-b) for b < \|z\| <= c psi(z) = 0 for \|z\| > c """ z = np.asarray(z) a = self.a; b = self.b; c = self.c t1, t2, t3 = self._subset(z) s = np.sign(z) z = np.fabs(z) v = s * (t1 * z + t2 * a*s + t3 * a*s * (c - z) / (c - b)) return v def weights(self, z): """ Hampel weighting function for the IRLS algorithm The psi function scaled by z Parameters ---------- z : array-like 1d array Returns ------- weights : array weights(z) = 1 for \|z\| <= a weights(z) = a/\|z\| for a < \|z\| <= b weights(z) = a*(c - \|z\|)/(\|z\|*(c-b)) for b < \|z\| <= c weights(z) = 0 for \|z\| > c """ z = np.asarray(z) a = self.a; b = self.b; c = self.c t1, t2, t3 = self._subset(z) v = (t1 + t2 * a/np.fabs(z) + t3 * a*(c-np.fabs(z))/(np.fabs(z)*(c-b))) v[np.where(np.isnan(v))]=1. # for some reason 0 returns a nan? return v def psi_deriv(self, z): t1, t2, t3 = self._subset(z) return t1 + t3 * (self.a*np.sign(z)*z)/(np.fabs(z)*(self.c-self.b)) class TukeyBiweight(RobustNorm): """ Tukey's biweight function for M-estimation. Parameters ---------- c : float, optional The tuning constant for Tukey's Biweight. The default value is c = 4.685. Notes ----- Tukey's biweight is sometime's called bisquare. """ def __init__(self, c = 4.685): self.c = c def _subset(self, z): """ Tukey's biweight is defined piecewise over the range of z """ z = np.fabs(np.asarray(z)) return np.less_equal(z, self.c) def rho(self, z): """ The robust criterion function for Tukey's biweight estimator Parameters ---------- z : array-like 1d array Returns ------- rho : array rho(z) = -(1 - (z/c)**2)**3 * c**2/6. for \|z\| <= R rho(z) = 0 for \|z\| > R """ subset = self._subset(z) return -(1 - (z / self.c)**2)**3 * subset * self.c**2 / 6. def psi(self, z): """ The psi function for Tukey's biweight estimator The analytic derivative of rho Parameters ---------- z : array-like 1d array Returns ------- psi : array psi(z) = z*(1 - (z/c)**2)**2 for \|z\| <= R psi(z) = 0 for \|z\| > R """ z = np.asarray(z) subset = self._subset(z) return z * (1 - (z / self.c)**2)**2 * subset def weights(self, z): """ Tukey's biweight weighting function for the IRLS algorithm The psi function scaled by z Parameters ---------- z : array-like 1d array Returns ------- weights : array psi(z) = (1 - (z/c)**2)**2 for \|z\| <= R psi(z) = 0 for \|z\| > R """ subset = self._subset(z) return (1 - (z / self.c)**2)**2 * subset def psi_deriv(self, z): """ The derivative of Tukey's biweight psi function Notes ----- Used to estimate the robust covariance matrix. """ subset = self._subset(z) return subset*((1 - (z/self.c)**2)**2 - (4*z**2/self.c**2) *\ (1-(z/self.c)**2)) def estimate_location(a, scale, norm=None, axis=0, initial=None, maxiter=30, tol=1.0e-06): """ M-estimator of location using self.norm and a current estimator of scale. This iteratively finds a solution to norm.psi((a-mu)/scale).sum() == 0 Parameters ---------- a : array Array over which the location parameter is to be estimated scale : array Scale parameter to be used in M-estimator norm : RobustNorm, optional Robust norm used in the M-estimator. The default is HuberT(). axis : int, optional Axis along which to estimate the location parameter. The default is 0. initial : array, optional Initial condition for the location parameter. Default is None, which uses the median of a. niter : int, optional Maximum number of iterations. The default is 30. tol : float, optional Toleration for convergence. The default is 1e-06. Returns -------- mu : array Estimate of location """ if norm is None: norm = HuberT() if initial is None: mu = np.median(a, axis) else: mu = initial for iter in range(maxiter): W = norm.weights((a-mu)/scale) nmu = np.sum(W*a, axis) / np.sum(W, axis) if np.alltrue(np.less(np.fabs(mu - nmu), scale * tol)): return nmu else: mu = nmu raise ValueError("location estimator failed to converge in %d iterations"\ % maxiter)
bsd-3-clause
8128919ac92397db89b56018b27892e0
22.09027
79
0.457098
3.745911
false
false
false
false
yarikoptic/pystatsmodels
statsmodels/sandbox/examples/example_maxent.py
6
1273
""" This is an example of using scipy.maxentropy to solve Jaynes' dice problem See Golan, Judge, and Miller Section 2.3 """ from scipy import maxentropy import numpy as np samplespace = [1., 2., 3., 4., 5., 6.] def sump(x): return x in samplespace def meanp(x): return np.mean(x) # Set the constraints # 1) We have a proper probability # 2) The mean is equal to... F = [sump, meanp] model = maxentropy.model(F, samplespace) # set the desired feature expectations K = np.ones((5,2)) K[:,1] = [2.,3.,3.5,4.,5.] model.verbose = False for i in range(K.shape[0]): model.fit(K[i]) # Output the distribution print "\nFitted model parameters are:\n" + str(model.params) print "\nFitted distribution is:" p = model.probdist() for j in range(len(model.samplespace)): x = model.samplespace[j] print "y = %-15s\tx = %-15s" %(str(K[i,1])+":",str(x) + ":") + \ " p(x) = "+str(p[j]) # Now show how well the constraints are satisfied: print print "Desired constraints:" print "\tsum_{i}p_{i}= 1" print "\tE[X] = %-15s" % str(K[i,1]) print print "Actual expectations under the fitted model:" print "\tsum_{i}p_{i} =", np.sum(p) print "\tE[X] = " + str(np.sum(p*np.arange(1,7)))
bsd-3-clause
eabe52655ab86ccb21019cfbd9d671e5
24.46
74
0.604085
2.873589
false
false
false
false
yarikoptic/pystatsmodels
statsmodels/tsa/tests/results/arima111_results.py
35
44167
import numpy as np llf = np.array([-241.75576160303]) nobs = np.array([ 202]) k = np.array([ 4]) k_exog = np.array([ 1]) sigma = np.array([ .79987660416529]) chi2 = np.array([ 342.91413339514]) df_model = np.array([ 2]) k_ar = np.array([ 1]) k_ma = np.array([ 1]) params = np.array([ .88084748605315, .93989719451385, -.7709851377434, .79987660416529]) cov_params = np.array([ .15020189867396, -.01642122563089, .01456018801049, -.00156750041014, -.01642122563089, .0032067778715, -.00350387326241, .00059634328354, .01456018801049, -.00350387326241, .00480028434835, -.00068065418463, -.00156750041014, .00059634328354, -.00068065418463, .00029322997097]).reshape(4,4) xb = np.array([ .88084751367569, .88084751367569, .65303039550781, .55365419387817, .45908725261688, .42810925841331, .37837743759155, .37686342000961, .35719576478004, .3220648765564, .31943875551224, .30907514691353, .30120712518692, .31383177638054, .29652059078217, .30856171250343, .30095273256302, .29171526432037, .31331890821457, .30463594198227, .31990340352058, .30126947164536, .29703867435455, .29884466528893, .31037190556526, .30912432074547, .32505416870117, .31537705659866, .33494210243225, .37874156236649, .37366089224815, .40859284996986, .37640652060509, .37692713737488, .39422073960304, .40755322575569, .43472331762314, .43878075480461, .47569087147713, .48725643754005, .49617394804955, .53683114051819, .55128628015518, .56243091821671, .58791494369507, .60756206512451, .58892780542374, .59145200252533, .59339815378189, .54422444105148, .55698639154434, .53304374217987, .51458370685577, .50035130977631, .48937830328941, .49780988693237, .52120143175125, .62369203567505, .6182547211647, .76608312129974, .84627467393875, .92499214410782, .96879118680954, 1.0870156288147, 1.1105998754501, 1.0274360179901, 1.013991355896, .98673474788666, .96571969985962, .84817039966583, .85888928174973, .86715340614319, .85663330554962, .93297851085663, .90738350152969, .88765007257462, .92311006784439, .96734017133713, 1.0690053701401, 1.1473876237869, 1.1740373373032, 1.3128218650818, 1.4704967737198, 1.5582785606384, 1.7273052930832, 1.8745132684708, 1.7853132486343, 1.7841064929962, 1.850741147995, 1.800768494606, 1.8466963768005, 1.7976499795914, 1.6078149080276, 1.3938897848129, 1.5498898029327, 1.3492304086685, 1.059396147728, 1.0217411518097, 1.0096007585526, 1.0002405643463, 1.0436969995499, 1.0603114366531, 1.0055546760559, .99712115526199, .92305397987366, .9841884970665, .92997401952744, .90506774187088, .9872123003006, .61137217283249, .65943044424057, .67959040403366, .77959072589874, .87357920408249, .91226226091385, .95897603034973, .96120971441269, .99671375751495, 1.0409790277481, 1.0919979810715, 1.1144404411316, 1.2330915927887, 1.2401138544083, 1.161071896553, 1.3028255701065, 1.2938764095306, 1.3207612037659, 1.5610725879669, 1.4760913848877, 1.258552312851, 1.2090681791306, 1.1540271043777, 1.12848341465, 1.1087870597839, 1.0936040878296, 1.0987877845764, 1.0858948230743, 1.0590622425079, .98770052194595, 1.0002481937408, .94235575199127, .93150353431702, .97381073236465, .9726470708847, .98864215612411, 1.0347559452057, .98585307598114, .96503925323486, .9996662735939, 1.0601476430893, 1.022319316864, 1.043828368187, 1.0604115724564, .95495897531509, .87365657091141, .91232192516327, .84078407287598, .73495537042618, .78849309682846, .77909576892853, .78874284029007, .8637443780899, .8540056347847, .94784545898438, .98641014099121, 1.0837067365646, 1.1925053596497, 1.1750392913818, 1.2460317611694, 1.1487410068512, 1.1075156927109, .94060403108597, .7950227856636, .93615245819092, .89293897151947, .94407802820206, 1.0172899961472, .93860250711441, .86104601621628, .91948908567429, .99833220243454, 1.008442401886, 1.1175880432129, 1.2017351388931, 1.1483734846115, 1.2761443853378, 1.188849568367, 1.7296310663223, 1.4202431440353, 1.3675138950348, 1.445098400116, 1.031960606575, 1.1313284635544, 1.3214453458786, 1.3112732172012, 1.367110490799, 1.674845457077, 1.5979281663895, 2.064112663269, 1.3536450862885, .30015936493874, .36831066012383, .64060544967651]) y = np.array([np.nan, 29.860847473145, 29.803030014038, 29.903654098511, 29.82908821106, 29.968111038208, 29.928377151489, 30.126863479614, 30.197195053101, 30.132064819336, 30.23943901062, 30.289073944092, 30.341207504272, 30.523830413818, 30.516519546509, 30.68856048584, 30.740953445435, 30.771715164185, 31.003318786621, 31.054636001587, 31.25990486145, 31.251270294189, 31.317039489746, 31.418846130371, 31.590372085571, 31.689123153687, 31.905054092407, 31.965375900269, 32.214942932129, 32.658740997314, 32.823661804199, 33.258590698242, 33.27640914917, 33.47692489624, 33.7942237854, 34.107555389404, 34.534721374512, 34.83878326416, 35.375694274902, 35.787254333496, 36.196174621582, 36.83683013916, 37.3512840271, 37.86243057251, 38.487915039063, 39.107563018799, 39.488929748535, 39.991455078125, 40.49340057373, 40.644222259521, 41.156986236572, 41.433044433594, 41.714584350586, 42.000350952148, 42.289379119873, 42.697811126709, 43.221202850342, 44.323692321777, 44.818256378174, 46.366081237793, 47.64627456665, 49.024990081787, 50.26879119873, 52.087017059326, 53.410598754883, 54.027435302734, 55.01399230957, 55.886737823486, 56.765720367432, 56.948169708252, 57.858890533447, 58.767154693604, 59.556632995605, 60.93297958374, 61.707382202148, 62.487648010254, 63.623111724854, 64.867340087891, 66.569007873535, 68.247383117676, 69.674034118652, 71.912818908691, 74.470497131348, 76.758277893066, 79.72730255127, 82.774513244629, 84.385314941406, 86.484100341797, 89.050735473633, 90.900764465332, 93.346694946289, 95.197654724121, 96.007820129395, 96.393890380859, 99.04988861084, 99.449226379395, 98.959396362305, 99.821746826172, 100.80960083008, 101.80024719238, 103.1436920166, 104.36031341553, 105.10555267334, 106.09712219238, 106.62305450439, 107.98419189453, 108.62997436523, 109.40506744385, 110.88721466064, 109.31137084961, 110.15943145752, 110.87958526611, 112.17959594727, 113.57357788086, 114.71226501465, 115.95897674561, 116.9612121582, 118.1967086792, 119.54097747803, 120.99199676514, 122.31443786621, 124.33309173584, 125.74011230469, 126.56107330322, 128.80282592773, 130.19386291504, 131.82075500488, 134.96105957031, 136.17608642578, 136.35855102539, 137.40905761719, 138.35401916504, 139.42848205566, 140.50877380371, 141.59359741211, 142.79878234863, 143.88589477539, 144.85906982422, 145.48770141602, 146.60025024414, 147.24235534668, 148.13150024414, 149.37380981445, 150.3726348877, 151.48864746094, 152.83476257324, 153.58586120605, 154.46504211426, 155.69966125488, 157.16015625, 158.0223236084, 159.24382019043, 160.46040344238, 160.85494995117, 161.27365112305, 162.41232299805, 162.84078979492, 162.93495178223, 163.98849487305, 164.67909240723, 165.48873901367, 166.76373291016, 167.55400085449, 169.0478515625, 170.2864074707, 171.98370361328, 173.89250183105, 175.07502746582, 176.84603881836, 177.54873657227, 178.50750732422, 178.5406036377, 178.49502563477, 180.23616027832, 180.89294433594, 182.14407348633, 183.61729431152, 184.13859558105, 184.56105041504, 185.81948852539, 187.29833984375, 188.40843200684, 190.21759033203, 192.00173950195, 192.9483795166, 195.07614135742, 195.88883972168, 200.92962646484, 200.82023620605, 202.06750488281, 204.1450958252, 202.93196105957, 204.70533752441, 207.24143981934, 208.6492767334, 210.50010681152, 214.16984558105, 215.59492492676, 220.67411804199, 218.24264526367, 212.47415161133, 213.03932189941, 215.10960388184]) resid = np.array([np.nan, -.71084743738174, -.45302960276604, -.5336537361145, -.28908717632294, -.41811093688011, -.17837668955326, -.28686326742172, -.38719645142555, -.21206425130367, -.25943928956985, -.24907378852367, -.13120894134045, -.3038315474987, -.13652075827122, -.24856032431126, -.26095372438431, -.0817142650485, -.25331944227219, -.11463540792465, -.30990317463875, -.2312697917223, -.19703827798367, -.13884480297565, -.21037344634533, -.10912357270718, -.25505447387695, -.08537751436234, .06505750864744, -.20873957872391, .02633681893349, -.35858979821205, -.1764095723629, -.07692407816648, -.09422151744366, -.00755552388728, -.13472028076649, .06121923774481, -.07569316774607, -.08725491166115, .10382451862097, -.03683112934232, -.05128625407815, .03757134452462, .0120835499838, -.20756052434444, -.08892779797316, -.09145200997591, -.3934012055397, -.04422445222735, -.25698333978653, -.23304453492165, -.21458448469639, -.2003520578146, -.08937677741051, .00219011562876, .47879853844643, -.12369203567505, .78174299001694, .43391767144203, .4537245631218, .27500861883163, .73120957612991, .21298357844353, -.41059911251068, -.02743596211076, -.11398979276419, -.08673703670502, -.66572046279907, .05183110013604, .04111221805215, -.06715416908264, .44336593151093, -.13297925889492, -.1073842421174, .21235218644142, .27689066529274, .63265830278397, .53099316358566, .25261387228966, .92596107721329, 1.0871796607971, .72950023412704, 1.2417244911194, 1.1726962327957, -.17451636493206, .31468516588211, .71589350700378, .04926039651036, .59923303127289, .05330519750714, -.79764997959137, -1.0078164339066, 1.1061102151871, -.94989138841629, -1.5492273569107, -.15939457714558, -.02174116671085, -.00960071571171, .29975482821465, .15630762279034, -.2603160738945, -.00555467186496, -.3971226811409, .37694907188416, -.28419154882431, -.12997098267078, .49493381381035, -2.1872169971466, .18863087892532, .04056651890278, .52041417360306, .52040469646454, .22642692923546, .28773468732834, .0410239957273, .2387872338295, .30328929424286, .35902243852615, .20799747109413, .78556102514267, .16690990328789, -.34011232852936, .93892657756805, .0971682742238, .30612966418266, 1.5792326927185, -.26106956601143, -1.0760822296143, -.15856145322323, -.2090682387352, -.05402099713683, -.02849259786308, -.00878097955137, .10639289021492, .00121826829854, -.08589478582144, -.35906526446342, .11230555176735, -.30025118589401, -.04236188530922, .26849341392517, .02618926763535, .12735903263092, .31136092543602, -.23475293815136, -.08585914969444, .23495768010616, .40034285187721, -.1601537913084, .17767761647701, .15616858005524, -.56041151285172, -.45495894551277, .2263495028019, -.41232195496559, -.64078712463379, .26504465937614, -.08849616348743, .02090725488961, .41125410795212, -.06374131888151, .54600352048874, .25215145945549, .61358070373535, .71629631519318, .00749156065285, .52497291564941, -.44604399800301, -.14874097704887, -.90750348567963, -.84061318635941, .80498331785202, -.23615552484989, .30705797672272, .45593112707138, -.41729912161827, -.43860253691673, .33895090222359, .48052009940147, .10165861994028, .69156980514526, .58240884542465, -.20173519849777, .85162657499313, -.37615045905113, 3.3111503124237, -1.5296341180801, -.12024004757404, .63248610496521, -2.2451014518738, .64205056428909, 1.2146645784378, .09655395895243, .48372489213943, 1.9948890209198, -.17284658551216, 3.0150785446167, -3.7851057052612, -6.0686569213867, .19684991240501, 1.4296782016754, 1.2753949165344]) yr = np.array([np.nan, -.71084743738174, -.45302960276604, -.5336537361145, -.28908717632294, -.41811093688011, -.17837668955326, -.28686326742172, -.38719645142555, -.21206425130367, -.25943928956985, -.24907378852367, -.13120894134045, -.3038315474987, -.13652075827122, -.24856032431126, -.26095372438431, -.0817142650485, -.25331944227219, -.11463540792465, -.30990317463875, -.2312697917223, -.19703827798367, -.13884480297565, -.21037344634533, -.10912357270718, -.25505447387695, -.08537751436234, .06505750864744, -.20873957872391, .02633681893349, -.35858979821205, -.1764095723629, -.07692407816648, -.09422151744366, -.00755552388728, -.13472028076649, .06121923774481, -.07569316774607, -.08725491166115, .10382451862097, -.03683112934232, -.05128625407815, .03757134452462, .0120835499838, -.20756052434444, -.08892779797316, -.09145200997591, -.3934012055397, -.04422445222735, -.25698333978653, -.23304453492165, -.21458448469639, -.2003520578146, -.08937677741051, .00219011562876, .47879853844643, -.12369203567505, .78174299001694, .43391767144203, .4537245631218, .27500861883163, .73120957612991, .21298357844353, -.41059911251068, -.02743596211076, -.11398979276419, -.08673703670502, -.66572046279907, .05183110013604, .04111221805215, -.06715416908264, .44336593151093, -.13297925889492, -.1073842421174, .21235218644142, .27689066529274, .63265830278397, .53099316358566, .25261387228966, .92596107721329, 1.0871796607971, .72950023412704, 1.2417244911194, 1.1726962327957, -.17451636493206, .31468516588211, .71589350700378, .04926039651036, .59923303127289, .05330519750714, -.79764997959137, -1.0078164339066, 1.1061102151871, -.94989138841629, -1.5492273569107, -.15939457714558, -.02174116671085, -.00960071571171, .29975482821465, .15630762279034, -.2603160738945, -.00555467186496, -.3971226811409, .37694907188416, -.28419154882431, -.12997098267078, .49493381381035, -2.1872169971466, .18863087892532, .04056651890278, .52041417360306, .52040469646454, .22642692923546, .28773468732834, .0410239957273, .2387872338295, .30328929424286, .35902243852615, .20799747109413, .78556102514267, .16690990328789, -.34011232852936, .93892657756805, .0971682742238, .30612966418266, 1.5792326927185, -.26106956601143, -1.0760822296143, -.15856145322323, -.2090682387352, -.05402099713683, -.02849259786308, -.00878097955137, .10639289021492, .00121826829854, -.08589478582144, -.35906526446342, .11230555176735, -.30025118589401, -.04236188530922, .26849341392517, .02618926763535, .12735903263092, .31136092543602, -.23475293815136, -.08585914969444, .23495768010616, .40034285187721, -.1601537913084, .17767761647701, .15616858005524, -.56041151285172, -.45495894551277, .2263495028019, -.41232195496559, -.64078712463379, .26504465937614, -.08849616348743, .02090725488961, .41125410795212, -.06374131888151, .54600352048874, .25215145945549, .61358070373535, .71629631519318, .00749156065285, .52497291564941, -.44604399800301, -.14874097704887, -.90750348567963, -.84061318635941, .80498331785202, -.23615552484989, .30705797672272, .45593112707138, -.41729912161827, -.43860253691673, .33895090222359, .48052009940147, .10165861994028, .69156980514526, .58240884542465, -.20173519849777, .85162657499313, -.37615045905113, 3.3111503124237, -1.5296341180801, -.12024004757404, .63248610496521, -2.2451014518738, .64205056428909, 1.2146645784378, .09655395895243, .48372489213943, 1.9948890209198, -.17284658551216, 3.0150785446167, -3.7851057052612, -6.0686569213867, .19684991240501, 1.4296782016754, 1.2753949165344]) mse = np.array([ .7963672876358, .7963672876358, .71457105875015, .67959600687027, .66207146644592, .65259438753128, .64725720882416, .644182741642, .64238852262497, .64133352041245, .64071041345596, .6403414607048, .64012265205383, .63999271392822, .63991558551788, .63986974954605, .63984251022339, .63982629776001, .6398167014122, .6398109793663, .63980758190155, .63980555534363, .63980436325073, .639803647995, .63980323076248, .63980293273926, .63980281352997, .63980269432068, .63980263471603, .63980263471603, .63980263471603, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139, .63980257511139]) stdp = np.array([ .88084751367569, .88084751367569, .65303039550781, .55365419387817, .45908725261688, .42810925841331, .37837743759155, .37686342000961, .35719576478004, .3220648765564, .31943875551224, .30907514691353, .30120712518692, .31383177638054, .29652059078217, .30856171250343, .30095273256302, .29171526432037, .31331890821457, .30463594198227, .31990340352058, .30126947164536, .29703867435455, .29884466528893, .31037190556526, .30912432074547, .32505416870117, .31537705659866, .33494210243225, .37874156236649, .37366089224815, .40859284996986, .37640652060509, .37692713737488, .39422073960304, .40755322575569, .43472331762314, .43878075480461, .47569087147713, .48725643754005, .49617394804955, .53683114051819, .55128628015518, .56243091821671, .58791494369507, .60756206512451, .58892780542374, .59145200252533, .59339815378189, .54422444105148, .55698639154434, .53304374217987, .51458370685577, .50035130977631, .48937830328941, .49780988693237, .52120143175125, .62369203567505, .6182547211647, .76608312129974, .84627467393875, .92499214410782, .96879118680954, 1.0870156288147, 1.1105998754501, 1.0274360179901, 1.013991355896, .98673474788666, .96571969985962, .84817039966583, .85888928174973, .86715340614319, .85663330554962, .93297851085663, .90738350152969, .88765007257462, .92311006784439, .96734017133713, 1.0690053701401, 1.1473876237869, 1.1740373373032, 1.3128218650818, 1.4704967737198, 1.5582785606384, 1.7273052930832, 1.8745132684708, 1.7853132486343, 1.7841064929962, 1.850741147995, 1.800768494606, 1.8466963768005, 1.7976499795914, 1.6078149080276, 1.3938897848129, 1.5498898029327, 1.3492304086685, 1.059396147728, 1.0217411518097, 1.0096007585526, 1.0002405643463, 1.0436969995499, 1.0603114366531, 1.0055546760559, .99712115526199, .92305397987366, .9841884970665, .92997401952744, .90506774187088, .9872123003006, .61137217283249, .65943044424057, .67959040403366, .77959072589874, .87357920408249, .91226226091385, .95897603034973, .96120971441269, .99671375751495, 1.0409790277481, 1.0919979810715, 1.1144404411316, 1.2330915927887, 1.2401138544083, 1.161071896553, 1.3028255701065, 1.2938764095306, 1.3207612037659, 1.5610725879669, 1.4760913848877, 1.258552312851, 1.2090681791306, 1.1540271043777, 1.12848341465, 1.1087870597839, 1.0936040878296, 1.0987877845764, 1.0858948230743, 1.0590622425079, .98770052194595, 1.0002481937408, .94235575199127, .93150353431702, .97381073236465, .9726470708847, .98864215612411, 1.0347559452057, .98585307598114, .96503925323486, .9996662735939, 1.0601476430893, 1.022319316864, 1.043828368187, 1.0604115724564, .95495897531509, .87365657091141, .91232192516327, .84078407287598, .73495537042618, .78849309682846, .77909576892853, .78874284029007, .8637443780899, .8540056347847, .94784545898438, .98641014099121, 1.0837067365646, 1.1925053596497, 1.1750392913818, 1.2460317611694, 1.1487410068512, 1.1075156927109, .94060403108597, .7950227856636, .93615245819092, .89293897151947, .94407802820206, 1.0172899961472, .93860250711441, .86104601621628, .91948908567429, .99833220243454, 1.008442401886, 1.1175880432129, 1.2017351388931, 1.1483734846115, 1.2761443853378, 1.188849568367, 1.7296310663223, 1.4202431440353, 1.3675138950348, 1.445098400116, 1.031960606575, 1.1313284635544, 1.3214453458786, 1.3112732172012, 1.367110490799, 1.674845457077, 1.5979281663895, 2.064112663269, 1.3536450862885, .30015936493874, .36831066012383, .64060544967651]) icstats = np.array([ 202, np.nan, -241.75576160303, 4, 491.51152320605, 504.74459399566]) class Bunch(dict): def __init__(self, **kw): dict.__init__(self, kw) self.__dict__ = self results = Bunch(llf=llf, nobs=nobs, k=k, k_exog=k_exog, sigma=sigma, chi2=chi2, df_model=df_model, k_ar=k_ar, k_ma=k_ma, params=params, cov_params=cov_params, xb=xb, y=y, resid=resid, yr=yr, mse=mse, stdp=stdp, icstats=icstats, )
bsd-3-clause
fbfd2756a96ed98fa7a522a30805cbe2
33.478532
229
0.400978
3.831945
false
false
false
false
yarikoptic/pystatsmodels
examples/example_wls.py
2
2809
"""Weighted Least Squares """ import numpy as np from scipy import stats import statsmodels.api as sm import matplotlib.pyplot as plt from statsmodels.sandbox.regression.predstd import wls_prediction_std from statsmodels.iolib.table import (SimpleTable, default_txt_fmt) np.random.seed(1024) # WLS Estimation # -------------- # Artificial data: Heteroscedasticity 2 groups # ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ # Model assumptions: # # * Misspecificaion: true model is quadratic, estimate only linear # * Independent noise/error term # * Two groups for error variance, low and high variance groups nsample = 50 x = np.linspace(0, 20, nsample) X = np.c_[x, (x - 5)**2, np.ones(nsample)] beta = [0.5, -0.01, 5.] sig = 0.5 w = np.ones(nsample) w[nsample * 6 / 10:] = 3 y_true = np.dot(X, beta) e = np.random.normal(size=nsample) y = y_true + sig * w * e X = X[:, [0, 2]] #WLS knowing the true variance ratio of heteroscedasticity #^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ mod_wls = sm.WLS(y, X, weights=1. / w) res_wls = mod_wls.fit() print res_wls.summary() #OLS vs. WLS #----------- # Estimate an OLS model for comparison res_ols = sm.OLS(y, X).fit() # Compare the estimated parameters in WLS and OLS print res_ols.params print res_wls.params # Compare the WLS standard errors to heteroscedasticity corrected OLS standard # errors: se = np.vstack([[res_wls.bse], [res_ols.bse], [res_ols.HC0_se], [res_ols.HC1_se], [res_ols.HC2_se], [res_ols.HC3_se]]) se = np.round(se, 4) colnames = 'x1', 'const' rownames = 'WLS', 'OLS', 'OLS_HC0', 'OLS_HC1', 'OLS_HC3', 'OLS_HC3' tabl = SimpleTable(se, colnames, rownames, txt_fmt=default_txt_fmt) print tabl # Calculate OLS prediction interval covb = res_ols.cov_params() prediction_var = res_ols.mse_resid + (X * np.dot(covb, X.T).T).sum(1) prediction_std = np.sqrt(prediction_var) tppf = stats.t.ppf(0.975, res_ols.df_resid) # Draw a plot to compare predicted values in WLS and OLS: prstd, iv_l, iv_u = wls_prediction_std(res_wls) plt.figure(); plt.plot(x, y, 'o', x, y_true, 'b-'); plt.plot(x, res_ols.fittedvalues, 'r--'); plt.plot(x, res_ols.fittedvalues + tppf * prediction_std, 'r--'); plt.plot(x, res_ols.fittedvalues - tppf * prediction_std, 'r--'); plt.plot(x, res_wls.fittedvalues, 'g--.'); plt.plot(x, iv_u, 'g--'); plt.plot(x, iv_l, 'g--'); #@savefig wls_ols_0.png plt.title('blue: true, red: OLS, green: WLS'); # Feasible Weighted Least Squares (2-stage FWLS) # ---------------------------------------------- resid1 = res_ols.resid[w == 1.] var1 = resid1.var(ddof=int(res_ols.df_model) + 1) resid2 = res_ols.resid[w != 1.] var2 = resid2.var(ddof=int(res_ols.df_model) + 1) w_est = w.copy() w_est[w != 1.] = np.sqrt(var2) / np.sqrt(var1) res_fwls = sm.WLS(y, X, 1. / w_est).fit() print res_fwls.summary()
bsd-3-clause
a6a01de18b895ef01fcb1ec34d62ef74
30.211111
79
0.632609
2.519283
false
false
false
false
yarikoptic/pystatsmodels
statsmodels/examples/ex_generic_mle_tdist.py
3
39633
# -*- coding: utf-8 -*- """ Created on Wed Jul 28 08:28:04 2010 Author: josef-pktd """ import numpy as np from scipy import stats, special, optimize import statsmodels.api as sm from statsmodels.base.model import GenericLikelihoodModel #redefine some shortcuts np_log = np.log np_pi = np.pi sps_gamln = special.gammaln def maxabs(arr1, arr2): return np.max(np.abs(arr1 - arr2)) def maxabsrel(arr1, arr2): return np.max(np.abs(arr2 / arr1 - 1)) #global store_params = [] class MyT(GenericLikelihoodModel): '''Maximum Likelihood Estimation of Linear Model with t-distributed errors This is an example for generic MLE which has the same statistical model as discretemod.Poisson. Except for defining the negative log-likelihood method, all methods and results are generic. Gradients and Hessian and all resulting statistics are based on numerical differentiation. ''' def loglike(self, params): return -self.nloglikeobs(params).sum(0) # copied from discretemod.Poisson def nloglikeobs(self, params): """ Loglikelihood of Poisson model Parameters ---------- params : array-like The parameters of the model. Returns ------- The log likelihood of the model evaluated at `params` Notes -------- .. math :: \\ln L=\\sum_{i=1}^{n}\\left[-\\lambda_{i}+y_{i}x_{i}^{\\prime}\\beta-\\ln y_{i}!\\right] """ #print len(params), store_params.append(params) if not self.fixed_params is None: #print 'using fixed' params = self.expandparams(params) beta = params[:-2] df = params[-2] scale = params[-1] loc = np.dot(self.exog, beta) endog = self.endog x = (endog - loc)/scale #next part is stats.t._logpdf lPx = sps_gamln((df+1)/2) - sps_gamln(df/2.) lPx -= 0.5*np_log(df*np_pi) + (df+1)/2.*np_log(1+(x**2)/df) lPx -= np_log(scale) # correction for scale return -lPx #Example: np.random.seed(98765678) nobs = 1000 nvars = 6 df = 5 rvs = np.random.randn(nobs, nvars-1) data_exog = sm.add_constant(rvs, prepend=False) xbeta = 0.9 + 0.1*rvs.sum(1) data_endog = xbeta + 0.1*np.random.standard_t(df, size=nobs) print data_endog.var() res_ols = sm.OLS(data_endog, data_exog).fit() print res_ols.scale print np.sqrt(res_ols.scale) print res_ols.params kurt = stats.kurtosis(res_ols.resid) df_fromkurt = 6./kurt + 4 print stats.t.stats(df_fromkurt, moments='mvsk') print stats.t.stats(df, moments='mvsk') modp = MyT(data_endog, data_exog) start_value = 0.1*np.ones(data_exog.shape[1]+2) #start_value = np.zeros(data_exog.shape[1]+2) #start_value[:nvars] = sm.OLS(data_endog, data_exog).fit().params start_value[:nvars] = res_ols.params start_value[-2] = df_fromkurt #10 start_value[-1] = np.sqrt(res_ols.scale) #0.5 modp.start_params = start_value #adding fixed parameters fixdf = np.nan * np.zeros(modp.start_params.shape) fixdf[-2] = 100 fixone = 0 if fixone: modp.fixed_params = fixdf modp.fixed_paramsmask = np.isnan(fixdf) modp.start_params = modp.start_params[modp.fixed_paramsmask] else: modp.fixed_params = None modp.fixed_paramsmask = None resp = modp.fit(start_params = modp.start_params, disp=1, method='nm')#'newton') #resp = modp.fit(start_params = modp.start_params, disp=1, method='newton') print '\nestimation results t-dist' print resp.params print resp.bse resp2 = modp.fit(start_params = resp.params, method='Newton') print 'using Newton' print resp2.params print resp2.bse from statsmodels.tools.numdiff import approx_fprime, approx_hess hb=-approx_hess(modp.start_params, modp.loglike, epsilon=-1e-4) tmp = modp.loglike(modp.start_params) print tmp.shape #np.linalg.eigh(np.linalg.inv(hb))[0] pp=np.array(store_params) print pp.min(0) print pp.max(0) ##################### Example: Pareto # estimating scale doesn't work yet, a bug somewhere ? # fit_ks works well, but no bse or other result statistics yet #import for kstest based estimation #should be replace import statsmodels.sandbox.distributions.sppatch class MyPareto(GenericLikelihoodModel): '''Maximum Likelihood Estimation pareto distribution first version: iid case, with constant parameters ''' #copied from stats.distribution def pdf(self, x, b): return b * x**(-b-1) def loglike(self, params): return -self.nloglikeobs(params).sum(0) def nloglikeobs(self, params): #print params.shape if not self.fixed_params is None: #print 'using fixed' params = self.expandparams(params) b = params[0] loc = params[1] scale = params[2] #loc = np.dot(self.exog, beta) endog = self.endog x = (endog - loc)/scale logpdf = np_log(b) - (b+1.)*np_log(x) #use np_log(1 + x) for Pareto II logpdf -= np.log(scale) #lb = loc + scale #logpdf[endog<lb] = -inf #import pdb; pdb.set_trace() logpdf[x<1] = -10000 #-np.inf return -logpdf def fit_ks(self): '''fit Pareto with nested optimization originally published on stackoverflow this doesn't trim lower values during ks optimization ''' rvs = self.endog rvsmin = rvs.min() fixdf = np.nan * np.ones(3) self.fixed_params = fixdf self.fixed_paramsmask = np.isnan(fixdf) def pareto_ks(loc, rvs): #start_scale = rvs.min() - loc # not used yet #est = self.fit_fr(rvs, 1., frozen=[np.nan, loc, np.nan]) self.fixed_params[1] = loc est = self.fit(start_params=self.start_params[self.fixed_paramsmask]).params #est = self.fit(start_params=self.start_params, method='nm').params args = (est[0], loc, est[1]) return stats.kstest(rvs,'pareto',args)[0] locest = optimize.fmin(pareto_ks, rvsmin - 1.5, (rvs,)) est = stats.pareto.fit_fr(rvs, 0., frozen=[np.nan, locest, np.nan]) args = (est[0], locest[0], est[1]) return args def fit_ks1_trim(self): '''fit Pareto with nested optimization originally published on stackoverflow ''' self.nobs = self.endog.shape[0] rvs = np.sort(self.endog) rvsmin = rvs.min() def pareto_ks(loc, rvs): #start_scale = rvs.min() - loc # not used yet est = stats.pareto.fit_fr(rvs, frozen=[np.nan, loc, np.nan]) args = (est[0], loc, est[1]) return stats.kstest(rvs,'pareto',args)[0] #locest = optimize.fmin(pareto_ks, rvsmin*0.7, (rvs,)) maxind = min(np.floor(self.nobs*0.95).astype(int), self.nobs-10) res = [] for trimidx in range(self.nobs//2, maxind): xmin = loc = rvs[trimidx] res.append([trimidx, pareto_ks(loc-1e-10, rvs[trimidx:])]) res = np.array(res) bestidx = res[np.argmin(res[:,1]),0].astype(int) print bestidx locest = rvs[bestidx] est = stats.pareto.fit_fr(rvs[bestidx:], 1., frozen=[np.nan, locest, np.nan]) args = (est[0], locest, est[1]) return args def fit_ks1(self): '''fit Pareto with nested optimization originally published on stackoverflow ''' rvs = self.endog rvsmin = rvs.min() def pareto_ks(loc, rvs): #start_scale = rvs.min() - loc # not used yet est = stats.pareto.fit_fr(rvs, 1., frozen=[np.nan, loc, np.nan]) args = (est[0], loc, est[1]) return stats.kstest(rvs,'pareto',args)[0] #locest = optimize.fmin(pareto_ks, rvsmin*0.7, (rvs,)) locest = optimize.fmin(pareto_ks, rvsmin - 1.5, (rvs,)) est = stats.pareto.fit_fr(rvs, 1., frozen=[np.nan, locest, np.nan]) args = (est[0], locest[0], est[1]) return args #y = stats.pareto.rvs(1, loc=10, scale=2, size=nobs) y = stats.pareto.rvs(1, loc=0, scale=2, size=nobs) par_start_params = np.array([1., 9., 2.]) mod_par = MyPareto(y) mod_par.start_params = np.array([1., 10., 2.]) mod_par.start_params = np.array([1., -9., 2.]) mod_par.fixed_params = None fixdf = np.nan * np.ones(mod_par.start_params.shape) fixdf[1] = 9.9 #fixdf[2] = 2. fixone = 0 if fixone: mod_par.fixed_params = fixdf mod_par.fixed_paramsmask = np.isnan(fixdf) mod_par.start_params = mod_par.start_params[mod_par.fixed_paramsmask] mod_par.df_model = 2 mod_par.df_resid = mod_par.endog.shape[0] - mod_par.df_model mod_par.data.xnames = ['shape', 'scale'] else: mod_par.fixed_params = None mod_par.fixed_paramsmask = None mod_par.df_model = 3 mod_par.df_resid = mod_par.endog.shape[0] - mod_par.df_model mod_par.data.xnames = ['shape', 'loc', 'scale'] res_par = mod_par.fit(start_params=mod_par.start_params, method='nm', maxfun=10000, maxiter=5000) #res_par2 = mod_par.fit(start_params=res_par.params, method='newton', maxfun=10000, maxiter=5000) res_parks = mod_par.fit_ks1() print res_par.params #print res_par2.params print res_parks print res_par.params[1:].sum(), sum(res_parks[1:]), mod_par.endog.min() #start new model, so we don't get two result instances with the same model instance mod_par = MyPareto(y) mod_par.fixed_params = fixdf mod_par.fixed_paramsmask = np.isnan(fixdf) mod_par.df_model = mod_par.fixed_paramsmask.sum() mod_par.df_resid = mod_par.endog.shape[0] - mod_par.df_model #mod_par.data.xnames = np.array(['shape', 'loc', 'scale'])[mod_par.fixed_paramsmask].tolist() # works also mod_par.data.xnames = [name for (name, incl) in zip(['shape', 'loc', 'scale'], mod_par.fixed_paramsmask) if incl] res_par3 = mod_par.start_params = par_start_params[mod_par.fixed_paramsmask] res5 = mod_par.fit(start_params=mod_par.start_params) ##res_parks2 = mod_par.fit_ks() ## ##res_parkst = mod_par.fit_ks1_trim() ##print res_parkst print res5.summary() print res5.t_test([[1,0]]) ''' C:\Programs\Python25\lib\site-packages\matplotlib-0.99.1-py2.5-win32.egg\matplotlib\rcsetup.py:117: UserWarning: rcParams key "numerix" is obsolete and has no effect; please delete it from your matplotlibrc file warnings.warn('rcParams key "numerix" is obsolete and has no effect;\n' 0.0686702747648 0.0164150896481 0.128121386381 [ 0.10370428 0.09921315 0.09676723 0.10457413 0.10201618 0.89964496] (array(0.0), array(1.4552599885729831), array(0.0), array(2.5072143354058238)) (array(0.0), array(1.6666666666666667), array(0.0), array(6.0)) repr(start_params) array([ 0.10370428, 0.09921315, 0.09676723, 0.10457413, 0.10201618, 0.89964496, 6.39309417, 0.12812139]) Optimization terminated successfully. Current function value: -679.951339 Iterations: 398 Function evaluations: 609 estimation results t-dist [ 0.10400826 0.10111893 0.09725133 0.10507788 0.10086163 0.8996041 4.72131318 0.09825355] [ 0.00365493 0.00356149 0.00349329 0.00362333 0.003732 0.00362716 0.7232824 0.00388829] repr(start_params) array([ 0.10400826, 0.10111893, 0.09725133, 0.10507788, 0.10086163, 0.8996041 , 4.72131318, 0.09825355]) Optimization terminated successfully. Current function value: -679.950443 Iterations 3 using Newton [ 0.10395383 0.10106762 0.09720665 0.10503384 0.10080599 0.89954546 4.70918964 0.09815885] [ 0.00365299 0.00355968 0.00349147 0.00362166 0.00373015 0.00362533 0.72014031 0.00388434] () [ 0.09992709 0.09786601 0.09387356 0.10229919 0.09756623 0.85466272 4.60459182 0.09661986] [ 0.11308292 0.10828401 0.1028508 0.11268895 0.10934726 0.94462721 7.15412655 0.13452746] repr(start_params) array([ 1., 2.]) Warning: Maximum number of function evaluations has been exceeded. >>> res_par.params array([ 7.42705803e+152, 2.17339053e+153]) >>> mod_par.loglike(mod_p.start_params) Traceback (most recent call last): File "<stdin>", line 1, in <module> NameError: name 'mod_p' is not defined >>> mod_par.loglike(mod_par.start_params) -1085.1993430947232 >>> np.log(mod_par.pdf(mod_par.start_params)) Traceback (most recent call last): File "<stdin>", line 1, in <module> TypeError: pdf() takes exactly 3 arguments (2 given) >>> np.log(mod_par.pdf(*mod_par.start_params)) 0.69314718055994529 >>> mod_par.loglike(*mod_par.start_params) Traceback (most recent call last): File "<stdin>", line 1, in <module> TypeError: loglike() takes exactly 2 arguments (3 given) >>> mod_par.loglike(mod_par.start_params) -1085.1993430947232 >>> np.log(stats.pareto.pdf(y[0],*mod_par.start_params)) -4.6414308627431353 >>> mod_par.loglike(mod_par.start_params) -1085.1993430947232 >>> mod_par.nloglikeobs(mod_par.start_params)[0] 0.29377232943845044 >>> mod_par.start_params array([ 1., 2.]) >>> np.log(stats.pareto.pdf(y[0],1,9.5,2)) -1.2806918394368461 >>> mod_par.fixed_params= None >>> mod_par.nloglikeobs(np.array([1., 10., 2.]))[0] 0.087533156771285828 >>> y[0] 12.182956907488885 >>> mod_para.endog[0] Traceback (most recent call last): File "<stdin>", line 1, in <module> NameError: name 'mod_para' is not defined >>> mod_par.endog[0] 12.182956907488885 >>> np.log(stats.pareto.pdf(y[0],1,10,2)) -0.86821349410251702 >>> np.log(stats.pareto.pdf(y[0],1.,10.,2.)) -0.86821349410251702 >>> stats.pareto.pdf(y[0],1.,10.,2.) 0.41970067762301644 >>> mod_par.loglikeobs(np.array([1., 10., 2.]))[0] -0.087533156771285828 >>> ''' ''' >>> mod_par.nloglikeobs(np.array([1., 10., 2.]))[0] 0.86821349410251691 >>> np.log(stats.pareto.pdf(y,1.,10.,2.)).sum() -2627.9403758026938 ''' #''' #C:\Programs\Python25\lib\site-packages\matplotlib-0.99.1-py2.5-win32.egg\matplotlib\rcsetup.py:117: UserWarning: rcParams key "numerix" is obsolete and has no effect; # please delete it from your matplotlibrc file # warnings.warn('rcParams key "numerix" is obsolete and has no effect;\n' #0.0686702747648 #0.0164150896481 #0.128121386381 #[ 0.10370428 0.09921315 0.09676723 0.10457413 0.10201618 0.89964496] #(array(0.0), array(1.4552599885729827), array(0.0), array(2.5072143354058203)) #(array(0.0), array(1.6666666666666667), array(0.0), array(6.0)) #repr(start_params) array([ 0.10370428, 0.09921315, 0.09676723, 0.10457413, 0.10201618, # 0.89964496, 6.39309417, 0.12812139]) #Optimization terminated successfully. # Current function value: -679.951339 # Iterations: 398 # Function evaluations: 609 # #estimation results t-dist #[ 0.10400826 0.10111893 0.09725133 0.10507788 0.10086163 0.8996041 # 4.72131318 0.09825355] #[ 0.00365493 0.00356149 0.00349329 0.00362333 0.003732 0.00362716 # 0.72325227 0.00388822] #repr(start_params) array([ 0.10400826, 0.10111893, 0.09725133, 0.10507788, 0.10086163, # 0.8996041 , 4.72131318, 0.09825355]) #Optimization terminated successfully. # Current function value: -679.950443 # Iterations 3 #using Newton #[ 0.10395383 0.10106762 0.09720665 0.10503384 0.10080599 0.89954546 # 4.70918964 0.09815885] #[ 0.00365299 0.00355968 0.00349147 0.00362166 0.00373015 0.00362533 # 0.72014669 0.00388436] #() #[ 0.09992709 0.09786601 0.09387356 0.10229919 0.09756623 0.85466272 # 4.60459182 0.09661986] #[ 0.11308292 0.10828401 0.1028508 0.11268895 0.10934726 0.94462721 # 7.15412655 0.13452746] #repr(start_params) array([ 1., 2.]) #Warning: Maximum number of function evaluations has been exceeded. #repr(start_params) array([ 3.06504406e+302, 3.29325579e+303]) #Traceback (most recent call last): # File "C:\Josef\eclipsegworkspace\statsmodels-josef-experimental-gsoc\scikits\statsmodels\examples\ex_generic_mle_tdist.py", line 222, in <module> # res_par2 = mod_par.fit(start_params=res_par.params, method='newton', maxfun=10000, maxiter=5000) # File "c:\josef\eclipsegworkspace\statsmodels-josef-experimental-gsoc\scikits\statsmodels\model.py", line 547, in fit # disp=disp, callback=callback, **kwargs) # File "c:\josef\eclipsegworkspace\statsmodels-josef-experimental-gsoc\scikits\statsmodels\model.py", line 262, in fit # newparams = oldparams - np.dot(np.linalg.inv(H), # File "C:\Programs\Python25\lib\site-packages\numpy\linalg\linalg.py", line 423, in inv # return wrap(solve(a, identity(a.shape[0], dtype=a.dtype))) # File "C:\Programs\Python25\lib\site-packages\numpy\linalg\linalg.py", line 306, in solve # raise LinAlgError, 'Singular matrix' #numpy.linalg.linalg.LinAlgError: Singular matrix # #>>> mod_par.fixed_params #array([ NaN, 10., NaN]) #>>> mod_par.start_params #array([ 1., 2.]) #>>> np.source(stats.pareto.fit_fr) #In file: c:\josef\eclipsegworkspace\statsmodels-josef-experimental-gsoc\scikits\statsmodels\sandbox\stats\distributions_patch.py # #def fit_fr(self, data, *args, **kwds): # '''estimate distribution parameters by MLE taking some parameters as fixed # # Parameters # ---------- # data : array, 1d # data for which the distribution parameters are estimated, # args : list ? check # starting values for optimization # kwds : # # - 'frozen' : array_like # values for frozen distribution parameters and, for elements with # np.nan, the corresponding parameter will be estimated # # Returns # ------- # argest : array # estimated parameters # # # Examples # -------- # generate random sample # >>> np.random.seed(12345) # >>> x = stats.gamma.rvs(2.5, loc=0, scale=1.2, size=200) # # estimate all parameters # >>> stats.gamma.fit(x) # array([ 2.0243194 , 0.20395655, 1.44411371]) # >>> stats.gamma.fit_fr(x, frozen=[np.nan, np.nan, np.nan]) # array([ 2.0243194 , 0.20395655, 1.44411371]) # # keep loc fixed, estimate shape and scale parameters # >>> stats.gamma.fit_fr(x, frozen=[np.nan, 0.0, np.nan]) # array([ 2.45603985, 1.27333105]) # # keep loc and scale fixed, estimate shape parameter # >>> stats.gamma.fit_fr(x, frozen=[np.nan, 0.0, 1.0]) # array([ 3.00048828]) # >>> stats.gamma.fit_fr(x, frozen=[np.nan, 0.0, 1.2]) # array([ 2.57792969]) # # estimate only scale parameter for fixed shape and loc # >>> stats.gamma.fit_fr(x, frozen=[2.5, 0.0, np.nan]) # array([ 1.25087891]) # # Notes # ----- # self is an instance of a distribution class. This can be attached to # scipy.stats.distributions.rv_continuous # # *Todo* # # * check if docstring is correct # * more input checking, args is list ? might also apply to current fit method # # ''' # loc0, scale0 = map(kwds.get, ['loc', 'scale'],[0.0, 1.0]) # Narg = len(args) # # if Narg == 0 and hasattr(self, '_fitstart'): # x0 = self._fitstart(data) # elif Narg > self.numargs: # raise ValueError, "Too many input arguments." # else: # args += (1.0,)*(self.numargs-Narg) # # location and scale are at the end # x0 = args + (loc0, scale0) # # if 'frozen' in kwds: # frmask = np.array(kwds['frozen']) # if len(frmask) != self.numargs+2: # raise ValueError, "Incorrect number of frozen arguments." # else: # # keep starting values for not frozen parameters # x0 = np.array(x0)[np.isnan(frmask)] # else: # frmask = None # # #print x0 # #print frmask # return optimize.fmin(self.nnlf_fr, x0, # args=(np.ravel(data), frmask), disp=0) # #>>> stats.pareto.fit_fr(y, 1., frozen=[np.nan, loc, np.nan]) #Traceback (most recent call last): # File "<stdin>", line 1, in <module> #NameError: name 'loc' is not defined # #>>> stats.pareto.fit_fr(y, 1., frozen=[np.nan, 10., np.nan]) #array([ 1.0346268 , 2.00184808]) #>>> stats.pareto.fit_fr(y, (1.,2), frozen=[np.nan, 10., np.nan]) #Traceback (most recent call last): # File "<stdin>", line 1, in <module> # File "c:\josef\eclipsegworkspace\statsmodels-josef-experimental-gsoc\scikits\statsmodels\sandbox\stats\distributions_patch.py", line 273, in fit_fr # x0 = np.array(x0)[np.isnan(frmask)] #ValueError: setting an array element with a sequence. # #>>> stats.pareto.fit_fr(y, [1.,2], frozen=[np.nan, 10., np.nan]) #Traceback (most recent call last): # File "<stdin>", line 1, in <module> # File "c:\josef\eclipsegworkspace\statsmodels-josef-experimental-gsoc\scikits\statsmodels\sandbox\stats\distributions_patch.py", line 273, in fit_fr # x0 = np.array(x0)[np.isnan(frmask)] #ValueError: setting an array element with a sequence. # #>>> stats.pareto.fit_fr(y, frozen=[np.nan, 10., np.nan]) #array([ 1.03463526, 2.00184809]) #>>> stats.pareto.pdf(y, 1.03463526, 10, 2.00184809).sum() #173.33947284555239 #>>> mod_par(1.03463526, 10, 2.00184809) #Traceback (most recent call last): # File "<stdin>", line 1, in <module> #TypeError: 'MyPareto' object is not callable # #>>> mod_par.loglike(1.03463526, 10, 2.00184809) #Traceback (most recent call last): # File "<stdin>", line 1, in <module> #TypeError: loglike() takes exactly 2 arguments (4 given) # #>>> mod_par.loglike((1.03463526, 10, 2.00184809)) #-962.21623668859741 #>>> np.log(stats.pareto.pdf(y, 1.03463526, 10, 2.00184809)).sum() #-inf #>>> np.log(stats.pareto.pdf(y, 1.03463526, 9, 2.00184809)).sum() #-3074.5947476137271 #>>> np.log(stats.pareto.pdf(y, 1.03463526, 10., 2.00184809)).sum() #-inf #>>> np.log(stats.pareto.pdf(y, 1.03463526, 9.9, 2.00184809)).sum() #-2677.3867091635661 #>>> y.min() #12.001848089426717 #>>> np.log(stats.pareto.pdf(y, 1.03463526, loc=9.9, scale=2.00184809)).sum() #-2677.3867091635661 #>>> np.log(stats.pareto.pdf(y, 1.03463526, loc=10., scale=2.00184809)).sum() #-inf #>>> stats.pareto.logpdf(y, 1.03463526, loc=10., scale=2.00184809).sum() #-inf #>>> stats.pareto.logpdf(y, 1.03463526, loc=9.99, scale=2.00184809).sum() #-2631.6120098202355 #>>> mod_par.loglike((1.03463526, 9.99, 2.00184809)) #-963.2513896113644 #>>> maxabs(y, mod_par.endog) #0.0 #>>> np.source(stats.pareto.logpdf) #In file: C:\Josef\_progs\Subversion\scipy-trunk_after\trunk\dist\scipy-0.9.0.dev6579.win32\Programs\Python25\Lib\site-packages\scipy\stats\distributions.py # # def logpdf(self, x, *args, **kwds): # """ # Log of the probability density function at x of the given RV. # # This uses more numerically accurate calculation if available. # # Parameters # ---------- # x : array-like # quantiles # arg1, arg2, arg3,... : array-like # The shape parameter(s) for the distribution (see docstring of the # instance object for more information) # loc : array-like, optional # location parameter (default=0) # scale : array-like, optional # scale parameter (default=1) # # Returns # ------- # logpdf : array-like # Log of the probability density function evaluated at x # # """ # loc,scale=map(kwds.get,['loc','scale']) # args, loc, scale = self._fix_loc_scale(args, loc, scale) # x,loc,scale = map(arr,(x,loc,scale)) # args = tuple(map(arr,args)) # x = arr((x-loc)*1.0/scale) # cond0 = self._argcheck(*args) & (scale > 0) # cond1 = (scale > 0) & (x >= self.a) & (x <= self.b) # cond = cond0 & cond1 # output = empty(shape(cond),'d') # output.fill(NINF) # putmask(output,(1-cond0)*array(cond1,bool),self.badvalue) # goodargs = argsreduce(cond, *((x,)+args+(scale,))) # scale, goodargs = goodargs[-1], goodargs[:-1] # place(output,cond,self._logpdf(*goodargs) - log(scale)) # if output.ndim == 0: # return output[()] # return output # #>>> np.source(stats.pareto._logpdf) #In file: C:\Josef\_progs\Subversion\scipy-trunk_after\trunk\dist\scipy-0.9.0.dev6579.win32\Programs\Python25\Lib\site-packages\scipy\stats\distributions.py # # def _logpdf(self, x, *args): # return log(self._pdf(x, *args)) # #>>> np.source(stats.pareto._pdf) #In file: C:\Josef\_progs\Subversion\scipy-trunk_after\trunk\dist\scipy-0.9.0.dev6579.win32\Programs\Python25\Lib\site-packages\scipy\stats\distributions.py # # def _pdf(self, x, b): # return b * x**(-b-1) # #>>> stats.pareto.a #1.0 #>>> (1-loc)/scale #Traceback (most recent call last): # File "<stdin>", line 1, in <module> #NameError: name 'loc' is not defined # #>>> b, loc, scale = (1.03463526, 9.99, 2.00184809) #>>> (1-loc)/scale #-4.4908502522786327 #>>> (x-loc)/scale == 1 #Traceback (most recent call last): # File "<stdin>", line 1, in <module> #NameError: name 'x' is not defined # #>>> (lb-loc)/scale == 1 #Traceback (most recent call last): # File "<stdin>", line 1, in <module> #NameError: name 'lb' is not defined # #>>> lb = scale + loc #>>> lb #11.991848090000001 #>>> (lb-loc)/scale == 1 #False #>>> (lb-loc)/scale #1.0000000000000004 #>>> #''' ''' repr(start_params) array([ 1., 10., 2.]) Optimization terminated successfully. Current function value: 2626.436870 Iterations: 102 Function evaluations: 210 Optimization terminated successfully. Current function value: 0.016555 Iterations: 16 Function evaluations: 35 [ 1.03482659 10.00737039 1.9944777 ] (1.0596088578825995, 9.9043376069230007, 2.0975104813987118) >>> 9.9043376069230007 + 2.0975104813987118 12.001848088321712 >>> y.min() 12.001848089426717 ''' ''' C:\Programs\Python25\lib\site-packages\matplotlib-0.99.1-py2.5-win32.egg\matplotlib\rcsetup.py:117: UserWarning: rcParams key "numerix" is obsolete and has no effect; please delete it from your matplotlibrc file warnings.warn('rcParams key "numerix" is obsolete and has no effect;\n' 0.0686702747648 0.0164150896481 0.128121386381 [ 0.10370428 0.09921315 0.09676723 0.10457413 0.10201618 0.89964496] (array(0.0), array(1.4552599885729829), array(0.0), array(2.5072143354058221)) (array(0.0), array(1.6666666666666667), array(0.0), array(6.0)) repr(start_params) array([ 0.10370428, 0.09921315, 0.09676723, 0.10457413, 0.10201618, 0.89964496, 6.39309417, 0.12812139]) Optimization terminated successfully. Current function value: -679.951339 Iterations: 398 Function evaluations: 609 estimation results t-dist [ 0.10400826 0.10111893 0.09725133 0.10507788 0.10086163 0.8996041 4.72131318 0.09825355] [ 0.00365493 0.00356149 0.00349329 0.00362333 0.003732 0.00362716 0.72329352 0.00388832] repr(start_params) array([ 0.10400826, 0.10111893, 0.09725133, 0.10507788, 0.10086163, 0.8996041 , 4.72131318, 0.09825355]) Optimization terminated successfully. Current function value: -679.950443 Iterations 3 using Newton [ 0.10395383 0.10106762 0.09720665 0.10503384 0.10080599 0.89954546 4.70918964 0.09815885] [ 0.00365299 0.00355968 0.00349147 0.00362166 0.00373015 0.00362533 0.7201488 0.00388437] () [ 0.09992709 0.09786601 0.09387356 0.10229919 0.09756623 0.85466272 4.60459182 0.09661986] [ 0.11308292 0.10828401 0.1028508 0.11268895 0.10934726 0.94462721 7.15412655 0.13452746] repr(start_params) array([ 1., 9., 2.]) Optimization terminated successfully. Current function value: 2636.129089 Iterations: 147 Function evaluations: 279 Optimization terminated successfully. Current function value: 0.016555 Iterations: 16 Function evaluations: 35 [ 0.84856418 10.2197801 1.78206799] (1.0596088578825995, 9.9043376069230007, 2.0975104813987118) 12.0018480891 12.0018480883 12.0018480894 repr(start_params) array([ 1., 2.]) Warning: Desired error not necessarily achieveddue to precision loss Current function value: 2643.549907 Iterations: 2 Function evaluations: 13 Gradient evaluations: 12 >>> res_parks2 = mod_par.fit_ks() repr(start_params) array([ 1., 2.]) Optimization terminated successfully. Current function value: 2642.465273 Iterations: 92 Function evaluations: 172 repr(start_params) array([ 1., 2.]) Optimization terminated successfully. Current function value: 2636.639863 Iterations: 73 Function evaluations: 136 repr(start_params) array([ 1., 2.]) Optimization terminated successfully. Current function value: 2631.568778 Iterations: 75 Function evaluations: 133 repr(start_params) array([ 1., 2.]) Optimization terminated successfully. Current function value: 2627.821044 Iterations: 75 Function evaluations: 135 repr(start_params) array([ 1., 2.]) Warning: Maximum number of function evaluations has been exceeded. repr(start_params) array([ 1., 2.]) Optimization terminated successfully. Current function value: 2631.568778 Iterations: 75 Function evaluations: 133 repr(start_params) array([ 1., 2.]) Optimization terminated successfully. Current function value: 2626.431596 Iterations: 58 Function evaluations: 109 repr(start_params) array([ 1., 2.]) Warning: Maximum number of function evaluations has been exceeded. repr(start_params) array([ 1., 2.]) Warning: Maximum number of function evaluations has been exceeded. repr(start_params) array([ 1., 2.]) Optimization terminated successfully. Current function value: 2626.737426 Iterations: 60 Function evaluations: 109 repr(start_params) array([ 1., 2.]) Optimization terminated successfully. Current function value: 2627.821044 Iterations: 75 Function evaluations: 135 repr(start_params) array([ 1., 2.]) Optimization terminated successfully. Current function value: 2626.471666 Iterations: 48 Function evaluations: 94 repr(start_params) array([ 1., 2.]) Optimization terminated successfully. Current function value: 2627.196314 Iterations: 66 Function evaluations: 119 repr(start_params) array([ 1., 2.]) Optimization terminated successfully. Current function value: 2626.578538 Iterations: 56 Function evaluations: 103 repr(start_params) array([ 1., 2.]) Optimization terminated successfully. Current function value: 2626.471666 Iterations: 48 Function evaluations: 94 repr(start_params) array([ 1., 2.]) Optimization terminated successfully. Current function value: 2626.651702 Iterations: 67 Function evaluations: 122 repr(start_params) array([ 1., 2.]) Optimization terminated successfully. Current function value: 2626.737426 Iterations: 60 Function evaluations: 109 repr(start_params) array([ 1., 2.]) Optimization terminated successfully. Current function value: 2626.613505 Iterations: 73 Function evaluations: 141 repr(start_params) array([ 1., 2.]) Optimization terminated successfully. Current function value: 2626.578538 Iterations: 56 Function evaluations: 103 repr(start_params) array([ 1., 2.]) Optimization terminated successfully. Current function value: 2626.632218 Iterations: 64 Function evaluations: 119 repr(start_params) array([ 1., 2.]) Optimization terminated successfully. Current function value: 2626.651702 Iterations: 67 Function evaluations: 122 repr(start_params) array([ 1., 2.]) Optimization terminated successfully. Current function value: 2626.622789 Iterations: 63 Function evaluations: 114 repr(start_params) array([ 1., 2.]) Optimization terminated successfully. Current function value: 2626.613505 Iterations: 73 Function evaluations: 141 repr(start_params) array([ 1., 2.]) Optimization terminated successfully. Current function value: 2626.627465 Iterations: 59 Function evaluations: 109 repr(start_params) array([ 1., 2.]) Optimization terminated successfully. Current function value: 2626.632218 Iterations: 64 Function evaluations: 119 repr(start_params) array([ 1., 2.]) Optimization terminated successfully. Current function value: 2626.625104 Iterations: 59 Function evaluations: 108 repr(start_params) array([ 1., 2.]) Optimization terminated successfully. Current function value: 2626.629829 Iterations: 66 Function evaluations: 118 repr(start_params) array([ 1., 2.]) Optimization terminated successfully. Current function value: 2626.632218 Iterations: 64 Function evaluations: 119 repr(start_params) array([ 1., 2.]) Optimization terminated successfully. Current function value: 2626.632218 Iterations: 64 Function evaluations: 119 repr(start_params) array([ 1., 2.]) Optimization terminated successfully. Current function value: 2626.628642 Iterations: 67 Function evaluations: 122 repr(start_params) array([ 1., 2.]) Optimization terminated successfully. Current function value: 2626.631023 Iterations: 68 Function evaluations: 129 repr(start_params) array([ 1., 2.]) Optimization terminated successfully. Current function value: 2626.630430 Iterations: 57 Function evaluations: 108 repr(start_params) array([ 1., 2.]) Optimization terminated successfully. Current function value: 2626.629598 Iterations: 60 Function evaluations: 112 repr(start_params) array([ 1., 2.]) Optimization terminated successfully. Current function value: 2626.630430 Iterations: 57 Function evaluations: 108 repr(start_params) array([ 1., 2.]) Optimization terminated successfully. Current function value: 2626.630130 Iterations: 65 Function evaluations: 122 repr(start_params) array([ 1., 2.]) Optimization terminated successfully. Current function value: 2626.629536 Iterations: 62 Function evaluations: 111 repr(start_params) array([ 1., 2.]) Optimization terminated successfully. Current function value: 2626.630130 Iterations: 65 Function evaluations: 122 repr(start_params) array([ 1., 2.]) Optimization terminated successfully. Current function value: 2626.629984 Iterations: 67 Function evaluations: 123 Optimization terminated successfully. Current function value: 0.016560 Iterations: 18 Function evaluations: 38 >>> res_parks2 (1.0592352626264809, 9.9051580457572399, 2.0966900385041591) >>> res_parks (1.0596088578825995, 9.9043376069230007, 2.0975104813987118) >>> res_par.params array([ 0.84856418, 10.2197801 , 1.78206799]) >>> np.sqrt(np.diag(mod_par.hessian(res_par.params))) array([ NaN, NaN, NaN]) >>> mod_par.hessian(res_par.params ... ) array([[ NaN, NaN, NaN], [ NaN, NaN, NaN], [ NaN, NaN, NaN]]) >>> mod_par.hessian(res_parks) Traceback (most recent call last): File "<stdin>", line 1, in <module> File "c:\josef\eclipsegworkspace\statsmodels-josef-experimental-gsoc\scikits\statsmodels\model.py", line 533, in hessian return approx_hess(params, self.loglike)[0] #need options for hess (epsilon) File "c:\josef\eclipsegworkspace\statsmodels-josef-experimental-gsoc\scikits\statsmodels\sandbox\regression\numdiff.py", line 118, in approx_hess xh = x + h TypeError: can only concatenate tuple (not "float") to tuple >>> mod_par.hessian(np.array(res_parks)) array([[ NaN, NaN, NaN], [ NaN, NaN, NaN], [ NaN, NaN, NaN]]) >>> mod_par.fixed_params array([ NaN, 9.90510677, NaN]) >>> mod_par.fixed_params=None >>> mod_par.hessian(np.array(res_parks)) array([[-890.48553491, NaN, NaN], [ NaN, NaN, NaN], [ NaN, NaN, NaN]]) >>> mod_par.loglike(np.array(res_parks)) -2626.6322080820569 >>> mod_par.bsejac Traceback (most recent call last): File "<stdin>", line 1, in <module> File "c:\josef\eclipsegworkspace\statsmodels-josef-experimental-gsoc\scikits\statsmodels\decorators.py", line 85, in __get__ _cachedval = self.fget(obj) File "c:\josef\eclipsegworkspace\statsmodels-josef-experimental-gsoc\scikits\statsmodels\model.py", line 592, in bsejac return np.sqrt(np.diag(self.covjac)) File "c:\josef\eclipsegworkspace\statsmodels-josef-experimental-gsoc\scikits\statsmodels\decorators.py", line 85, in __get__ _cachedval = self.fget(obj) File "c:\josef\eclipsegworkspace\statsmodels-josef-experimental-gsoc\scikits\statsmodels\model.py", line 574, in covjac jacv = self.jacv File "c:\josef\eclipsegworkspace\statsmodels-josef-experimental-gsoc\scikits\statsmodels\decorators.py", line 85, in __get__ _cachedval = self.fget(obj) File "c:\josef\eclipsegworkspace\statsmodels-josef-experimental-gsoc\scikits\statsmodels\model.py", line 557, in jacv return self.jac(self._results.params) File "c:\josef\eclipsegworkspace\statsmodels-josef-experimental-gsoc\scikits\statsmodels\model.py", line 530, in jac return approx_fprime1(params, self.loglikeobs, **kwds) File "c:\josef\eclipsegworkspace\statsmodels-josef-experimental-gsoc\scikits\statsmodels\sandbox\regression\numdiff.py", line 80, in approx_fprime1 f0 = f(*((xk,)+args)) File "c:\josef\eclipsegworkspace\statsmodels-josef-experimental-gsoc\scikits\statsmodels\model.py", line 522, in loglikeobs return -self.nloglikeobs(params) File "C:\Josef\eclipsegworkspace\statsmodels-josef-experimental-gsoc\scikits\statsmodels\examples\ex_generic_mle_tdist.py", line 184, in nloglikeobs scale = params[2] IndexError: index out of bounds >>> hasattr(self, 'start_params') Traceback (most recent call last): File "<stdin>", line 1, in <module> NameError: name 'self' is not defined >>> hasattr(mod_par, 'start_params') True >>> mod_par.start_params array([ 1., 2.]) >>> stats.pareto.stats(1., 9., 2., moments='mvsk') (array(1.#INF), array(1.#INF), array(1.#QNAN), array(1.#QNAN)) >>> stats.pareto.stats(1., 8., 2., moments='mvsk') (array(1.#INF), array(1.#INF), array(1.#QNAN), array(1.#QNAN)) >>> stats.pareto.stats(1., 8., 1., moments='mvsk') (array(1.#INF), array(1.#INF), array(1.#QNAN), array(1.#QNAN)) >>> stats.pareto.stats(1., moments='mvsk') (array(1.#INF), array(1.#INF), array(1.#QNAN), array(1.#QNAN)) >>> stats.pareto.stats(0.5., moments='mvsk') File "<stdin>", line 1 stats.pareto.stats(0.5., moments='mvsk') ^ SyntaxError: invalid syntax >>> stats.pareto.stats(0.5, moments='mvsk') (array(1.#INF), array(1.#INF), array(1.#QNAN), array(1.#QNAN)) >>> stats.pareto.stats(2, moments='mvsk') (array(2.0), array(1.#INF), array(1.#QNAN), array(1.#QNAN)) >>> stats.pareto.stats(10, moments='mvsk') (array(1.1111111111111112), array(0.015432098765432098), array(2.8110568859997356), array(14.828571428571429)) >>> stats.pareto.rvs(10, size=10) array([ 1.07716265, 1.18977526, 1.07093 , 1.05157081, 1.15991232, 1.31015589, 1.06675107, 1.08082475, 1.19501243, 1.34967158]) >>> r = stats.pareto.rvs(10, size=1000) >>> plt Traceback (most recent call last): File "<stdin>", line 1, in <module> NameError: name 'plt' is not defined >>> import matplotlib.pyplot as plt >>> plt.hist(r) (array([962, 32, 3, 2, 0, 0, 0, 0, 0, 1]), array([ 1.00013046, 1.3968991 , 1.79366773, 2.19043637, 2.587205 , 2.98397364, 3.38074227, 3.77751091, 4.17427955, 4.57104818, 4.96781682]), <a list of 10 Patch objects>) >>> plt.show() '''
bsd-3-clause
d38cb1c1bbfa390d53cd5e06b9508159
35.26075
167
0.658542
2.877169
false
false
false
false
yarikoptic/pystatsmodels
statsmodels/tools/parallel.py
3
1838
'''Parallel utility function using joblib copied from https://github.com/mne-tools/mne-python Author: Alexandre Gramfort <gramfort@nmr.mgh.harvard.edu> License: Simplified BSD changes for statsmodels (Josef Perktold) - try import from joblib directly, (doesn't import all of sklearn) ''' def parallel_func(func, n_jobs, verbose=5): """Return parallel instance with delayed function Util function to use joblib only if available Parameters ---------- func: callable A function n_jobs: int Number of jobs to run in parallel verbose: int Verbosity level Returns ------- parallel: instance of joblib.Parallel or list The parallel object my_func: callable func if not parallel or delayed(func) n_jobs: int Number of jobs >= 0 Examples -------- >>> from math import sqrt >>> from statsmodels.tools.parallel import parallel_func >>> parallel, p_func, n_jobs = parallel_func(sqrt, n_jobs=-1, verbose=0) >>> print n_jobs >>> parallel(p_func(i**2) for i in range(10)) """ try: try: from joblib import Parallel, delayed except ImportError: from sklearn.externals.joblib import Parallel, delayed parallel = Parallel(n_jobs, verbose=verbose) my_func = delayed(func) if n_jobs == -1: try: import multiprocessing n_jobs = multiprocessing.cpu_count() except (ImportError, NotImplementedError): print "multiprocessing not installed. Cannot run in parallel." n_jobs = 1 except ImportError: print "joblib not installed. Cannot run in parallel." n_jobs = 1 my_func = func parallel = list return parallel, my_func, n_jobs
bsd-3-clause
e242c7789a3a528df7103b06f08c2f5f
26.432836
78
0.615343
4.284382
false
false
false
false
yarikoptic/pystatsmodels
statsmodels/sandbox/regression/tools.py
6
13069
'''gradient/Jacobian of normal and t loglikelihood use chain rule normal derivative wrt mu, sigma and beta new version: loc-scale distributions, derivative wrt loc, scale also includes "standardized" t distribution (for use in GARCH) TODO: * use sympy for derivative of loglike wrt shape parameters it works for df of t distribution dlog(gamma(a))da = polygamma(0,a) check polygamma is available in scipy.special * get loc-scale example to work with mean = X*b * write some full unit test examples A: josef-pktd ''' import numpy as np from scipy import special from scipy.special import gammaln def norm_lls(y, params): '''normal loglikelihood given observations and mean mu and variance sigma2 Parameters ---------- y : array, 1d normally distributed random variable params: array, (nobs, 2) array of mean, variance (mu, sigma2) with observations in rows Returns ------- lls : array contribution to loglikelihood for each observation ''' mu, sigma2 = params.T lls = -0.5*(np.log(2*np.pi) + np.log(sigma2) + (y-mu)**2/sigma2) return lls def norm_lls_grad(y, params): '''Jacobian of normal loglikelihood wrt mean mu and variance sigma2 Parameters ---------- y : array, 1d normally distributed random variable params: array, (nobs, 2) array of mean, variance (mu, sigma2) with observations in rows Returns ------- grad : array (nobs, 2) derivative of loglikelihood for each observation wrt mean in first column, and wrt variance in second column Notes ----- this is actually the derivative wrt sigma not sigma**2, but evaluated with parameter sigma2 = sigma**2 ''' mu, sigma2 = params.T dllsdmu = (y-mu)/sigma2 dllsdsigma2 = ((y-mu)**2/sigma2 - 1)/np.sqrt(sigma2) return np.column_stack((dllsdmu, dllsdsigma2)) def mean_grad(x, beta): '''gradient/Jacobian for d (x*beta)/ d beta ''' return x def normgrad(y, x, params): '''Jacobian of normal loglikelihood wrt mean mu and variance sigma2 Parameters ---------- y : array, 1d normally distributed random variable with mean x*beta, and variance sigma2 x : array, 2d explanatory variables, observation in rows, variables in columns params: array_like, (nvars + 1) array of coefficients and variance (beta, sigma2) Returns ------- grad : array (nobs, 2) derivative of loglikelihood for each observation wrt mean in first column, and wrt scale (sigma) in second column assume params = (beta, sigma2) Notes ----- TODO: for heteroscedasticity need sigma to be a 1d array ''' beta = params[:-1] sigma2 = params[-1]*np.ones((len(y),1)) dmudbeta = mean_grad(x, beta) mu = np.dot(x, beta) #print beta, sigma2 params2 = np.column_stack((mu,sigma2)) dllsdms = norm_lls_grad(y,params2) grad = np.column_stack((dllsdms[:,:1]*dmudbeta, dllsdms[:,:1])) return grad def tstd_lls(y, params, df): '''t loglikelihood given observations and mean mu and variance sigma2 = 1 Parameters ---------- y : array, 1d normally distributed random variable params: array, (nobs, 2) array of mean, variance (mu, sigma2) with observations in rows df : integer degrees of freedom of the t distribution Returns ------- lls : array contribution to loglikelihood for each observation Notes ----- parameterized for garch ''' mu, sigma2 = params.T df = df*1.0 #lls = gammaln((df+1)/2.) - gammaln(df/2.) - 0.5*np.log((df-2)*np.pi) #lls -= (df+1)/2. * np.log(1. + (y-mu)**2/(df-2.)/sigma2) + 0.5 * np.log(sigma2) lls = gammaln((df+1)/2.) - gammaln(df/2.) - 0.5*np.log((df-2)*np.pi) lls -= (df+1)/2. * np.log(1. + (y-mu)**2/(df-2)/sigma2) + 0.5 * np.log(sigma2) return lls def norm_dlldy(y): '''derivative of log pdf of standard normal with respect to y ''' return -y def ts_dlldy(y, df): '''derivative of log pdf of standardized (?) t with respect to y Notes ----- parameterized for garch, with mean 0 and variance 1 ''' #(df+1)/2. / (1 + y**2/(df-2.)) * 2.*y/(df-2.) #return -(df+1)/(df-2.) / (1 + y**2/(df-2.)) * y return -(df+1)/(df) / (1 + y**2/(df)) * y def tstd_pdf(x, df): '''pdf for standardized (not standard) t distribution, variance is one ''' r = np.array(df*1.0) Px = np.exp(special.gammaln((r+1)/2.)-special.gammaln(r/2.))/np.sqrt((r-2)*pi) Px /= (1+(x**2)/(r-2))**((r+1)/2.) return Px def ts_lls(y, params, df): '''t loglikelihood given observations and mean mu and variance sigma2 = 1 Parameters ---------- y : array, 1d normally distributed random variable params: array, (nobs, 2) array of mean, variance (mu, sigma2) with observations in rows df : integer degrees of freedom of the t distribution Returns ------- lls : array contribution to loglikelihood for each observation Notes ----- parameterized for garch normalized/rescaled so that sigma2 is the variance >>> df = 10; sigma = 1. >>> stats.t.stats(df, loc=0., scale=sigma.*np.sqrt((df-2.)/df)) (array(0.0), array(1.0)) >>> sigma = np.sqrt(2.) >>> stats.t.stats(df, loc=0., scale=sigma*np.sqrt((df-2.)/df)) (array(0.0), array(2.0)) ''' print y, params, df mu, sigma2 = params.T df = df*1.0 #lls = gammaln((df+1)/2.) - gammaln(df/2.) - 0.5*np.log((df-2)*np.pi) #lls -= (df+1)/2. * np.log(1. + (y-mu)**2/(df-2.)/sigma2) + 0.5 * np.log(sigma2) lls = gammaln((df+1)/2.) - gammaln(df/2.) - 0.5*np.log((df)*np.pi) lls -= (df+1.)/2. * np.log(1. + (y-mu)**2/(df)/sigma2) + 0.5 * np.log(sigma2) return lls def ts_dlldy(y, df): '''derivative of log pdf of standard t with respect to y Parameters ---------- y : array_like data points of random variable at which loglike is evaluated df : array_like degrees of freedom,shape parameters of log-likelihood function of t distribution Returns ------- dlldy : array derivative of loglikelihood wrt random variable y evaluated at the points given in y Notes ----- with mean 0 and scale 1, but variance is df/(df-2) ''' df = df*1. #(df+1)/2. / (1 + y**2/(df-2.)) * 2.*y/(df-2.) #return -(df+1)/(df-2.) / (1 + y**2/(df-2.)) * y return -(df+1)/(df) / (1 + y**2/(df)) * y def tstd_dlldy(y, df): '''derivative of log pdf of standardized t with respect to y Parameters ---------- y : array_like data points of random variable at which loglike is evaluated df : array_like degrees of freedom,shape parameters of log-likelihood function of t distribution Returns ------- dlldy : array derivative of loglikelihood wrt random variable y evaluated at the points given in y Notes ----- parameterized for garch, standardized to variance=1 ''' #(df+1)/2. / (1 + y**2/(df-2.)) * 2.*y/(df-2.) return -(df+1)/(df-2.) / (1 + y**2/(df-2.)) * y #return (df+1)/(df) / (1 + y**2/(df)) * y def locscale_grad(y, loc, scale, dlldy, *args): '''derivative of log-likelihood with respect to location and scale Parameters ---------- y : array_like data points of random variable at which loglike is evaluated loc : float location parameter of distribution scale : float scale parameter of distribution dlldy : function derivative of loglikelihood fuction wrt. random variable x args : array_like shape parameters of log-likelihood function Returns ------- dlldloc : array derivative of loglikelihood wrt location evaluated at the points given in y dlldscale : array derivative of loglikelihood wrt scale evaluated at the points given in y ''' yst = (y-loc)/scale #ystandardized dlldloc = -dlldy(yst, *args) / scale dlldscale = -1./scale - dlldy(yst, *args) * (y-loc)/scale**2 return dlldloc, dlldscale if __name__ == '__main__': verbose = 0 if verbose: sig = 0.1 beta = np.ones(2) rvs = np.random.randn(10,3) x = rvs[:,1:] y = np.dot(x,beta) + sig*rvs[:,0] params = [1,1,1] print normgrad(y, x, params) dllfdbeta = (y-np.dot(x, beta))[:,None]*x #for sigma = 1 print dllfdbeta print locscale_grad(y, np.dot(x, beta), 1, norm_dlldy) print (y-np.dot(x, beta)) from scipy import stats, misc def llt(y,loc,scale,df): return np.log(stats.t.pdf(y, df, loc=loc, scale=scale)) def lltloc(loc,y,scale,df): return np.log(stats.t.pdf(y, df, loc=loc, scale=scale)) def lltscale(scale,y,loc,df): return np.log(stats.t.pdf(y, df, loc=loc, scale=scale)) def llnorm(y,loc,scale): return np.log(stats.norm.pdf(y, loc=loc, scale=scale)) def llnormloc(loc,y,scale): return np.log(stats.norm.pdf(y, loc=loc, scale=scale)) def llnormscale(scale,y,loc): return np.log(stats.norm.pdf(y, loc=loc, scale=scale)) if verbose: print '\ngradient of t' print misc.derivative(llt, 1, dx=1e-6, n=1, args=(0,1,10), order=3) print 't ', locscale_grad(1, 0, 1, tstd_dlldy, 10) print 'ts', locscale_grad(1, 0, 1, ts_dlldy, 10) print misc.derivative(llt, 1.5, dx=1e-10, n=1, args=(0,1,20), order=3), print 'ts', locscale_grad(1.5, 0, 1, ts_dlldy, 20) print misc.derivative(llt, 1.5, dx=1e-10, n=1, args=(0,2,20), order=3), print 'ts', locscale_grad(1.5, 0, 2, ts_dlldy, 20) print misc.derivative(llt, 1.5, dx=1e-10, n=1, args=(1,2,20), order=3), print 'ts', locscale_grad(1.5, 1, 2, ts_dlldy, 20) print misc.derivative(lltloc, 1, dx=1e-10, n=1, args=(1.5,2,20), order=3), print misc.derivative(lltscale, 2, dx=1e-10, n=1, args=(1.5,1,20), order=3) y,loc,scale,df = 1.5, 1, 2, 20 print 'ts', locscale_grad(y,loc,scale, ts_dlldy, 20) print misc.derivative(lltloc, loc, dx=1e-10, n=1, args=(y,scale,df), order=3), print misc.derivative(lltscale, scale, dx=1e-10, n=1, args=(y,loc,df), order=3) print '\ngradient of norm' print misc.derivative(llnorm, 1, dx=1e-6, n=1, args=(0,1), order=3) print locscale_grad(1, 0, 1, norm_dlldy) y,loc,scale = 1.5, 1, 2 print 'ts', locscale_grad(y,loc,scale, norm_dlldy) print misc.derivative(llnormloc, loc, dx=1e-10, n=1, args=(y,scale), order=3), print misc.derivative(llnormscale, scale, dx=1e-10, n=1, args=(y,loc), order=3) y,loc,scale = 1.5, 0, 1 print 'ts', locscale_grad(y,loc,scale, norm_dlldy) print misc.derivative(llnormloc, loc, dx=1e-10, n=1, args=(y,scale), order=3), print misc.derivative(llnormscale, scale, dx=1e-10, n=1, args=(y,loc), order=3) #print 'still something wrong with handling of scale and variance' #looks ok now print '\nloglike of t' print tstd_lls(1, np.array([0,1]), 100), llt(1,0,1,100), 'differently standardized' print tstd_lls(1, np.array([0,1]), 10), llt(1,0,1,10), 'differently standardized' print ts_lls(1, np.array([0,1]), 10), llt(1,0,1,10) print tstd_lls(1, np.array([0,1.*10./8.]), 10), llt(1.,0,1.,10) print ts_lls(1, np.array([0,1]), 100), llt(1,0,1,100) print tstd_lls(1, np.array([0,1]), 10), llt(1,0,1.*np.sqrt(8/10.),10) from numpy.testing import assert_almost_equal params =[(0, 1), (1.,1.), (0.,2.), ( 1., 2.)] yt = np.linspace(-2.,2.,11) for loc,scale in params: dlldlo = misc.derivative(llnormloc, loc, dx=1e-10, n=1, args=(yt,scale), order=3) dlldsc = misc.derivative(llnormscale, scale, dx=1e-10, n=1, args=(yt,loc), order=3) gr = locscale_grad(yt, loc, scale, norm_dlldy) assert_almost_equal(dlldlo, gr[0], 5, err_msg='deriv loc') assert_almost_equal(dlldsc, gr[1], 5, err_msg='deriv scale') for df in [3, 10, 100]: for loc,scale in params: dlldlo = misc.derivative(lltloc, loc, dx=1e-10, n=1, args=(yt,scale,df), order=3) dlldsc = misc.derivative(lltscale, scale, dx=1e-10, n=1, args=(yt,loc,df), order=3) gr = locscale_grad(yt, loc, scale, ts_dlldy, df) assert_almost_equal(dlldlo, gr[0], 4, err_msg='deriv loc') assert_almost_equal(dlldsc, gr[1], 4, err_msg='deriv scale') assert_almost_equal(ts_lls(yt, np.array([loc, scale**2]), df), llt(yt,loc,scale,df), 5, err_msg='loglike') assert_almost_equal(tstd_lls(yt, np.array([loc, scale**2]), df), llt(yt,loc,scale*np.sqrt((df-2.)/df),df), 5, err_msg='loglike')
bsd-3-clause
ea65577f9c78d9838e6f180f097c2456
31.591022
95
0.587038
3.061373
false
false
false
false
yarikoptic/pystatsmodels
statsmodels/sandbox/tsa/examples/example_var.py
36
1218
""" Look at some macro plots, then do some VARs and IRFs. """ import numpy as np import statsmodels.api as sm import scikits.timeseries as ts import scikits.timeseries.lib.plotlib as tplt from matplotlib import pyplot as plt data = sm.datasets.macrodata.load() data = data.data ### Create Timeseries Representations of a few vars dates = ts.date_array(start_date=ts.Date('Q', year=1959, quarter=1), end_date=ts.Date('Q', year=2009, quarter=3)) ts_data = data[['realgdp','realcons','cpi']].view(float).reshape(-1,3) ts_data = np.column_stack((ts_data, (1 - data['unemp']/100) * data['pop'])) ts_series = ts.time_series(ts_data, dates) fig = tplt.tsfigure() fsp = fig.add_tsplot(221) fsp.tsplot(ts_series[:,0],'-') fsp.set_title("Real GDP") fsp = fig.add_tsplot(222) fsp.tsplot(ts_series[:,1],'r-') fsp.set_title("Real Consumption") fsp = fig.add_tsplot(223) fsp.tsplot(ts_series[:,2],'g-') fsp.set_title("CPI") fsp = fig.add_tsplot(224) fsp.tsplot(ts_series[:,3],'y-') fsp.set_title("Employment") # Plot real GDP #plt.subplot(221) #plt.plot(data['realgdp']) #plt.title("Real GDP") # Plot employment #plt.subplot(222) # Plot cpi #plt.subplot(223) # Plot real consumption #plt.subplot(224) #plt.show()
bsd-3-clause
f25410803ba2cdbd647124a012046f54
21.145455
75
0.690476
2.490798
false
false
true
false
yarikoptic/pystatsmodels
statsmodels/tsa/tests/results/arima111_css_results.py
35
44167
import numpy as np llf = np.array([-242.06033399744]) nobs = np.array([ 202]) k = np.array([ 4]) k_exog = np.array([ 1]) sigma = np.array([ .80201496146073]) chi2 = np.array([ 348.43324197088]) df_model = np.array([ 2]) k_ar = np.array([ 1]) k_ma = np.array([ 1]) params = np.array([ .82960638524364, .93479332833705, -.75728342544279, .64322799840686]) cov_params = np.array([ .14317811930738, -.01646077810033, .01510986837498, -.00280799533479, -.01646077810033, .00321032468661, -.00353027620719, .00097645385252, .01510986837498, -.00353027620719, .00484312817753, -.00112050648944, -.00280799533479, .00097645385252, -.00112050648944, .0007715609499]).reshape(4,4) xb = np.array([ .82960641384125, .82960641384125, .697261095047, .61113905906677, .51607495546341, .47362637519836, .41342103481293, .40238001942635, .37454023957253, .33222004771233, .32514902949333, .31093680858612, .30019253492355, .31159669160843, .29182952642441, .30349296331406, .29457464814186, .28427124023438, .30664679408073, .29696446657181, .31270903348923, .29268020391464, .28816330432892, .29006817936897, .30216124653816, .30066826939583, .31728908419609, .30679926276207, .3272570669651, .37292611598969, .36668366193771, .40278288722038, .36799272894859, .36827209591866, .38623574376106, .39983862638474, .42789059877396, .43138384819031, .46953064203262, .48066720366478, .48910140991211, .53098994493484, .54496067762375, .55554050207138, .58130383491516, .60081332921982, .58008605241776, .58214038610458, .58369606733322, .53162068128586, .54543834924698, .52040082216263, .50143963098526, .48708060383797, .47620677947998, .48572361469269, .51068127155304, .61833620071411, .61110657453537, .76539021730423, .84672522544861, .92606955766678, .96840506792068, 1.0892199277878, 1.1097067594528, 1.0187155008316, 1.0030621290207, .97345739603043, .95103752613068, .82755368947983, .84054774045944, .85038793087006, .84008830785751, .92104357481003, .89359468221664, .87280809879303, .91032028198242, .95647835731506, 1.0624366998672, 1.1426770687103, 1.1679404973984, 1.311328291893, 1.473167181015, 1.5602221488953, 1.7326545715332, 1.8809853792191, 1.7803012132645, 1.7750589847565, 1.8420933485031, 1.7863517999649, 1.8328944444656, 1.7793855667114, 1.5791050195694, 1.3564316034317, 1.5250737667084, 1.3155146837234, 1.014811873436, .98235523700714, .97552710771561, .97035628557205, 1.0196926593781, 1.0393049716949, .98315137624741, .97613000869751, .89980864524841, .96626943349838, .91009211540222, .88530200719833, .97303456068039, .57794612646103, .63377332687378, .65829831361771, .76562696695328, .86465454101563, .90414637327194, .95180231332779, .95238989591599, .98833626508713, 1.0333099365234, 1.0851185321808, 1.1066001653671, 1.2293750047684, 1.233595252037, 1.1480363607407, 1.2962552309036, 1.2842413187027, 1.3106474876404, 1.5614050626755, 1.4672855138779, 1.2362524271011, 1.1855486631393, 1.1294020414352, 1.1046353578568, 1.0858771800995, 1.0716745853424, 1.0786685943604, 1.0662157535553, 1.0390332937241, .96519494056702, .9802839756012, .92070508003235, .91108840703964, .95705932378769, .95637094974518, .97360169887543, 1.0221517086029, .9701629281044, .94854199886322, .98542231321335, 1.048855304718, 1.0081344842911, 1.0305507183075, 1.0475262403488, .93612504005432, .85176283121109, .89438372850418, .820152759552, .71068543195724, .76979607343674, .76130604743958, .77262878417969, .85220617055893, .84146595001221, .93983960151672, .97883212566376, 1.0793634653091, 1.1909983158112, 1.1690304279327, 1.2411522865295, 1.1360056400299, 1.0918840169907, .9164656996727, .76586949825287, .918093085289, .87360894680023, .92867678403854, 1.00588285923, .92233866453171, .84132260084152, .90422683954239, .9873673915863, .99707210063934, 1.1109310388565, 1.1971517801285, 1.138188958168, 1.2710473537445, 1.1763968467712, 1.7437561750412, 1.4101150035858, 1.3527159690857, 1.4335050582886, .99765706062317, 1.1067585945129, 1.3086627721786, 1.2968333959579, 1.3547962903976, 1.6768488883972, 1.5905654430389, 2.0774590969086, 1.3218278884888, .21813294291496, .30750840902328, .60612773895264]) y = np.array([np.nan, 29.809606552124, 29.847261428833, 29.961139678955, 29.886075973511, 30.013628005981, 29.96342086792, 30.152379989624, 30.214540481567, 30.142219543457, 30.245149612427, 30.290935516357, 30.3401927948, 30.521595001221, 30.511829376221, 30.683492660522, 30.734575271606, 30.764270782471, 30.996646881104, 31.046964645386, 31.252710342407, 31.242681503296, 31.308164596558, 31.410068511963, 31.582162857056, 31.680667877197, 31.897289276123, 31.956798553467, 32.207256317139, 32.652923583984, 32.8166847229, 33.252780914307, 33.267993927002, 33.468269348145, 33.786235809326, 34.099838256836, 34.527889251709, 34.831386566162, 35.369533538818, 35.780666351318, 36.189102172852, 36.830989837646, 37.344959259033, 37.855541229248, 38.481304168701, 39.100814819336, 39.480087280273, 39.9821434021, 40.483695983887, 40.631618499756, 41.145435333252, 41.420402526855, 41.701438903809, 41.987079620361, 42.276206970215, 42.685726165771, 43.210681915283, 44.318336486816, 44.811107635498, 46.365386962891, 47.646724700928, 49.026069641113, 50.268405914307, 52.089218139648, 53.409706115723, 54.018714904785, 55.003063201904, 55.873458862305, 56.751037597656, 56.927551269531, 57.840549468994, 58.750389099121, 59.540088653564, 60.921043395996, 61.693592071533, 62.472805023193, 63.610321044922, 64.856483459473, 66.562438964844, 68.24267578125, 69.667938232422, 71.911323547363, 74.473167419434, 76.760215759277, 79.732650756836, 82.780990600586, 84.380302429199, 86.475059509277, 89.042091369629, 90.886352539063, 93.332893371582, 95.179389953613, 95.979103088379, 96.356430053711, 99.02507019043, 99.415512084961, 98.914810180664, 99.782356262207, 100.7755279541, 101.770362854, 103.11968994141, 104.33930969238, 105.083152771, 106.07612609863, 106.59980773926, 107.96627044678, 108.61009216309, 109.38529968262, 110.87303924561, 109.27794647217, 110.13377380371, 110.85829162598, 112.16562652588, 113.56465148926, 114.70414733887, 115.95180511475, 116.95239257813, 118.188331604, 119.53330993652, 120.98512268066, 122.30659484863, 124.3293762207, 125.73359680176, 126.54803466797, 128.79624938965, 130.18423461914, 131.81065368652, 134.96139526367, 136.16728210449, 136.33625793457, 137.38554382324, 138.32939147949, 139.40463256836, 140.48587036133, 141.57167053223, 142.77867126465, 143.86622619629, 144.83903503418, 145.46519470215, 146.58029174805, 147.220703125, 148.11108398438, 149.35705566406, 150.35636901855, 151.47360229492, 152.82215881348, 153.5701751709, 154.44854736328, 155.68542480469, 157.14886474609, 158.00813293457, 159.23054504395, 160.44752502441, 160.83612060547, 161.25175476074, 162.39437866211, 162.82015991211, 162.91067504883, 163.96978759766, 164.66130065918, 165.47262573242, 166.75219726563, 167.54145812988, 169.03984069824, 170.27883911133, 171.9793548584, 173.89099121094, 175.06903076172, 176.84115600586, 177.5359954834, 178.49188232422, 178.5164642334, 178.46586608887, 180.21809387207, 180.8736114502, 182.12867736816, 183.60589599609, 184.12232971191, 184.54132080078, 185.80421447754, 187.28736877441, 188.39706420898, 190.2109375, 191.99716186523, 192.93818664551, 195.07104492188, 195.8763885498, 200.94375610352, 200.81010437012, 202.05271911621, 204.13349914551, 202.89764404297, 204.68077087402, 207.22866821289, 208.63482666016, 210.48779296875, 214.17184448242, 215.58755493164, 220.68745422363, 218.21083068848, 212.39213562012, 212.978515625, 215.07511901855]) resid = np.array([np.nan, -.6596063375473, -.49726036190987, -.5911386013031, -.34607490897179, -.46362805366516, -.21342028677464, -.31237986683846, -.40454092621803, -.22221945226192, -.26514956355095, -.2509354352951, -.13019436597824, -.30159646272659, -.1318296790123, -.24349159002304, -.25457563996315, -.07427024841309, -.24664734303951, -.10696394741535, -.30270880460739, -.22268049418926, -.18816292285919, -.13006833195686, -.20216277241707, -.10066751390696, -.24728938937187, -.07679972797632, .07274255156517, -.20292413234711, .03331403434277, -.35277983546257, -.16799576580524, -.06826904416084, -.08623649924994, .00015908146452, -.12788754701614, .06861615926027, -.06953293830156, -.08066567778587, .11089706420898, -.03098993562162, -.04496069997549, .04446176066995, .01869462057948, -.20081178843975, -.08008606731892, -.08214038610458, -.38369914889336, -.03162068501115, -.24543529748917, -.22040157020092, -.20144037902355, -.18708138167858, -.07620526105165, .01427639275789, .48931872844696, -.11833623051643, .78889113664627, .43461054563522, .45327401161194, .27393117547035, .73159569501877, .21077930927277, -.40970605611801, -.01871551014483, -.10306061804295, -.0734596773982, -.65103828907013, .0724478662014, .05945380032063, -.05038867890835, .45991089940071, -.12104434520006, -.09359546005726, .22719417512417, .28968048095703, .64352011680603, .53756183385849, .25732442736626, .93205803632736, 1.0886732339859, .72682982683182, 1.2397809028625, 1.1673469543457, -.18098846077919, .31969723105431, .72494095563889, .05790812522173, .61364978551865, .06710703670979, -.77938556671143, -.97910648584366, 1.1435683965683, -.92507529258728, -1.5155116319656, -.11481033265591, .01764474436641, .02447287365794, .32963913679123, .18031190335751, -.23930950462818, .01684862375259, -.37613153457642, .40019443631172, -.2662724852562, -.11008904129267, .51469951868057, -2.1730391979218, .22205695509911, .06622361391783, .54170626401901, .53436845541, .2353515625, .29585054516792, .04819770529866, .24760706722736, .31166675686836, .36669155955315, .21487690508366, .79340130090714, .17062658071518, -.33359375596046, .95196217298508, .10373862832785, .31576481461525, 1.589346408844, -.26140204071999, -1.0672763586044, -.13626158237457, -.18554861843586, -.02939598634839, -.00464448658749, .01412893645465, .1283223181963, .02133745700121, -.06621573865414, -.33903631567955, .13481116294861, -.28028702735901, -.02071117423475, .28890857100487, .04294065013528, .14363515377045, .32640132308006, -.22214868664742, -.0701690018177, .25145494937897, .41458681225777, -.14886146783829, .19186246395111, .16944620013237, -.54752624034882, -.43612506985664, .2482432872057, -.39438369870186, -.62015581130981, .28931456804276, -.06979911774397, .03869699314237, .4273681640625, -.05220314115286, .55854320526123, .26015737652779, .62115871906281, .72063958644867, .00899865385145, .53098171949387, -.44116449356079, -.13600566983223, -.89187180995941, -.81647485494614, .83413660526276, -.21809615194798, .32638800144196, .47133237123489, -.4058920443058, -.42233863472939, .35867437720299, .49578228592873, .11262346804142, .70294010639191, .58906590938568, -.19715182483196, .86181098222733, -.37105345726013, 3.3236031532288, -1.543759226799, -.11011194437742, .64728397130966, -2.2335081100464, .67635416984558, 1.2392344474792, .10933646559715, .49816474318504, 2.0072033405304, -.17484994232655, 3.0224411487579, -3.7984521389008, -6.0368394851685, .27887633442879, 1.4904805421829, 1.3098726272583]) yr = np.array([np.nan, -.6596063375473, -.49726036190987, -.5911386013031, -.34607490897179, -.46362805366516, -.21342028677464, -.31237986683846, -.40454092621803, -.22221945226192, -.26514956355095, -.2509354352951, -.13019436597824, -.30159646272659, -.1318296790123, -.24349159002304, -.25457563996315, -.07427024841309, -.24664734303951, -.10696394741535, -.30270880460739, -.22268049418926, -.18816292285919, -.13006833195686, -.20216277241707, -.10066751390696, -.24728938937187, -.07679972797632, .07274255156517, -.20292413234711, .03331403434277, -.35277983546257, -.16799576580524, -.06826904416084, -.08623649924994, .00015908146452, -.12788754701614, .06861615926027, -.06953293830156, -.08066567778587, .11089706420898, -.03098993562162, -.04496069997549, .04446176066995, .01869462057948, -.20081178843975, -.08008606731892, -.08214038610458, -.38369914889336, -.03162068501115, -.24543529748917, -.22040157020092, -.20144037902355, -.18708138167858, -.07620526105165, .01427639275789, .48931872844696, -.11833623051643, .78889113664627, .43461054563522, .45327401161194, .27393117547035, .73159569501877, .21077930927277, -.40970605611801, -.01871551014483, -.10306061804295, -.0734596773982, -.65103828907013, .0724478662014, .05945380032063, -.05038867890835, .45991089940071, -.12104434520006, -.09359546005726, .22719417512417, .28968048095703, .64352011680603, .53756183385849, .25732442736626, .93205803632736, 1.0886732339859, .72682982683182, 1.2397809028625, 1.1673469543457, -.18098846077919, .31969723105431, .72494095563889, .05790812522173, .61364978551865, .06710703670979, -.77938556671143, -.97910648584366, 1.1435683965683, -.92507529258728, -1.5155116319656, -.11481033265591, .01764474436641, .02447287365794, .32963913679123, .18031190335751, -.23930950462818, .01684862375259, -.37613153457642, .40019443631172, -.2662724852562, -.11008904129267, .51469951868057, -2.1730391979218, .22205695509911, .06622361391783, .54170626401901, .53436845541, .2353515625, .29585054516792, .04819770529866, .24760706722736, .31166675686836, .36669155955315, .21487690508366, .79340130090714, .17062658071518, -.33359375596046, .95196217298508, .10373862832785, .31576481461525, 1.589346408844, -.26140204071999, -1.0672763586044, -.13626158237457, -.18554861843586, -.02939598634839, -.00464448658749, .01412893645465, .1283223181963, .02133745700121, -.06621573865414, -.33903631567955, .13481116294861, -.28028702735901, -.02071117423475, .28890857100487, .04294065013528, .14363515377045, .32640132308006, -.22214868664742, -.0701690018177, .25145494937897, .41458681225777, -.14886146783829, .19186246395111, .16944620013237, -.54752624034882, -.43612506985664, .2482432872057, -.39438369870186, -.62015581130981, .28931456804276, -.06979911774397, .03869699314237, .4273681640625, -.05220314115286, .55854320526123, .26015737652779, .62115871906281, .72063958644867, .00899865385145, .53098171949387, -.44116449356079, -.13600566983223, -.89187180995941, -.81647485494614, .83413660526276, -.21809615194798, .32638800144196, .47133237123489, -.4058920443058, -.42233863472939, .35867437720299, .49578228592873, .11262346804142, .70294010639191, .58906590938568, -.19715182483196, .86181098222733, -.37105345726013, 3.3236031532288, -1.543759226799, -.11011194437742, .64728397130966, -2.2335081100464, .67635416984558, 1.2392344474792, .10933646559715, .49816474318504, 2.0072033405304, -.17484994232655, 3.0224411487579, -3.7984521389008, -6.0368394851685, .27887633442879, 1.4904805421829, 1.3098726272583]) mse = np.array([ 1.0121052265167, .66349595785141, .65449619293213, .64957880973816, .64683443307877, .64528465270996, .64440369606018, .64390099048615, .64361357688904, .64344894886017, .64335465431213, .64330065250397, .64326965808868, .64325189590454, .64324170351028, .6432358622551, .64323252439499, .64323055744171, .64322948455811, .64322882890701, .64322847127914, .64322829246521, .64322817325592, .64322811365128, .64322805404663, .64322805404663, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199, .64322799444199]) stdp = np.array([ .82960641384125, .82960641384125, .697261095047, .61113905906677, .51607495546341, .47362637519836, .41342103481293, .40238001942635, .37454023957253, .33222004771233, .32514902949333, .31093680858612, .30019253492355, .31159669160843, .29182952642441, .30349296331406, .29457464814186, .28427124023438, .30664679408073, .29696446657181, .31270903348923, .29268020391464, .28816330432892, .29006817936897, .30216124653816, .30066826939583, .31728908419609, .30679926276207, .3272570669651, .37292611598969, .36668366193771, .40278288722038, .36799272894859, .36827209591866, .38623574376106, .39983862638474, .42789059877396, .43138384819031, .46953064203262, .48066720366478, .48910140991211, .53098994493484, .54496067762375, .55554050207138, .58130383491516, .60081332921982, .58008605241776, .58214038610458, .58369606733322, .53162068128586, .54543834924698, .52040082216263, .50143963098526, .48708060383797, .47620677947998, .48572361469269, .51068127155304, .61833620071411, .61110657453537, .76539021730423, .84672522544861, .92606955766678, .96840506792068, 1.0892199277878, 1.1097067594528, 1.0187155008316, 1.0030621290207, .97345739603043, .95103752613068, .82755368947983, .84054774045944, .85038793087006, .84008830785751, .92104357481003, .89359468221664, .87280809879303, .91032028198242, .95647835731506, 1.0624366998672, 1.1426770687103, 1.1679404973984, 1.311328291893, 1.473167181015, 1.5602221488953, 1.7326545715332, 1.8809853792191, 1.7803012132645, 1.7750589847565, 1.8420933485031, 1.7863517999649, 1.8328944444656, 1.7793855667114, 1.5791050195694, 1.3564316034317, 1.5250737667084, 1.3155146837234, 1.014811873436, .98235523700714, .97552710771561, .97035628557205, 1.0196926593781, 1.0393049716949, .98315137624741, .97613000869751, .89980864524841, .96626943349838, .91009211540222, .88530200719833, .97303456068039, .57794612646103, .63377332687378, .65829831361771, .76562696695328, .86465454101563, .90414637327194, .95180231332779, .95238989591599, .98833626508713, 1.0333099365234, 1.0851185321808, 1.1066001653671, 1.2293750047684, 1.233595252037, 1.1480363607407, 1.2962552309036, 1.2842413187027, 1.3106474876404, 1.5614050626755, 1.4672855138779, 1.2362524271011, 1.1855486631393, 1.1294020414352, 1.1046353578568, 1.0858771800995, 1.0716745853424, 1.0786685943604, 1.0662157535553, 1.0390332937241, .96519494056702, .9802839756012, .92070508003235, .91108840703964, .95705932378769, .95637094974518, .97360169887543, 1.0221517086029, .9701629281044, .94854199886322, .98542231321335, 1.048855304718, 1.0081344842911, 1.0305507183075, 1.0475262403488, .93612504005432, .85176283121109, .89438372850418, .820152759552, .71068543195724, .76979607343674, .76130604743958, .77262878417969, .85220617055893, .84146595001221, .93983960151672, .97883212566376, 1.0793634653091, 1.1909983158112, 1.1690304279327, 1.2411522865295, 1.1360056400299, 1.0918840169907, .9164656996727, .76586949825287, .918093085289, .87360894680023, .92867678403854, 1.00588285923, .92233866453171, .84132260084152, .90422683954239, .9873673915863, .99707210063934, 1.1109310388565, 1.1971517801285, 1.138188958168, 1.2710473537445, 1.1763968467712, 1.7437561750412, 1.4101150035858, 1.3527159690857, 1.4335050582886, .99765706062317, 1.1067585945129, 1.3086627721786, 1.2968333959579, 1.3547962903976, 1.6768488883972, 1.5905654430389, 2.0774590969086, 1.3218278884888, .21813294291496, .30750840902328, .60612773895264]) icstats = np.array([ 202, np.nan, -242.06033399744, 4, 492.12066799488, 505.35373878448]) class Bunch(dict): def __init__(self, **kw): dict.__init__(self, kw) self.__dict__ = self results = Bunch(llf=llf, nobs=nobs, k=k, k_exog=k_exog, sigma=sigma, chi2=chi2, df_model=df_model, k_ar=k_ar, k_ma=k_ma, params=params, cov_params=cov_params, xb=xb, y=y, resid=resid, yr=yr, mse=mse, stdp=stdp, icstats=icstats, )
bsd-3-clause
acee6294eb035e11edce5cac636ecd48
33.478532
229
0.401431
3.829287
false
false
false
false
yarikoptic/pystatsmodels
statsmodels/tsa/tests/results/arima211nc_results.py
35
36847
import numpy as np llf = np.array([-241.25977940638]) nobs = np.array([ 202]) k = np.array([ 4]) k_exog = np.array([ 1]) sigma = np.array([ .79533686587485]) chi2 = np.array([ 48655.961417345]) df_model = np.array([ 3]) k_ar = np.array([ 2]) k_ma = np.array([ 1]) params = np.array([ 1.1870704073154, -.19095698898571, -.90853757573555, .79533686587485]) cov_params = np.array([ .00204336743511, -.00177522179187, -.00165894353702, -.00031352141782, -.00177522179187, .00157376214003, .00132907629148, .00030367391511, -.00165894353702, .00132907629148, .00210988984438, .00024199988464, -.00031352141782, .00030367391511, .00024199988464, .00027937875185]).reshape(4,4) xb = np.array([ 0, 0, .11248598247766, .14283391833305, .0800810828805, .12544548511505, .07541109621525, .1297073662281, .10287435352802, .06303016841412, .09501431882381, .08120259642601, .07862555980682, .10874316096306, .06787430495024, .10527064651251, .08142036944628, .07337106764317, .11828763782978, .08380854874849, .11801292747259, .07338324189186, .0842502862215, .09106454998255, .10832596570253, .09570593386889, .1236881390214, .09362822026014, .13587079942226, .19111332297325, .14459040760994, .21043147146702, .12866979837418, .16308072209358, .19356986880302, .20215991139412, .23782986402512, .22326464951038, .28485587239265, .27474755048752, .28465977311134, .34938132762909, .3421268761158, .35463020205498, .39384591579437, .41037485003471, .36968034505844, .39875456690788, .40607318282127, .32915702462196, .40012913942337, .35161358118057, .34572568535805, .34037715196609, .3355179131031, .35895752906799, .38901025056839, .53648668527603, .43572762608528, .69034379720688, .69410443305969, .76356476545334, .77972346544266, .95276647806168, .9030898809433, .76722019910812, .84191131591797, .82463103532791, .82802563905716, .66399103403091, .79665386676788, .80260843038559, .78016436100006, .91813576221466, .80874294042587, .80483394861221, .8848432302475, .92809981107712, 1.0597171783447, 1.1029140949249, 1.0864543914795, 1.3046631813049, 1.4528053998947, 1.4744025468826, 1.6993381977081, 1.816978096962, 1.5705223083496, 1.6871707439423, 1.8281806707382, 1.7127912044525, 1.8617957830429, 1.7624272108078, 1.5169456005096, 1.3543643951416, 1.8122490644455, 1.3362231254578, 1.0437293052673, 1.2371381521225, 1.2306576967239, 1.2056746482849, 1.2665351629257, 1.2366921901703, 1.1172571182251, 1.1408381462097, 1.0126565694809, 1.1675561666489, 1.0074961185455, 1.0045058727264, 1.1498116254807, .44306626915932, .85451871156693, .81856834888458, .94427144527435, .99084824323654, .95836746692657, .994897544384, .95328682661057, 1.0093784332275, 1.0500040054321, 1.0956697463989, 1.090208530426, 1.2714649438858, 1.1823015213013, 1.0575052499771, 1.373840212822, 1.2371203899384, 1.3022859096527, 1.6853868961334, 1.3395566940308, 1.0802086591721, 1.2114092111588, 1.1690926551819, 1.1775953769684, 1.1662193536758, 1.1558910608292, 1.1743551492691, 1.1441857814789, 1.1080147027969, 1.0106881856918, 1.0909667015076, .97610247135162, 1.0038343667984, 1.0743995904922, 1.0255174636841, 1.0471519231796, 1.1034165620804, .97707790136337, .9856236577034, 1.0578545331955, 1.1219012737274, 1.0026258230209, 1.0733016729355, 1.0802255868912, .89154416322708, .85378932952881, .98660898208618, .82558387517929, .71030122041702, .88567733764648, .80868631601334, .82387971878052, .92999804019928, .83861750364304, .99909782409668, .97461491823196, 1.1019765138626, 1.1970175504684, 1.0780508518219, 1.2238110303879, 1.0100719928741, 1.0434579849243, .81277370452881, .72809249162674, 1.0880596637726, .87798285484314, .99824965000153, 1.0677480697632, .86986482143402, .81499886512756, .97921711206436, 1.0504562854767, .99342101812363, 1.1660186052322, 1.208247423172, 1.0516448020935, 1.3215674161911, 1.0694575309753, 2.0531799793243, 1.0617904663086, 1.2885792255402, 1.4795436859131, .73947989940643, 1.290878534317, 1.506583571434, 1.3157633543015, 1.424609541893, 1.8879710435867, 1.4916514158249, 2.3532779216766, .77780252695084, -.27798706293106, .7862361073494, 1.1202166080475]) y = np.array([np.nan, 28.979999542236, 29.26248550415, 29.492834091187, 29.450082778931, 29.665447235107, 29.625410079956, 29.879707336426, 29.942874908447, 29.873029708862, 30.015014648438, 30.06120300293, 30.118625640869, 30.318742752075, 30.287874221802, 30.485269546509, 30.521421432495, 30.553371429443, 30.808288574219, 30.833808898926, 31.058013916016, 31.023384094238, 31.104249954224, 31.211065292358, 31.388326644897, 31.475704193115, 31.703687667847, 31.743627548218, 32.015869140625, 32.471111297607, 32.594593048096, 33.060428619385, 33.028671264648, 33.263080596924, 33.593570709229, 33.902160644531, 34.337829589844, 34.623264312744, 35.184856414795, 35.574745178223, 35.984661102295, 36.649379730225, 37.142127990723, 37.654628753662, 38.293846130371, 38.910373687744, 39.269680023193, 39.798755645752, 40.306076049805, 40.42915725708, 41.00012588501, 41.251613616943, 41.545726776123, 41.840377807617, 42.135517120361, 42.558959960938, 43.089012145996, 44.236488342285, 44.635726928711, 46.290340423584, 47.494102478027, 48.863563537598, 50.079723358154, 51.952766418457, 53.203090667725, 53.767219543457, 54.841911315918, 55.724632263184, 56.628025054932, 56.763988494873, 57.796653747559, 58.702610015869, 59.480163574219, 60.91813659668, 61.608741760254, 62.404830932617, 63.584842681885, 64.828102111816, 66.559715270996, 68.202911376953, 69.586456298828, 71.904663085938, 74.45280456543, 76.67440032959, 79.699340820313, 82.716979980469, 84.170516967773, 86.387168884277, 89.028175354004, 90.812789916992, 93.361793518066, 95.16242980957, 95.916946411133, 96.354362487793, 99.31224822998, 99.436218261719, 98.943733215332, 100.03713989258, 101.03066253662, 102.00567626953, 103.36653137207, 104.5366973877, 105.21725463867, 106.24083709717, 106.71265411377, 108.1675567627, 108.70749664307, 109.50450897217, 111.04981231689, 109.14306640625, 110.35451507568, 111.01856231689, 112.34427642822, 113.6908416748, 114.7583694458, 115.99489593506, 116.95328521729, 118.20937347412, 119.55000305176, 120.9956741333, 122.29020690918, 124.37145996094, 125.68230438232, 126.45750427246, 128.87384033203, 130.13711547852, 131.80229187012, 135.08538818359, 136.03955078125, 136.18022155762, 137.4114074707, 138.36909484863, 139.47760009766, 140.56620788574, 141.65588378906, 142.87435913086, 143.94418334961, 144.90802001953, 145.51068115234, 146.69097900391, 147.27610778809, 148.2038269043, 149.47439575195, 150.4255065918, 151.5471496582, 152.90342712402, 153.57708740234, 154.4856262207, 155.75785827637, 157.22190856934, 158.00262451172, 159.2733001709, 160.48022460938, 160.79153442383, 161.25378417969, 162.4866027832, 162.82557678223, 162.9102935791, 164.08567810059, 164.70867919922, 165.52388000488, 166.82998657227, 167.53861999512, 169.09910583496, 170.27461242676, 172.00196838379, 173.89701843262, 174.97804260254, 176.82382202148, 177.41006469727, 178.44345092773, 178.41278076172, 178.42808532715, 180.38806152344, 180.87797546387, 182.1982421875, 183.66775512695, 184.06985473633, 184.51499938965, 185.87921142578, 187.35046386719, 188.3934173584, 190.26602172852, 192.00825500488, 192.85165405273, 195.12156677246, 195.76945495605, 201.25317382813, 200.4617767334, 201.98857116699, 204.17953491211, 202.63948059082, 204.86488342285, 207.42657470703, 208.65376281738, 210.55760192871, 214.38296508789, 215.48864746094, 220.96327209473, 217.66680908203, 211.89601135254, 213.45724487305, 215.58921813965]) resid = np.array([np.nan, .17000007629395, .08751478046179, -.12283346056938, .08991899341345, -.11544716358185, .12458966672421, -.03970721364021, -.13287504017353, .04697044193745, -.03501485288143, -.02120122499764, .09137260913849, -.09874293208122, .09212554246187, -.04526927694678, -.04142136126757, .13662992417812, -.0582881718874, .10619198530912, -.10801269859076, -.00338354869746, .01575009897351, .06893529742956, -.00832748971879, .10429482907057, -.05368844047189, .13637132942677, .26412883400917, -.02111134678125, .2554073035717, -.16042841970921, .07132714986801, .13692232966423, .1064293757081, .19783779978752, .0621731877327, .27673536539078, .11514183133841, .12525399029255, .31533870100975, .15061867237091, .15787313878536, .24537208676338, .20615255832672, -.01037331111729, .13031965494156, .10124543309212, -.20607624948025, .1708429902792, -.10012608766556, -.05161434784532, -.04572645947337, -.04037792980671, .06448362022638, .14104247093201, .61098974943161, -.03648666664958, .96427005529404, .50965696573257, .60589480400085, .43643599748611, .9202772974968, .34723278880119, -.20308908820152, .23277981579304, .05809023976326, .07536666095257, -.52802640199661, .23601049184799, .1033476293087, -.00260917330161, .51983487606049, -.11813650280237, -.00874368380755, .29516834020615, .31515756249428, .67189866304398, .54028129577637, .29708743095398, 1.0135440826416, 1.095338344574, .74719160795212, 1.3256005048752, 1.2006633281708, -.11698111891747, .52947622537613, .81282931566238, .07182084023952, .68721032142639, .03820572793484, -.7624272108078, -.91694712638855, 1.1456356048584, -1.2122505903244, -1.5362200737, -.14372782409191, -.23713812232018, -.2306577116251, .09432080388069, -.06653053313494, -.43669676780701, -.11725706607103, -.54083967208862, .28734645247459, -.467559248209, -.20749309659004, .39549562335014, -2.3498160839081, .3569367825985, -.15452179312706, .38143622875214, .35572397708893, .1091578528285, .24162948131561, .00510244909674, .24671010673046, .29062458872795, .34999752044678, .20432561635971, .80979299545288, .12853652238846, -.28230002522469, 1.042493224144, .02615367434919, .36288577318192, 1.5977079868317, -.38538381457329, -.93954759836197, .01978221163154, -.21140915155411, -.06908652186394, -.07760456204414, -.06621328741312, .0441059358418, -.07434900850058, -.14418575167656, -.40801778435707, .08931794017553, -.39096972346306, -.07610860466957, .19616261124611, -.07439963519573, .07448863238096, .25285106897354, -.30341354012489, -.07708399742842, .21437329053879, .34215462207794, -.22190742194653, .19737112522125, .12669529020786, -.58022564649582, -.3915441930294, .24621678888798, -.48660898208618, -.6255869269371, .28969877958298, -.18568041920662, -.00868325773627, .37611722946167, -.12999498844147, .5613916516304, .20089910924435, .6253759264946, .69802659749985, .00297940592282, .621961414814, -.42382326722145, -.01007199659944, -.84344571828842, -.71278285980225, .87191361188889, -.38806268572807, .32201409339905, .40175950527191, -.4677571952343, -.36986482143402, .38499811291695, .42079201340675, .04953457415104, .70659118890762, .53397834300995, -.20824746787548, .94835525751114, -.42157354950905, 3.4305424690247, -1.8531830310822, .23821261525154, .71142077445984, -2.2795467376709, .93453133106232, 1.0551145076752, -.08858433365822, .47923478484154, 1.9373899698257, -.38597220182419, 3.1213552951813, -4.0742712020874, -5.4928140640259, .77499634027481, 1.0117527246475, .79578375816345]) yr = np.array([np.nan, .17000007629395, .08751478046179, -.12283346056938, .08991899341345, -.11544716358185, .12458966672421, -.03970721364021, -.13287504017353, .04697044193745, -.03501485288143, -.02120122499764, .09137260913849, -.09874293208122, .09212554246187, -.04526927694678, -.04142136126757, .13662992417812, -.0582881718874, .10619198530912, -.10801269859076, -.00338354869746, .01575009897351, .06893529742956, -.00832748971879, .10429482907057, -.05368844047189, .13637132942677, .26412883400917, -.02111134678125, .2554073035717, -.16042841970921, .07132714986801, .13692232966423, .1064293757081, .19783779978752, .0621731877327, .27673536539078, .11514183133841, .12525399029255, .31533870100975, .15061867237091, .15787313878536, .24537208676338, .20615255832672, -.01037331111729, .13031965494156, .10124543309212, -.20607624948025, .1708429902792, -.10012608766556, -.05161434784532, -.04572645947337, -.04037792980671, .06448362022638, .14104247093201, .61098974943161, -.03648666664958, .96427005529404, .50965696573257, .60589480400085, .43643599748611, .9202772974968, .34723278880119, -.20308908820152, .23277981579304, .05809023976326, .07536666095257, -.52802640199661, .23601049184799, .1033476293087, -.00260917330161, .51983487606049, -.11813650280237, -.00874368380755, .29516834020615, .31515756249428, .67189866304398, .54028129577637, .29708743095398, 1.0135440826416, 1.095338344574, .74719160795212, 1.3256005048752, 1.2006633281708, -.11698111891747, .52947622537613, .81282931566238, .07182084023952, .68721032142639, .03820572793484, -.7624272108078, -.91694712638855, 1.1456356048584, -1.2122505903244, -1.5362200737, -.14372782409191, -.23713812232018, -.2306577116251, .09432080388069, -.06653053313494, -.43669676780701, -.11725706607103, -.54083967208862, .28734645247459, -.467559248209, -.20749309659004, .39549562335014, -2.3498160839081, .3569367825985, -.15452179312706, .38143622875214, .35572397708893, .1091578528285, .24162948131561, .00510244909674, .24671010673046, .29062458872795, .34999752044678, .20432561635971, .80979299545288, .12853652238846, -.28230002522469, 1.042493224144, .02615367434919, .36288577318192, 1.5977079868317, -.38538381457329, -.93954759836197, .01978221163154, -.21140915155411, -.06908652186394, -.07760456204414, -.06621328741312, .0441059358418, -.07434900850058, -.14418575167656, -.40801778435707, .08931794017553, -.39096972346306, -.07610860466957, .19616261124611, -.07439963519573, .07448863238096, .25285106897354, -.30341354012489, -.07708399742842, .21437329053879, .34215462207794, -.22190742194653, .19737112522125, .12669529020786, -.58022564649582, -.3915441930294, .24621678888798, -.48660898208618, -.6255869269371, .28969877958298, -.18568041920662, -.00868325773627, .37611722946167, -.12999498844147, .5613916516304, .20089910924435, .6253759264946, .69802659749985, .00297940592282, .621961414814, -.42382326722145, -.01007199659944, -.84344571828842, -.71278285980225, .87191361188889, -.38806268572807, .32201409339905, .40175950527191, -.4677571952343, -.36986482143402, .38499811291695, .42079201340675, .04953457415104, .70659118890762, .53397834300995, -.20824746787548, .94835525751114, -.42157354950905, 3.4305424690247, -1.8531830310822, .23821261525154, .71142077445984, -2.2795467376709, .93453133106232, 1.0551145076752, -.08858433365822, .47923478484154, 1.9373899698257, -.38597220182419, 3.1213552951813, -4.0742712020874, -5.4928140640259, .77499634027481, 1.0117527246475, .79578375816345]) mse = np.array([ 1.4402351379395, 1.4402351379395, .80966705083847, .74677377939224, .71241801977158, .69108927249908, .67678099870682, .66667699813843, .6592805981636, .6537224650383, .64946305751801, .64614951610565, .64354157447815, .64147007465363, .639812707901, .63847899436951, .63740062713623, .63652545213699, .63581293821335, .63523155450821, .63475602865219, .63436657190323, .63404709100723, .63378477096558, .63356912136078, .63339179754257, .63324582576752, .63312560319901, .63302659988403, .63294500112534, .63287770748138, .63282227516174, .63277649879456, .63273876905441, .63270765542984, .63268196582794, .63266080617905, .63264334201813, .63262891769409, .63261699676514, .63260716199875, .63259905576706, .63259238004684, .63258683681488, .63258230686188, .63257849216461, .63257539272308, .63257282972336, .63257074356079, .63256901502609, .63256752490997, .63256633281708, .63256537914276, .63256454467773, .63256388902664, .63256335258484, .63256287574768, .63256251811981, .63256222009659, .63256192207336, .63256174325943, .6325615644455, .63256138563156, .63256126642227, .63256120681763, .63256108760834, .63256102800369, .63256096839905, .63256096839905, .6325609087944, .63256084918976, .63256084918976, .63256084918976, .63256078958511, .63256078958511, .63256078958511, .63256078958511, .63256078958511, .63256078958511, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047, .63256072998047]) icstats = np.array([ 202, np.nan, -241.25977940638, 4, 490.51955881276, 503.75262960236]) class Bunch(dict): def __init__(self, **kw): dict.__init__(self, kw) self.__dict__ = self results = Bunch(llf=llf, nobs=nobs, k=k, k_exog=k_exog, sigma=sigma, chi2=chi2, df_model=df_model, k_ar=k_ar, k_ma=k_ma, params=params, cov_params=cov_params, xb=xb, y=y, resid=resid, yr=yr, mse=mse, icstats=icstats, )
bsd-3-clause
dd5fda91562c6151b5a22216ec0c8f9e
33.212628
218
0.403126
3.808869
false
false
false
false
yarikoptic/pystatsmodels
statsmodels/datasets/cpunish/data.py
3
2548
"""US Capital Punishment dataset.""" __docformat__ = 'restructuredtext' COPYRIGHT = """Used with express permission from the original author, who retains all rights.""" TITLE = __doc__ SOURCE = """ Jeff Gill's `Generalized Linear Models: A Unified Approach` http://jgill.wustl.edu/research/books.html """ DESCRSHORT = """Number of state executions in 1997""" DESCRLONG = """This data describes the number of times capital punishment is implemented at the state level for the year 1997. The outcome variable is the number of executions. There were executions in 17 states. Included in the data are explanatory variables for median per capita income in dollars, the percent of the population classified as living in poverty, the percent of Black citizens in the population, the rate of violent crimes per 100,000 residents for 1996, a dummy variable indicating whether the state is in the South, and (an estimate of) the proportion of the population with a college degree of some kind. """ NOTE = """ Number of Observations - 17 Number of Variables - 7 Variable name definitions:: EXECUTIONS - Executions in 1996 INCOME - Median per capita income in 1996 dollars PERPOVERTY - Percent of the population classified as living in poverty PERBLACK - Percent of black citizens in the population VC100k96 - Rate of violent crimes per 100,00 residents for 1996 SOUTH - SOUTH == 1 indicates a state in the South DEGREE - An esimate of the proportion of the state population with a college degree of some kind State names are included in the data file, though not returned by load. """ from numpy import recfromtxt, column_stack, array from statsmodels.datasets import utils as du from os.path import dirname, abspath def load(): """ Load the cpunish data and return a Dataset class. Returns ------- Dataset instance: See DATASET_PROPOSAL.txt for more information. """ data = _get_data() return du.process_recarray(data, endog_idx=0, dtype=float) def load_pandas(): """ Load the cpunish data and return a Dataset class. Returns ------- Dataset instance: See DATASET_PROPOSAL.txt for more information. """ data = _get_data() return du.process_recarray_pandas(data, endog_idx=0, dtype=float) def _get_data(): filepath = dirname(abspath(__file__)) data = recfromtxt(open(filepath + '/cpunish.csv', 'rb'), delimiter=",", names=True, dtype=float, usecols=(1,2,3,4,5,6,7)) return data
bsd-3-clause
d5f92df7b4e92ce0abdff6f9720424f0
31.666667
90
0.706829
3.65043
false
false
false
false
yarikoptic/pystatsmodels
statsmodels/sandbox/stats/tests/__init__.py
218
6354
''' Econometrics for a Datarich Environment ======================================= Introduction ------------ In many cases we are performing statistical analysis when many observed variables are available, when we are in a data rich environment. Machine learning has a wide variety of tools for dimension reduction and penalization when there are many varibles compared to the number of observation. Chemometrics has a long tradition of using Partial Least Squares, NIPALS and similar in these cases. In econometrics the same problem shows up when there are either many possible regressors, many (weak) instruments or when there are a large number of moment conditions in GMM. This section is intended to collect some models and tools in this area that are relevant for the statical analysis and econometrics. Covariance Matrices =================== Several methods are available to reduce the small sample noise in estimated covariance matrices with many variable. Some applications: weighting matrix with many moments, covariance matrix for portfolio choice Dimension Reduction =================== Principal Component and Partial Least Squares try to extract the important low dimensional factors from the data with many variables. Regression with many regressors =============================== Factor models, selection of regressors and shrinkage and penalization are used to improve the statistical properties, when the presence of too many regressors leads to over-fitting and too noisy small sample estimators and statistics. Regression with many moments or many instruments ================================================ The same tools apply and can be used in these two cases. e.g. Tychonov regularization of weighting matrix in GMM, similar to Ridge regression, the weighting matrix can be shrunk towards the identity matrix. Simplest case will be part of GMM. I don't know how much will be standalone functions. Intended Content ================ PLS --- what should be available in class? Factormodel and supporting helper functions ------------------------------------------- PCA based ~~~~~~~~~ First version based PCA on Stock/Watson and Bai/Ng, and recent papers on the selection of the number of factors. Not sure about Forni et al. in approach. Basic support of this needs additional results for PCA, error covariance matrix of data on reduced factors, required for criteria in Bai/Ng. Selection criteria based on eigenvalue cutoffs. Paper on PCA and structural breaks. Could add additional results during find_nfact to test for parameter stability. I haven't read the paper yet. Idea: for forecasting, use up to h-step ahead endogenous variables to directly get the forecasts. Asymptotic results and distribution: not too much idea yet. Standard OLS results are conditional on factors, paper by Haerdle (abstract seems to suggest that this is ok, Park 2009). Simulation: add function to simulate DGP of Bai/Ng and recent extension. Sensitivity of selection criteria to heteroscedasticity and autocorrelation. Bai, J. & Ng, S., 2002. Determining the Number of Factors in Approximate Factor Models. Econometrica, 70(1), pp.191-221. Kapetanios, G., 2010. A Testing Procedure for Determining the Number of Factors in Approximate Factor Models With Large Datasets. Journal of Business and Economic Statistics, 28(3), pp.397-409. Onatski, A., 2010. Determining the Number of Factors from Empirical Distribution of Eigenvalues. Review of Economics and Statistics, 92(4), pp.1004-1016. Alessi, L., Barigozzi, M. & Capasso, M., 2010. Improved penalization for determining the number of factors in approximate factor models. Statistics & Probability Letters, 80(23-24), pp.1806-1813. Breitung, J. & Eickmeier, S., Testing for structural breaks in dynamic factor models. Journal of Econometrics, In Press, Accepted Manuscript. Available at: http://www.sciencedirect.com/science/article/B6VC0-51G3W92-1/2/f45ce2332443374fd770e42e5a68ddb4 [Accessed November 15, 2010]. Croux, C., Renault, E. & Werker, B., 2004. Dynamic factor models. Journal of Econometrics, 119(2), pp.223-230. Forni, M. et al., 2009. Opening the Black Box: Structural Factor Models with Large Cross Sections. Econometric Theory, 25(05), pp.1319-1347. Forni, M. et al., 2000. The Generalized Dynamic-Factor Model: Identification and Estimation. Review of Economics and Statistics, 82(4), pp.540-554. Forni, M. & Lippi, M., The general dynamic factor model: One-sided representation results. Journal of Econometrics, In Press, Accepted Manuscript. Available at: http://www.sciencedirect.com/science/article/B6VC0-51FNPJN-1/2/4fcdd0cfb66e3050ff5d19bf2752ed19 [Accessed November 15, 2010]. Kapetanios, G., 2010. A Testing Procedure for Determining the Number of Factors in Approximate Factor Models With Large Datasets. Journal of Business and Economic Statistics, 28(3), pp.397-409. Onatski, A., 2010. Determining the Number of Factors from Empirical Distribution of Eigenvalues. Review of Economics and Statistics, 92(4), pp.1004-1016. Park, B.U. et al., 2009. Time Series Modelling With Semiparametric Factor Dynamics. Journal of the American Statistical Association, 104(485), pp.284-298. other factor algorithm ~~~~~~~~~~~~~~~~~~~~~~ PLS should fit in reasonably well. Bai/Ng have a recent paper, where they compare LASSO, PCA, and similar, individual and in combination. Check how much we can use scikits.learn for this. miscellaneous ~~~~~~~~~~~~~ Time series modeling of factors for prediction, ARMA, VARMA. SUR and correlation structure What about sandwich estimation, robust covariance matrices? Similarity to Factor-Garch and Go-Garch Updating: incremental PCA, ...? TODO next ========= MVOLS : OLS with multivariate endogenous and identical exogenous variables. rewrite and expand current varma_process.VAR PCA : write a class after all, and/or adjust the current donated class and keep adding required statistics, e.g. residual variance, projection of X on k-factors, ... updating ? FactorModelUnivariate : started, does basic principal component regression, based on standard information criteria, not Bai/Ng adjusted FactorModelMultivariate : follow pattern for univariate version and use MVOLS '''
bsd-3-clause
a9e0fe95e0e7e82118585fa9e0049469
37.509091
99
0.748033
3.766449
false
false
false
false
yarikoptic/pystatsmodels
statsmodels/sandbox/distributions/tests/_est_fit.py
5
2608
# NOTE: contains only one test, _est_cont_fit, that is renamed so that # nose doesn't run it # I put this here for the record and for the case when someone wants to # verify the quality of fit # with current parameters: relatively small sample size, default starting values # Ran 84 tests in 401.797s # FAILED (failures=15) import numpy.testing as npt import numpy as np from scipy import stats from distparams import distcont # this is not a proper statistical test for convergence, but only # verifies that the estimate and true values don't differ by too much n_repl1 = 1000 # sample size for first run n_repl2 = 5000 # sample size for second run, if first run fails thresh_percent = 0.25 # percent of true parameters for fail cut-off thresh_min = 0.75 # minimum difference estimate - true to fail test #distcont = [['genextreme', (3.3184017469423535,)]] def _est_cont_fit(): # this tests the closeness of the estimated parameters to the true # parameters with fit method of continuous distributions # Note: is slow, some distributions don't converge with sample size <= 10000 for distname, arg in distcont: yield check_cont_fit, distname,arg def check_cont_fit(distname,arg): distfn = getattr(stats, distname) rvs = distfn.rvs(size=n_repl1,*arg) est = distfn.fit(rvs) #,*arg) # start with default values truearg = np.hstack([arg,[0.0,1.0]]) diff = est-truearg txt = '' diffthreshold = np.max(np.vstack([truearg*thresh_percent, np.ones(distfn.numargs+2)*thresh_min]),0) # threshold for location diffthreshold[-2] = np.max([np.abs(rvs.mean())*thresh_percent,thresh_min]) if np.any(np.isnan(est)): raise AssertionError('nan returned in fit') else: if np.any((np.abs(diff) - diffthreshold) > 0.0): ## txt = 'WARNING - diff too large with small sample' ## print 'parameter diff =', diff - diffthreshold, txt rvs = np.concatenate([rvs,distfn.rvs(size=n_repl2-n_repl1,*arg)]) est = distfn.fit(rvs) #,*arg) truearg = np.hstack([arg,[0.0,1.0]]) diff = est-truearg if np.any((np.abs(diff) - diffthreshold) > 0.0): txt = 'parameter: %s\n' % str(truearg) txt += 'estimated: %s\n' % str(est) txt += 'diff : %s\n' % str(diff) raise AssertionError('fit not very good in %s\n' % distfn.name + txt) if __name__ == "__main__": import nose #nose.run(argv=['', __file__]) nose.runmodule(argv=[__file__,'-s'], exit=False)
bsd-3-clause
87d733bfb0e8ae73b9a1300acfe29427
36.257143
85
0.634202
3.445178
false
true
false
false
yarikoptic/pystatsmodels
statsmodels/sandbox/tests/maketests_mlabwrap.py
36
9022
'''generate py modules with test cases and results from mlabwrap currently matlab: princomp, garchar, garchma ''' import numpy as np from numpy.testing import assert_array_almost_equal from numpy import array xo = array([[ -419, -731, -1306, -1294], [ 6, 529, -200, -437], [ -27, -833, -6, -564], [ -304, -273, -502, -739], [ 1377, -912, 927, 280], [ -375, -517, -514, 49], [ 247, -504, 123, -259], [ 712, 534, -773, 286], [ 195, -1080, 3256, -178], [ -854, 75, -706, -1084], [-1219, -612, -15, -203], [ 550, -628, -483, -2686], [ -365, 1376, -1266, 317], [ -489, 544, -195, 431], [ -656, 854, 840, -723], [ 16, -1385, -880, -460], [ 258, -2252, 96, 54], [ 2049, -750, -1115, 381], [ -65, 280, -777, 416], [ 755, 82, -806, 1027], [ -39, -170, -2134, 743], [ -859, 780, 746, -133], [ 762, 252, -450, -459], [ -941, -202, 49, -202], [ -54, 115, 455, 388], [-1348, 1246, 1430, -480], [ 229, -535, -1831, 1524], [ -651, -167, 2116, 483], [-1249, -1373, 888, -1092], [ -75, -2162, 486, -496], [ 2436, -1627, -1069, 162], [ -63, 560, -601, 587], [ -60, 1051, -277, 1323], [ 1329, -1294, 68, 5], [ 1532, -633, -923, 696], [ 669, 895, -1762, -375], [ 1129, -548, 2064, 609], [ 1320, 573, 2119, 270], [ -213, -412, -2517, 1685], [ 73, -979, 1312, -1220], [-1360, -2107, -237, 1522], [ -645, 205, -543, -169], [ -212, 1072, 543, -128], [ -352, -129, -605, -904], [ 511, 85, 167, -1914], [ 1515, 1862, 942, 1622], [ -465, 623, -495, -89], [-1396, -979, 1758, 128], [ -255, -47, 980, 501], [-1282, -58, -49, -610], [ -889, -1177, -492, 494], [ 1415, 1146, 696, -722], [ 1237, -224, -1609, -64], [ -528, -1625, 231, 883], [ -327, 1636, -476, -361], [ -781, 793, 1882, 234], [ -506, -561, 1988, -810], [-1233, 1467, -261, 2164], [ 53, 1069, 824, 2123], [-1200, -441, -321, 339], [ 1606, 298, -995, 1292], [-1740, -672, -1628, -129], [-1450, -354, 224, -657], [-2556, 1006, -706, -1453], [ -717, -463, 345, -1821], [ 1056, -38, -420, -455], [ -523, 565, 425, 1138], [-1030, -187, 683, 78], [ -214, -312, -1171, -528], [ 819, 736, -265, 423], [ 1339, 351, 1142, 579], [ -387, -126, -1573, 2346], [ 969, 2, 327, -134], [ 163, 227, 90, 2021], [ 1022, -1076, 174, 304], [ 1042, 1317, 311, 880], [ 2018, -840, 295, 2651], [ -277, 566, 1147, -189], [ 20, 467, 1262, 263], [ -663, 1061, -1552, -1159], [ 1830, 391, 2534, -199], [ -487, 752, -1061, 351], [-2138, -556, -367, -457], [ -868, -411, -559, 726], [ 1770, 819, -892, -363], [ 553, -736, -169, -490], [ 388, -503, 809, -821], [ -516, -1452, -192, 483], [ 493, 2904, 1318, 2591], [ 175, 584, -1001, 1675], [ 1316, -1596, -460, 1500], [ 1212, 214, -644, -696], [ -501, 338, 1197, -841], [ -587, -469, -1101, 24], [-1205, 1910, 659, 1232], [ -150, 398, 594, 394], [ 34, -663, 235, -334], [-1580, 647, 239, -351], [-2177, -345, 1215, -1494], [ 1923, 329, -152, 1128]]) x = xo/1000. class HoldIt(object): def __init__(self, name): self.name = name def save(self, what=None, filename=None, header=True, useinstant=True, comment=None): if what is None: what = (i for i in self.__dict__ if i[0] != '_') if header: txt = ['import numpy as np\nfrom numpy import array\n\n'] if useinstant: txt.append('class Holder(object):\n pass\n\n') else: txt = [] if useinstant: txt.append('%s = Holder()' % self.name) prefix = '%s.' % self.name else: prefix = '' if not comment is None: txt.append("%scomment = '%s'" % (prefix, comment)) for x in what: txt.append('%s%s = %s' % (prefix, x, repr(getattr(self,x)))) txt.extend(['','']) #add empty lines at end if not filename is None: file(filename, 'a+').write('\n'.join(txt)) return txt def generate_princomp(xo, filen='testsave.py'): # import mlabwrap only when run as script import mlabwrap from mlabwrap import mlab np.set_printoptions(precision=14, linewidth=100) data = HoldIt('data') data.xo = xo data.save(filename='testsave.py', comment='generated data, divide by 1000') res_princomp = HoldIt('princomp1') res_princomp.coef, res_princomp.factors, res_princomp.values = \ mlab.princomp(x, nout=3) res_princomp.save(filename=filen, header=False, comment='mlab.princomp(x, nout=3)') res_princomp = HoldIt('princomp2') res_princomp.coef, res_princomp.factors, res_princomp.values = \ mlab.princomp(x[:20,], nout=3) np.set_printoptions(precision=14, linewidth=100) res_princomp.save(filename=filen, header=False, comment='mlab.princomp(x[:20,], nout=3)') res_princomp = HoldIt('princomp3') res_princomp.coef, res_princomp.factors, res_princomp.values = \ mlab.princomp(x[:20,]-x[:20,].mean(0), nout=3) np.set_printoptions(precision=14, linewidth=100) res_princomp.save(filename=filen, header=False, comment='mlab.princomp(x[:20,]-x[:20,].mean(0), nout=3)') def generate_armarep(filen='testsave.py'): # import mlabwrap only when run as script import mlabwrap from mlabwrap import mlab res_armarep = HoldIt('armarep') res_armarep.ar = np.array([1., -0.5, +0.8]) res_armarep.ma = np.array([1., -0.6, 0.08]) res_armarep.marep = mlab.garchma(-res_armarep.ar[1:], res_armarep.ma[1:], 20) res_armarep.arrep = mlab.garchar(-res_armarep.ar[1:], res_armarep.ma[1:], 20) res_armarep.save(filename=filen, header=False, comment=("''mlab.garchma(-res_armarep.ar[1:], res_armarep.ma[1:], 20)\n" + "mlab.garchar(-res_armarep.ar[1:], res_armarep.ma[1:], 20)''")) def exampletest(): from statsmodels.sandbox import tsa arrep = tsa.arma_impulse_response(res_armarep.ma, res_armarep.ar, nobs=21)[1:] marep = tsa.arma_impulse_response(res_armarep.ar, res_armarep.ma, nobs=21)[1:] assert_array_almost_equal(res_armarep.marep.ravel(), marep, 14) #difference in sign convention to matlab for AR term assert_array_almost_equal(-res_armarep.arrep.ravel(), arrep, 14) if __name__ == '__main__': import mlabwrap from mlabwrap import mlab import savedrvs xo = savedrvs.rvsdata.xar2 x100 = xo[-100:]/1000. x1000 = xo/1000. filen = 'testsavetls.py' res_pacf = HoldIt('mlpacf') res_pacf.comment = 'mlab.parcorr(x, [], 2, nout=3)' res_pacf.pacf100, res_pacf.lags100, res_pacf.bounds100 = \ mlab.parcorr(x100, [], 2, nout=3) res_pacf.pacf1000, res_pacf.lags1000, res_pacf.bounds1000 = \ mlab.parcorr(x1000, [], 2, nout=3) res_pacf.save(filename=filen, header=True) res_acf = HoldIt('mlacf') res_acf.comment = 'mlab.autocorr(x, [], 2, nout=3)' res_acf.acf100, res_acf.lags100, res_acf.bounds100 = \ mlab.autocorr(x100, [], 2, nout=3) res_acf.acf1000, res_acf.lags1000, res_acf.bounds1000 = \ mlab.autocorr(x1000, [], 2, nout=3) res_acf.save(filename=filen, header=False) res_ccf = HoldIt('mlccf') res_ccf.comment = 'mlab.crosscorr(x[4:], x[:-4], [], 2, nout=3)' res_ccf.ccf100, res_ccf.lags100, res_ccf.bounds100 = \ mlab.crosscorr(x100[4:], x100[:-4], [], 2, nout=3) res_ccf.ccf1000, res_ccf.lags1000, res_ccf.bounds1000 = \ mlab.crosscorr(x1000[4:], x1000[:-4], [], 2, nout=3) res_ccf.save(filename=filen, header=False) res_ywar = HoldIt('mlywar') res_ywar.comment = "mlab.ar(x100-x100.mean(), 10, 'yw').a.ravel()" mbaryw = mlab.ar(x100-x100.mean(), 10, 'yw') res_ywar.arcoef100 = np.array(mbaryw.a.ravel()) mbaryw = mlab.ar(x1000-x1000.mean(), 20, 'yw') res_ywar.arcoef1000 = np.array(mbaryw.a.ravel()) res_ywar.save(filename=filen, header=False)
bsd-3-clause
0376429c335875feee6649ab1b5b10a6
35.674797
86
0.495345
2.551471
false
false
false
false
yarikoptic/pystatsmodels
statsmodels/examples/ex_feasible_gls_het_0.py
3
6390
# -*- coding: utf-8 -*- """Examples for linear model with heteroscedasticity estimated by feasible GLS These are examples to check the results during developement. The assumptions: We have a linear model y = X*beta where the variance of an observation depends on some explanatory variable Z (`exog_var`). linear_model.WLS estimated the model for a given weight matrix here we want to estimate also the weight matrix by two step or iterative WLS Created on Wed Dec 21 12:28:17 2011 Author: Josef Perktold """ import numpy as np from numpy.testing import assert_almost_equal from statsmodels.regression.linear_model import OLS, WLS, GLS from statsmodels.regression.feasible_gls import GLSHet, GLSHet2 from statsmodels.tools.tools import add_constant examples = ['ex1'] if 'ex1' in examples: nsample = 300 #different pattern last graph with 100 or 200 or 500 sig = 0.5 np.random.seed(9876789) #9876543) X = np.random.randn(nsample, 3) X = np.column_stack((np.ones((nsample,1)), X)) beta = [1, 0.5, -0.5, 1.] y_true2 = np.dot(X, beta) x1 = np.linspace(0, 1, nsample) gamma = np.array([1, 3.]) #with slope 3 instead of two, I get negative weights, Not correct # - was misspecified, but the negative weights are still possible with identity link #gamma /= gamma.sum() #normalize assuming x1.max is 1 z_true = add_constant(x1) winv = np.dot(z_true, gamma) het_params = sig**2 * np.array([1, 3.]) # for squared sig2_het = sig**2 * winv weights_dgp = 1/winv weights_dgp /= weights_dgp.max() #should be already normalized - NOT check normalization #y2[:nsample*6/10] = y_true2[:nsample*6/10] + sig*1. * np.random.normal(size=nsample*6/10) z0 = np.zeros(nsample) z0[(nsample * 5)//10:] = 1 #dummy for 2 halfs of sample z0 = add_constant(z0) z1 = add_constant(x1) noise = np.sqrt(sig2_het) * np.random.normal(size=nsample) y2 = y_true2 + noise X2 = X[:,[0,2]] #misspecified, missing regressor in main equation X2 = X #correctly specigied res_ols = OLS(y2, X2).fit() print 'OLS beta estimates' print res_ols.params print 'OLS stddev of beta' print res_ols.bse print '\nWLS' mod0 = GLSHet2(y2, X2, exog_var=winv) res0 = mod0.fit() print 'new version' mod1 = GLSHet(y2, X2, exog_var=winv) res1 = mod1.iterative_fit(2) print 'WLS beta estimates' print res1.params print res0.params print 'WLS stddev of beta' print res1.bse #compare with previous version GLSHet2, refactoring check #assert_almost_equal(res1.params, np.array([ 0.37642521, 1.51447662])) #this fails ??? more iterations? different starting weights? print res1.model.weights/res1.model.weights.max() #why is the error so small in the estimated weights ? assert_almost_equal(res1.model.weights/res1.model.weights.max(), weights_dgp, 14) print 'residual regression params' print res1.results_residual_regression.params print 'scale of model ?' print res1.scale print 'unweighted residual variance, note unweighted mean is not zero' print res1.resid.var() #Note weighted mean is zero: #(res1.model.weights * res1.resid).mean() doplots = True #False if doplots: import matplotlib.pyplot as plt plt.figure() plt.plot(x1, y2, 'o') plt.plot(x1, y_true2, 'b-', label='true') plt.plot(x1, res1.fittedvalues, 'r-', label='fwls') plt.plot(x1, res_ols.fittedvalues, '--', label='ols') plt.legend() #the next only works if w has finite support, discrete/categorical #z = (w[:,None] == [1,4]).astype(float) #dummy variable #z = (w0[:,None] == np.unique(w0)).astype(float) #dummy variable #changed z0 contains dummy and constant mod2 = GLSHet(y2, X2, exog_var=z0) res2 = mod2.iterative_fit(3) print res2.params import statsmodels.api as sm #z = sm.add_constant(w, prepend=True) z = sm.add_constant(x1/x1.max()) mod3 = GLSHet(y2, X2, exog_var=z1)#, link=sm.families.links.log()) res3 = mod3.iterative_fit(20) error_var_3 = res3.mse_resid/res3.model.weights print res3.params print "np.array(res3.model.history['ols_params'])" print np.array(res3.model.history['ols_params']) print "np.array(res3.model.history['self_params'])" print np.array(res3.model.history['self_params']) #Models 2 and 3 are equivalent with different parameterization of Z print np.unique(res2.model.weights) #for discrete z only, only a few uniques print np.unique(res3.model.weights) print res3.summary() print '\n\nResults of estimation of weights' print '--------------------------------' print res3.results_residual_regression.summary() if doplots: plt.figure() plt.plot(x1, y2, 'o') plt.plot(x1, y_true2, 'b-', label='true') plt.plot(x1, res1.fittedvalues, '-', label='fwls1') plt.plot(x1, res2.fittedvalues, '-', label='fwls2') plt.plot(x1, res3.fittedvalues, '-', label='fwls3') plt.plot(x1, res_ols.fittedvalues, '--', label='ols') plt.legend() plt.figure() plt.ylim(0, 5) res_e2 = OLS(noise**2, z).fit() plt.plot(noise**2, 'bo', alpha=0.5, label='dgp error**2') plt.plot(res_e2.fittedvalues, lw=2, label='ols for noise**2') #plt.plot(res3.model.weights, label='GLSHet weights') plt.plot(error_var_3, lw=2, label='GLSHet error var') plt.plot(res3.resid**2, 'ro', alpha=0.5, label='resid squared') #plt.plot(weights_dgp, label='DGP weights') plt.plot(sig**2 * winv, lw=2, label='DGP error var') plt.legend() plt.show() '''Note these are close but maybe biased because of skewed distribution >>> res3.mse_resid/res3.model.weights[-10:] array([ 1.03115871, 1.03268209, 1.03420547, 1.03572885, 1.03725223, 1.03877561, 1.04029899, 1.04182237, 1.04334575, 1.04486913]) >>> res_e2.fittedvalues[-10:] array([ 1.0401953 , 1.04171386, 1.04323242, 1.04475098, 1.04626954, 1.0477881 , 1.04930666, 1.05082521, 1.05234377, 1.05386233]) >>> sig**2 * w[-10:] array([ 0.98647295, 0.98797595, 0.98947896, 0.99098196, 0.99248497, 0.99398798, 0.99549098, 0.99699399, 0.99849699, 1. ]) '''
bsd-3-clause
d64a9fcc5f4274e674cf489e4e53d9cb
35.936416
94
0.641158
2.966574
false
false
false
false
yarikoptic/pystatsmodels
statsmodels/tsa/tests/results/results_ar.py
35
9626
import numpy as np import os class ARLagResults(object): """ Results are from R vars::VARselect for sunspot data. Comands run were var_select <- VARselect(SUNACTIVITY, lag.max=16, type=c("const")) """ def __init__(self, type="const"): # order of results is AIC, HQ, SC, FPE if type == "const": ic = [6.311751824815273, 6.321813007357017, 6.336872456958734, 551.009492543133547, 5.647615009344886, 5.662706783157502, 5.685295957560077, 283.614444209634655, 5.634199640773091, 5.654322005856580, 5.684440905060013, 279.835333966272003, 5.639415797766900, 5.664568754121261, 5.702217378125553, 281.299267441683185, 5.646102475432464, 5.676286023057697, 5.721464371862848, 283.187210932784524, 5.628416873122441, 5.663631012018546, 5.716339085624555, 278.223839284844701, 5.584204185137150, 5.624448915304128, 5.684686713710994, 266.191975554941564, 5.541163244029505, 5.586438565467356, 5.654206088675081, 254.979353737235556, 5.483155367013447, 5.533461279722170, 5.608758527730753, 240.611088468544949, 5.489939895595428, 5.545276399575022, 5.628103372384465, 242.251199397394288, 5.496713895370946, 5.557080990621412, 5.647437688231713, 243.900349905069504, 5.503539311586831, 5.568936998108170, 5.666823420519329, 245.573823561989144, 5.510365149977393, 5.580793427769605, 5.686209574981622, 247.259396991133599, 5.513740912139918, 5.589199781203001, 5.702145653215877, 248.099655693709479, 5.515627471325321, 5.596116931659277, 5.716592528473011, 248.572915484827206, 5.515935627515806, 5.601455679120634, 5.729461000735226, 248.654927915301300] self.ic = np.asarray(ic).reshape(4,-1, order='F') class ARResultsOLS(object): """ Results of fitting an AR(9) model to the sunspot data. Results were taken from Stata using the var command. """ def __init__(self, constant=True): self.avobs = 300. if constant: self.params = [ 6.7430535917332, 1.1649421971129, -.40535742259304, -.16653934246587, .14980629416032, -.09462417064796, .00491001240749, .0504665930841, -.08635349190816, .25349103194757] # These are returned by stata VAR, using the (V)AR scale/sigma # we return the true OLS bse by default # the stata residuals can be achived by np.sqrt(np.diag(res1.cov_params())) self.bse_stata = [2.413485601, .0560359041, .0874490762, .0900894414, .0899348339, .0900100797, .0898385666, .0896997939, .0869773089, .0559505756] # The below are grom gretl's ARIMA command with conditional maxium likelihood self.bse_gretl = [2.45474, 0.0569939, 0.0889440, 0.0916295, 0.0914723, 0.0915488, 0.0913744, 0.0912332, 0.0884642, 0.0569071] self.rmse = 15.1279294937327 self.fpe = 236.4827257929261 self.llf = -1235.559128419549 #NOTE: we use a different definition of these ic than Stata # but our order selection results agree with R VARselect # close to Stata for Lutkepohl but we penalize the ic for the trend terms # self.bic = 8.427186938618863 # self.aic = 8.30372752279699 # self.hqic = 8.353136159250697 #NOTE: predictions were taken from gretl, but agree with Stata # test predict #TODO: remove one of the files filename = os.path.join(os.path.dirname(os.path.abspath(__file__)), "AROLSConstantPredict.csv") predictresults = np.loadtxt(filename) fv = predictresults[:300,0] pv = predictresults[300:,1] pv_lb = predictresults[300:,2] pv_ub = predictresults[300:,3] pv_se = predictresults[300:,4] del predictresults # cases - in sample predict # n = -1, start = 0 (fitted values) self.FVOLSnneg1start0 = fv # n=-1, start=9 self.FVOLSnneg1start9 = fv # n=-1, start=100 self.FVOLSnneg1start100 = fv[100-9:] # n = 200, start = 0 self.FVOLSn200start0 = fv[:192] # n = 200, start = 200 self.FVOLSn200start200 = np.hstack((fv[200-9:],pv[:101-9])) # n = 200, start = -109 use above self.FVOLSn200startneg109 = self.FVOLSn200start200 # n = 100, start = 325, post-sample forecasting self.FVOLSn100start325 = np.hstack((fv[-1],pv)) # n = 301, start = 9 self.FVOLSn301start9 = np.hstack((fv,pv[:2])) # n = 301, start = 0 self.FVOLSdefault = fv # n = 4, start = 312 self.FVOLSn4start312 = np.hstack((fv[-1],pv[:8])) # n = 15, start = 312 self.FVOLSn15start312 = np.hstack((fv[-1],pv[:19])) elif not constant: self.params = [1.19582389902985, -0.40591818219637, -0.15813796884843, 0.16620079925202, -0.08570200254617, 0.01876298948686, 0.06130211910707, -0.08461507700047, 0.27995084653313] self.bse_stata = [.055645055, .088579237, .0912031179, .0909032462, .0911161784, .0908611473, .0907743174, .0880993504, .0558560278] self.bse_gretl = [0.0564990, 0.0899386, 0.0926027, 0.0922983, 0.0925145, 0.0922555, 0.0921674, 0.0894513, 0.0567132] self.rmse = 15.29712618677774 self.sigma = 226.9820074869752 self.llf = -1239.41217278661 # See note above # self.bic = 8.433861292817106 # self.hqic = 8.367215591385756 # self.aic = 8.322747818577421 self.fpe = 241.0221316614273 filename = os.path.join(os.path.dirname(os.path.abspath(__file__)), "AROLSNoConstantPredict.csv") predictresults = np.loadtxt(filename) fv = predictresults[:300,0] pv = predictresults[300:,1] pv_lb = predictresults[300:,2] pv_ub = predictresults[300:,3] pv_se = predictresults[300:,4] del predictresults # cases - in sample predict # n = -1, start = 0 (fitted values) self.FVOLSnneg1start0 = fv # n=-1, start=9 self.FVOLSnneg1start9 = fv # n=-1, start=100 self.FVOLSnneg1start100 = fv[100-9:] # n = 200, start = 0 self.FVOLSn200start0 = fv[:192] # n = 200, start = 200 self.FVOLSn200start200 = np.hstack((fv[200-9:],pv[:101-9])) # n = 200, start = -109 use above self.FVOLSn200startneg109 = self.FVOLSn200start200 # n = 100, start = 325, post-sample forecasting self.FVOLSn100start325 = np.hstack((fv[-1],pv)) # n = 301, start = 9 self.FVOLSn301start9 = np.hstack((fv,pv[:2])) # n = 301, start = 0 self.FVOLSdefault = fv # n = 4, start = 312 self.FVOLSn4start312 = np.hstack((fv[-1],pv[:8])) # n = 15, start = 312 self.FVOLSn15start312 = np.hstack((fv[-1],pv[:19])) class ARResultsMLE(object): """ Results of fitting an AR(9) model to the sunspot data using exact MLE. Results were taken from gretl. """ def __init__(self, constant=True): self.avobs = 300 if constant: # NOTE: Stata's estimated parameters differ from gretl filename = os.path.join(os.path.dirname(os.path.abspath(__file__)), "ARMLEConstantPredict.csv") filename2 = os.path.join(os.path.dirname(os.path.abspath(__file__)), 'results_ar_forecast_mle_dynamic.csv') predictresults = np.loadtxt(filename, delimiter=",") year = predictresults[:,0] pv = predictresults[:,1] dynamicpv = np.genfromtxt(filename2, delimiter=",", skip_header=1) # cases - in sample predict # start = 0 (fitted values) self.FVMLEdefault = pv[:309] # start=9 self.FVMLEstart9end308 = pv[9:309] # start=100, end=309 self.FVMLEstart100end308 = pv[100:309] # start = 0, end self.FVMLEstart0end200 = pv[:201] # n = 200, start = 200 self.FVMLEstart200end334 = pv[200:] # start = 309, end=334 post-sample forecasting self.FVMLEstart308end334 = pv[308:] # end = 310, start = 9 self.FVMLEstart9end309 = pv[9:310] # end = 301, start = 0 self.FVMLEstart0end301 = pv[:302] # end = 312, start = 4 self.FVMLEstart4end312 = pv[4:313] # end = 7, start = 2 self.FVMLEstart2end7 = pv[2:8] self.fcdyn = dynamicpv[:,0] self.fcdyn2 = dynamicpv[:,1] self.fcdyn3 = dynamicpv[:,2] self.fcdyn4 = dynamicpv[:,3] else: pass
bsd-3-clause
08fae5bdf524f435559e28955c2a0122
43.155963
80
0.554436
3.159173
false
false
false
false
yarikoptic/pystatsmodels
statsmodels/regression/tests/results/results_regression.py
3
8038
""" Hard-coded results for test_regression """ ### REGRESSION MODEL RESULTS : OLS, GLS, WLS, AR### import numpy as np class Longley(object): ''' The results for the Longley dataset were obtained from NIST http://www.itl.nist.gov/div898/strd/general/dataarchive.html Other results were obtained from Stata ''' def __init__(self): self.params = ( 15.0618722713733, -0.358191792925910E-01, -2.02022980381683, -1.03322686717359, -0.511041056535807E-01, 1829.15146461355, -3482258.63459582) self.bse = (84.9149257747669, 0.334910077722432E-01, 0.488399681651699, 0.214274163161675, 0.226073200069370, 455.478499142212, 890420.383607373) self.scale = 92936.0061673238 self.rsquared = 0.995479004577296 self.rsquared_adj = 0.99246501 self.df_model = 6 self.df_resid = 9 self.ess = 184172401.944494 self.ssr = 836424.055505915 self.mse_model = 30695400.3240823 self.mse_resid = 92936.0061673238 self.mse_total = (self.ess + self.ssr) / (self.df_model + self.df_resid) self.fvalue = 330.285339234588 self.llf = -109.6174 self.aic = 233.2349 self.bic = 238.643 self.pvalues = np.array([ 0.86314083, 0.31268106, 0.00253509, 0.00094437, 0.8262118 , 0.0030368 , 0.0035604 ]) #pvalues from rmodelwrap self.resid = np.array((267.34003, -94.01394, 46.28717, -410.11462, 309.71459, -249.31122, -164.04896, -13.18036, 14.30477, 455.39409, -17.26893, -39.05504, -155.54997, -85.67131, 341.93151, -206.75783)) def conf_int(self): # a method to be consistent with sm return [(-177.0291,207.1524), (-.111581,.0399428),(-3.125065, -.9153928),(-1.517948,-.5485049),(-.5625173,.4603083), (798.7873,2859.515),(-5496529,-1467987)] HC0_se=(51.22035, 0.02458, 0.38324, 0.14625, 0.15821, 428.38438, 832212) HC1_se=(68.29380, 0.03277, 0.51099, 0.19499, 0.21094, 571.17917, 1109615) HC2_se=(67.49208, 0.03653, 0.55334, 0.20522, 0.22324, 617.59295, 1202370) HC3_se=(91.11939, 0.05562, 0.82213, 0.29879, 0.32491, 922.80784, 1799477) class LongleyGls(object): ''' The following results were obtained from running the test script with R. ''' def __init__(self): self.params = (6.73894832e-02, -4.74273904e-01, 9.48988771e+04) self.bse = (1.07033903e-02, 1.53385472e-01, 1.39447723e+04) self.llf = -121.4294962954981 self.fittedvalues = [59651.8255, 60860.1385, 60226.5336, 61467.1268, 63914.0846, 64561.9553, 64935.9028, 64249.1684, 66010.0426, 66834.7630, 67612.9309, 67018.8998, 68918.7758, 69310.1280, 69181.4207, 70598.8734] self.resid = [671.174465, 261.861502, -55.533603, -280.126803, -693.084618, -922.955349, 53.097212, -488.168351, 8.957367, 1022.236970, 556.069099, -505.899787, -263.775842, 253.871965, 149.579309, -47.873374] self.scale = 542.443043098**2 self.tvalues = [6.296088, -3.092039, 6.805337] self.pvalues = [2.761673e-05, 8.577197e-03, 1.252284e-05] self.bic = 253.118790021 self.aic = 250.858992591 class CCardWLS(object): def __init__(self): self.params = [-2.6941851611, 158.426977524, -7.24928987289, 60.4487736936, -114.10886935] self.bse = [3.807306306, 76.39115431, 9.724337321, 58.55088753, 139.6874965] #NOTE: we compute the scale differently than they do for analytic #weights self.scale = 189.0025755829012 ** 2 self.rsquared = .2549143871187359 self.rsquared_adj = .2104316639616448 self.df_model = 4 self.df_resid = 67 self.ess = 818838.8079468152 self.ssr = 2393372.229657007 self.mse_model = 818838.8079468152 / 4 self.mse_resid = 2393372.229657007 / 67 self.mse_total = (self.ess + self.ssr) / 71. self.fvalue = 5.730638077585917 self.llf = -476.9792946562806 self.aic = 963.95858931256 self.bic = 975.34191990764 # pvalues from R self.pvalues = [0.4816259843354, 0.0419360764848, 0.4585895209814, 0.3055904431658, 0.4168883565685] self.resid = [-286.964904785, -128.071563721, -405.860900879, -20.1363945007, -169.824432373, -82.6842575073, -283.314300537, -52.1719360352, 433.822174072, -190.607543945, -118.839683533, -133.97076416, -85.5728149414, 66.8180847168, -107.571769714, -149.883285522, -140.972610474, 75.9255981445, -135.979736328, -415.701263428, 130.080032349, 25.2313785553, 1042.14013672, -75.6622238159, 177.336639404, 315.870544434, -8.72801017761, 240.823760986, 54.6106033325, 65.6312484741, -40.9218444824, 24.6115856171, -131.971786499, 36.1587944031, 92.5052108765, -136.837036133, 242.73274231, -65.0315093994, 20.1536407471, -15.8874826431, 27.3513431549, -173.861785889, -113.121154785, -37.1303443909, 1510.31530762, 582.916931152, -17.8628063202, -132.77381897, -108.896934509, 12.4665794373, -122.014572144, -158.986968994, -175.798873901, 405.886505127, 99.3692703247, 85.3450698853, -179.15007019, -34.1245117188, -33.4909172058, -20.7287139893, -116.217689514, 53.8837738037, -52.1533050537, -100.632293701, 34.9342498779, -96.6685943604, -367.32925415, -40.1300048828, -72.8692245483, -60.8728256226, -35.9937324524, -222.944747925] def conf_int(self): # a method to be consistent with sm return [( -10.2936, 4.90523), ( 5.949595, 310.9044), (-26.65915, 12.16057), (-56.41929, 177.3168), (-392.9263, 164.7085)] class LongleyRTO(object): def __init__(self): # Regression Through the Origin model # from Stata, make sure you force double to replicate self.params = [-52.993523, .07107319, -.42346599, -.57256869, -.41420348, 48.417859] self.bse = [129.5447812, .0301663805, .4177363573, .2789908665, .3212848136, 17.68947719] self.scale = 475.1655079819532**2 self.rsquared = .9999670130705958 self.rsquared_adj = .9999472209129532 self.df_model = 6 self.df_resid = 10 self.ess = 68443718827.40025 self.ssr = 2257822.599757476 self.mse_model = 68443718827.40025 / 6 self.mse_resid = 2257822.599757476 / 10 self.mse_total = (self.ess + self.ssr) / 16. self.fvalue = 50523.39573737409 self.llf = -117.5615983965251 self.aic = 247.123196793 self.bic = 251.758729126 self.pvalues = [0.6911082828354, 0.0402241925699, 0.3346175334102, 0.0672506018552, 0.2263470345100, 0.0209367642585] self.resid = [279.902740479, -130.324661255, 90.7322845459, -401.312530518, -440.467681885, -543.54510498, 201.321121216, 215.908889771, 73.0936813354, 913.216918945, 424.824859619, -8.56475830078, -361.329742432, 27.3456058502, 151.28956604, -492.499359131] def conf_int(self): return [(-341.6373, 235.6502), ( .0038583, .1382881), (-1.354241, .5073086), (-1.194199, .0490617), (-1.130071, .3016637), ( 9.003248, 87.83247)]
bsd-3-clause
7148b2425383849980cc9356a254c87b
45.732558
80
0.573277
2.711876
false
false
false
false
yarikoptic/pystatsmodels
statsmodels/sandbox/examples/thirdparty/ex_ratereturn.py
4
4353
# -*- coding: utf-8 -*- """Playing with correlation of DJ-30 stock returns this uses pickled data that needs to be created with findow.py to see graphs, uncomment plt.show() Created on Sat Jan 30 16:30:18 2010 Author: josef-pktd """ import numpy as np import matplotlib.finance as fin import matplotlib.pyplot as plt import datetime as dt import pandas as pa import pickle import statsmodels.api as sm import statsmodels.sandbox as sb import statsmodels.sandbox.tools as sbtools from statsmodels.graphics.correlation import plot_corr, plot_corr_grid try: rrdm = pickle.load(file('dj30rr','rb')) except Exception: #blanket for any unpickling error print "Error with unpickling, a new pickle file can be created with findow_1" raise ticksym = rrdm.columns.tolist() rr = rrdm.values[1:400] rrcorr = np.corrcoef(rr, rowvar=0) plot_corr(rrcorr, xnames=ticksym) nvars = rrcorr.shape[0] plt.figure() plt.hist(rrcorr[np.triu_indices(nvars,1)]) plt.title('Correlation Coefficients') xreda, facta, evaa, evea = sbtools.pcasvd(rr) evallcs = (evaa).cumsum() print evallcs/evallcs[-1] xred, fact, eva, eve = sbtools.pcasvd(rr, keepdim=4) pcacorr = np.corrcoef(xred, rowvar=0) plot_corr(pcacorr, xnames=ticksym, title='Correlation PCA') resid = rr-xred residcorr = np.corrcoef(resid, rowvar=0) plot_corr(residcorr, xnames=ticksym, title='Correlation Residuals') plt.matshow(residcorr) plt.imshow(residcorr, cmap=plt.cm.jet, interpolation='nearest', extent=(0,30,0,30), vmin=-1.0, vmax=1.0) plt.colorbar() normcolor = (0,1) #False #True fig = plt.figure() ax = fig.add_subplot(2,2,1) plot_corr(rrcorr, xnames=ticksym, normcolor=normcolor, ax=ax) ax2 = fig.add_subplot(2,2,3) #pcacorr = np.corrcoef(xred, rowvar=0) plot_corr(pcacorr, xnames=ticksym, title='Correlation PCA', normcolor=normcolor, ax=ax2) ax3 = fig.add_subplot(2,2,4) plot_corr(residcorr, xnames=ticksym, title='Correlation Residuals', normcolor=normcolor, ax=ax3) import matplotlib as mpl images = [c for ax in fig.axes for c in ax.get_children() if isinstance(c, mpl.image.AxesImage)] print images print ax.get_children() #cax = fig.add_subplot(2,2,2) #[0.85, 0.1, 0.075, 0.8] fig. subplots_adjust(bottom=0.1, right=0.9, top=0.9) cax = fig.add_axes([0.9, 0.1, 0.025, 0.8]) fig.colorbar(images[0], cax=cax) fig.savefig('corrmatrixgrid.png', dpi=120) has_sklearn = True try: import sklearn except ImportError: has_sklearn = False print 'sklearn not available' def cov2corr(cov): std_ = np.sqrt(np.diag(cov)) corr = cov / np.outer(std_, std_) return corr if has_sklearn: from sklearn.covariance import LedoitWolf, OAS, MCD lw = LedoitWolf(store_precision=False) lw.fit(rr, assume_centered=False) cov_lw = lw.covariance_ corr_lw = cov2corr(cov_lw) oas = OAS(store_precision=False) oas.fit(rr, assume_centered=False) cov_oas = oas.covariance_ corr_oas = cov2corr(cov_oas) mcd = MCD()#.fit(rr, reweight=None) mcd.fit(rr, assume_centered=False) cov_mcd = mcd.covariance_ corr_mcd = cov2corr(cov_mcd) titles = ['raw correlation', 'lw', 'oas', 'mcd'] normcolor = None fig = plt.figure() for i, c in enumerate([rrcorr, corr_lw, corr_oas, corr_mcd]): #for i, c in enumerate([np.cov(rr, rowvar=0), cov_lw, cov_oas, cov_mcd]): ax = fig.add_subplot(2,2,i+1) plot_corr(c, xnames=None, title=titles[i], normcolor=normcolor, ax=ax) images = [c for ax in fig.axes for c in ax.get_children() if isinstance(c, mpl.image.AxesImage)] fig. subplots_adjust(bottom=0.1, right=0.9, top=0.9) cax = fig.add_axes([0.9, 0.1, 0.025, 0.8]) fig.colorbar(images[0], cax=cax) corrli = [rrcorr, corr_lw, corr_oas, corr_mcd, pcacorr] diffssq = np.array([[((ci-cj)**2).sum() for ci in corrli] for cj in corrli]) diffsabs = np.array([[np.max(np.abs(ci-cj)) for ci in corrli] for cj in corrli]) print diffssq print '\nmaxabs' print diffsabs fig.savefig('corrmatrix_sklearn.png', dpi=120) fig2 = plot_corr_grid(corrli+[residcorr], ncols=3, titles=titles+['pca', 'pca-residual'], xnames=[], ynames=[]) fig2.savefig('corrmatrix_sklearn_2.png', dpi=120) #plt.show() #plt.close('all')
bsd-3-clause
8d6faab6e3e932b3c98ed65d86eed32a
28.815068
100
0.668504
2.769084
false
false
false
false
yarikoptic/pystatsmodels
statsmodels/stats/power.py
3
47401
# -*- coding: utf-8 -*- #pylint: disable-msg=W0142 """Statistical power, solving for nobs, ... - trial version Created on Sat Jan 12 21:48:06 2013 Author: Josef Perktold Example roundtrip - root with respect to all variables calculated, desired nobs 33.367204205 33.367204205 effect 0.5 0.5 alpha 0.05 0.05 power 0.8 0.8 TODO: refactoring - rename beta -> power, beta (type 2 error is beta = 1-power) DONE - I think the current implementation can handle any kinds of extra keywords (except for maybe raising meaningful exceptions - streamline code, I think internally classes can be merged how to extend to k-sample tests? user interface for different tests that map to the same (internal) test class - sequence of arguments might be inconsistent, arg and/or kwds so python checks what's required and what can be None. - templating for docstrings ? """ import numpy as np from scipy import stats, optimize from statsmodels.tools.rootfinding import brentq_expanding def ttest_power(effect_size, nobs, alpha, df=None, alternative='two-sided'): '''Calculate power of a ttest ''' d = effect_size if df is None: df = nobs - 1 if alternative in ['two-sided', '2s']: alpha_ = alpha / 2. #no inplace changes, doesn't work elif alternative in ['smaller', 'larger']: alpha_ = alpha else: raise ValueError("alternative has to be 'two-sided', 'larger' " + "or 'smaller'") pow_ = 0 if alternative in ['two-sided', '2s', 'larger']: crit_upp = stats.t.isf(alpha_, df) #print crit_upp, df, d*np.sqrt(nobs) # use private methods, generic methods return nan with negative d if np.any(np.isnan(crit_upp)): # avoid endless loop, https://github.com/scipy/scipy/issues/2667 pow_ = np.nan else: pow_ = stats.nct._sf(crit_upp, df, d*np.sqrt(nobs)) if alternative in ['two-sided', '2s', 'smaller']: crit_low = stats.t.ppf(alpha_, df) #print crit_low, df, d*np.sqrt(nobs) if np.any(np.isnan(crit_low)): pow_ = np.nan else: pow_ += stats.nct._cdf(crit_low, df, d*np.sqrt(nobs)) return pow_ def normal_power(effect_size, nobs, alpha, alternative='two-sided', sigma=1.): '''Calculate power of a normal distributed test statistic ''' d = effect_size if alternative in ['two-sided', '2s']: alpha_ = alpha / 2. #no inplace changes, doesn't work elif alternative in ['smaller', 'larger']: alpha_ = alpha else: raise ValueError("alternative has to be 'two-sided', 'larger' " + "or 'smaller'") pow_ = 0 if alternative in ['two-sided', '2s', 'larger']: crit = stats.norm.isf(alpha_) pow_ = stats.norm.sf(crit - d*np.sqrt(nobs)/sigma) if alternative in ['two-sided', '2s', 'smaller']: crit = stats.norm.ppf(alpha_) pow_ += stats.norm.cdf(crit - d*np.sqrt(nobs)/sigma) return pow_ def ftest_anova_power(effect_size, nobs, alpha, k_groups=2, df=None): '''power for ftest for one way anova with k equal sized groups nobs total sample size, sum over all groups should be general nobs observations, k_groups restrictions ??? ''' df_num = nobs - k_groups df_denom = k_groups - 1 crit = stats.f.isf(alpha, df_denom, df_num) pow_ = stats.ncf.sf(crit, df_denom, df_num, effect_size**2 * nobs) return pow_#, crit def ftest_power(effect_size, df_num, df_denom, alpha, ncc=1): '''Calculate the power of a F-test. Parameters ---------- effect_size : float standardized effect size, mean divided by the standard deviation. effect size has to be positive. df_num : int or float numerator degrees of freedom. df_denom : int or float denominator degrees of freedom. alpha : float in interval (0,1) significance level, e.g. 0.05, is the probability of a type I error, that is wrong rejections if the Null Hypothesis is true. ncc : int degrees of freedom correction for non-centrality parameter. see Notes Returns ------- power : float Power of the test, e.g. 0.8, is one minus the probability of a type II error. Power is the probability that the test correctly rejects the Null Hypothesis if the Alternative Hypothesis is true. Notes ----- sample size is given implicitly by df_num set ncc=0 to match t-test, or f-test in LikelihoodModelResults. ncc=1 matches the non-centrality parameter in R::pwr::pwr.f2.test ftest_power with ncc=0 should also be correct for f_test in regression models, with df_num and d_denom as defined there. (not verified yet) ''' nc = effect_size**2 * (df_denom + df_num + ncc) crit = stats.f.isf(alpha, df_denom, df_num) pow_ = stats.ncf.sf(crit, df_denom, df_num, nc) return pow_ #, crit, nc #class based implementation #-------------------------- class Power(object): '''Statistical Power calculations, Base Class so far this could all be class methods ''' def __init__(self, **kwds): self.__dict__.update(kwds) # used only for instance level start values self.start_ttp = dict(effect_size=0.01, nobs=10., alpha=0.15, power=0.6, nobs1=10., ratio=1, df_num=10, df_denom=3 # for FTestPower ) # TODO: nobs1 and ratio are for ttest_ind, # need start_ttp for each test/class separately, # possible rootfinding problem for effect_size, starting small seems to # work from collections import defaultdict self.start_bqexp = defaultdict(dict) for key in ['nobs', 'nobs1', 'df_num', 'df_denom']: self.start_bqexp[key] = dict(low=2., start_upp=50.) for key in ['df_denom']: self.start_bqexp[key] = dict(low=1., start_upp=50.) for key in ['ratio']: self.start_bqexp[key] = dict(low=1e-8, start_upp=2) for key in ['alpha']: self.start_bqexp[key] = dict(low=1e-12, upp=1 - 1e-12) def power(self, *args, **kwds): raise NotImplementedError def _power_identity(self, *args, **kwds): power_ = kwds.pop('power') return self.power(*args, **kwds) - power_ def solve_power(self, **kwds): '''solve for any one of the parameters of a t-test for t-test the keywords are: effect_size, nobs, alpha, power exactly one needs to be ``None``, all others need numeric values *attaches* cache_fit_res : list Cache of the result of the root finding procedure for the latest call to ``solve_power``, mainly for debugging purposes. The first element is the success indicator, one if successful. The remaining elements contain the return information of the up to three solvers that have been tried. ''' #TODO: maybe use explicit kwds, # nicer but requires inspect? and not generic across tests # I'm duplicating this in the subclass to get informative docstring key = [k for k,v in kwds.iteritems() if v is None] #print kwds, key; if len(key) != 1: raise ValueError('need exactly one keyword that is None') key = key[0] if key == 'power': del kwds['power'] return self.power(**kwds) self._counter = 0 def func(x): kwds[key] = x fval = self._power_identity(**kwds) self._counter += 1 #print self._counter, if self._counter > 500: raise RuntimeError('possible endless loop (500 NaNs)') if np.isnan(fval): return np.inf else: return fval #TODO: I'm using the following so I get a warning when start_ttp is not defined try: start_value = self.start_ttp[key] except KeyError: start_value = 0.9 print 'Warning: using default start_value for', key fit_kwds = self.start_bqexp[key] fit_res = [] #print vars() try: val, res = brentq_expanding(func, full_output=True, **fit_kwds) failed = False fit_res.append(res) except ValueError: failed = True fit_res.append(None) success = None if (not failed) and res.converged: success = 1 else: # try backup #TODO: check more cases to make this robust val, infodict, ier, msg = optimize.fsolve(func, start_value, full_output=True) #scalar #val = optimize.newton(func, start_value) #scalar fval = infodict['fvec'] fit_res.append(infodict) if ier == 1 and np.abs(fval) < 1e-4 : success = 1 else: #print infodict if key in ['alpha', 'power', 'effect_size']: val, r = optimize.brentq(func, 1e-8, 1-1e-8, full_output=True) #scalar success = 1 if r.converged else 0 fit_res.append(r) else: success = 0 if not success == 1: import warnings from statsmodels.tools.sm_exceptions import ConvergenceWarning warnings.warn('finding solution failed', ConvergenceWarning) #attach fit_res, for reading only, should be needed only for debugging fit_res.insert(0, success) self.cache_fit_res = fit_res return val def plot_power(self, dep_var='nobs', nobs=None, effect_size=None, alpha=0.05, ax=None, title=None, plt_kwds=None, **kwds): '''plot power with number of observations or effect size on x-axis Parameters ---------- dep_var : string in ['nobs', 'effect_size', 'alpha'] This specifies which variable is used for the horizontal axis. If dep_var='nobs' (default), then one curve is created for each value of ``effect_size``. If dep_var='effect_size' or alpha, then one curve is created for each value of ``nobs``. nobs : scalar or array_like specifies the values of the number of observations in the plot effect_size : scalar or array_like specifies the values of the effect_size in the plot alpha : float or array_like The significance level (type I error) used in the power calculation. Can only be more than a scalar, if ``dep_var='alpha'`` ax : None or axis instance If ax is None, than a matplotlib figure is created. If ax is a matplotlib axis instance, then it is reused, and the plot elements are created with it. title : string title for the axis. Use an empty string, ``''``, to avoid a title. plt_kwds : None or dict not used yet kwds : optional keywords for power function These remaining keyword arguments are used as arguments to the power function. Many power function support ``alternative`` as a keyword argument, two-sample test support ``ratio``. Returns ------- fig : matplotlib figure instance Notes ----- This works only for classes where the ``power`` method has ``effect_size``, ``nobs`` and ``alpha`` as the first three arguments. If the second argument is ``nobs1``, then the number of observations in the plot are those for the first sample. TODO: fix this for FTestPower and GofChisquarePower TODO: maybe add line variable, if we want more than nobs and effectsize ''' #if pwr_kwds is None: # pwr_kwds = {} from statsmodels.graphics import utils from statsmodels.graphics.plottools import rainbow fig, ax = utils.create_mpl_ax(ax) import matplotlib.pyplot as plt colormap = plt.cm.Dark2 #pylint: disable-msg=E1101 plt_alpha = 1 #0.75 lw = 2 if dep_var == 'nobs': colors = rainbow(len(effect_size)) colors = [colormap(i) for i in np.linspace(0, 0.9, len(effect_size))] for ii, es in enumerate(effect_size): power = self.power(es, nobs, alpha, **kwds) ax.plot(nobs, power, lw=lw, alpha=plt_alpha, color=colors[ii], label='es=%4.2F' % es) xlabel = 'Number of Observations' elif dep_var in ['effect size', 'effect_size', 'es']: colors = rainbow(len(nobs)) colors = [colormap(i) for i in np.linspace(0, 0.9, len(nobs))] for ii, n in enumerate(nobs): power = self.power(effect_size, n, alpha, **kwds) ax.plot(effect_size, power, lw=lw, alpha=plt_alpha, color=colors[ii], label='N=%4.2F' % n) xlabel = 'Effect Size' elif dep_var in ['alpha']: # experimental nobs as defining separate lines colors = rainbow(len(nobs)) for ii, n in enumerate(nobs): power = self.power(effect_size, n, alpha, **kwds) ax.plot(alpha, power, lw=lw, alpha=plt_alpha, color=colors[ii], label='N=%4.2F' % n) xlabel = 'alpha' else: raise ValueError('depvar not implemented') if title is None: title = 'Power of Test' ax.set_xlabel(xlabel) ax.set_title(title) ax.legend(loc='lower right') return fig class TTestPower(Power): '''Statistical Power calculations for one sample or paired sample t-test ''' def power(self, effect_size, nobs, alpha, df=None, alternative='two-sided'): '''Calculate the power of a t-test for one sample or paired samples. Parameters ---------- effect_size : float standardized effect size, mean divided by the standard deviation. effect size has to be positive. nobs : int or float sample size, number of observations. alpha : float in interval (0,1) significance level, e.g. 0.05, is the probability of a type I error, that is wrong rejections if the Null Hypothesis is true. df : int or float degrees of freedom. By default this is None, and the df from the one sample or paired ttest is used, ``df = nobs1 - 1`` alternative : string, 'two-sided' (default), 'larger', 'smaller' extra argument to choose whether the power is calculated for a two-sided (default) or one sided test. The one-sided test can be either 'larger', 'smaller'. . Returns ------- power : float Power of the test, e.g. 0.8, is one minus the probability of a type II error. Power is the probability that the test correctly rejects the Null Hypothesis if the Alternative Hypothesis is true. ''' # for debugging #print 'calling ttest power with', (effect_size, nobs, alpha, df, alternative) return ttest_power(effect_size, nobs, alpha, df=df, alternative=alternative) #method is only added to have explicit keywords and docstring def solve_power(self, effect_size=None, nobs=None, alpha=None, power=None, alternative='two-sided'): '''solve for any one parameter of the power of a one sample t-test for the one sample t-test the keywords are: effect_size, nobs, alpha, power Exactly one needs to be ``None``, all others need numeric values. This test can also be used for a paired t-test, where effect size is defined in terms of the mean difference, and nobs is the number of pairs. Parameters ---------- effect_size : float standardized effect size, mean divided by the standard deviation. effect size has to be positive. nobs : int or float sample size, number of observations. alpha : float in interval (0,1) significance level, e.g. 0.05, is the probability of a type I error, that is wrong rejections if the Null Hypothesis is true. power : float in interval (0,1) power of the test, e.g. 0.8, is one minus the probability of a type II error. Power is the probability that the test correctly rejects the Null Hypothesis if the Alternative Hypothesis is true. alternative : string, 'two-sided' (default) or 'one-sided' extra argument to choose whether the power is calculated for a two-sided (default) or one sided test. 'one-sided' assumes we are in the relevant tail. Returns ------- value : float The value of the parameter that was set to None in the call. The value solves the power equation given the remaining parameters. *attaches* cache_fit_res : list Cache of the result of the root finding procedure for the latest call to ``solve_power``, mainly for debugging purposes. The first element is the success indicator, one if successful. The remaining elements contain the return information of the up to three solvers that have been tried. Notes ----- The function uses scipy.optimize for finding the value that satisfies the power equation. It first uses ``brentq`` with a prior search for bounds. If this fails to find a root, ``fsolve`` is used. If ``fsolve`` also fails, then, for ``alpha``, ``power`` and ``effect_size``, ``brentq`` with fixed bounds is used. However, there can still be cases where this fails. ''' # for debugging #print 'calling ttest solve with', (effect_size, nobs, alpha, power, alternative) return super(TTestPower, self).solve_power(effect_size=effect_size, nobs=nobs, alpha=alpha, power=power, alternative=alternative) class TTestIndPower(Power): '''Statistical Power calculations for t-test for two independent sample currently only uses pooled variance ''' def power(self, effect_size, nobs1, alpha, ratio=1, df=None, alternative='two-sided'): '''Calculate the power of a t-test for two independent sample Parameters ---------- effect_size : float standardized effect size, difference between the two means divided by the standard deviation. `effect_size` has to be positive. nobs1 : int or float number of observations of sample 1. The number of observations of sample two is ratio times the size of sample 1, i.e. ``nobs2 = nobs1 * ratio`` alpha : float in interval (0,1) significance level, e.g. 0.05, is the probability of a type I error, that is wrong rejections if the Null Hypothesis is true. ratio : float ratio of the number of observations in sample 2 relative to sample 1. see description of nobs1 The default for ratio is 1; to solve for ratio given the other arguments, it has to be explicitly set to None. df : int or float degrees of freedom. By default this is None, and the df from the ttest with pooled variance is used, ``df = (nobs1 - 1 + nobs2 - 1)`` alternative : string, 'two-sided' (default), 'larger', 'smaller' extra argument to choose whether the power is calculated for a two-sided (default) or one sided test. The one-sided test can be either 'larger', 'smaller'. Returns ------- power : float Power of the test, e.g. 0.8, is one minus the probability of a type II error. Power is the probability that the test correctly rejects the Null Hypothesis if the Alternative Hypothesis is true. ''' nobs2 = nobs1*ratio #pooled variance if df is None: df = (nobs1 - 1 + nobs2 - 1) nobs = 1./ (1. / nobs1 + 1. / nobs2) #print 'calling ttest power with', (effect_size, nobs, alpha, df, alternative) return ttest_power(effect_size, nobs, alpha, df=df, alternative=alternative) #method is only added to have explicit keywords and docstring def solve_power(self, effect_size=None, nobs1=None, alpha=None, power=None, ratio=1., alternative='two-sided'): '''solve for any one parameter of the power of a two sample t-test for t-test the keywords are: effect_size, nobs1, alpha, power, ratio exactly one needs to be ``None``, all others need numeric values Parameters ---------- effect_size : float standardized effect size, difference between the two means divided by the standard deviation. `effect_size` has to be positive. nobs1 : int or float number of observations of sample 1. The number of observations of sample two is ratio times the size of sample 1, i.e. ``nobs2 = nobs1 * ratio`` alpha : float in interval (0,1) significance level, e.g. 0.05, is the probability of a type I error, that is wrong rejections if the Null Hypothesis is true. power : float in interval (0,1) power of the test, e.g. 0.8, is one minus the probability of a type II error. Power is the probability that the test correctly rejects the Null Hypothesis if the Alternative Hypothesis is true. ratio : float ratio of the number of observations in sample 2 relative to sample 1. see description of nobs1 The default for ratio is 1; to solve for ratio given the other arguments it has to be explicitly set to None. alternative : string, 'two-sided' (default), 'larger', 'smaller' extra argument to choose whether the power is calculated for a two-sided (default) or one sided test. The one-sided test can be either 'larger', 'smaller'. Returns ------- value : float The value of the parameter that was set to None in the call. The value solves the power equation given the remaining parameters. Notes ----- The function uses scipy.optimize for finding the value that satisfies the power equation. It first uses ``brentq`` with a prior search for bounds. If this fails to find a root, ``fsolve`` is used. If ``fsolve`` also fails, then, for ``alpha``, ``power`` and ``effect_size``, ``brentq`` with fixed bounds is used. However, there can still be cases where this fails. ''' return super(TTestIndPower, self).solve_power(effect_size=effect_size, nobs1=nobs1, alpha=alpha, power=power, ratio=ratio, alternative=alternative) class NormalIndPower(Power): '''Statistical Power calculations for z-test for two independent samples. currently only uses pooled variance ''' def __init__(self, ddof=0, **kwds): self.ddof = ddof super(NormalIndPower, self).__init__(**kwds) def power(self, effect_size, nobs1, alpha, ratio=1, alternative='two-sided'): '''Calculate the power of a t-test for two independent sample Parameters ---------- effect_size : float standardized effect size, difference between the two means divided by the standard deviation. effect size has to be positive. nobs1 : int or float number of observations of sample 1. The number of observations of sample two is ratio times the size of sample 1, i.e. ``nobs2 = nobs1 * ratio`` ``ratio`` can be set to zero in order to get the power for a one sample test. alpha : float in interval (0,1) significance level, e.g. 0.05, is the probability of a type I error, that is wrong rejections if the Null Hypothesis is true. ratio : float ratio of the number of observations in sample 2 relative to sample 1. see description of nobs1 The default for ratio is 1; to solve for ratio given the other arguments it has to be explicitly set to None. alternative : string, 'two-sided' (default), 'larger', 'smaller' extra argument to choose whether the power is calculated for a two-sided (default) or one sided test. The one-sided test can be either 'larger', 'smaller'. Returns ------- power : float Power of the test, e.g. 0.8, is one minus the probability of a type II error. Power is the probability that the test correctly rejects the Null Hypothesis if the Alternative Hypothesis is true. ''' ddof = self.ddof # for correlation, ddof=3 # get effective nobs, factor for std of test statistic if ratio > 0: nobs2 = nobs1*ratio #equivalent to nobs = n1*n2/(n1+n2)=n1*ratio/(1+ratio) nobs = 1./ (1. / (nobs1 - ddof) + 1. / (nobs2 - ddof)) else: nobs = nobs1 - ddof return normal_power(effect_size, nobs, alpha, alternative=alternative) #method is only added to have explicit keywords and docstring def solve_power(self, effect_size=None, nobs1=None, alpha=None, power=None, ratio=1., alternative='two-sided'): '''solve for any one parameter of the power of a two sample z-test for z-test the keywords are: effect_size, nobs1, alpha, power, ratio exactly one needs to be ``None``, all others need numeric values Parameters ---------- effect_size : float standardized effect size, difference between the two means divided by the standard deviation. If ratio=0, then this is the standardized mean in the one sample test. nobs1 : int or float number of observations of sample 1. The number of observations of sample two is ratio times the size of sample 1, i.e. ``nobs2 = nobs1 * ratio`` ``ratio`` can be set to zero in order to get the power for a one sample test. alpha : float in interval (0,1) significance level, e.g. 0.05, is the probability of a type I error, that is wrong rejections if the Null Hypothesis is true. power : float in interval (0,1) power of the test, e.g. 0.8, is one minus the probability of a type II error. Power is the probability that the test correctly rejects the Null Hypothesis if the Alternative Hypothesis is true. ratio : float ratio of the number of observations in sample 2 relative to sample 1. see description of nobs1 The default for ratio is 1; to solve for ration given the other arguments it has to be explicitly set to None. alternative : string, 'two-sided' (default), 'larger', 'smaller' extra argument to choose whether the power is calculated for a two-sided (default) or one sided test. The one-sided test can be either 'larger', 'smaller'. Returns ------- value : float The value of the parameter that was set to None in the call. The value solves the power equation given the remaining parameters. Notes ----- The function uses scipy.optimize for finding the value that satisfies the power equation. It first uses ``brentq`` with a prior search for bounds. If this fails to find a root, ``fsolve`` is used. If ``fsolve`` also fails, then, for ``alpha``, ``power`` and ``effect_size``, ``brentq`` with fixed bounds is used. However, there can still be cases where this fails. ''' return super(NormalIndPower, self).solve_power(effect_size=effect_size, nobs1=nobs1, alpha=alpha, power=power, ratio=ratio, alternative=alternative) class FTestPower(Power): '''Statistical Power calculations for generic F-test ''' def power(self, effect_size, df_num, df_denom, alpha, ncc=1): '''Calculate the power of a F-test. Parameters ---------- effect_size : float standardized effect size, mean divided by the standard deviation. effect size has to be positive. df_num : int or float numerator degrees of freedom. df_denom : int or float denominator degrees of freedom. alpha : float in interval (0,1) significance level, e.g. 0.05, is the probability of a type I error, that is wrong rejections if the Null Hypothesis is true. ncc : int degrees of freedom correction for non-centrality parameter. see Notes Returns ------- power : float Power of the test, e.g. 0.8, is one minus the probability of a type II error. Power is the probability that the test correctly rejects the Null Hypothesis if the Alternative Hypothesis is true. Notes ----- sample size is given implicitly by df_num set ncc=0 to match t-test, or f-test in LikelihoodModelResults. ncc=1 matches the non-centrality parameter in R::pwr::pwr.f2.test ftest_power with ncc=0 should also be correct for f_test in regression models, with df_num and d_denom as defined there. (not verified yet) ''' pow_ = ftest_power(effect_size, df_num, df_denom, alpha, ncc=ncc) #print effect_size, df_num, df_denom, alpha, pow_ return pow_ #method is only added to have explicit keywords and docstring def solve_power(self, effect_size=None, df_num=None, df_denom=None, nobs=None, alpha=None, power=None, ncc=1): '''solve for any one parameter of the power of a F-test for the one sample F-test the keywords are: effect_size, df_num, df_denom, alpha, power Exactly one needs to be ``None``, all others need numeric values. Parameters ---------- effect_size : float standardized effect size, mean divided by the standard deviation. effect size has to be positive. nobs : int or float sample size, number of observations. alpha : float in interval (0,1) significance level, e.g. 0.05, is the probability of a type I error, that is wrong rejections if the Null Hypothesis is true. power : float in interval (0,1) power of the test, e.g. 0.8, is one minus the probability of a type II error. Power is the probability that the test correctly rejects the Null Hypothesis if the Alternative Hypothesis is true. alternative : string, 'two-sided' (default) or 'one-sided' extra argument to choose whether the power is calculated for a two-sided (default) or one sided test. 'one-sided' assumes we are in the relevant tail. Returns ------- value : float The value of the parameter that was set to None in the call. The value solves the power equation given the remainding parameters. Notes ----- The function uses scipy.optimize for finding the value that satisfies the power equation. It first uses ``brentq`` with a prior search for bounds. If this fails to find a root, ``fsolve`` is used. If ``fsolve`` also fails, then, for ``alpha``, ``power`` and ``effect_size``, ``brentq`` with fixed bounds is used. However, there can still be cases where this fails. ''' return super(FTestPower, self).solve_power(effect_size=effect_size, df_num=df_num, df_denom=df_denom, alpha=alpha, power=power, ncc=ncc) class FTestAnovaPower(Power): '''Statistical Power calculations F-test for one factor balanced ANOVA ''' def power(self, effect_size, nobs, alpha, k_groups=2): '''Calculate the power of a F-test for one factor ANOVA. Parameters ---------- effect_size : float standardized effect size, mean divided by the standard deviation. effect size has to be positive. nobs : int or float sample size, number of observations. alpha : float in interval (0,1) significance level, e.g. 0.05, is the probability of a type I error, that is wrong rejections if the Null Hypothesis is true. k_groups : int or float number of groups in the ANOVA or k-sample comparison. Default is 2. Returns ------- power : float Power of the test, e.g. 0.8, is one minus the probability of a type II error. Power is the probability that the test correctly rejects the Null Hypothesis if the Alternative Hypothesis is true. ''' return ftest_anova_power(effect_size, nobs, alpha, k_groups=k_groups) #method is only added to have explicit keywords and docstring def solve_power(self, effect_size=None, nobs=None, alpha=None, power=None, k_groups=2): '''solve for any one parameter of the power of a F-test for the one sample F-test the keywords are: effect_size, nobs, alpha, power Exactly one needs to be ``None``, all others need numeric values. Parameters ---------- effect_size : float standardized effect size, mean divided by the standard deviation. effect size has to be positive. nobs : int or float sample size, number of observations. alpha : float in interval (0,1) significance level, e.g. 0.05, is the probability of a type I error, that is wrong rejections if the Null Hypothesis is true. power : float in interval (0,1) power of the test, e.g. 0.8, is one minus the probability of a type II error. Power is the probability that the test correctly rejects the Null Hypothesis if the Alternative Hypothesis is true. Returns ------- value : float The value of the parameter that was set to None in the call. The value solves the power equation given the remainding parameters. Notes ----- The function uses scipy.optimize for finding the value that satisfies the power equation. It first uses ``brentq`` with a prior search for bounds. If this fails to find a root, ``fsolve`` is used. If ``fsolve`` also fails, then, for ``alpha``, ``power`` and ``effect_size``, ``brentq`` with fixed bounds is used. However, there can still be cases where this fails. ''' # update start values for root finding if not k_groups is None: self.start_ttp['nobs'] = k_groups * 10 self.start_bqexp['nobs'] = dict(low=k_groups * 2, start_upp=k_groups * 10) # first attempt at special casing if effect_size is None: return self._solve_effect_size(effect_size=effect_size, nobs=nobs, alpha=alpha, k_groups=k_groups, power=power) return super(FTestAnovaPower, self).solve_power(effect_size=effect_size, nobs=nobs, alpha=alpha, k_groups=k_groups, power=power) def _solve_effect_size(self, effect_size=None, nobs=None, alpha=None, power=None, k_groups=2): '''experimental, test failure in solve_power for effect_size ''' def func(x): effect_size = x return self._power_identity(effect_size=effect_size, nobs=nobs, alpha=alpha, k_groups=k_groups, power=power) val, r = optimize.brentq(func, 1e-8, 1-1e-8, full_output=True) if not r.converged: print r return val class GofChisquarePower(Power): '''Statistical Power calculations for one sample chisquare test ''' def power(self, effect_size, nobs, alpha, n_bins, ddof=0): #alternative='two-sided'): '''Calculate the power of a chisquare test for one sample Only two-sided alternative is implemented Parameters ---------- effect_size : float standardized effect size, according to Cohen's definition. see :func:`statsmodels.stats.gof.chisquare_effectsize` nobs : int or float sample size, number of observations. alpha : float in interval (0,1) significance level, e.g. 0.05, is the probability of a type I error, that is wrong rejections if the Null Hypothesis is true. n_bins : int number of bins or cells in the distribution. Returns ------- power : float Power of the test, e.g. 0.8, is one minus the probability of a type II error. Power is the probability that the test correctly rejects the Null Hypothesis if the Alternative Hypothesis is true. ''' from statsmodels.stats.gof import chisquare_power return chisquare_power(effect_size, nobs, n_bins, alpha, ddof=0) #method is only added to have explicit keywords and docstring def solve_power(self, effect_size=None, nobs=None, alpha=None, power=None, n_bins=2): '''solve for any one parameter of the power of a one sample chisquare-test for the one sample chisquare-test the keywords are: effect_size, nobs, alpha, power Exactly one needs to be ``None``, all others need numeric values. n_bins needs to be defined, a default=2 is used. Parameters ---------- effect_size : float standardized effect size, according to Cohen's definition. see :func:`statsmodels.stats.gof.chisquare_effectsize` nobs : int or float sample size, number of observations. alpha : float in interval (0,1) significance level, e.g. 0.05, is the probability of a type I error, that is wrong rejections if the Null Hypothesis is true. power : float in interval (0,1) power of the test, e.g. 0.8, is one minus the probability of a type II error. Power is the probability that the test correctly rejects the Null Hypothesis if the Alternative Hypothesis is true. n_bins : int number of bins or cells in the distribution Returns ------- value : float The value of the parameter that was set to None in the call. The value solves the power equation given the remaining parameters. Notes ----- The function uses scipy.optimize for finding the value that satisfies the power equation. It first uses ``brentq`` with a prior search for bounds. If this fails to find a root, ``fsolve`` is used. If ``fsolve`` also fails, then, for ``alpha``, ``power`` and ``effect_size``, ``brentq`` with fixed bounds is used. However, there can still be cases where this fails. ''' return super(GofChisquarePower, self).solve_power(effect_size=effect_size, nobs=nobs, n_bins=n_bins, alpha=alpha, power=power) class _GofChisquareIndPower(Power): '''Statistical Power calculations for chisquare goodness-of-fit test TODO: this is not working yet for 2sample case need two nobs in function no one-sided chisquare test, is there one? use normal distribution? -> drop one-sided options? ''' def power(self, effect_size, nobs1, alpha, ratio=1, alternative='two-sided'): '''Calculate the power of a chisquare for two independent sample Parameters ---------- effect_size : float standardize effect size, difference between the two means divided by the standard deviation. effect size has to be positive. nobs1 : int or float number of observations of sample 1. The number of observations of sample two is ratio times the size of sample 1, i.e. ``nobs2 = nobs1 * ratio`` alpha : float in interval (0,1) significance level, e.g. 0.05, is the probability of a type I error, that is wrong rejections if the Null Hypothesis is true. ratio : float ratio of the number of observations in sample 2 relative to sample 1. see description of nobs1 The default for ratio is 1; to solve for ration given the other arguments it has to be explicitely set to None. alternative : string, 'two-sided' (default) or 'one-sided' extra argument to choose whether the power is calculated for a two-sided (default) or one sided test. 'one-sided' assumes we are in the relevant tail. Returns ------- power : float Power of the test, e.g. 0.8, is one minus the probability of a type II error. Power is the probability that the test correctly rejects the Null Hypothesis if the Alternative Hypothesis is true. ''' from statsmodels.stats.gof import chisquare_power nobs2 = nobs1*ratio #equivalent to nobs = n1*n2/(n1+n2)=n1*ratio/(1+ratio) nobs = 1./ (1. / nobs1 + 1. / nobs2) return chisquare_power(effect_size, nobs, alpha) #method is only added to have explicit keywords and docstring def solve_power(self, effect_size=None, nobs1=None, alpha=None, power=None, ratio=1., alternative='two-sided'): '''solve for any one parameter of the power of a two sample z-test for z-test the keywords are: effect_size, nobs1, alpha, power, ratio exactly one needs to be ``None``, all others need numeric values Parameters ---------- effect_size : float standardize effect size, difference between the two means divided by the standard deviation. nobs1 : int or float number of observations of sample 1. The number of observations of sample two is ratio times the size of sample 1, i.e. ``nobs2 = nobs1 * ratio`` alpha : float in interval (0,1) significance level, e.g. 0.05, is the probability of a type I error, that is wrong rejections if the Null Hypothesis is true. power : float in interval (0,1) power of the test, e.g. 0.8, is one minus the probability of a type II error. Power is the probability that the test correctly rejects the Null Hypothesis if the Alternative Hypothesis is true. ratio : float ratio of the number of observations in sample 2 relative to sample 1. see description of nobs1 The default for ratio is 1; to solve for ration given the other arguments it has to be explicitely set to None. alternative : string, 'two-sided' (default) or 'one-sided' extra argument to choose whether the power is calculated for a two-sided (default) or one sided test. 'one-sided' assumes we are in the relevant tail. Returns ------- value : float The value of the parameter that was set to None in the call. The value solves the power equation given the remainding parameters. Notes ----- The function uses scipy.optimize for finding the value that satisfies the power equation. It first uses ``brentq`` with a prior search for bounds. If this fails to find a root, ``fsolve`` is used. If ``fsolve`` also fails, then, for ``alpha``, ``power`` and ``effect_size``, ``brentq`` with fixed bounds is used. However, there can still be cases where this fails. ''' return super(_GofChisquareIndPower, self).solve_power(effect_size=effect_size, nobs1=nobs1, alpha=alpha, power=power, ratio=ratio, alternative=alternative) #shortcut functions tt_solve_power = TTestPower().solve_power tt_ind_solve_power = TTestIndPower().solve_power zt_ind_solve_power = NormalIndPower().solve_power
bsd-3-clause
ba0eb96387632c056cf7ceb9440d1fb3
40.799824
89
0.577203
4.337573
false
true
false
false
yarikoptic/pystatsmodels
statsmodels/stats/tests/test_corrpsd.py
3
7727
# -*- coding: utf-8 -*- """Tests for findind a positive semi-definite correlation of covariance matrix Created on Mon May 27 12:07:02 2013 Author: Josef Perktold """ import numpy as np from numpy.testing import assert_almost_equal, assert_allclose from statsmodels.stats.correlation_tools import ( corr_nearest, corr_clipped, cov_nearest) def norm_f(x, y): '''Frobenious norm (squared sum) of difference between two arrays ''' d = ((x - y)**2).sum() return np.sqrt(d) class Holder(object): pass # R library Matrix results cov1_r = Holder() #> nc <- nearPD(pr, conv.tol = 1e-7, keepDiag = TRUE, doDykstra =FALSE, corr=TRUE) #> cat_items(nc, prefix="cov1_r.") cov1_r.mat = '''<S4 object of class structure("dpoMatrix", package = "Matrix")>''' cov1_r.eigenvalues = np.array([ 4.197315628646795, 0.7540460243978023, 0.5077608149667492, 0.3801267599652769, 0.1607508970775889, 4.197315628646795e-08 ]) cov1_r.corr = '''TRUE''' cov1_r.normF = 0.0743805226512533 cov1_r.iterations = 11 cov1_r.rel_tol = 8.288594638441735e-08 cov1_r.converged = '''TRUE''' #> mkarray2(as.matrix(nc$mat), name="cov1_r.mat") cov1_r.mat = np.array([ 1, 0.487968018215892, 0.642651880010906, 0.4906386709070835, 0.6440990530811909, 0.8087111845493985, 0.487968018215892, 1, 0.5141147294352735, 0.2506688108312097, 0.672351311297074, 0.725832055882795, 0.642651880010906, 0.5141147294352735, 1, 0.596827778712154, 0.5821917790519067, 0.7449631633814129, 0.4906386709070835, 0.2506688108312097, 0.596827778712154, 1, 0.729882058012399, 0.772150225146826, 0.6440990530811909, 0.672351311297074, 0.5821917790519067, 0.729882058012399, 1, 0.813191720191944, 0.8087111845493985, 0.725832055882795, 0.7449631633814129, 0.772150225146826, 0.813191720191944, 1 ]).reshape(6,6, order='F') cov_r = Holder() #nc <- nearPD(pr+0.01*diag(6), conv.tol = 1e-7, keepDiag = TRUE, doDykstra =FALSE, corr=FALSE) #> cat_items(nc, prefix="cov_r.") #cov_r.mat = '''<S4 object of class structure("dpoMatrix", package = "Matrix")>''' cov_r.eigenvalues = np.array([ 4.209897516692652, 0.7668341923072066, 0.518956980021938, 0.390838551407132, 0.1734728460460068, 4.209897516692652e-08 ]) cov_r.corr = '''FALSE''' cov_r.normF = 0.0623948693159157 cov_r.iterations = 11 cov_r.rel_tol = 5.83987595937896e-08 cov_r.converged = '''TRUE''' #> mkarray2(as.matrix(nc$mat), name="cov_r.mat") cov_r.mat = np.array([ 1.01, 0.486207476951913, 0.6428524769306785, 0.4886092840296514, 0.645175579158233, 0.811533860074678, 0.486207476951913, 1.01, 0.514394615153752, 0.2478398278204047, 0.673852495852274, 0.7297661648968664, 0.6428524769306785, 0.514394615153752, 1.01, 0.5971503271420517, 0.582018469844712, 0.7445177382760834, 0.4886092840296514, 0.2478398278204047, 0.5971503271420517, 1.01, 0.73161232298669, 0.7766852947049376, 0.645175579158233, 0.673852495852274, 0.582018469844712, 0.73161232298669, 1.01, 0.8107916469252828, 0.811533860074678, 0.7297661648968664, 0.7445177382760834, 0.7766852947049376, 0.8107916469252828, 1.01 ]).reshape(6,6, order='F') def test_corr_psd(): # test positive definite matrix is unchanged x = np.array([[1, -0.2, -0.9], [-0.2, 1, -0.2], [-0.9, -0.2, 1]]) y = corr_nearest(x, n_fact=100) #print np.max(np.abs(x - y)) assert_almost_equal(x, y, decimal=14) y = corr_clipped(x) assert_almost_equal(x, y, decimal=14) y = cov_nearest(x, n_fact=100) assert_almost_equal(x, y, decimal=14) x2 = x + 0.001 * np.eye(3) y = cov_nearest(x2, n_fact=100) assert_almost_equal(x2, y, decimal=14) class CheckCorrPSDMixin(object): def test_nearest(self): x = self.x res_r = self.res y = corr_nearest(x, threshold=1e-7, n_fact=100) #print np.max(np.abs(x - y)) assert_almost_equal(y, res_r.mat, decimal=3) d = norm_f(x, y) assert_allclose(d, res_r.normF, rtol=0.0015) evals = np.linalg.eigvalsh(y) #print 'evals', evals / res_r.eigenvalues[::-1] - 1 assert_allclose(evals, res_r.eigenvalues[::-1], rtol=0.003, atol=1e-7) #print evals[0] / 1e-7 - 1 assert_allclose(evals[0], 1e-7, rtol=1e-6) def test_clipped(self): x = self.x res_r = self.res y = corr_clipped(x, threshold=1e-7) #print np.max(np.abs(x - y)), np.max(np.abs((x - y) / y)) assert_almost_equal(y, res_r.mat, decimal=1) d = norm_f(x, y) assert_allclose(d, res_r.normF, rtol=0.15) evals = np.linalg.eigvalsh(y) assert_allclose(evals, res_r.eigenvalues[::-1], rtol=0.1, atol=1e-7) assert_allclose(evals[0], 1e-7, rtol=0.02) def test_cov_nearest(self): x = self.x res_r = self.res y = cov_nearest(x, method='nearest', threshold=1e-7) #print np.max(np.abs(x - y)) assert_almost_equal(y, res_r.mat, decimal=2) d = norm_f(x, y) assert_allclose(d, res_r.normF, rtol=0.0015) class TestCovPSD(object): @classmethod def setup_class(cls): x = np.array([ 1, 0.477, 0.644, 0.478, 0.651, 0.826, 0.477, 1, 0.516, 0.233, 0.682, 0.75, 0.644, 0.516, 1, 0.599, 0.581, 0.742, 0.478, 0.233, 0.599, 1, 0.741, 0.8, 0.651, 0.682, 0.581, 0.741, 1, 0.798, 0.826, 0.75, 0.742, 0.8, 0.798, 1]).reshape(6,6) cls.x = x + 0.01 * np.eye(6) cls.res = cov_r def test_cov_nearest(self): x = self.x res_r = self.res y = cov_nearest(x, method='nearest') #print np.max(np.abs(x - y)) assert_almost_equal(y, res_r.mat, decimal=3) d = norm_f(x, y) assert_allclose(d, res_r.normF, rtol=0.001) y = cov_nearest(x, method='clipped') #print np.max(np.abs(x - y)) assert_almost_equal(y, res_r.mat, decimal=2) d = norm_f(x, y) assert_allclose(d, res_r.normF, rtol=0.15) class TestCorrPSD1(CheckCorrPSDMixin): @classmethod def setup_class(cls): x = np.array([ 1, 0.477, 0.644, 0.478, 0.651, 0.826, 0.477, 1, 0.516, 0.233, 0.682, 0.75, 0.644, 0.516, 1, 0.599, 0.581, 0.742, 0.478, 0.233, 0.599, 1, 0.741, 0.8, 0.651, 0.682, 0.581, 0.741, 1, 0.798, 0.826, 0.75, 0.742, 0.8, 0.798, 1]).reshape(6,6) cls.x = x cls.res = cov1_r def test_corrpsd_threshold(): x = np.array([[1, -0.9, -0.9], [-0.9, 1, -0.9], [-0.9, -0.9, 1]]) #print np.linalg.eigvalsh(x) for threshold in [0, 1e-15, 1e-10, 1e-6]: y = corr_nearest(x, n_fact=100, threshold=threshold) evals = np.linalg.eigvalsh(y) #print 'evals', evals, threshold assert_allclose(evals[0], threshold, rtol=1e-6, atol=1e-15) y = corr_clipped(x, threshold=threshold) evals = np.linalg.eigvalsh(y) #print 'evals', evals, threshold assert_allclose(evals[0], threshold, rtol=0.25, atol=1e-15) y = cov_nearest(x, method='nearest', n_fact=100, threshold=threshold) evals = np.linalg.eigvalsh(y) #print 'evals', evals, threshold #print evals[0] / threshold - 1 assert_allclose(evals[0], threshold, rtol=1e-6, atol=1e-15) y = cov_nearest(x, n_fact=100, threshold=threshold) evals = np.linalg.eigvalsh(y) #print 'evals', evals, threshold #print evals[0] / threshold - 1 assert_allclose(evals[0], threshold, rtol=0.25, atol=1e-15)
bsd-3-clause
fe218c4077d81fc429eaa7c246a8d0c8
36.328502
95
0.603598
2.518579
false
true
false
false
yarikoptic/pystatsmodels
statsmodels/graphics/tests/test_boxplots.py
4
1261
import numpy as np from numpy.testing import dec from statsmodels.graphics.boxplots import violinplot, beanplot from statsmodels.datasets import anes96 try: import matplotlib.pyplot as plt have_matplotlib = True except: have_matplotlib = False @dec.skipif(not have_matplotlib) def test_violinplot_beanplot(): """Test violinplot and beanplot with the same dataset.""" data = anes96.load_pandas() party_ID = np.arange(7) labels = ["Strong Democrat", "Weak Democrat", "Independent-Democrat", "Independent-Independent", "Independent-Republican", "Weak Republican", "Strong Republican"] age = [data.exog['age'][data.endog == id] for id in party_ID] fig = plt.figure() ax = fig.add_subplot(111) violinplot(age, ax=ax, labels=labels, plot_opts={'cutoff_val':5, 'cutoff_type':'abs', 'label_fontsize':'small', 'label_rotation':30}) plt.close(fig) fig = plt.figure() ax = fig.add_subplot(111) beanplot(age, ax=ax, labels=labels, plot_opts={'cutoff_val':5, 'cutoff_type':'abs', 'label_fontsize':'small', 'label_rotation':30}) plt.close(fig)
bsd-3-clause
4cfae03f27be64d658afb8cb8d8382eb
28.325581
73
0.604282
3.552113
false
true
false
false
yarikoptic/pystatsmodels
statsmodels/tsa/tests/results/arima112nc_css_results.py
35
44167
import numpy as np llf = np.array([-239.75290561974]) nobs = np.array([ 202]) k = np.array([ 4]) k_exog = np.array([ 1]) sigma = np.array([ .79291203639424]) chi2 = np.array([ 35036.682546665]) df_model = np.array([ 3]) k_ar = np.array([ 1]) k_ma = np.array([ 2]) params = np.array([ .99954097483478, -.69022779461512, -.20477682541104, .62870949745886]) cov_params = np.array([ .00007344276568, -.00016074342677, -.00018478942445, 8.040251506e-06, -.00016074342677, .00094099304774, .00017233777676, -.0000145011098, -.00018478942445, .00017233777676, .00103352686916, .00030686101903, 8.040251506e-06, -.0000145011098, .00030686101903, .00067796985496]).reshape(4,4) xb = np.array([ 0, 0, .05104803293943, .06663129478693, .02164618112147, .0773858949542, .02606418170035, .09391833096743, .05710592120886, .03083370067179, .07319989800453, .05287836492062, .05776296555996, .09105986356735, .04293738678098, .09576436132193, .06068528071046, .06157376244664, .11172580718994, .06527806818485, .11443704366684, .05653077363968, .08205550909042, .08481238037348, .10436166077852, .0875685736537, .12320486456156, .08366665989161, .13979130983353, .1902572363615, .1306214183569, .21803694963455, .11079790443182, .17274764180183, .1937662512064, .20047917962074, .24034893512726, .21783453226089, .29279819130898, .26804205775261, .28678458929062, .35651323199272, .33659368753433, .35760068893433, .39895334839821, .41131839156151, .36645981669426, .40991494059563, .41024547815323, .32657703757286, .42312324047089, .34933325648308, .35912537574768, .35077446699142, .34701564908028, .37364318966866, .40170526504517, .56070649623871, .41915491223335, .73478156328201, .67748892307281, .7744625210762, .77825599908829, .97586625814438, .88692498207092, .76232481002808, .87376874685287, .83281141519547, .84783887863159, .66423743963242, .84904235601425, .81613594293594, .80033475160599, .95782464742661, .80624777078629, .83626395463943, .91873735189438, .95130664110184, 1.0939226150513, 1.1171194314957, 1.1004731655121, 1.3512066602707, 1.4703129529953, 1.4805699586868, 1.7385860681534, 1.8268398046494, 1.5489361286163, 1.7446503639221, 1.864644408226, 1.7200467586517, 1.9223358631134, 1.775306224823, 1.5392524003983, 1.4067870378494, 1.9366238117218, 1.2984343767166, 1.1080636978149, 1.3500427007675, 1.2837564945221, 1.2670782804489, 1.3347851037979, 1.2857422828674, 1.1625040769577, 1.2111755609512, 1.0548515319824, 1.2553508281708, 1.0327949523926, 1.0740388631821, 1.222040772438, .40555971860886, 1.0233588218689, .84209614992142, 1.0186324119568, 1.0319027900696, .99487775564194, 1.0439211130142, .98785293102264, 1.0620124340057, 1.0916963815689, 1.1378232240677, 1.1243290901184, 1.3305295705795, 1.1925677061081, 1.0872994661331, 1.4599523544312, 1.2333589792252, 1.3584797382355, 1.7595859766006, 1.3009568452835, 1.1157965660095, 1.2948887348175, 1.2063180208206, 1.2332669496536, 1.2132470607758, 1.2049551010132, 1.2260574102402, 1.1875206232071, 1.1547852754593, 1.0519831180573, 1.1594845056534, 1.0069926977158, 1.0675266981125, 1.1299223899841, 1.0620901584625, 1.0999356508255, 1.1535499095917, 1.0026944875717, 1.0428657531738, 1.1120204925537, 1.1684119701385, 1.0258769989014, 1.1342295408249, 1.1183958053589, .91313683986664, .91156214475632, 1.0540328025818, .84359037876129, .75758427381516, .96401190757751, .83226495981216, .8759680390358, .98239886760712, .85917687416077, 1.0634194612503, .99442666769028, 1.153311252594, 1.2288066148758, 1.0869039297104, 1.281947016716, 1.0067318677902, 1.1028815507889, .82448446750641, .78489726781845, 1.1850204467773, .86753690242767, 1.0692945718765, 1.1030179262161, .8791960477829, .86451041698456, 1.0455346107483, 1.085998415947, 1.0172398090363, 1.2250980138779, 1.2316122055054, 1.062157869339, 1.3991860151291, 1.0520887374878, 2.2203133106232, .88833123445511, 1.4289729595184, 1.5206423997879, .68520504236221, 1.4659557342529, 1.5350053310394, 1.3178979158401, 1.4888265132904, 1.9698411226273, 1.4406447410583, 2.517040014267, .55537897348404, -.20722626149654, 1.0899519920349, 1.164245724678]) y = np.array([np.nan, 28.979999542236, 29.201047897339, 29.416631698608, 29.391647338867, 29.617385864258, 29.576063156128, 29.84391784668, 29.897106170654, 29.84083366394, 29.993200302124, 30.032876968384, 30.097763061523, 30.3010597229, 30.262937545776, 30.475763320923, 30.500686645508, 30.541572570801, 30.801725387573, 30.815279006958, 31.054437637329, 31.006530761719, 31.102056503296, 31.20481300354, 31.384363174438, 31.467567443848, 31.703205108643, 31.733665466309, 32.019790649414, 32.47025680542, 32.580623626709, 33.068035125732, 33.010799407959, 33.272747039795, 33.593769073486, 33.900478363037, 34.340347290039, 34.617835998535, 35.192798614502, 35.568042755127, 35.986785888672, 36.656513214111, 37.13659286499, 37.657600402832, 38.29895401001, 38.911319732666, 39.266460418701, 39.809917449951, 40.310245513916, 40.426574707031, 41.023120880127, 41.249336242676, 41.559127807617, 41.850772857666, 42.14701461792, 42.573642730713, 43.101707458496, 44.260707855225, 44.619155883789, 46.334781646729, 47.477489471436, 48.874462127686, 50.078254699707, 51.9758644104, 53.186923980713, 53.762325286865, 54.873767852783, 55.732814788818, 56.647838592529, 56.764236450195, 57.849040985107, 58.716136932373, 59.500335693359, 60.957824707031, 61.606246948242, 62.436264038086, 63.61873626709, 64.85131072998, 66.593925476074, 68.21711730957, 69.600471496582, 71.951202392578, 74.470314025879, 76.680564880371, 79.738586425781, 82.726844787598, 84.148933410645, 86.444648742676, 89.064643859863, 90.820045471191, 93.422332763672, 95.175308227539, 95.939254760742, 96.406784057617, 99.436622619629, 99.398429870605, 99.00806427002, 100.15004730225, 101.08376312256, 102.06707763672, 103.43478393555, 104.58574676514, 105.26250457764, 106.31117248535, 106.75485229492, 108.25534820557, 108.73279571533, 109.57403564453, 111.12203979492, 109.10556030273, 110.52336120605, 111.04209136963, 112.41863250732, 113.73190307617, 114.79488372803, 116.04392242432, 116.98785400391, 118.26200866699, 119.59169769287, 121.03782653809, 122.32432556152, 124.4305267334, 125.69256591797, 126.4873046875, 128.95994567871, 130.13334655762, 131.85847473145, 135.15957641602, 136.00096130371, 136.21580505371, 137.49488830566, 138.40631103516, 139.53326416016, 140.61323547363, 141.70495605469, 142.9260559082, 143.98751831055, 144.95478820801, 145.55198669434, 146.7594909668, 147.30699157715, 148.26751708984, 149.52992248535, 150.46208190918, 151.59992980957, 152.95355224609, 153.60270690918, 154.54286193848, 155.81201171875, 157.2684173584, 158.02587890625, 159.33422851563, 160.51838684082, 160.81312561035, 161.31155395508, 162.55403137207, 162.84359741211, 162.95758056641, 164.16400146484, 164.73225402832, 165.57595825195, 166.88238525391, 167.55917358398, 169.16342163086, 170.29443359375, 172.0532989502, 173.92880249023, 174.9868927002, 176.88195800781, 177.40672302246, 178.50286865234, 178.42448425293, 178.48489379883, 180.48501586914, 180.86753845215, 182.26928710938, 183.70301818848, 184.07919311523, 184.56451416016, 185.94552612305, 187.38600158691, 188.41723632813, 190.32510375977, 192.03161621094, 192.8621673584, 195.19918823242, 195.7520904541, 201.42030334473, 200.28833007813, 202.12896728516, 204.22064208984, 202.58520507813, 205.03996276855, 207.45500183105, 208.65589904785, 210.62182617188, 214.46484375, 215.4376373291, 221.12704467773, 217.44438171387, 211.96676635742, 213.76095581055, 215.63323974609]) resid = np.array([np.nan, .17000007629395, .14895272254944, -.04663083329797, .14835388958454, -.067387573421, .17393657565117, -.00391817837954, -.08710660785437, .07916691154242, -.01320043392479, .00712300790474, .11223520338535, -.08105963468552, .11706246435642, -.03576298803091, -.0206862706691, .14842723309994, -.05172633752227, .12472246587276, -.104436814785, .01346892025322, .01794487424195, .07518746703863, -.00436318712309, .11243218928576, -.05320516973734, .14633287489414, .26020830869675, -.02025525644422, .26937630772591, -.16803389787674, .08919904381037, .12725540995598, .10623299330473, .1995185315609, .05965411290526, .28216546773911, .10719951242208, .13195945322514, .31321388483047, .14348678290844, .16340629756451, .24240159988403, .20104511082172, -.01131687406451, .13354018330574, .09008505195379, -.21024852991104, .1734229773283, -.12312019616365, -.04933403432369, -.05912613496184, -.05077522993088, .05298588424921, .12635681033134, .59829473495483, -.06070650368929, .98084276914597, .46521919965744, .62251031398773, .42553821206093, .92174476385117, .32413294911385, -.18692423403263, .23767517507076, .02623280882835, .06718628853559, -.54783964157104, .23576405644417, .05095916613936, -.01613672077656, .49966445565224, -.1578253954649, -.00624855514616, .26373836398125, .28126338124275, .64869183301926, .50607579946518, .28288212418556, .99952530860901, 1.0487948656082, .72968405485153, 1.319433093071, 1.1614154577255, -.126842841506, .55106240510941, .75534957647324, .03535716608167, .67995470762253, -.02233435213566, -.775306224823, -.93925392627716, 1.0932129621506, -1.3366253376007, -1.4984313249588, -.20806217193604, -.35004270076752, -.28375649452209, .03291719779372, -.13478049635887, -.48574689030647, -.16250413656235, -.61117708683014, .24515150487423, -.55535387992859, -.23279193043709, .32596263289452, -2.4220454692841, .39444333314896, -.32336190342903, .3579084277153, .28136301040649, .06810333579779, .20511920750141, -.04392115399241, .2121440321207, .23799060285091, .30830511450768, .16217225790024, .77567237615585, .06947190314531, -.29256621003151, 1.012699007988, -.05995841696858, .36664715409279, 1.5415141582489, -.45958289504051, -.90094763040543, -.01580573618412, -.29488867521286, -.10631188750267, -.13327611982822, -.11324090510607, -.00495810527354, -.12605135142803, -.18752062320709, -.45478835701942, .04802301898599, -.45948752760887, -.1069987937808, .13247023522854, -.12992236018181, .0379159450531, .20006743073463, -.35354685783386, -.10270062834024, .15713113546371, .28798860311508, -.26841807365417, .17411996424198, .06576746702194, -.61839586496353, -.41313683986664, .18844394385815, -.55403280258179, -.6435934305191, .24241572618484, -.26401495933533, -.03226193040609, .32402887940407, -.1823958158493, .54083228111267, .13657745718956, .60556417703629, .64669179916382, -.02880969829857, .61310833692551, -.48195922374725, -.00673185009509, -.90286934375763, -.72449362277985, .81510883569717, -.48502343893051, .33246004581451, .33071464300156, -.50302714109421, -.3791960477829, .33548650145531, .35447460412979, .01399240642786, .682772397995, .474898904562, -.23161220550537, .93784213066101, -.49919214844704, 3.4479112625122, -2.020316362381, .41167178750038, .57102704048157, -2.3206453323364, .98880618810654, .8800373673439, -.11700607091188, .47710022330284, 1.8731729984283, -.46784225106239, 3.1723618507385, -4.2380332946777, -5.2703905105591, .70423555374146, .70803683996201, .7517546415329]) yr = np.array([np.nan, .17000007629395, .14895272254944, -.04663083329797, .14835388958454, -.067387573421, .17393657565117, -.00391817837954, -.08710660785437, .07916691154242, -.01320043392479, .00712300790474, .11223520338535, -.08105963468552, .11706246435642, -.03576298803091, -.0206862706691, .14842723309994, -.05172633752227, .12472246587276, -.104436814785, .01346892025322, .01794487424195, .07518746703863, -.00436318712309, .11243218928576, -.05320516973734, .14633287489414, .26020830869675, -.02025525644422, .26937630772591, -.16803389787674, .08919904381037, .12725540995598, .10623299330473, .1995185315609, .05965411290526, .28216546773911, .10719951242208, .13195945322514, .31321388483047, .14348678290844, .16340629756451, .24240159988403, .20104511082172, -.01131687406451, .13354018330574, .09008505195379, -.21024852991104, .1734229773283, -.12312019616365, -.04933403432369, -.05912613496184, -.05077522993088, .05298588424921, .12635681033134, .59829473495483, -.06070650368929, .98084276914597, .46521919965744, .62251031398773, .42553821206093, .92174476385117, .32413294911385, -.18692423403263, .23767517507076, .02623280882835, .06718628853559, -.54783964157104, .23576405644417, .05095916613936, -.01613672077656, .49966445565224, -.1578253954649, -.00624855514616, .26373836398125, .28126338124275, .64869183301926, .50607579946518, .28288212418556, .99952530860901, 1.0487948656082, .72968405485153, 1.319433093071, 1.1614154577255, -.126842841506, .55106240510941, .75534957647324, .03535716608167, .67995470762253, -.02233435213566, -.775306224823, -.93925392627716, 1.0932129621506, -1.3366253376007, -1.4984313249588, -.20806217193604, -.35004270076752, -.28375649452209, .03291719779372, -.13478049635887, -.48574689030647, -.16250413656235, -.61117708683014, .24515150487423, -.55535387992859, -.23279193043709, .32596263289452, -2.4220454692841, .39444333314896, -.32336190342903, .3579084277153, .28136301040649, .06810333579779, .20511920750141, -.04392115399241, .2121440321207, .23799060285091, .30830511450768, .16217225790024, .77567237615585, .06947190314531, -.29256621003151, 1.012699007988, -.05995841696858, .36664715409279, 1.5415141582489, -.45958289504051, -.90094763040543, -.01580573618412, -.29488867521286, -.10631188750267, -.13327611982822, -.11324090510607, -.00495810527354, -.12605135142803, -.18752062320709, -.45478835701942, .04802301898599, -.45948752760887, -.1069987937808, .13247023522854, -.12992236018181, .0379159450531, .20006743073463, -.35354685783386, -.10270062834024, .15713113546371, .28798860311508, -.26841807365417, .17411996424198, .06576746702194, -.61839586496353, -.41313683986664, .18844394385815, -.55403280258179, -.6435934305191, .24241572618484, -.26401495933533, -.03226193040609, .32402887940407, -.1823958158493, .54083228111267, .13657745718956, .60556417703629, .64669179916382, -.02880969829857, .61310833692551, -.48195922374725, -.00673185009509, -.90286934375763, -.72449362277985, .81510883569717, -.48502343893051, .33246004581451, .33071464300156, -.50302714109421, -.3791960477829, .33548650145531, .35447460412979, .01399240642786, .682772397995, .474898904562, -.23161220550537, .93784213066101, -.49919214844704, 3.4479112625122, -2.020316362381, .41167178750038, .57102704048157, -2.3206453323364, .98880618810654, .8800373673439, -.11700607091188, .47710022330284, 1.8731729984283, -.46784225106239, 3.1723618507385, -4.2380332946777, -5.2703905105591, .70423555374146, .70803683996201, .7517546415329]) mse = np.array([ .95459979772568, .71522510051727, .63122135400772, .6314896941185, .63029319047928, .63014930486679, .62988424301147, .62969470024109, .62953060865402, .6293950676918, .62928181886673, .62918740510941, .62910866737366, .62904292345047, .62898802757263, .6289421916008, .62890386581421, .62887191772461, .62884521484375, .6288229227066, .62880426645279, .6287887096405, .6287756562233, .62876480817795, .62875574827194, .62874811887741, .62874180078506, .62873649597168, .62873202562332, .62872833013535, .62872523069382, .62872266769409, .62872052192688, .62871867418289, .62871718406677, .62871593236923, .62871485948563, .62871396541595, .62871325016022, .62871265411377, .62871211767197, .62871170043945, .62871134281158, .62871104478836, .62871074676514, .6287105679512, .62871038913727, .62871026992798, .62871015071869, .6287100315094, .62870991230011, .62870985269547, .62870979309082, .62870973348618, .62870973348618, .62870967388153, .62870967388153, .62870961427689, .62870961427689, .62870955467224, .62870955467224, .62870955467224, .62870955467224, .62870955467224, .62870955467224, .62870955467224, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676, .6287094950676]) stdp = np.array([ 0, 0, .05104803293943, .06663129478693, .02164618112147, .0773858949542, .02606418170035, .09391833096743, .05710592120886, .03083370067179, .07319989800453, .05287836492062, .05776296555996, .09105986356735, .04293738678098, .09576436132193, .06068528071046, .06157376244664, .11172580718994, .06527806818485, .11443704366684, .05653077363968, .08205550909042, .08481238037348, .10436166077852, .0875685736537, .12320486456156, .08366665989161, .13979130983353, .1902572363615, .1306214183569, .21803694963455, .11079790443182, .17274764180183, .1937662512064, .20047917962074, .24034893512726, .21783453226089, .29279819130898, .26804205775261, .28678458929062, .35651323199272, .33659368753433, .35760068893433, .39895334839821, .41131839156151, .36645981669426, .40991494059563, .41024547815323, .32657703757286, .42312324047089, .34933325648308, .35912537574768, .35077446699142, .34701564908028, .37364318966866, .40170526504517, .56070649623871, .41915491223335, .73478156328201, .67748892307281, .7744625210762, .77825599908829, .97586625814438, .88692498207092, .76232481002808, .87376874685287, .83281141519547, .84783887863159, .66423743963242, .84904235601425, .81613594293594, .80033475160599, .95782464742661, .80624777078629, .83626395463943, .91873735189438, .95130664110184, 1.0939226150513, 1.1171194314957, 1.1004731655121, 1.3512066602707, 1.4703129529953, 1.4805699586868, 1.7385860681534, 1.8268398046494, 1.5489361286163, 1.7446503639221, 1.864644408226, 1.7200467586517, 1.9223358631134, 1.775306224823, 1.5392524003983, 1.4067870378494, 1.9366238117218, 1.2984343767166, 1.1080636978149, 1.3500427007675, 1.2837564945221, 1.2670782804489, 1.3347851037979, 1.2857422828674, 1.1625040769577, 1.2111755609512, 1.0548515319824, 1.2553508281708, 1.0327949523926, 1.0740388631821, 1.222040772438, .40555971860886, 1.0233588218689, .84209614992142, 1.0186324119568, 1.0319027900696, .99487775564194, 1.0439211130142, .98785293102264, 1.0620124340057, 1.0916963815689, 1.1378232240677, 1.1243290901184, 1.3305295705795, 1.1925677061081, 1.0872994661331, 1.4599523544312, 1.2333589792252, 1.3584797382355, 1.7595859766006, 1.3009568452835, 1.1157965660095, 1.2948887348175, 1.2063180208206, 1.2332669496536, 1.2132470607758, 1.2049551010132, 1.2260574102402, 1.1875206232071, 1.1547852754593, 1.0519831180573, 1.1594845056534, 1.0069926977158, 1.0675266981125, 1.1299223899841, 1.0620901584625, 1.0999356508255, 1.1535499095917, 1.0026944875717, 1.0428657531738, 1.1120204925537, 1.1684119701385, 1.0258769989014, 1.1342295408249, 1.1183958053589, .91313683986664, .91156214475632, 1.0540328025818, .84359037876129, .75758427381516, .96401190757751, .83226495981216, .8759680390358, .98239886760712, .85917687416077, 1.0634194612503, .99442666769028, 1.153311252594, 1.2288066148758, 1.0869039297104, 1.281947016716, 1.0067318677902, 1.1028815507889, .82448446750641, .78489726781845, 1.1850204467773, .86753690242767, 1.0692945718765, 1.1030179262161, .8791960477829, .86451041698456, 1.0455346107483, 1.085998415947, 1.0172398090363, 1.2250980138779, 1.2316122055054, 1.062157869339, 1.3991860151291, 1.0520887374878, 2.2203133106232, .88833123445511, 1.4289729595184, 1.5206423997879, .68520504236221, 1.4659557342529, 1.5350053310394, 1.3178979158401, 1.4888265132904, 1.9698411226273, 1.4406447410583, 2.517040014267, .55537897348404, -.20722626149654, 1.0899519920349, 1.164245724678]) icstats = np.array([ 202, np.nan, -239.75290561974, 4, 487.50581123949, 500.73888202909]) class Bunch(dict): def __init__(self, **kw): dict.__init__(self, kw) self.__dict__ = self results = Bunch(llf=llf, nobs=nobs, k=k, k_exog=k_exog, sigma=sigma, chi2=chi2, df_model=df_model, k_ar=k_ar, k_ma=k_ma, params=params, cov_params=cov_params, xb=xb, y=y, resid=resid, yr=yr, mse=mse, stdp=stdp, icstats=icstats, )
bsd-3-clause
38e1b631a2ad894d093af386f39961a5
33.478532
229
0.396767
3.829287
false
false
false
false
rootpy/rootpy
rootpy/plotting/contrib/quantiles.py
3
3602
""" Taken from example by Zhiyi Liu, zhiyil@fnal.gov here: http://root.cern.ch/phpBB3/viewtopic.php?f=3&t=6865 and converted into Python """ from __future__ import absolute_import import ROOT from math import sqrt from array import array from .. import Graph from ...extern.six.moves import range __all__ = [ 'qqgraph', ] def qqgraph(h1, h2, quantiles=None): """ Return a Graph of a quantile-quantile (QQ) plot and confidence band """ if quantiles is None: quantiles = max(min(len(h1), len(h2)) / 2, 1) nq = quantiles # position where to compute the quantiles in [0, 1] xq = array('d', [0.] * nq) # array to contain the quantiles yq1 = array('d', [0.] * nq) # array to contain the quantiles yq2 = array('d', [0.] * nq) for i in range(nq): xq[i] = float(i + 1) / nq h1.GetQuantiles(nq, yq1, xq) h2.GetQuantiles(nq, yq2, xq) xq_plus = array('d', [0.] * nq) xq_minus = array('d', [0.] * nq) yq2_plus = array('d', [0.] * nq) yq2_minus = array('d', [0.] * nq) """ KS_cv: KS critical value 1.36 KS_cv = ----------- sqrt( N ) Where 1.36 is for alpha = 0.05 (confidence level 1-5%=95%, about 2 sigma) For 1 sigma (alpha=0.32, CL=68%), the value in the nominator is 0.9561, it is gotten by GetCriticalValue(1, 1 - 0.68). Notes ----- * For 1-sample KS test (data and theoretic), N should be n * For 2-sample KS test (2 data set), N should be sqrt(m*n/(m+n))! Here is the case m or n (size of samples) should be effective size for a histogram * Critical value here is valid for only for sample size >= 80 (some references say 35) which means, for example, for a unweighted histogram, it must have more than 80 (or 35) entries filled and then confidence band is reliable. """ esum1 = effective_sample_size(h1) esum2 = effective_sample_size(h2) # one sigma band KS_cv = (critical_value(1, 1 - 0.68) / sqrt((esum1 * esum2) / (esum1 + esum2))) for i in range(nq): # upper limit xq_plus[i] = float(xq[i] + KS_cv) # lower limit xq_minus[i] = float(xq[i] - KS_cv) h2.GetQuantiles(nq, yq2_plus, xq_plus) h2.GetQuantiles(nq, yq2_minus, xq_minus) yq2_err_plus = array('d', [0.] * nq) yq2_err_minus = array('d', [0.] * nq) for i in range(nq): yq2_err_plus[i] = yq2_plus[i] - yq2[i] yq2_err_minus[i] = yq2[i] - yq2_minus[i] # forget the last point, so number of points: (nq - 1) gr = Graph(nq - 1) for i in range(nq - 1): gr[i] = (yq1[i], yq2[i]) # confidence level band gr.SetPointEYlow(i, yq2_err_minus[i]) gr.SetPointEYhigh(i, yq2_err_plus[i]) return gr def effective_sample_size(h): """ Calculate the effective sample size for a histogram the same way as ROOT does. """ sum = 0 ew = 0 w = 0 for bin in h.bins(overflow=False): sum += bin.value ew = bin.error w += ew * ew esum = sum * sum / w return esum def critical_value(n, p): """ This function calculates the critical value given n and p, and confidence level = 1 - p. """ dn = 1 delta = 0.5 res = ROOT.TMath.KolmogorovProb(dn * sqrt(n)) while res > 1.0001 * p or res < 0.9999 * p: if (res > 1.0001 * p): dn = dn + delta if (res < 0.9999 * p): dn = dn - delta delta = delta / 2. res = ROOT.TMath.KolmogorovProb(dn * sqrt(n)) return dn
bsd-3-clause
18d1f25931dd5e300dbc305b432cf04b
25.101449
78
0.564409
2.8542
false
false
false
false
rootpy/rootpy
examples/tree/chain_overwrite.py
7
2603
#!/usr/bin/env python """ ============================================ Copy a tree chain while overwriting branches ============================================ This is an example showing how to copy a tree chain while overwriting one or more of its branches with new values. """ print(__doc__) from rootpy.tree import Tree, TreeModel, TreeChain, FloatCol, IntCol from rootpy.io import root_open from random import gauss """ This first section of code only creates an example tree chain. """ class Event(TreeModel): """Event model definition""" x = FloatCol() y = FloatCol() z = FloatCol() i = IntCol() # first create several example trees in separate files fnames = ["test_{0:d}.root".format(i) for i in range(5)] for fname in fnames: with root_open(fname, "recreate") as f: tree = Tree("test", model=Event) # fill the tree for i in range(100): tree.x = gauss(.5, 1.) tree.y = gauss(.3, 2.) tree.z = gauss(13., 42.) tree.i = i tree.fill() tree.write() """ This section below takes the example trees and copies it while overwriting a branch with new values. """ # first define the chain of trees chain = TreeChain(name="test", files=fnames) # Now we want to copy the tree above into a new file while overwriting a branch # First create a new file to save the new tree in: f_copy = root_open("test_copy.root", "recreate") # You may not know the entire model of the original tree but only the branches # you intend to overwrite, so I am not specifying the model=Event below as an # example of how to deal with this in general: tree_copy = Tree("test_copy") # If the original tree was not handed to you through rootpy don't forget to: # >>> from rootpy import asrootpy # >>> tree = asrootpy(tree) # Here we specify the buffer for the new tree to use. We use the same buffer as # the original tree. This creates all the same branches in the new tree but # their addresses point to the same memory used by the original tree. tree_copy.set_buffer( chain._buffer, create_branches=True) # Now loop over the original tree and fill the new tree for entry in chain: # Overwrite a branch value. This changes the value that will be written to # the new tree but leaves the value unchanged in the original tree on disk. entry.x = 3.141 # "entry" is actually the buffer, which is shared between both trees. tree_copy.Fill() # tree_copy is now a copy of tree where the "x" branch has been overwritten # with new values tree_copy.Write() f_copy.Close()
bsd-3-clause
2a26b1156f389b01d1bf097fe6f96344
30.743902
79
0.663081
3.788937
false
true
false
false
rootpy/rootpy
examples/plotting/plot_matplotlib_graph.py
7
2423
#!/usr/bin/env python """ ===================================== Plot a ROOT graph with matplotlib ===================================== This example demonstrates how a ROOT graph can be styled with simple attributes and displayed via ROOT or matplotlib. """ print(__doc__) import ROOT import numpy as np from rootpy.plotting import Canvas, Graph from rootpy.plotting.style import get_style, set_style from rootpy.interactive import wait import rootpy.plotting.root2matplotlib as rplt import matplotlib.pyplot as plt from matplotlib.ticker import AutoMinorLocator, MultipleLocator # set the random seed ROOT.gRandom.SetSeed(42) np.random.seed(42) # points x = np.sort(np.random.random(10)) * 3500 y = np.random.random(10) # set style for ROOT set_style('ATLAS') # create graph graph = Graph(x.shape[0]) for i, (xx, yy) in enumerate(zip(x, y)): graph.SetPoint(i, xx, yy) # set visual attributes graph.linecolor = 'blue' graph.markercolor = 'blue' graph.xaxis.SetTitle("E_{T} [GeV]") graph.yaxis.SetTitle("d#sigma_{jet}/dE_{T,jet} [fb/GeV]") graph.xaxis.SetRangeUser(0, 3500) graph.yaxis.SetRangeUser(0, 1) # plot with ROOT canvas = Canvas() graph.Draw("APL") label = ROOT.TText(0.4, 0.8, "ROOT") label.SetTextFont(43) label.SetTextSize(25) label.SetNDC() label.Draw() canvas.Modified() canvas.Update() # plot with matplotlib def plot_with_matplotlib(): fig, axes = plt.subplots() axes.plot(x, y, 'o-', markeredgewidth=0) axes.set_xlabel(r"$E_T$ [GeV]", horizontalalignment="right", x=1, labelpad=20) axes.set_ylabel(r"$d\sigma_{jet}/dE_{T,jet}$ [fb/GeV]", horizontalalignment="right", y=1, labelpad=32) axes.set_xlim(0, 3500) axes.set_ylim(0, 1) return fig, axes # plot without style fig1, axes1 = plot_with_matplotlib() axes1.text(0.4, 0.8, 'matplotlib (no style)', verticalalignment='center', horizontalalignment='center', transform=axes1.transAxes, fontsize=20) # plot with ATLAS style set_style('ATLAS', mpl=True) fig2, axes2 = plot_with_matplotlib() axes2.text(0.4, 0.8, 'matplotlib', verticalalignment='center', horizontalalignment='center', transform=axes2.transAxes, fontsize=20) axes2.xaxis.set_minor_locator(AutoMinorLocator()) axes2.yaxis.set_minor_locator(AutoMinorLocator()) if not ROOT.gROOT.IsBatch(): plt.show() # wait for you to close the canvas before exiting wait(True)
bsd-3-clause
435357524153c214c2bf7a3f179d56b8
25.922222
68
0.680149
3.082697
false
false
false
false
rootpy/rootpy
rootpy/utils/tests/test_cpp.py
3
1191
from __future__ import print_function import sys from ROOT import MethodProxy import inspect from rootpy.utils.cpp import CPPGrammar from rootpy.utils.extras import iter_ROOT_classes from nose.plugins.attrib import attr @attr('slow') def test_cpp(): i = 0 num_methods = 0 for cls in iter_ROOT_classes(): members = inspect.getmembers(cls) # filter out those starting with "_" or "operator " # and non-method members # also split overloaded methods methods = {} for name, func in members: if name.startswith('_') or name.startswith('operator'): continue if not isinstance(func, MethodProxy): continue methods[name] = (func, func.func_doc.split('\n')) for name, (func, sigs) in methods.items(): for sig in sigs: num_methods += 1 if CPPGrammar.parse_method(sig, silent=False): i += 1 print("{0} / {1}".format(i, num_methods), end='\r') sys.stdout.flush() print("{0} / {1}".format(i, num_methods)) if __name__ == "__main__": import nose nose.runmodule()
bsd-3-clause
0097066d9aba4a3a9e15c39579844e63
28.775
67
0.570949
4.078767
false
false
false
false
rootpy/rootpy
rootpy/tree/chain.py
3
10692
from __future__ import absolute_import import multiprocessing import time from .. import log; log = log[__name__] from .. import QROOT from ..io import root_open, DoesNotExist from ..utils.extras import humanize_bytes from ..context import preserve_current_directory from ..plotting.graph import _GraphBase from ..extern.six import string_types from .filtering import EventFilterList __all__ = [ 'TreeChain', 'TreeQueue', ] class BaseTreeChain(object): def __init__(self, name, treebuffer=None, branches=None, ignore_branches=None, events=-1, onfilechange=None, read_branches_on_demand=False, cache=False, # 30 MB cache by default cache_size=30000000, learn_entries=10, always_read=None, ignore_unsupported=False, filters=None): self._name = name self._buffer = treebuffer self._branches = branches self._ignore_branches = ignore_branches self._tree = None self._file = None self._events = events self._total_events = 0 self._ignore_unsupported = ignore_unsupported self._initialized = False if filters is None: self._filters = EventFilterList([]) else: self._filters = filters if onfilechange is None: onfilechange = [] self._filechange_hooks = onfilechange self._read_branches_on_demand = read_branches_on_demand self._use_cache = cache self._cache_size = cache_size self._learn_entries = learn_entries self.weight = 1. self.userdata = {} if not self._rollover(): raise RuntimeError("unable to initialize TreeChain") if always_read is None: self._always_read = [] elif isinstance(always_read, string_types): if '*' in always_read: always_read = self._tree.glob(always_read) else: always_read = [always_read] self.always_read(always_read) else: branches = [] for branch in always_read: if '*' in branch: branches += self._tree.glob(branch) else: branches.append(branch) self.always_read(branches) def __nonzero__(self): return len(self) > 0 __bool__ = __nonzero__ def _next_file(self): """ Override in subclasses """ return None def always_read(self, branches): self._always_read = branches self._tree.always_read(branches) def reset(self): if self._tree is not None: self._tree = None if self._file is not None: self._file.Close() self._file = None def Draw(self, *args, **kwargs): """ Loop over subfiles, draw each, and sum the output into a single histogram. """ self.reset() output = None while self._rollover(): if output is None: # Make our own copy of the drawn histogram output = self._tree.Draw(*args, **kwargs) if output is not None: output = output.Clone() # Make it memory resident (histograms) if hasattr(output, 'SetDirectory'): output.SetDirectory(0) else: newoutput = self._tree.Draw(*args, **kwargs) if newoutput is not None: if isinstance(output, _GraphBase): output.Append(newoutput) else: # histogram output += newoutput return output draw = Draw def __getattr__(self, attr): try: return getattr(self._tree, attr) except AttributeError: raise AttributeError("{0} instance has no attribute '{1}'".format( self.__class__.__name__, attr)) def __getitem__(self, item): return self._tree.__getitem__(item) def __contains__(self, branch): return self._tree.__contains__(branch) def __iter__(self): passed_events = 0 self.reset() while self._rollover(): entries = 0 total_entries = float(self._tree.GetEntries()) t1 = time.time() t2 = t1 for entry in self._tree: entries += 1 self.userdata = {} if self._filters(entry): yield entry passed_events += 1 if self._events == passed_events: break if time.time() - t2 > 60: entry_rate = int(entries / (time.time() - t1)) log.info( "{0:d} entr{1} per second. " "{2:.0f}% done current tree.".format( entry_rate, 'ies' if entry_rate != 1 else 'y', 100 * entries / total_entries)) t2 = time.time() if self._events == passed_events: break log.info("{0:d} entries per second".format( int(entries / (time.time() - t1)))) log.info("read {0:d} bytes in {1:d} transactions".format( self._file.GetBytesRead(), self._file.GetReadCalls())) self._total_events += entries self._filters.finalize() def _rollover(self): filename = self._next_file() if filename is None: return False log.info("current file: {0}".format(filename)) try: with preserve_current_directory(): if self._file is not None: self._file.Close() self._file = root_open(filename) except IOError: self._file = None log.warning("could not open file {0} (skipping)".format(filename)) return self._rollover() try: self._tree = self._file.Get(self._name) except DoesNotExist: log.warning( "tree {0} does not exist in file {1} (skipping)".format( self._name, filename)) return self._rollover() if len(self._tree.GetListOfBranches()) == 0: log.warning("tree with no branches in file {0} (skipping)".format( filename)) return self._rollover() if self._branches is not None: self._tree.activate(self._branches, exclusive=True) if self._ignore_branches is not None: self._tree.deactivate(self._ignore_branches, exclusive=False) if self._buffer is None: self._tree.create_buffer(self._ignore_unsupported) self._buffer = self._tree._buffer else: self._tree.set_buffer( self._buffer, ignore_missing=True, transfer_objects=True) self._buffer = self._tree._buffer if self._use_cache: # enable TTreeCache for this tree log.info( "enabling a {0} TTreeCache for the current tree " "({1:d} learning entries)".format( humanize_bytes(self._cache_size), self._learn_entries)) self._tree.SetCacheSize(self._cache_size) self._tree.SetCacheLearnEntries(self._learn_entries) self._tree.read_branches_on_demand = self._read_branches_on_demand self._tree.always_read(self._always_read) self.weight = self._tree.GetWeight() for target, args in self._filechange_hooks: # run any user-defined functions target(*args, name=self._name, file=self._file, tree=self._tree) return True class TreeChain(BaseTreeChain): """ A ROOT.TChain replacement """ def __init__(self, name, files, **kwargs): if isinstance(files, tuple): files = list(files) elif not isinstance(files, list): files = [files] else: files = files[:] if not files: raise RuntimeError( "unable to initialize TreeChain: no files") self._files = files self.curr_file_idx = 0 super(TreeChain, self).__init__(name, **kwargs) self._tchain = QROOT.TChain(name) for filename in self._files: self._tchain.Add(filename) def GetEntries(self, *args, **kwargs): return self._tchain.GetEntries(*args, **kwargs) def GetEntriesFast(self, *args, **kwargs): return self._tchain.GetEntriesFast(*args, **kwargs) def reset(self): """ Reset the chain to the first file Note: not valid when in queue mode """ super(TreeChain, self).reset() self.curr_file_idx = 0 def __len__(self): return len(self._files) def _next_file(self): if self.curr_file_idx >= len(self._files): return None filename = self._files[self.curr_file_idx] nfiles_remaining = len(self._files) - self.curr_file_idx log.info("{0:d} file{1} remaining".format( nfiles_remaining, 's' if nfiles_remaining > 1 else '')) self.curr_file_idx += 1 return filename class TreeQueue(BaseTreeChain): """ A chain of files in a multiprocessing Queue. Note that asking for the number of files in the queue with len(treequeue) can be unreliable. Also, methods not overridden by TreeQueue will always be called on the current tree, so GetEntries will give you the number of entries in the current tree. """ SENTINEL = None def __init__(self, name, files, **kwargs): # multiprocessing.queues d.n.e. until one has been created multiprocessing.Queue() if not isinstance(files, multiprocessing.queues.Queue): raise TypeError("files must be a multiprocessing.Queue") self._files = files super(TreeQueue, self).__init__(name, **kwargs) def __len__(self): # not reliable return self._files.qsize() def __nonzero__(self): # not reliable return not self._files.empty() __bool__ = __nonzero__ def _next_file(self): filename = self._files.get() if filename == self.SENTINEL: return None return filename
bsd-3-clause
24da8dd1b849509f3d9ffedd9fe9a6c8
32.517241
79
0.528713
4.421836
false
false
false
false
rootpy/rootpy
rootpy/stats/pdf.py
3
1145
from __future__ import absolute_import import ROOT from . import log; log = log[__name__] from .. import QROOT, asrootpy from ..base import NamedObject from .value import AbsArg __all__ = [ 'Simultaneous', 'AddPdf', 'ProdPdf', ] class Simultaneous(NamedObject, AbsArg, QROOT.RooSimultaneous): _ROOT = QROOT.RooSimultaneous def __iter__(self): iterator = self.indexCat().typeIterator() category = iterator.Next() while category: yield asrootpy(category) category = iterator.Next() def getPdf(self, category): if isinstance(category, ROOT.RooCatType): category = category.GetName() return asrootpy(super(Simultaneous, self).getPdf(category)) def pdf(self, category): return self.getPdf(category) def indexCat(self): return asrootpy(super(Simultaneous, self).indexCat()) @property def index_category(self): return self.indexCat() class AddPdf(NamedObject, AbsArg, QROOT.RooAddPdf): _ROOT = QROOT.RooAddPdf class ProdPdf(NamedObject, AbsArg, QROOT.RooProdPdf): _ROOT = QROOT.RooProdPdf
bsd-3-clause
0214bee32504f13e7685316718940531
22.367347
67
0.655895
3.480243
false
false
false
false
rootpy/rootpy
rootpy/io/tests/test_pickler.py
3
1942
""" Tests for the file module. """ from rootpy.io import root_open, TemporaryFile from rootpy.io.pickler import load, dump from rootpy.plotting import Hist import random import tempfile from nose.tools import assert_equal, assert_true, assert_false def test_pickler(): hlist = list() for i in range(10): hlist.append(Hist(10, 0, 10)) with TemporaryFile() as tmpfile: dump(hlist, tmpfile) hlist_out = load(tmpfile) assert_equal([h.name for h in hlist_out], [h.name for h in hlist]) hdict = dict() for i in range(100): hist = Hist(10, 0, 1, type=random.choice('CSIFD')) hdict[hist.name] = hist with TemporaryFile() as tmpfile: rdir = tmpfile.mkdir('pickle') dump(hdict, rdir) hdict_out = load(rdir) assert_equal(len(hdict_out), 100) for name, hist in hdict_out.items(): assert_equal(name, hist.name) assert_equal(hist.TYPE, hdict[hist.name].TYPE) def test_pickler_proxy(): h = Hist(5, 0, 1, name='hist') f = tempfile.NamedTemporaryFile(suffix='.root') with root_open(f.name, 'recreate') as outfile: dump([h], outfile) class IsCalled(object): def __init__(self, func): self.func = func self.called = False def __call__(self, path): if path != '_pickle;1': self.called = True return self.func(path) with root_open(f.name) as infile: infile.Get = IsCalled(infile.Get) hlist = load(infile, use_proxy=False) assert_true(infile.Get.called) with root_open(f.name) as infile: infile.Get = IsCalled(infile.Get) hlist = load(infile, use_proxy=True) assert_false(infile.Get.called) assert_equal(hlist[0].name, 'hist') assert_true(infile.Get.called) f.close() if __name__ == "__main__": import nose nose.runmodule()
bsd-3-clause
3b1711f7bc8506bdbc7cd18940b2c10b
25.972222
74
0.595778
3.425044
false
false
false
false
rootpy/rootpy
rootpy/plotting/base.py
3
37647
""" This module contains base classes defining core funcionality """ from __future__ import absolute_import from functools import wraps import warnings import sys from .. import asrootpy, ROOT from ..decorators import chainable from ..memory.keepalive import keepalive from ..extern.six import string_types __all__ = [ 'dim', 'Plottable', ] def dim(thing): if hasattr(thing.__class__, 'DIM'): return thing.__class__.DIM elif hasattr(thing, '__dim__'): return thing.__dim__() elif hasattr(thing, 'GetDimension'): return thing.GetDimension() else: raise TypeError( "Unable to determine dimensionality of " "object of type {0}".format(type(thing))) class Plottable(object): """ This is a mixin to provide additional attributes for plottable classes and to override ROOT TAttXXX and Draw methods. """ EXTRA_ATTRS = { 'norm': None, 'drawstyle': '', 'legendstyle': 'P', 'integermode': False, 'visible': True, 'inlegend': True, } EXTRA_ATTRS_DEPRECATED = { 'format': 'drawstyle', 'intmode': 'integermode', } EXTRA_SETTERS = [ 'color', ] # TODO: respect current TStyle DEFAULT_DECOR = { 'markerstyle': 'circle', 'markercolor': 'black', 'markersize': 1, 'fillcolor': 'white', 'fillstyle': 'hollow', 'linecolor': 'black', 'linestyle': 'solid', 'linewidth': 1, } @classmethod def _get_attr_depr(cls, depattr, newattr): def f(self): warnings.warn( "`{0}` is deprecated and will be removed in " "future versions. Use `{1}` instead".format( depattr, newattr), DeprecationWarning) return getattr(self, newattr) return f @classmethod def _set_attr_depr(cls, depattr, newattr): def f(self, value): warnings.warn( "`{0}` is deprecated and will be removed in " "future versions. Use `{1}` instead".format( depattr, newattr), DeprecationWarning) setattr(self, newattr, value) return f def _post_init(self, **kwargs): self._clone_post_init(**kwargs) def _clone_post_init(self, obj=None, **kwargs): """ obj must be another Plottable instance. obj is used by Clone to properly transfer all attributes onto this object. """ # Initialize the extra attributes if obj is None or obj is self: # We must be asrootpy-ing a ROOT object # or freshly init-ing a rootpy object for attr, value in Plottable.EXTRA_ATTRS.items(): # Use the default value setattr(self, attr, value) else: for attr, value in Plottable.EXTRA_ATTRS.items(): setattr(self, attr, getattr(obj, attr)) # Create aliases from deprecated to current attributes for depattr, newattr in Plottable.EXTRA_ATTRS_DEPRECATED.items(): setattr(Plottable, depattr, property( fget=Plottable._get_attr_depr(depattr, newattr), fset=Plottable._set_attr_depr(depattr, newattr))) if obj is None or obj is self: # We must be asrootpy-ing a ROOT object # or freshly init-ing a rootpy object # Initialize style attrs to style of TObject if isinstance(self, ROOT.TAttLine): self._linecolor = Color(ROOT.TAttLine.GetLineColor(self)) self._linestyle = LineStyle(ROOT.TAttLine.GetLineStyle(self)) self._linewidth = ROOT.TAttLine.GetLineWidth(self) else: # HistStack self._linecolor = Color(Plottable.DEFAULT_DECOR['linecolor']) self._linestyle = LineStyle(Plottable.DEFAULT_DECOR['linestyle']) self._linewidth = Plottable.DEFAULT_DECOR['linewidth'] if isinstance(self, ROOT.TAttFill): self._fillcolor = Color(ROOT.TAttFill.GetFillColor(self)) self._fillstyle = FillStyle(ROOT.TAttFill.GetFillStyle(self)) else: # HistStack self._fillcolor = Color(Plottable.DEFAULT_DECOR['fillcolor']) self._fillstyle = FillStyle(Plottable.DEFAULT_DECOR['fillstyle']) if isinstance(self, ROOT.TAttMarker): self._markercolor = Color(ROOT.TAttMarker.GetMarkerColor(self)) self._markerstyle = MarkerStyle(ROOT.TAttMarker.GetMarkerStyle(self)) self._markersize = ROOT.TAttMarker.GetMarkerSize(self) else: # HistStack self._markercolor = Color(Plottable.DEFAULT_DECOR['markercolor']) self._markerstyle = MarkerStyle(Plottable.DEFAULT_DECOR['markerstyle']) self._markersize = Plottable.DEFAULT_DECOR['markersize'] if obj is None: # Populate defaults if we are not asrootpy-ing existing object decor = dict(**Plottable.DEFAULT_DECOR) decor.update(Plottable.EXTRA_ATTRS) if 'color' in kwargs: decor.pop('linecolor', None) decor.pop('fillcolor', None) decor.pop('markercolor', None) decor.update(kwargs) self.decorate(**decor) else: # Initialize style attrs to style of the other object if isinstance(obj, ROOT.TAttLine): self.SetLineColor(obj.GetLineColor()) self.SetLineStyle(obj.GetLineStyle()) self.SetLineWidth(obj.GetLineWidth()) if isinstance(obj, ROOT.TAttFill): self.SetFillColor(obj.GetFillColor()) self.SetFillStyle(obj.GetFillStyle()) if isinstance(obj, ROOT.TAttMarker): self.SetMarkerColor(obj.GetMarkerColor()) self.SetMarkerStyle(obj.GetMarkerStyle()) self.SetMarkerSize(obj.GetMarkerSize()) if kwargs: self.decorate(**kwargs) #TODO: @chainable def decorate(self, other=None, **kwargs): """ Apply style options to a Plottable object. Returns a reference to self. """ if 'color' in kwargs: incompatible = [] for othercolor in ('linecolor', 'fillcolor', 'markercolor'): if othercolor in kwargs: incompatible.append(othercolor) if incompatible: raise ValueError( "Setting both the `color` and the `{0}` attribute{1} " "is ambiguous. Please set only one.".format( ', '.join(incompatible), 's' if len(incompatible) != 1 else '')) if other is not None: decor = other.decorators if 'color' in kwargs: decor.pop('linecolor', None) decor.pop('fillcolor', None) decor.pop('markercolor', None) decor.update(kwargs) kwargs = decor for key, value in kwargs.items(): if key in Plottable.EXTRA_ATTRS_DEPRECATED: newkey = Plottable.EXTRA_ATTRS_DEPRECATED[key] warnings.warn( "`{0}` is deprecated and will be removed in " "future versions. Use `{1}` instead".format( key, newkey), DeprecationWarning) key = newkey if key in Plottable.EXTRA_ATTRS: setattr(self, key, value) elif key == 'markerstyle': self.SetMarkerStyle(value) elif key == 'markercolor': self.SetMarkerColor(value) elif key == 'markersize': self.SetMarkerSize(value) elif key == 'fillcolor': self.SetFillColor(value) elif key == 'fillstyle': self.SetFillStyle(value) elif key == 'linecolor': self.SetLineColor(value) elif key == 'linestyle': self.SetLineStyle(value) elif key == 'linewidth': self.SetLineWidth(value) elif key == 'color': self.SetColor(value) else: raise AttributeError( "unknown decoration attribute: `{0}`".format(key)) return self @property def decorators(self): return { "norm": self.norm, "drawstyle": self.drawstyle, "legendstyle": self.legendstyle, "integermode": self.integermode, "visible": self.visible, "inlegend": self.inlegend, "markercolor": self.GetMarkerColor(), "markerstyle": self.GetMarkerStyle(), "markersize": self.GetMarkerSize(), "fillcolor": self.GetFillColor(), "fillstyle": self.GetFillStyle(), "linecolor": self.GetLineColor(), "linestyle": self.GetLineStyle(), "linewidth": self.GetLineWidth(), } def SetLineColor(self, color): """ *color* may be any color understood by ROOT or matplotlib. For full documentation of accepted *color* arguments, see :class:`rootpy.plotting.style.Color`. """ self._linecolor = Color(color) if isinstance(self, ROOT.TAttLine): ROOT.TAttLine.SetLineColor(self, self._linecolor('root')) def GetLineColor(self, mode=None): """ *mode* may be 'root', 'mpl', or None to return the ROOT, matplotlib, or input value. """ return self._linecolor(mode) @property def linecolor(self): return self.GetLineColor() @linecolor.setter def linecolor(self, color): self.SetLineColor(color) def SetLineStyle(self, style): """ *style* may be any line style understood by ROOT or matplotlib. For full documentation of accepted *style* arguments, see :class:`rootpy.plotting.style.LineStyle`. """ self._linestyle = LineStyle(style) if isinstance(self, ROOT.TAttLine): ROOT.TAttLine.SetLineStyle(self, self._linestyle('root')) def GetLineStyle(self, mode=None): """ *mode* may be 'root', 'mpl', or None to return the ROOT, matplotlib, or input value. """ return self._linestyle(mode) @property def linestyle(self): return self.GetLineStyle() @linestyle.setter def linestyle(self, style): self.SetLineStyle(style) def SetLineWidth(self, width): if isinstance(self, ROOT.TAttLine): ROOT.TAttLine.SetLineWidth(self, width) else: self._linewidth = width def GetLineWidth(self): if isinstance(self, ROOT.TAttLine): return ROOT.TAttLine.GetLineWidth(self) else: return self._linewidth @property def linewidth(self): return self.GetLineWidth() @linewidth.setter def linewidth(self, width): self.SetLineWidth(width) def SetFillColor(self, color): """ *color* may be any color understood by ROOT or matplotlib. For full documentation of accepted *color* arguments, see :class:`rootpy.plotting.style.Color`. """ self._fillcolor = Color(color) if isinstance(self, ROOT.TAttFill): ROOT.TAttFill.SetFillColor(self, self._fillcolor('root')) def GetFillColor(self, mode=None): """ *mode* may be 'root', 'mpl', or None to return the ROOT, matplotlib, or input value. """ return self._fillcolor(mode) @property def fillcolor(self): return self.GetFillColor() @fillcolor.setter def fillcolor(self, color): self.SetFillColor(color) def SetFillStyle(self, style): """ *style* may be any fill style understood by ROOT or matplotlib. For full documentation of accepted *style* arguments, see :class:`rootpy.plotting.style.FillStyle`. """ self._fillstyle = FillStyle(style) if isinstance(self, ROOT.TAttFill): ROOT.TAttFill.SetFillStyle(self, self._fillstyle('root')) def GetFillStyle(self, mode=None): """ *mode* may be 'root', 'mpl', or None to return the ROOT, matplotlib, or input value. """ return self._fillstyle(mode) @property def fillstyle(self): return self.GetFillStyle() @fillstyle.setter def fillstyle(self, style): self.SetFillStyle(style) def SetMarkerColor(self, color): """ *color* may be any color understood by ROOT or matplotlib. For full documentation of accepted *color* arguments, see :class:`rootpy.plotting.style.Color`. """ self._markercolor = Color(color) if isinstance(self, ROOT.TAttMarker): ROOT.TAttMarker.SetMarkerColor(self, self._markercolor('root')) def GetMarkerColor(self, mode=None): """ *mode* may be 'root', 'mpl', or None to return the ROOT, matplotlib, or input value. """ return self._markercolor(mode) @property def markercolor(self): return self.GetMarkerColor() @markercolor.setter def markercolor(self, color): self.SetMarkerColor(color) def SetMarkerStyle(self, style): """ *style* may be any marker style understood by ROOT or matplotlib. For full documentation of accepted *style* arguments, see :class:`rootpy.plotting.style.MarkerStyle`. """ self._markerstyle = MarkerStyle(style) if isinstance(self, ROOT.TAttMarker): ROOT.TAttMarker.SetMarkerStyle(self, self._markerstyle('root')) def GetMarkerStyle(self, mode=None): """ *mode* may be 'root', 'mpl', or None to return the ROOT, matplotlib, or input value. """ return self._markerstyle(mode) @property def markerstyle(self): return self.GetMarkerStyle() @markerstyle.setter def markerstyle(self, style): self.SetMarkerStyle(style) def SetMarkerSize(self, size): if isinstance(self, ROOT.TAttMarker): ROOT.TAttMarker.SetMarkerSize(self, size) else: self._markersize = size def GetMarkerSize(self): if isinstance(self, ROOT.TAttMarker): return ROOT.TAttMarker.GetMarkerSize(self) else: return self._markersize @property def markersize(self): return self.GetMarkerSize() @markersize.setter def markersize(self, size): self.SetMarkerSize(size) def SetColor(self, color): """ *color* may be any color understood by ROOT or matplotlib. Set all color attributes with one method call. For full documentation of accepted *color* arguments, see :class:`rootpy.plotting.style.Color`. """ self.SetFillColor(color) self.SetLineColor(color) self.SetMarkerColor(color) def GetColor(self): return self.GetMarkerColor(), self.GetLineColor(), self.GetFillColor() @property def color(self): return self.GetColor() @color.setter def color(self, color): self.SetColor(color) @property def xaxis(self): return asrootpy(self.GetXaxis()) @property def yaxis(self): return asrootpy(self.GetYaxis()) @property def zaxis(self): return asrootpy(self.GetZaxis()) def Draw(self, *args, **kwargs): """ Parameters ---------- args : positional arguments Positional arguments are passed directly to ROOT's Draw kwargs : keyword arguments If keyword arguments are present, then a clone is drawn instead with DrawCopy, where the name, title, and style attributes are taken from ``kwargs``. Returns ------- If ``kwargs`` is not empty and a clone is drawn, then the clone is returned, otherwise None is returned. """ if kwargs: return self.DrawCopy(*args, **kwargs) pad = ROOT.gPad own_pad = False if not pad: # avoid circular import by delaying import until needed here from .canvas import Canvas pad = Canvas() own_pad = True if self.visible: if self.drawstyle: self.__class__.__bases__[-1].Draw(self, " ".join((self.drawstyle, ) + args)) else: self.__class__.__bases__[-1].Draw(self, " ".join(args)) pad.Modified() pad.Update() if own_pad: keepalive(self, pad) def DrawCopy(self, *args, **kwargs): """ Parameters ---------- args : positional arguments Positional arguments are passed directly to ROOT's Draw kwargs : keyword arguments The name, title, and style attributes of the clone are taken from ``kwargs``. Returns ------- The clone. """ copy = self.Clone(**kwargs) copy.Draw(*args) return copy class _StyleContainer(object): """ Base class for grouping together an input style with ROOT and matplotlib styles. """ def __init__(self, value, function): self._input = value self._root = function(value, 'root') try: self._mpl = function(value, 'mpl') except ValueError: self._mpl = self._root def __call__(self, output_type=None): if not output_type: output_type = 'input' return getattr(self, '_' + output_type) def __repr__(self): return str(self._input) ############################## #### Markers ################# markerstyles_root2mpl = { 1: '.', 2: '+', 3: '*', 4: 'o', 5: 'x', 20: 'o', 21: 's', 22: '^', 23: 'v', 24: 'o', 25: 's', 26: '^', 27: 'd', 28: '+', 29: '*', 30: '*', 31: '*', 32: 'v', 33: 'D', 34: '+', } for i in range(6, 20): markerstyles_root2mpl[i] = '.' markerstyles_mpl2root = { '.': 1, ',': 1, 'o': 4, 'v': 23, '^': 22, '<': 23, '>': 22, '1': 23, '2': 22, '3': 23, '4': 22, 's': 25, 'p': 25, '*': 3, 'h': 25, 'H': 25, '+': 2, 'x': 5, 'D': 33, 'd': 27, '|': 2, '_': 2, 0: 1, # TICKLEFT 1: 1, # TICKRIGHT 2: 1, # TICKUP 3: 1, # TICKDOWN 4: 1, # CARETLEFT 5: 1, # CARETRIGHT 6: 1, # CARETUP 7: 1, # CARETDOWN 'None': '.', ' ': '.', '': '.', } markerstyles_text2root = { "smalldot": 6, "mediumdot": 7, "largedot": 8, "dot": 9, "circle": 20, "square": 21, "triangle": 22, "triangleup": 22, "triangledown": 23, "opencircle": 24, "opensquare": 25, "opentriangle": 26, "opendiamond": 27, "diamond": 33, "opencross": 28, "cross": 34, "openstar": 29, "fullstar": 30, "star": 29, } def convert_markerstyle(inputstyle, mode, inputmode=None): """ Convert *inputstyle* to ROOT or matplotlib format. Output format is determined by *mode* ('root' or 'mpl'). The *inputstyle* may be a ROOT marker style, a matplotlib marker style, or a description such as 'star' or 'square'. """ mode = mode.lower() if mode not in ('mpl', 'root'): raise ValueError("`{0}` is not valid `mode`".format(mode)) if inputmode is None: if inputstyle in markerstyles_root2mpl: inputmode = 'root' elif inputstyle in markerstyles_mpl2root or '$' in str(inputstyle): inputmode = 'mpl' elif inputstyle in markerstyles_text2root: inputmode = 'root' inputstyle = markerstyles_text2root[inputstyle] else: raise ValueError( "`{0}` is not a valid `markerstyle`".format(inputstyle)) if inputmode == 'root': if inputstyle not in markerstyles_root2mpl: raise ValueError( "`{0}` is not a valid ROOT `markerstyle`".format( inputstyle)) if mode == 'root': return inputstyle return markerstyles_root2mpl[inputstyle] else: if '$' in str(inputstyle): if mode == 'root': return 1 else: return inputstyle if inputstyle not in markerstyles_mpl2root: raise ValueError( "`{0}` is not a valid matplotlib `markerstyle`".format( inputstyle)) if mode == 'mpl': return inputstyle return markerstyles_mpl2root[inputstyle] class MarkerStyle(_StyleContainer): """ Container for grouping together ROOT and matplotlib marker styles. The *style* argument to the constructor may be a ROOT marker style, a matplotlib marker style, or one of the following descriptions: """ __doc__ = __doc__[:__doc__.rfind('\n') + 1] __doc__ += '\n'.join([" '{0}'".format(x) for x in markerstyles_text2root]) if sys.version_info[0] < 3: del x __doc__ += """ Examples -------- >>> style = MarkerStyle('opentriangle') >>> style('root') 26 >>> style('mpl') '^' """ def __init__(self, style): _StyleContainer.__init__(self, style, convert_markerstyle) ############################## #### Lines ################### linestyles_root2mpl = { 1: 'solid', 2: 'dashed', 3: 'dotted', 4: 'dashdot', 5: 'dashdot', 6: 'dashdot', 7: 'dashed', 8: 'dashdot', 9: 'dashed', 10: 'dashdot', } linestyles_mpl2root = { 'solid': 1, 'dashed': 2, 'dotted': 3, 'dashdot': 4, } linestyles_text2root = { 'solid': 1, 'dashed': 2, 'dotted': 3, 'dashdot': 4, 'longdashdot': 5, 'longdashdotdotdot': 6, 'longdash': 7, 'longdashdotdot': 8, 'verylongdash': 9, 'verylongdashdot': 10 } def convert_linestyle(inputstyle, mode, inputmode=None): """ Convert *inputstyle* to ROOT or matplotlib format. Output format is determined by *mode* ('root' or 'mpl'). The *inputstyle* may be a ROOT line style, a matplotlib line style, or a description such as 'solid' or 'dotted'. """ mode = mode.lower() if mode not in ('mpl', 'root'): raise ValueError( "`{0}` is not a valid `mode`".format(mode)) try: inputstyle = int(inputstyle) if inputstyle < 1: inputstyle = 1 except (TypeError, ValueError): pass if inputmode is None: if inputstyle in linestyles_root2mpl: inputmode = 'root' elif inputstyle in linestyles_mpl2root: inputmode = 'mpl' elif inputstyle in linestyles_text2root: inputmode = 'root' inputstyle = linestyles_text2root[inputstyle] else: raise ValueError( "`{0}` is not a valid `linestyle`".format( inputstyle)) if inputmode == 'root': if inputstyle not in linestyles_root2mpl: raise ValueError( "`{0}` is not a valid ROOT `linestyle`".format( inputstyle)) if mode == 'root': return inputstyle return linestyles_root2mpl[inputstyle] else: if inputstyle not in linestyles_mpl2root: raise ValueError( "`{0}` is not a valid matplotlib `linestyle`".format( inputstyle)) if mode == 'mpl': return inputstyle return linestyles_mpl2root[inputstyle] class LineStyle(_StyleContainer): """ Container for grouping together ROOT and matplotlib line styles. The *style* argument to the constructor may be a ROOT line style, a matplotlib line style, or one of the following descriptions: """ __doc__ = __doc__[:__doc__.rfind('\n') + 1] __doc__ += '\n'.join([" '{0}'".format(x) for x in linestyles_text2root]) if sys.version_info[0] < 3: del x __doc__ += """ Examples -------- >>> style = LineStyle('verylongdashdot') >>> style('root') 10 >>> style('mpl') 'dashdot' """ def __init__(self, style): _StyleContainer.__init__(self, style, convert_linestyle) ############################## #### Fills ################### fillstyles_root2mpl = { 0: None, 1001: None, 3003: '.', 3345: '\\', 3354: '/', 3006: '|', 3007: '-', 3011: '*', 3012: 'o', 3013: 'x', 3019: 'O', } fillstyles_mpl2root = {} for key, value in fillstyles_root2mpl.items(): fillstyles_mpl2root[value] = key fillstyles_mpl2root[None] = 0 fillstyles_text2root = { 'hollow': 0, 'none': 0, 'solid': 1001, } def convert_fillstyle(inputstyle, mode, inputmode=None): """ Convert *inputstyle* to ROOT or matplotlib format. Output format is determined by *mode* ('root' or 'mpl'). The *inputstyle* may be a ROOT fill style, a matplotlib hatch style, None, 'none', 'hollow', or 'solid'. """ mode = mode.lower() if mode not in ('mpl', 'root'): raise ValueError("`{0}` is not a valid `mode`".format(mode)) if inputmode is None: try: # inputstyle is a ROOT linestyle inputstyle = int(inputstyle) inputmode = 'root' except (TypeError, ValueError): if inputstyle is None: inputmode = 'mpl' elif inputstyle in fillstyles_text2root: inputmode = 'root' inputstyle = fillstyles_text2root[inputstyle] elif inputstyle[0] in fillstyles_mpl2root: inputmode = 'mpl' else: raise ValueError( "`{0}` is not a valid `fillstyle`".format(inputstyle)) if inputmode == 'root': if mode == 'root': return inputstyle if inputstyle in fillstyles_root2mpl: return fillstyles_root2mpl[inputstyle] raise ValueError( "`{0}` is not a valid `fillstyle`".format(inputstyle)) else: if inputstyle is not None and inputstyle[0] not in fillstyles_mpl2root: raise ValueError( "`{0}` is not a valid matplotlib `fillstyle`".format( inputstyle)) if mode == 'mpl': return inputstyle if inputstyle is None: return fillstyles_mpl2root[inputstyle] return fillstyles_mpl2root[inputstyle[0]] class FillStyle(_StyleContainer): """ Container for grouping together ROOT and matplotlib fill styles. The *style* argument to the constructor may be a ROOT fill style, a matplotlib fill style, or one of the following descriptions: """ __doc__ = __doc__[:__doc__.rfind('\n') + 1] __doc__ += '\n'.join([" '{0}'".format(x) for x in fillstyles_text2root]) if sys.version_info[0] < 3: del x __doc__ += """ For an input value of 'solid', the matplotlib hatch value will be set to None, which is the same value as for 'hollow'. The root2matplotlib functions will all check the ROOT value to see whether to make the fill solid or hollow. Examples -------- >>> style = FillStyle('hollow') >>> style('root') 0 >>> print style('mpl') None """ def __init__(self, style): _StyleContainer.__init__(self, style, convert_fillstyle) ############################## #### Colors ################## _cnames = { 'r' : '#FF0000', #@IgnorePep8 'g' : '#00FF00', 'b' : '#0000FF', 'c' : '#00BFBF', 'm' : '#BF00BF', 'y' : '#BFBF00', 'k' : '#000000', 'w' : '#FFFFFF', 'aliceblue' : '#F0F8FF', 'antiquewhite' : '#FAEBD7', 'aqua' : '#00FFFF', 'aquamarine' : '#7FFFD4', 'azure' : '#F0FFFF', 'beige' : '#F5F5DC', 'bisque' : '#FFE4C4', 'black' : '#000000', 'blanchedalmond' : '#FFEBCD', 'blue' : '#0000FF', 'blueviolet' : '#8A2BE2', 'brown' : '#A52A2A', 'burlywood' : '#DEB887', 'cadetblue' : '#5F9EA0', 'chartreuse' : '#7FFF00', 'chocolate' : '#D2691E', 'coral' : '#FF7F50', 'cornflowerblue' : '#6495ED', 'cornsilk' : '#FFF8DC', 'crimson' : '#DC143C', 'cyan' : '#00FFFF', 'darkblue' : '#00008B', 'darkcyan' : '#008B8B', 'darkgoldenrod' : '#B8860B', 'darkgray' : '#A9A9A9', 'darkgreen' : '#006400', 'darkkhaki' : '#BDB76B', 'darkmagenta' : '#8B008B', 'darkolivegreen' : '#556B2F', 'darkorange' : '#FF8C00', 'darkorchid' : '#9932CC', 'darkred' : '#8B0000', 'darksalmon' : '#E9967A', 'darkseagreen' : '#8FBC8F', 'darkslateblue' : '#483D8B', 'darkslategray' : '#2F4F4F', 'darkturquoise' : '#00CED1', 'darkviolet' : '#9400D3', 'deeppink' : '#FF1493', 'deepskyblue' : '#00BFFF', 'dimgray' : '#696969', 'dodgerblue' : '#1E90FF', 'firebrick' : '#B22222', 'floralwhite' : '#FFFAF0', 'forestgreen' : '#228B22', 'fuchsia' : '#FF00FF', 'gainsboro' : '#DCDCDC', 'ghostwhite' : '#F8F8FF', 'gold' : '#FFD700', 'goldenrod' : '#DAA520', 'gray' : '#808080', 'green' : '#008000', 'greenyellow' : '#ADFF2F', 'honeydew' : '#F0FFF0', 'hotpink' : '#FF69B4', 'indianred' : '#CD5C5C', 'indigo' : '#4B0082', 'ivory' : '#FFFFF0', 'khaki' : '#F0E68C', 'lavender' : '#E6E6FA', 'lavenderblush' : '#FFF0F5', 'lawngreen' : '#7CFC00', 'lemonchiffon' : '#FFFACD', 'lightblue' : '#ADD8E6', 'lightcoral' : '#F08080', 'lightcyan' : '#E0FFFF', 'lightgoldenrodyellow' : '#FAFAD2', 'lightgreen' : '#90EE90', 'lightgrey' : '#D3D3D3', 'lightpink' : '#FFB6C1', 'lightsalmon' : '#FFA07A', 'lightseagreen' : '#20B2AA', 'lightskyblue' : '#87CEFA', 'lightslategray' : '#778899', 'lightsteelblue' : '#B0C4DE', 'lightyellow' : '#FFFFE0', 'lime' : '#00FF00', 'limegreen' : '#32CD32', 'linen' : '#FAF0E6', 'magenta' : '#FF00FF', 'maroon' : '#800000', 'mediumaquamarine' : '#66CDAA', 'mediumblue' : '#0000CD', 'mediumorchid' : '#BA55D3', 'mediumpurple' : '#9370DB', 'mediumseagreen' : '#3CB371', 'mediumslateblue' : '#7B68EE', 'mediumspringgreen' : '#00FA9A', 'mediumturquoise' : '#48D1CC', 'mediumvioletred' : '#C71585', 'midnightblue' : '#191970', 'mintcream' : '#F5FFFA', 'mistyrose' : '#FFE4E1', 'moccasin' : '#FFE4B5', 'navajowhite' : '#FFDEAD', 'navy' : '#000080', 'oldlace' : '#FDF5E6', 'olive' : '#808000', 'olivedrab' : '#6B8E23', 'orange' : '#FFA500', 'orangered' : '#FF4500', 'orchid' : '#DA70D6', 'palegoldenrod' : '#EEE8AA', 'palegreen' : '#98FB98', 'palevioletred' : '#AFEEEE', 'papayawhip' : '#FFEFD5', 'peachpuff' : '#FFDAB9', 'peru' : '#CD853F', 'pink' : '#FFC0CB', 'plum' : '#DDA0DD', 'powderblue' : '#B0E0E6', 'purple' : '#800080', 'red' : '#FF0000', 'rosybrown' : '#BC8F8F', 'royalblue' : '#4169E1', 'saddlebrown' : '#8B4513', 'salmon' : '#FA8072', 'sandybrown' : '#FAA460', 'seagreen' : '#2E8B57', 'seashell' : '#FFF5EE', 'sienna' : '#A0522D', 'silver' : '#C0C0C0', 'skyblue' : '#87CEEB', 'slateblue' : '#6A5ACD', 'slategray' : '#708090', 'snow' : '#FFFAFA', 'springgreen' : '#00FF7F', 'steelblue' : '#4682B4', 'tan' : '#D2B48C', 'teal' : '#008080', 'thistle' : '#D8BFD8', 'tomato' : '#FF6347', 'turquoise' : '#40E0D0', 'violet' : '#EE82EE', 'wheat' : '#F5DEB3', 'white' : '#FFFFFF', 'whitesmoke' : '#F5F5F5', 'yellow' : '#FFFF00', 'yellowgreen' : '#9ACD32', } def convert_color(color, mode): """ Convert *color* to a TColor if *mode='root'* or to (r,g,b) if 'mpl'. The *color* argument can be a ROOT TColor or color index, an *RGB* or *RGBA* sequence or a string in any of several forms: 1) a letter from the set 'rgbcmykw' 2) a hex color string, like '#00FFFF' 3) a standard name, like 'aqua' 4) a float, like '0.4', indicating gray on a 0-1 scale if *arg* is *RGBA*, the transparency value will be ignored. """ mode = mode.lower() if mode not in ('mpl', 'root'): raise ValueError( "`{0}` is not a valid `mode`".format(mode)) try: # color is an r,g,b tuple color = tuple([float(x) for x in color[:3]]) if max(color) > 1.: color = tuple([x / 255. for x in color]) if mode == 'root': return ROOT.TColor.GetColor(*color) return color except (ValueError, TypeError): pass if isinstance(color, string_types): if color in _cnames: # color is a matplotlib letter or an html color name color = _cnames[color] if color[0] == '#': # color is a hex value color = color.lstrip('#') lv = len(color) color = tuple(int(color[i:i + lv // 3], 16) for i in range(0, lv, lv // 3)) if lv == 3: color = tuple(x * 16 + x for x in color) return convert_color(color, mode) # color is a shade of gray, i.e. '0.3' return convert_color((color, color, color), mode) try: # color is a TColor color = ROOT.TColor(color) color = color.GetRed(), color.GetGreen(), color.GetBlue() return convert_color(color, mode) except (TypeError, ReferenceError): pass try: # color is a ROOT color index if color < 0: color = 0 color = ROOT.gROOT.GetColor(color) # Protect against the case a histogram with a custom color # is saved in a ROOT file if not color: # Just return black color = ROOT.gROOT.GetColor(1) color = color.GetRed(), color.GetGreen(), color.GetBlue() return convert_color(color, mode) except (TypeError, ReferenceError): pass raise ValueError("'{0!s}' is not a valid `color`".format(color)) class Color(_StyleContainer): """ Container for grouping together ROOT and matplotlib colors. The *color* argument to the constructor can be a ROOT TColor or color index. If matplotlib is available, it can also accept an *RGB* or *RGBA* sequence, or a string in any of several forms: 1) a letter from the set 'rgbcmykw' 2) a hex color string, like '#00FFFF' 3) a standard name, like 'aqua' 4) a float, like '0.4', indicating gray on a 0-1 scale if *color* is *RGBA*, the *A* will simply be discarded. Examples -------- >>> color = Color(2) >>> color() 2 >>> color('mpl') (1.0, 0.0, 0.0) >>> color = Color('blue') >>> color('root') 4 >>> color('mpl') (0.0, 0.0, 1.0) >>> color = Color('0.25') >>> color('mpl') (0.25, 0.25, 0.25) >>> color('root') 924 """ def __init__(self, color): _StyleContainer.__init__(self, color, convert_color)
bsd-3-clause
f1a2eee04b07016a556fcb350ea80b9c
30.3725
87
0.51789
3.812354
false
false
false
false
rootpy/rootpy
rootpy/extern/byteplay3/wbyteplay.py
5
31871
# byteplay: CPython assembler/disassembler # Copyright (C) 2006 Noam Raphael | Version: http://code.google.com/p/byteplay # Rewritten 2009 Demur Rumed | Version: http://github.com/serprex/byteplay # Screwed the style over, modified stack logic to be more flexible, updated to Python 3 # # This library is free software; you can redistribute it and/or # modify it under the terms of the GNU Lesser General Public # License as published by the Free Software Foundation; either # version 2.1 of the License, or (at your option) any later version. # # This library is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU # Lesser General Public License for more details. # # You should have received a copy of the GNU Lesser General Public # License along with this library; if not, write to the Free Software # Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA __version__ = '1.0' __all__ = [ 'opmap', 'opname', 'opcodes', 'hasflow', 'stack_effect', 'cmp_op', 'hasarg', 'hasname', 'hasjrel', 'hasjabs', 'hasjump', 'haslocal', 'hascompare', 'hasfree', 'hasconst', 'hascode', 'Opcode', 'SetLineno', 'Label', 'isopcode', 'Code'] from sys import version_info if version_info < (3, 6): raise NotImplementedError("Currently only Python versions >3.5 are supported!") import opcode from dis import findlabels from types import CodeType from enum import Enum class Opcode(int): __str__ = __repr__ = lambda s: opname[s] opmap = {name.replace('+', '_'): Opcode(code) for name, code in opcode.opmap.items()} opname = {code: name for name, code in opmap.items()} opcodes = set(opname) for cmp_op, hasname in opmap.items(): globals()[cmp_op] = hasname __all__.append(cmp_op) cmp_op = opcode.cmp_op hasarg = {x for x in opcodes if x >= opcode.HAVE_ARGUMENT} hasconst = {Opcode(x) for x in opcode.hasconst} hasname = {Opcode(x) for x in opcode.hasname} hasjrel = {Opcode(x) for x in opcode.hasjrel} hasjabs = {Opcode(x) for x in opcode.hasjabs} hasjump = hasjabs | hasjrel haslocal = {Opcode(x) for x in opcode.haslocal} hascompare = {Opcode(x) for x in opcode.hascompare} hasfree = {Opcode(x) for x in opcode.hasfree} hascode = {MAKE_FUNCTION} STOP_CODE = -1 import dis # Fix bug in Python 3.6.0 (fixed in 3.6.1) if (3, 6, 0) <= version_info < (3, 6, 1): def stack_effect(o, arg): return (dis.stack_effect(o, arg) if o != CALL_FUNCTION_EX else -2 if arg else -1) else: stack_effect = dis.stack_effect hasflow = hasjump | { POP_BLOCK, END_FINALLY, BREAK_LOOP, RETURN_VALUE, RAISE_VARARGS, STOP_CODE, POP_EXCEPT, WITH_CLEANUP_START, WITH_CLEANUP_FINISH, SETUP_ASYNC_WITH} coroutine_opcodes = {GET_AWAITABLE, GET_AITER, GET_ANEXT, BEFORE_ASYNC_WITH, SETUP_ASYNC_WITH} class Label: pass class SetLinenoType: def __repr__(self): return 'SetLineno' SetLineno = SetLinenoType() def isopcode(x): return x is not SetLineno and not isinstance(x, Label) # Flags for codeobject.co_flags, taken from Include/code.h, other flags are no longer used CO_OPTIMIZED = 0x0001 CO_NEWLOCALS = 0x0002 CO_VARARGS = 0x0004 CO_VARKEYWORDS = 0x0008 CO_NESTED = 0x0010 CO_GENERATOR = 0x0020 CO_NOFREE = 0x0040 CO_COROUTINE = 0x0080 CO_ITERABLE_COROUTINE = 0x0100 CO_ASYNC_GENERATOR = 0x0200 CO_FUTURE_BARRY_AS_BDFL = 0x40000 CO_FUTURE_GENERATOR_STOP = 0x80000 class Code(object): """An object which holds all the information which a Python code object holds, but in an easy-to-play-with representation The attributes are: Affecting action code - list of 2-tuples: the code freevars - list of strings: the free vars of the code (those are names of variables created in outer functions and used in the function) args - list of strings: the arguments of the code kwonly - number of keyword only arguments varargs - boolean: Does args end with a '*args' argument varkwargs - boolean: Does args end with a '**kwargs' argument newlocals - boolean: Should a new local namespace be created (True in functions, False for module and exec code) force_generator - set CO_GENERATOR in co_flags for generator Code objects without generator-specific code Python 3.5: force_coroutine - set CO_COROUTINE in co_flags for coroutine Code objects (native coroutines) without coroutine-specific code force_iterable_coroutine - set CO_ITERABLE_COROUTINE in co_flags for generator-based coroutine Code objects future_generator_stop - set CO_FUTURE_GENERATOR_STOP flag (see PEP-479) Python 3.6: force_async_generator - set CO_ASYNC_GENERATOR in co_flags Not affecting action name - string: the name of the code (co_name) filename - string: the file name of the code (co_filename) firstlineno - int: the first line number (co_firstlineno) docstring - string or None: the docstring (the first item of co_consts, if it's str) code is a list of 2-tuples. The first item is an opcode, or SetLineno, or a Label instance. The second item is the argument, if applicable, or None""" def __init__(self, code, freevars, args, kwonly, varargs, varkwargs, newlocals, name, filename, firstlineno, docstring, force_generator=False, *, force_coroutine=False, force_iterable_coroutine=False, force_async_generator=False, future_generator_stop=False): self.code = code self.freevars = freevars self.args = args self.kwonly = kwonly self.varargs = varargs self.varkwargs = varkwargs self.newlocals = newlocals self.name = name self.filename = filename self.firstlineno = firstlineno self.docstring = docstring self.force_generator = force_generator self.force_coroutine = force_coroutine self.force_iterable_coroutine = force_iterable_coroutine self.force_async_generator = force_async_generator self.future_generator_stop = future_generator_stop @staticmethod def _findlinestarts(code): """Find the offsets in a byte code which are start of lines in the source Generate pairs offset,lineno as described in Python/compile.c This is a modified version of dis.findlinestarts, which allows multiplelinestarts with the same line number""" lineno = code.co_firstlineno addr = 0 for byte_incr, line_incr in zip(code.co_lnotab[0::2], code.co_lnotab[1::2]): if byte_incr: yield addr, lineno addr += byte_incr lineno += line_incr yield addr, lineno @classmethod def from_code(cls, co): """Disassemble a Python code object into a Code object""" free_cell_isection = set(co.co_cellvars) & set(co.co_freevars) if free_cell_isection: print(co.co_name + ': has non-empty co.co_cellvars & co.co_freevars', free_cell_isection) return None co_code = co.co_code labels = {addr: Label() for addr in findlabels(co_code)} linestarts = dict(cls._findlinestarts(co)) cellfree = co.co_cellvars + co.co_freevars code = [] extended_arg = 0 is_generator = False is_coroutine = False for i in range(0, len(co_code), 2): if i in labels: code.append((labels[i], None)) if i in linestarts: code.append((SetLineno, linestarts[i])) op = Opcode(co_code[i]) arg = co_code[i+1] | extended_arg if op in hascode: lastop, lastarg = code[-2] if lastop != LOAD_CONST: raise ValueError("%s should be preceded by LOAD_CONST" % op) sub_code = Code.from_code(lastarg) if sub_code is None: print(co.co_name + ': has unexpected subcode block') return None code[-2] = (LOAD_CONST, sub_code) if op == opcode.EXTENDED_ARG: extended_arg = arg << 8 else: if op not in hasarg: code.append((op, None)) continue extended_arg = 0 byteplay_arg = co.co_consts[arg] if op in hasconst else \ co.co_names[arg] if op in hasname else \ labels[arg] if op in hasjabs else \ labels[i + 2 + arg] if op in hasjrel else \ co.co_varnames[arg] if op in haslocal else \ cmp_op[arg] if op in hascompare else \ cellfree[arg] if op in hasfree else \ arg code.append((op, byteplay_arg)) if op == YIELD_VALUE or op == YIELD_FROM: is_generator = True if op in coroutine_opcodes: is_coroutine = True varargs = not not co.co_flags & CO_VARARGS varkwargs = not not co.co_flags & CO_VARKEYWORDS force_coroutine = not is_coroutine and (co.co_flags & CO_COROUTINE) force_iterable_coroutine = co.co_flags & CO_ITERABLE_COROUTINE force_async_generator = co.co_flags & CO_ASYNC_GENERATOR is_generator = False if force_async_generator else is_generator force_generator = not is_generator and (co.co_flags & CO_GENERATOR) assert not (force_coroutine and force_iterable_coroutine) assert not (force_coroutine and force_async_generator) assert not (force_iterable_coroutine and force_async_generator) future_generator_stop = co.co_flags & CO_FUTURE_GENERATOR_STOP return cls(code=code, freevars=co.co_freevars, args=co.co_varnames[:co.co_argcount + varargs + varkwargs + co.co_kwonlyargcount], kwonly=co.co_kwonlyargcount, varargs=varargs, varkwargs=varkwargs, newlocals=not not co.co_flags & CO_NEWLOCALS, name=co.co_name, filename=co.co_filename, firstlineno=co.co_firstlineno, docstring=co.co_consts[0] if co.co_consts and isinstance(co.co_consts[0], str) else None, force_generator=force_generator, force_coroutine=force_coroutine, force_iterable_coroutine=force_iterable_coroutine, force_async_generator=force_async_generator, future_generator_stop=future_generator_stop) def __eq__(self, other): try: if (self.freevars != other.freevars or self.args != other.args or self.kwonly != other.kwonly or self.varargs != other.varargs or self.varkwargs != other.varkwargs or self.newlocals != other.newlocals or self.name != other.name or self.filename != other.filename or self.firstlineno != other.firstlineno or self.docstring != other.docstring or self.force_generator != other.force_generator or len(self.code) != len(other.code)): return False else: if (self.force_coroutine != other.force_coroutine or self.force_iterable_coroutine != other.force_iterable_coroutine or self.future_generator_stop != other.future_generator_stop or self.force_async_generator != other.force_async_generator): return False # This isn't trivial due to labels lmap = {} for (op1, arg1), (op2, arg2) in zip(self.code, other.code): if isinstance(op1, Label): if lmap.setdefault(arg1, arg2) is not arg2: return False else: if op1 != op2: return False if op1 in hasjump: if lmap.setdefault(arg1, arg2) is not arg2: return False elif arg1 != arg2: return False return True except: return False def _compute_stacksize(self, logging=False): code = self.code label_pos = {op[0]: pos for pos, op in enumerate(code) if isinstance(op[0], Label)} # sf_targets are the targets of SETUP_FINALLY opcodes. They are recorded # because they have special stack behaviour. If an exception was raised # in the block pushed by a SETUP_FINALLY opcode, the block is popped # and 3 objects are pushed. On return or continue, the block is popped # and 2 objects are pushed. If nothing happened, the block is popped by # a POP_BLOCK opcode and 1 object is pushed by a (LOAD_CONST, None) # operation # Our solution is to record the stack state of SETUP_FINALLY targets # as having 3 objects pushed, which is the maximum. However, to make # stack recording consistent, the get_next_stacks function will always # yield the stack state of the target as if 1 object was pushed, but # this will be corrected in the actual stack recording sf_targets = {label_pos[arg] for op, arg in code if (op == SETUP_FINALLY or op == SETUP_WITH or op == SETUP_ASYNC_WITH)} states = [None] * len(code) maxsize = 0 class BlockType(Enum): DEFAULT = 0, TRY_FINALLY = 1, TRY_EXCEPT = 2, LOOP_BODY = 3, WITH_BLOCK = 4, EXCEPTION = 5, SILENCED_EXCEPTION_BLOCK = 6, class State: def __init__(self, pos=0, stack=(0,), block_stack=(BlockType.DEFAULT,), log=[]): self._pos = pos self._stack = stack self._block_stack = block_stack self._log = log @property def pos(self): return self._pos @property def stack(self): return self._stack @stack.setter def stack(self, val): self._stack = val def newstack(self, n): if self._stack[-1] < -n: raise ValueError("Popped a non-existing element at %s %s" % (self._pos, code[self._pos - 4: self._pos + 3])) return self._stack[:-1] + (self._stack[-1] + n,) @property def block_stack(self): return self._block_stack @property def log(self): return self._log def newlog(self, msg): if not logging: return None log_msg = str(self._pos) + ": " + msg if self._stack: log_msg += " (on stack: " log_depth = 2 log_depth = min(log_depth, len(self._stack)) for pos in range(-1, -log_depth, -1): log_msg += str(self._stack[pos]) + ", " log_msg += str(self._stack[-log_depth]) log_msg += ")" else: log_msg += " (empty stack)" return [log_msg] + self._log op = [State()] while op: cur_state = op.pop() o = sum(cur_state.stack) if o > maxsize: maxsize = o o, arg = code[cur_state.pos] if isinstance(o, Label): if cur_state.pos in sf_targets: cur_state.stack = cur_state.newstack(5) if states[cur_state.pos] is None: states[cur_state.pos] = cur_state elif states[cur_state.pos].stack != cur_state.stack: check_pos = cur_state.pos + 1 while code[check_pos][0] not in hasflow: check_pos += 1 if code[check_pos][0] not in (RETURN_VALUE, RAISE_VARARGS, STOP_CODE): if cur_state.pos not in sf_targets: raise ValueError("Inconsistent code at %s %s %s\n%s" % (cur_state.pos, cur_state.stack, states[cur_state.pos].stack, code[cur_state.pos - 5:cur_state.pos + 4])) else: # SETUP_FINALLY target inconsistent code! # # Since Python 3.2 assigned exception is cleared at the end of # the except clause (named exception handler). # To perform this CPython (checked in version 3.4.3) adds special # bytecode in exception handler which currently breaks 'regularity' of bytecode. # Exception handler is wrapped in try/finally block and POP_EXCEPT opcode # is inserted before END_FINALLY, as a result cleanup-finally block is executed outside # except handler. It's not a bug, as it doesn't cause any problems during execution, but # it breaks 'regularity' and we can't check inconsistency here. Maybe issue should be # posted to Python bug tracker. pass continue else: continue if o not in (BREAK_LOOP, RETURN_VALUE, RAISE_VARARGS, STOP_CODE): next_pos = cur_state.pos + 1 if not isopcode(o): op += State(next_pos, cur_state.stack, cur_state.block_stack, cur_state.log), elif o not in hasflow: if o in hasarg and not isinstance(arg, int): se = stack_effect(o, 0) else: se = stack_effect(o, arg) log = cur_state.newlog("non-flow command (" + str(o) + ", se = " + str(se) + ")") op += State(next_pos, cur_state.newstack(se), cur_state.block_stack, log), elif o == FOR_ITER: inside_for_log = cur_state.newlog("FOR_ITER (+1)") op += State(label_pos[arg], cur_state.newstack(-1), cur_state.block_stack, cur_state.log),\ State(next_pos, cur_state.newstack(1), cur_state.block_stack, inside_for_log) elif o in (JUMP_FORWARD, JUMP_ABSOLUTE): after_jump_log = cur_state.newlog(str(o)) op += State(label_pos[arg], cur_state.stack, cur_state.block_stack, after_jump_log), elif o in (JUMP_IF_FALSE_OR_POP, JUMP_IF_TRUE_OR_POP): after_jump_log = cur_state.newlog(str(o) + ", jumped") log = cur_state.newlog(str(o) + ", not jumped (-1)") op += State(label_pos[arg], cur_state.stack, cur_state.block_stack, after_jump_log),\ State(next_pos, cur_state.newstack(-1), cur_state.block_stack, log) elif o in {POP_JUMP_IF_TRUE, POP_JUMP_IF_FALSE}: after_jump_log = cur_state.newlog(str(o) + ", jumped (-1)") log = cur_state.newlog(str(o) + ", not jumped (-1)") op += State(label_pos[arg], cur_state.newstack(-1), cur_state.block_stack, after_jump_log),\ State(next_pos, cur_state.newstack(-1), cur_state.block_stack, log) elif o == CONTINUE_LOOP: next_stack, next_block_stack = cur_state.stack, cur_state.block_stack last_popped_block = None while next_block_stack[-1] != BlockType.LOOP_BODY: last_popped_block = next_block_stack[-1] next_stack, next_block_stack = next_stack[:-1], next_block_stack[:-1] if next_stack != cur_state.stack: log = cur_state.newlog("CONTINUE_LOOP, from non-loop block") else: log = cur_state.newlog("CONTINUE_LOOP") jump_to_pos = label_pos[arg] if last_popped_block == BlockType.WITH_BLOCK: next_stack = next_stack[:-1] + (next_stack[-1] - 1,) op += State(jump_to_pos, next_stack, next_block_stack, log), elif o == SETUP_LOOP: inside_loop_log = cur_state.newlog("SETUP_LOOP (+block)") op += State(label_pos[arg], cur_state.stack, cur_state.block_stack, cur_state.log),\ State(next_pos, cur_state.stack + (0,), cur_state.block_stack + (BlockType.LOOP_BODY,), inside_loop_log) elif o == SETUP_EXCEPT: inside_except_log = cur_state.newlog("SETUP_EXCEPT, exception (+6, +block)") inside_try_log = cur_state.newlog("SETUP_EXCEPT, try-block (+block)") op += State(label_pos[arg], cur_state.stack + (6,), cur_state.block_stack + (BlockType.EXCEPTION,), inside_except_log),\ State(next_pos, cur_state.stack + (0,), cur_state.block_stack + (BlockType.TRY_EXCEPT,), inside_try_log) elif o == SETUP_FINALLY: inside_finally_block = cur_state.newlog("SETUP_FINALLY (+1)") inside_try_log = cur_state.newlog("SETUP_FINALLY try-block (+block)") op += State(label_pos[arg], cur_state.newstack(1), cur_state.block_stack, inside_finally_block),\ State(next_pos, cur_state.stack + (0,), cur_state.block_stack + (BlockType.TRY_FINALLY,), inside_try_log) elif o == POP_BLOCK: log = cur_state.newlog("POP_BLOCK (-block)") op += State(next_pos, cur_state.stack[:-1], cur_state.block_stack[:-1], log), elif o == POP_EXCEPT: log = cur_state.newlog("POP_EXCEPT (-block)") op += State(next_pos, cur_state.stack[:-1], cur_state.block_stack[:-1], log), elif o == END_FINALLY: if cur_state.block_stack[-1] == BlockType.SILENCED_EXCEPTION_BLOCK: log = cur_state.newlog("END_FINALLY pop silenced exception block (-block)") op += State(next_pos, cur_state.stack[:-1], cur_state.block_stack[:-1], log), elif cur_state.block_stack[-1] == BlockType.EXCEPTION: # Reraise exception pass else: log = cur_state.newlog("END_FINALLY (-6)") op += State(next_pos, cur_state.newstack(-6), cur_state.block_stack, log), elif o == SETUP_WITH or o == SETUP_ASYNC_WITH: inside_with_block = cur_state.newlog("SETUP_WITH, with-block (+1, +block)") inside_finally_block = cur_state.newlog("SETUP_WITH, finally (+1)") op += State(label_pos[arg], cur_state.newstack(1), cur_state.block_stack, inside_finally_block),\ State(next_pos, cur_state.stack + (1,), cur_state.block_stack + (BlockType.WITH_BLOCK,), inside_with_block) elif o == WITH_CLEANUP_START: # There is special case when 'with' __exit__ function returns True, # that's the signal to silence exception, in this case additional element is pushed # and next END_FINALLY command won't reraise exception. # Emulate this situation on WITH_CLEANUP_START with creating special block which will be # handled differently by WITH_CLEANUP_FINISH and will cause END_FINALLY not to reraise exception. log = cur_state.newlog("WITH_CLEANUP_START (+1)") silenced_exception_log = cur_state.newlog("WITH_CLEANUP_START silenced_exception (+block)") op += State(next_pos, cur_state.newstack(1), cur_state.block_stack, log),\ State(next_pos, cur_state.newstack(-7) + (9,), cur_state.block_stack + (BlockType.SILENCED_EXCEPTION_BLOCK,), silenced_exception_log) elif o == WITH_CLEANUP_FINISH: if cur_state.block_stack[-1] == BlockType.SILENCED_EXCEPTION_BLOCK: # See comment in WITH_CLEANUP_START handler log = cur_state.newlog("WITH_CLEANUP_FINISH silenced_exception (-1)") op += State(next_pos, cur_state.newstack(-1), cur_state.block_stack, log), else: log = cur_state.newlog("WITH_CLEANUP_FINISH (-2)") op += State(next_pos, cur_state.newstack(-2), cur_state.block_stack, log), else: raise ValueError("Unhandled opcode %s" % o) return maxsize + 6 # for exception raise in deepest place def to_code(self, from_function=False): """Assemble a Python code object from a Code object""" num_fastnames = sum(1 for op, arg in self.code if isopcode(op) and op in haslocal) is_function = self.newlocals or num_fastnames > 0 or len(self.args) > 0 nested = is_function and from_function co_flags = {op[0] for op in self.code} if not self.force_async_generator: is_generator = (self.force_generator or (YIELD_VALUE in co_flags or YIELD_FROM in co_flags) ) else: is_generator = False no_free = (not self.freevars) and (not co_flags & hasfree) is_native_coroutine = bool(self.force_coroutine or (co_flags & coroutine_opcodes)) assert not (is_native_coroutine and self.force_iterable_coroutine) assert not (is_native_coroutine and self.force_async_generator) co_flags =\ (not(STORE_NAME in co_flags or LOAD_NAME in co_flags or DELETE_NAME in co_flags)) |\ (self.newlocals and CO_NEWLOCALS) |\ (self.varargs and CO_VARARGS) |\ (self.varkwargs and CO_VARKEYWORDS) |\ (is_generator and CO_GENERATOR) |\ (no_free and CO_NOFREE) |\ (nested and CO_NESTED) co_flags |= (is_native_coroutine and CO_COROUTINE) |\ (self.force_iterable_coroutine and CO_ITERABLE_COROUTINE) |\ (self.future_generator_stop and CO_FUTURE_GENERATOR_STOP) |\ (self.force_async_generator and CO_ASYNC_GENERATOR) co_consts = [self.docstring] co_names = [] co_varnames = list(self.args) co_freevars = tuple(self.freevars) # Find all cellvars beforehand for two reasons # Need the number of them to construct the numeric arg for ops in hasfree # Need to put args which are cells in the beginning of co_cellvars cellvars = {arg for op, arg in self.code if isopcode(op) and op in hasfree and arg not in co_freevars} co_cellvars = [jumps for jumps in self.args if jumps in cellvars] def index(seq, item, eq=True, can_append=True): for i, x in enumerate(seq): if x == item if eq else x is item: return i if can_append: seq.append(item) return len(seq) - 1 else: raise IndexError("Item not found") jumps = [] label_pos = {} lastlineno = self.firstlineno lastlinepos = 0 co_code = bytearray() co_lnotab = bytearray() for i, (op, arg) in enumerate(self.code): if isinstance(op, Label): label_pos[op] = len(co_code) elif op is SetLineno: incr_lineno = arg - lastlineno incr_pos = len(co_code) - lastlinepos lastlineno = arg lastlinepos += incr_pos if incr_lineno != 0 or incr_pos != 0: while incr_pos > 255: co_lnotab += b"\xFF\0" incr_pos -= 255 while incr_lineno > 255: co_lnotab += bytes((incr_pos, 255)) incr_pos = 0 incr_lineno -= 255 if incr_pos or incr_lineno: co_lnotab += bytes((incr_pos, incr_lineno)) elif op == opcode.EXTENDED_ARG: self.code[i + 1][1] |= 1 << 32 else: if op in hasconst: if (isinstance(arg, Code) and i + 2 < len(self.code) and self.code[i + 2][0] in hascode): arg = arg.to_code(from_function=is_function) assert arg is not None arg = index(co_consts, arg, 0) elif op in hasname: arg = index(co_names, arg) elif op in hasjump: jumps.append((len(co_code), arg)) co_code += bytes((0x90, 0, op, 0)) continue elif op in haslocal: arg = index(co_varnames, arg) elif op in hascompare: arg = index(cmp_op, arg, can_append=False) elif op in hasfree: try: arg = index(co_freevars, arg, can_append=False) + len(cellvars) except IndexError: arg = index(co_cellvars, arg) if arg is None: arg = 0 if arg > 0xFFFFFF: co_code += (opcode.EXTENDED_ARG | (arg >> 16 & 0xFF00)).to_bytes(2, "little") if arg > 0xFFFF: co_code += (opcode.EXTENDED_ARG | (arg >> 8 & 0xFF00)).to_bytes(2, "little") if arg > 0xFF: co_code += (opcode.EXTENDED_ARG | (arg & 0xFF00)).to_bytes(2, "little") co_code += (op | (arg & 0xFF) << 8).to_bytes(2, "little") for pos, label in jumps: jump = label_pos[label] if co_code[pos+2] in hasjrel: jump -= pos + 4 if jump > 0xFFFF: raise NotImplementedError("Multiple EXTENDED_ARG jumps not implemented") co_code[pos + 3] = jump & 0xFF co_code[pos + 1] = jump >> 8 & 0xFF co_argcount = len(self.args) - self.varargs - self.varkwargs - self.kwonly co_stacksize = self._compute_stacksize() return CodeType(co_argcount, self.kwonly, len(co_varnames), co_stacksize, co_flags, bytes(co_code), tuple(co_consts), tuple(co_names), tuple(co_varnames), self.filename, self.name, self.firstlineno, bytes(co_lnotab), co_freevars, tuple(co_cellvars))
bsd-3-clause
7be44803d9895f01d70fd61d897f530a
43.637255
159
0.540303
3.980889
false
false
false
false
rootpy/rootpy
rootpy/logger/formatter.py
3
1941
""" Provides a ``CustomFormatter`` and ``CustomColoredFormatter`` which are enable to insert ANSI color codes. """ from __future__ import absolute_import import logging __all__ = [ 'CustomFormatter', 'CustomColoredFormatter', ] # The background is set with 40 plus the number of the color, and the foreground with 30 RED, YELLOW, BLUE, WHITE = 1, 3, 4, 7 # These are the sequences need to get colored ouput RESET_SEQ = "\033[0m" COLOR_SEQ = "\033[1;%dm" BOLD_SEQ = "\033[1m" FORMAT = "{color}{levelname}$RESET:$BOLD{name}$RESET] {message}" def insert_seqs(message): return message.replace("$RESET", RESET_SEQ).replace("$BOLD", BOLD_SEQ) def remove_seqs(message): return message.replace("$RESET", "").replace("$BOLD", "") COLORS = { 'DEBUG' : BLUE, 'INFO' : WHITE, 'WARNING' : YELLOW, 'ERROR' : RED, 'CRITICAL' : RED, } class CustomFormatter(logging.Formatter): def __init__(self, fmt=remove_seqs(FORMAT), datefmt=None): logging.Formatter.__init__(self, fmt, datefmt) def format(self, record): if not hasattr(record, "message"): record.message = record.getMessage() record.asctime = self.formatTime(record, self.datefmt) return self._fmt.format(color="", **record.__dict__) class CustomColoredFormatter(CustomFormatter): def __init__(self, fmt=insert_seqs(FORMAT), datefmt=None, use_color=True): CustomFormatter.__init__(self, fmt, datefmt) self.use_color = use_color def format(self, record): levelname = record.levelname if self.use_color and levelname in COLORS: record.color = COLOR_SEQ % (30 + COLORS[levelname]) else: record.color = "" if not hasattr(record, "message"): record.message = record.getMessage() record.asctime = self.formatTime(record, self.datefmt) return self._fmt.format(**record.__dict__)
bsd-3-clause
9915ed5d2732bba751a6b995e97e4190
29.809524
88
0.634724
3.581181
false
false
false
false
rootpy/rootpy
examples/plotting/plot_matplotlib_hist.py
7
2749
#!/usr/bin/env python """ ===================================== Plot a ROOT histogram with matplotlib ===================================== This example demonstrates how a ROOT histogram can be styled with simple attributes and displayed via ROOT or matplotlib. """ print(__doc__) import ROOT import numpy as np from rootpy.plotting import Hist, HistStack, Legend, Canvas from rootpy.plotting.style import get_style, set_style from rootpy.plotting.utils import draw from rootpy.interactive import wait import rootpy.plotting.root2matplotlib as rplt import matplotlib.pyplot as plt from matplotlib.ticker import AutoMinorLocator, MultipleLocator # set the style style = get_style('ATLAS') style.SetEndErrorSize(3) set_style(style) # set the random seed ROOT.gRandom.SetSeed(42) np.random.seed(42) # signal distribution signal = 126 + 10 * np.random.randn(100) signal_obs = 126 + 10 * np.random.randn(100) # create histograms h1 = Hist(30, 40, 200, title='Background', markersize=0, legendstyle='F') h2 = h1.Clone(title='Signal') h3 = h1.Clone(title='Data', drawstyle='E1 X0', legendstyle='LEP') h3.markersize = 1.2 # fill the histograms with our distributions h1.FillRandom('landau', 1000) map(h2.Fill, signal) h3.FillRandom('landau', 1000) map(h3.Fill, signal_obs) # set visual attributes h1.fillstyle = 'solid' h1.fillcolor = 'green' h1.linecolor = 'green' h1.linewidth = 0 h2.fillstyle = 'solid' h2.fillcolor = 'red' h2.linecolor = 'red' h2.linewidth = 0 stack = HistStack([h1, h2], drawstyle='HIST E1 X0') # plot with ROOT canvas = Canvas(width=700, height=500) draw([stack, h3], xtitle='Mass', ytitle='Events', pad=canvas) # set the number of expected legend entries legend = Legend([h1, h2, h3], leftmargin=0.45, margin=0.3) legend.Draw() label = ROOT.TText(0.3, 0.8, 'ROOT') label.SetTextFont(43) label.SetTextSize(25) label.SetNDC() label.Draw() canvas.Modified() canvas.Update() # plot with matplotlib set_style('ATLAS', mpl=True) fig = plt.figure(figsize=(7, 5), dpi=100) axes = plt.axes() axes.xaxis.set_minor_locator(AutoMinorLocator()) axes.yaxis.set_minor_locator(AutoMinorLocator()) axes.yaxis.set_major_locator(MultipleLocator(20)) rplt.bar(stack, stacked=True, axes=axes) rplt.errorbar(h3, xerr=False, emptybins=False, axes=axes) plt.xlabel('Mass', position=(1., 0.), va='bottom', ha='right') plt.ylabel('Events', position=(0., 1.), va='top', ha='right') axes.xaxis.set_label_coords(1., -0.20) axes.yaxis.set_label_coords(-0.18, 1.) leg = plt.legend() axes.text(0.3, 0.8, 'matplotlib', verticalalignment='center', horizontalalignment='center', transform=axes.transAxes, fontsize=20) if not ROOT.gROOT.IsBatch(): plt.show() # wait for you to close the ROOT canvas before exiting wait(True)
bsd-3-clause
d67ea756524d27d42e049db00decedb6
28.244681
73
0.708985
2.994553
false
false
false
false
rootpy/rootpy
rootpy/plotting/utils.py
3
14154
from __future__ import absolute_import from math import log import operator from .. import ROOT from .canvas import _PadBase from .hist import _Hist, Hist, HistStack from .graph import _Graph1DBase, Graph from ..context import preserve_current_canvas, do_nothing from ..extern.six.moves import range __all__ = [ 'draw', 'get_limits', 'get_band', 'canvases_with', 'find_all_primitives', 'tick_length_pixels', ] def draw(plottables, pad=None, same=False, xaxis=None, yaxis=None, xtitle=None, ytitle=None, xlimits=None, ylimits=None, xdivisions=None, ydivisions=None, logx=False, logy=False, **kwargs): """ Draw a list of histograms, stacks, and/or graphs. Parameters ---------- plottables : Hist, Graph, HistStack, or list of such objects List of objects to draw. pad : Pad or Canvas, optional (default=None) The pad to draw onto. If None then use the current global pad. same : bool, optional (default=False) If True then use 'SAME' draw option for all objects instead of all but the first. Use this option if you are drawing onto a pad that already holds drawn objects. xaxis : TAxis, optional (default=None) Use this x-axis or use the x-axis of the first plottable if None. yaxis : TAxis, optional (default=None) Use this y-axis or use the y-axis of the first plottable if None. xtitle : str, optional (default=None) Set the x-axis title. ytitle : str, optional (default=None) Set the y-axis title. xlimits : tuple, optional (default=None) Set the x-axis limits with a 2-tuple of (min, max) ylimits : tuple, optional (default=None) Set the y-axis limits with a 2-tuple of (min, max) xdivisions : int, optional (default=None) Set the number of divisions for the x-axis ydivisions : int, optional (default=None) Set the number of divisions for the y-axis logx : bool, optional (default=False) If True, then set the x-axis to log scale. logy : bool, optional (default=False) If True, then set the y-axis to log scale. kwargs : dict All extra arguments are passed to get_limits when determining the axis limits. Returns ------- (xaxis, yaxis), (xmin, xmax, ymin, ymax) : tuple The axes and axes bounds. See Also -------- get_limits """ context = preserve_current_canvas if pad else do_nothing if not isinstance(plottables, (tuple, list)): plottables = [plottables] elif not plottables: raise ValueError("plottables is empty") with context(): if pad is not None: pad.cd() # get the axes limits xmin, xmax, ymin, ymax = get_limits(plottables, logx=logx, logy=logy, **kwargs) if xlimits is not None: xmin, xmax = xlimits if ylimits is not None: ymin, ymax = ylimits if not same: obj = plottables.pop(0) if isinstance(obj, ROOT.THStack): obj.SetMinimum(ymin) obj.SetMaximum(ymax) obj.Draw() xaxis = obj.xaxis yaxis = obj.yaxis # draw the plottables for i, obj in enumerate(plottables): if i == 0 and isinstance(obj, ROOT.THStack): # use SetMin/Max for y-axis obj.SetMinimum(ymin) obj.SetMaximum(ymax) # ROOT: please fix this... obj.Draw('SAME') # set the axes limits and titles if xaxis is not None: xaxis.SetLimits(xmin, xmax) xaxis.SetRangeUser(xmin, xmax) if xtitle is not None: xaxis.SetTitle(xtitle) if xdivisions is not None: xaxis.SetNdivisions(xdivisions) if yaxis is not None: yaxis.SetLimits(ymin, ymax) yaxis.SetRangeUser(ymin, ymax) if ytitle is not None: yaxis.SetTitle(ytitle) if ydivisions is not None: yaxis.SetNdivisions(ydivisions) if pad is None: pad = ROOT.gPad pad.SetLogx(bool(logx)) pad.SetLogy(bool(logy)) # redraw axes on top # axes ticks sometimes get hidden by filled histograms pad.RedrawAxis() return (xaxis, yaxis), (xmin, xmax, ymin, ymax) multiadd = lambda a, b: map(operator.add, a, b) multisub = lambda a, b: map(operator.sub, a, b) def _limits_helper(x1, x2, a, b, snap=False): """ Given x1, x2, a, b, where: x1 - x0 x3 - x2 a = ------- , b = ------- x3 - x0 x3 - x0 determine the points x0 and x3: x0 x1 x2 x3 |----------|-----------------|--------| """ if x2 < x1: raise ValueError("x2 < x1") if a + b >= 1: raise ValueError("a + b >= 1") if a < 0: raise ValueError("a < 0") if b < 0: raise ValueError("b < 0") if snap: if x1 >= 0: x1 = 0 a = 0 elif x2 <= 0: x2 = 0 b = 0 if x1 == x2 == 0: # garbage in garbage out return 0., 1. elif x1 == x2: # garbage in garbage out return x1 - 1., x1 + 1. if a == 0 and b == 0: return x1, x2 elif a == 0: return x1, (x2 - b * x1) / (1 - b) elif b == 0: return (x1 - a * x2) / (1 - a), x2 x0 = ((b / a) * x1 + x2 - (x2 - x1) / (1 - a - b)) / (1 + b / a) x3 = (x2 - x1) / (1 - a - b) + x0 return x0, x3 def get_limits(plottables, xpadding=0, ypadding=0.1, xerror_in_padding=True, yerror_in_padding=True, snap=True, logx=False, logy=False, logx_crop_value=1E-5, logy_crop_value=1E-5, logx_base=10, logy_base=10): """ Get the axes limits that should be used for a 1D histogram, graph, or stack of histograms. Parameters ---------- plottables : Hist, Graph, HistStack, or list of such objects The object(s) for which visually pleasing plot boundaries are requested. xpadding : float or 2-tuple, optional (default=0) The horizontal padding as a fraction of the final plot width. ypadding : float or 2-tuple, optional (default=0.1) The vertical padding as a fraction of the final plot height. xerror_in_padding : bool, optional (default=True) If False then exclude the x error bars from the calculation of the plot width. yerror_in_padding : bool, optional (default=True) If False then exclude the y error bars from the calculation of the plot height. snap : bool, optional (default=True) Make the minimum or maximum of the vertical range the x-axis depending on if the plot maximum and minimum are above or below the x-axis. If the plot maximum is above the x-axis while the minimum is below the x-axis, then this option will have no effect. logx : bool, optional (default=False) If True, then the x-axis is log scale. logy : bool, optional (default=False) If True, then the y-axis is log scale. logx_crop_value : float, optional (default=1E-5) If an x-axis is using a logarithmic scale then crop all non-positive values with this value. logy_crop_value : float, optional (default=1E-5) If the y-axis is using a logarithmic scale then crop all non-positive values with this value. logx_base : float, optional (default=10) The base used for the logarithmic scale of the x-axis. logy_base : float, optional (default=10) The base used for the logarithmic scale of the y-axis. Returns ------- xmin, xmax, ymin, ymax : tuple of plot boundaries The computed x and y-axis ranges. """ try: import numpy as np use_numpy = True except ImportError: use_numpy = False if not isinstance(plottables, (list, tuple)): plottables = [plottables] xmin = float('+inf') xmax = float('-inf') ymin = float('+inf') ymax = float('-inf') for h in plottables: if isinstance(h, HistStack): h = h.sum if not isinstance(h, (_Hist, _Graph1DBase)): raise TypeError( "unable to determine plot axes ranges " "from object of type `{0}`".format( type(h))) if use_numpy: y_array_min = y_array_max = np.array(list(h.y())) if yerror_in_padding: y_array_min = y_array_min - np.array(list(h.yerrl())) y_array_max = y_array_max + np.array(list(h.yerrh())) _ymin = y_array_min.min() _ymax = y_array_max.max() else: y_array_min = y_array_max = list(h.y()) if yerror_in_padding: y_array_min = multisub(y_array_min, list(h.yerrl())) y_array_max = multiadd(y_array_max, list(h.yerrh())) _ymin = min(y_array_min) _ymax = max(y_array_max) if isinstance(h, _Graph1DBase): if use_numpy: x_array_min = x_array_max = np.array(list(h.x())) if xerror_in_padding: x_array_min = x_array_min - np.array(list(h.xerrl())) x_array_max = x_array_max + np.array(list(h.xerrh())) _xmin = x_array_min.min() _xmax = x_array_max.max() else: x_array_min = x_array_max = list(h.x()) if xerror_in_padding: x_array_min = multisub(x_array_min, list(h.xerrl())) x_array_max = multiadd(x_array_max, list(h.xerrh())) _xmin = min(x_array_min) _xmax = max(x_array_max) else: _xmin = h.xedgesl(1) _xmax = h.xedgesh(h.nbins(0)) if logy: _ymin = max(logy_crop_value, _ymin) _ymax = max(logy_crop_value, _ymax) if logx: _xmin = max(logx_crop_value, _xmin) _xmax = max(logx_crop_value, _xmax) if _xmin < xmin: xmin = _xmin if _xmax > xmax: xmax = _xmax if _ymin < ymin: ymin = _ymin if _ymax > ymax: ymax = _ymax if isinstance(xpadding, (list, tuple)): if len(xpadding) != 2: raise ValueError("xpadding must be of length 2") xpadding_left = xpadding[0] xpadding_right = xpadding[1] else: xpadding_left = xpadding_right = xpadding if isinstance(ypadding, (list, tuple)): if len(ypadding) != 2: raise ValueError("ypadding must be of length 2") ypadding_top = ypadding[0] ypadding_bottom = ypadding[1] else: ypadding_top = ypadding_bottom = ypadding if logx: x0, x3 = _limits_helper( log(xmin, logx_base), log(xmax, logx_base), xpadding_left, xpadding_right) xmin = logx_base ** x0 xmax = logx_base ** x3 else: xmin, xmax = _limits_helper( xmin, xmax, xpadding_left, xpadding_right) if logy: y0, y3 = _limits_helper( log(ymin, logy_base), log(ymax, logy_base), ypadding_bottom, ypadding_top, snap=False) ymin = logy_base ** y0 ymax = logy_base ** y3 else: ymin, ymax = _limits_helper( ymin, ymax, ypadding_bottom, ypadding_top, snap=snap) return xmin, xmax, ymin, ymax def get_band(low_hist, high_hist, middle_hist=None): """ Convert the low and high histograms into a TGraphAsymmErrors centered at the middle histogram if not None otherwise the middle between the low and high points, to be used to draw a (possibly asymmetric) error band. """ npoints = low_hist.nbins(0) band = Graph(npoints) for i in range(npoints): center = low_hist.x(i + 1) width = low_hist.xwidth(i + 1) low, high = low_hist.y(i + 1), high_hist.y(i + 1) if middle_hist is not None: middle = middle_hist.y(i + 1) else: middle = (low + high) / 2. yerrh = max(high - middle, low - middle, 0) yerrl = abs(min(high - middle, low - middle, 0)) band.SetPoint(i, center, middle) band.SetPointError(i, width / 2., width / 2., yerrl, yerrh) return band def canvases_with(drawable): """ Return a list of all canvases where `drawable` has been painted. Note: This function is inefficient because it inspects all objects on all canvases, recursively. Avoid calling it if you have a large number of canvases and primitives. """ return [c for c in ROOT.gROOT.GetListOfCanvases() if drawable in find_all_primitives(c)] def find_all_primitives(pad): """ Recursively find all primities on a pad, even those hiding behind a GetListOfFunctions() of a primitive """ result = [] for primitive in pad.GetListOfPrimitives(): result.append(primitive) if hasattr(primitive, "GetListOfFunctions"): result.extend(primitive.GetListOfFunctions()) if hasattr(primitive, "GetHistogram"): p = primitive.GetHistogram() if p: result.append(p) if isinstance(primitive, ROOT.TPad): result.extend(find_all_primitives(primitive)) return result def tick_length_pixels(pad, xaxis, yaxis, xlength, ylength=None): """ Set the axes tick lengths in pixels """ if ylength is None: ylength = xlength xaxis.SetTickLength(xlength / float(pad.height_pixels)) yaxis.SetTickLength(ylength / float(pad.width_pixels))
bsd-3-clause
a7db7f629c313c8fd999fb1b8ed53821
30.950339
79
0.554896
3.646059
false
false
false
false
pybrain2/pybrain2
pybrain/structure/modules/gaussianlayer.py
26
1579
__author__ = 'Thomas Rueckstiess, ruecksti@in.tum.de' from scipy import random from pybrain.structure.modules.neuronlayer import NeuronLayer from pybrain.tools.functions import expln, explnPrime from pybrain.structure.parametercontainer import ParameterContainer class GaussianLayer(NeuronLayer, ParameterContainer): """ A layer implementing a gaussian interpretation of the input. The mean is the input, the sigmas are stored in the module parameters.""" def __init__(self, dim, name=None): NeuronLayer.__init__(self, dim, name) # initialize sigmas to 0 ParameterContainer.__init__(self, dim, stdParams = 0) # if autoalpha is set to True, alpha_sigma = alpha_mu = alpha*sigma^2 self.autoalpha = False self.enabled = True def setSigma(self, sigma): """Wrapper method to set the sigmas (the parameters of the module) to a certain value. """ assert len(sigma) == self.indim self._params *= 0 self._params += sigma def _forwardImplementation(self, inbuf, outbuf): if not self.enabled: outbuf[:] = inbuf else: outbuf[:] = random.normal(inbuf, expln(self.params)) def _backwardImplementation(self, outerr, inerr, outbuf, inbuf): expln_params = expln(self.params) self._derivs += ((outbuf - inbuf)**2 - expln_params**2) / expln_params * explnPrime(self.params) inerr[:] = (outbuf - inbuf) if not self.autoalpha: inerr /= expln_params**2 self._derivs /= expln_params**2
bsd-3-clause
4db5f901acdb186b4c4a909dff056452
37.512195
104
0.647878
3.750594
false
false
false
false
pybrain2/pybrain2
pybrain/rl/environments/ode/tasks/acrobot.py
31
1235
__author__ = 'Thomas Rueckstiess, ruecksti@in.tum.de' from pybrain.rl.environments import EpisodicTask from scipy import pi class GradualRewardTask(EpisodicTask): ''' task gives more reward, the higher the bar is.''' def __init__(self, environment): EpisodicTask.__init__(self, environment) self.maxPower = 0.5 self.reward_history = [] self.count = 0 # normalize to (-1, 1) self.sensor_limits = [(-pi, pi), (-20, 20)] #self.actor_limits = [(-1, 1)] self.actor_limits = None def isFinished(self): if self.count > 1000: self.count = 0 self.reward_history.append(self.getTotalReward()) return True else: self.count += 1 return False def getReward(self): # calculate reward and return reward jointSense = self.env.getSensorByName('JointSensor') veloSense = self.env.getSensorByName('JointVelocitySensor') j = jointSense[0] v = veloSense[0] reward = (abs(j)) ** 2 - 0.2 * abs(v) # time.sleep(0.001) return reward def performAction(self, action): EpisodicTask.performAction(self, action*self.maxPower)
bsd-3-clause
76edc76322217c90919d92729803e27e
29.875
67
0.591093
3.548851
false
false
false
false
pybrain2/pybrain2
examples/supervised/backprop/parityrnn.py
26
2046
from __future__ import print_function #!/usr/bin/env python """ A simple recurrent neural network that detects parity for arbitrary sequences. """ __author__ = 'Tom Schaul (tom@idsia.ch)' from datasets import ParityDataSet #@UnresolvedImport from pybrain.supervised.trainers.backprop import BackpropTrainer from pybrain.structure import RecurrentNetwork, LinearLayer, TanhLayer, BiasUnit, FullConnection def buildParityNet(): net = RecurrentNetwork() net.addInputModule(LinearLayer(1, name = 'i')) net.addModule(TanhLayer(2, name = 'h')) net.addModule(BiasUnit('bias')) net.addOutputModule(TanhLayer(1, name = 'o')) net.addConnection(FullConnection(net['i'], net['h'])) net.addConnection(FullConnection(net['bias'], net['h'])) net.addConnection(FullConnection(net['bias'], net['o'])) net.addConnection(FullConnection(net['h'], net['o'])) net.addRecurrentConnection(FullConnection(net['o'], net['h'])) net.sortModules() p = net.params p[:] = [-0.5, -1.5, 1, 1, -1, 1, 1, -1, 1] p *= 10. return net def evalRnnOnSeqDataset(net, DS, verbose = False, silent = False): """ evaluate the network on all the sequences of a dataset. """ r = 0. samples = 0. for seq in DS: net.reset() for i, t in seq: res = net.activate(i) if verbose: print(t, res) r += sum((t-res)**2) samples += 1 if verbose: print('-'*20) r /= samples if not silent: print('MSE:', r) return r if __name__ == "__main__": N = buildParityNet() DS = ParityDataSet() evalRnnOnSeqDataset(N, DS, verbose = True) print('(preset weights)') N.randomize() evalRnnOnSeqDataset(N, DS) print('(random weights)') # Backprop improves the network performance, and sometimes even finds the global optimum. N.reset() bp = BackpropTrainer(N, DS, verbose = True) bp.trainEpochs(5000) evalRnnOnSeqDataset(N, DS) print('(backprop-trained weights)')
bsd-3-clause
9bdaa7f3ee8ab20ea71827163ae785a1
30
96
0.627566
3.376238
false
false
false
false
pybrain2/pybrain2
pybrain/structure/modules/lstm.py
26
5992
__author__ = 'Daan Wierstra and Tom Schaul' from scipy import tanh from pybrain.structure.modules.neuronlayer import NeuronLayer from pybrain.structure.modules.module import Module from pybrain.structure.parametercontainer import ParameterContainer from pybrain.tools.functions import sigmoid, sigmoidPrime, tanhPrime class LSTMLayer(NeuronLayer, ParameterContainer): """Long short-term memory cell layer. The input consists of 4 parts, in the following order: - input gate - forget gate - cell input - output gate """ sequential = True peepholes = False maxoffset = 0 # Transfer functions and their derivatives f = lambda _, x: sigmoid(x) fprime = lambda _, x: sigmoidPrime(x) g = lambda _, x: tanh(x) gprime = lambda _, x: tanhPrime(x) h = lambda _, x: tanh(x) hprime = lambda _, x: tanhPrime(x) def __init__(self, dim, peepholes = False, name = None): """ :arg dim: number of cells :key peepholes: enable peephole connections (from state to gates)? """ self.setArgs(dim = dim, peepholes = peepholes) # Internal buffers, created dynamically: self.bufferlist = [ ('ingate', dim), ('outgate', dim), ('forgetgate', dim), ('ingatex', dim), ('outgatex', dim), ('forgetgatex', dim), ('state', dim), ('ingateError', dim), ('outgateError', dim), ('forgetgateError', dim), ('stateError', dim), ] Module.__init__(self, 4*dim, dim, name) if self.peepholes: ParameterContainer.__init__(self, dim*3) self._setParameters(self.params) self._setDerivatives(self.derivs) def _setParameters(self, p, owner = None): ParameterContainer._setParameters(self, p, owner) dim = self.outdim self.ingatePeepWeights = self.params[:dim] self.forgetgatePeepWeights = self.params[dim:dim*2] self.outgatePeepWeights = self.params[dim*2:] def _setDerivatives(self, d, owner = None): ParameterContainer._setDerivatives(self, d, owner) dim = self.outdim self.ingatePeepDerivs = self.derivs[:dim] self.forgetgatePeepDerivs = self.derivs[dim:dim*2] self.outgatePeepDerivs = self.derivs[dim*2:] def _isLastTimestep(self): """Tell wether the current offset is the maximum offset.""" return self.maxoffset == self.offset def _forwardImplementation(self, inbuf, outbuf): self.maxoffset = max(self.offset + 1, self.maxoffset) dim = self.outdim # slicing the input buffer into the 4 parts try: self.ingatex[self.offset] = inbuf[:dim] except IndexError: raise str((self.offset, self.ingatex.shape)) self.forgetgatex[self.offset] = inbuf[dim:dim*2] cellx = inbuf[dim*2:dim*3] self.outgatex[self.offset] = inbuf[dim*3:] # peephole treatment if self.peepholes and self.offset > 0: self.ingatex[self.offset] += self.ingatePeepWeights * self.state[self.offset-1] self.forgetgatex[self.offset] += self.forgetgatePeepWeights * self.state[self.offset-1] self.ingate[self.offset] = self.f(self.ingatex[self.offset]) self.forgetgate[self.offset] = self.f(self.forgetgatex[self.offset]) self.state[self.offset] = self.ingate[self.offset] * self.g(cellx) if self.offset > 0: self.state[self.offset] += self.forgetgate[self.offset] * self.state[self.offset-1] if self.peepholes: self.outgatex[self.offset] += self.outgatePeepWeights * self.state[self.offset] self.outgate[self.offset] = self.f(self.outgatex[self.offset]) outbuf[:] = self.outgate[self.offset] * self.h(self.state[self.offset]) def _backwardImplementation(self, outerr, inerr, outbuf, inbuf): dim = self.outdim cellx = inbuf[dim*2:dim*3] self.outgateError[self.offset] = self.fprime(self.outgatex[self.offset]) * outerr * self.h(self.state[self.offset]) self.stateError[self.offset] = outerr * self.outgate[self.offset] * self.hprime(self.state[self.offset]) if not self._isLastTimestep(): self.stateError[self.offset] += self.stateError[self.offset+1] * self.forgetgate[self.offset+1] if self.peepholes: self.stateError[self.offset] += self.ingateError[self.offset+1] * self.ingatePeepWeights self.stateError[self.offset] += self.forgetgateError[self.offset+1] * self.forgetgatePeepWeights if self.peepholes: self.stateError[self.offset] += self.outgateError[self.offset] * self.outgatePeepWeights cellError = self.ingate[self.offset] * self.gprime(cellx) * self.stateError[self.offset] if self.offset > 0: self.forgetgateError[self.offset] = self.fprime(self.forgetgatex[self.offset]) * self.stateError[self.offset] * self.state[self.offset-1] self.ingateError[self.offset] = self.fprime(self.ingatex[self.offset]) * self.stateError[self.offset] * self.g(cellx) # compute derivatives if self.peepholes: self.outgatePeepDerivs += self.outgateError[self.offset] * self.state[self.offset] if self.offset > 0: self.ingatePeepDerivs += self.ingateError[self.offset] * self.state[self.offset-1] self.forgetgatePeepDerivs += self.forgetgateError[self.offset] * self.state[self.offset-1] inerr[:dim] = self.ingateError[self.offset] inerr[dim:dim*2] = self.forgetgateError[self.offset] inerr[dim*2:dim*3] = cellError inerr[dim*3:] = self.outgateError[self.offset] def whichNeuron(self, inputIndex = None, outputIndex = None): if inputIndex != None: return inputIndex % self.dim if outputIndex != None: return outputIndex
bsd-3-clause
0c475615b54b4946f96111227ba02541
39.214765
149
0.633845
3.4417
false
false
false
false
pybrain2/pybrain2
pybrain/rl/environments/functions/transformations.py
26
9364
__author__ = 'Tom Schaul, tom@idsia.ch' from scipy import rand, dot, power, diag, eye, sqrt, sin, log, exp, ravel, clip, arange from scipy.linalg import orth, norm, inv from random import shuffle, random, gauss from pybrain.rl.environments.functions.function import FunctionEnvironment from pybrain.structure.parametercontainer import ParameterContainer from pybrain.rl.environments.fitnessevaluator import FitnessEvaluator from pybrain.utilities import sparse_orth, dense_orth from pybrain.rl.environments.functions.multiobjective import MultiObjectiveFunction def oppositeFunction(basef): """ the opposite of a function """ if isinstance(basef, FitnessEvaluator): if isinstance(basef, FunctionEnvironment): ''' added by JPQ ''' if isinstance(basef, MultiObjectiveFunction): res = MultiObjectiveFunction() else: # --- res = FunctionEnvironment(basef.xdim, basef.xopt) else: res = FitnessEvaluator() res.f = lambda x:-basef.f(x) if not basef.desiredValue is None: res.desiredValue = -basef.desiredValue res.toBeMinimized = not basef.toBeMinimized return res else: return lambda x:-basef(x) class TranslateFunction(FunctionEnvironment): """ change the position of the optimum """ def __init__(self, basef, distance=0.1, offset=None): """ by default the offset is random, with a distance of 0.1 to the old one """ FunctionEnvironment.__init__(self, basef.xdim, basef.xopt) if offset == None: self._offset = rand(basef.xdim) self._offset *= distance / norm(self._offset) else: self._offset = offset self.xopt += self._offset self.desiredValue = basef.desiredValue self.toBeMinimized = basef.toBeMinimized def tf(x): if isinstance(x, ParameterContainer): x = x.params return basef.f(x - self._offset) self.f = tf class RotateFunction(FunctionEnvironment): """ make the dimensions non-separable, by applying a matrix transformation to x before it is given to the function """ def __init__(self, basef, rotMat=None): """ by default the rotation matrix is random. """ FunctionEnvironment.__init__(self, basef.xdim, basef.xopt) if rotMat == None: # make a random orthogonal rotation matrix self._M = orth(rand(basef.xdim, basef.xdim)) else: self._M = rotMat self.desiredValue = basef.desiredValue self.toBeMinimized = basef.toBeMinimized self.xopt = dot(inv(self._M), self.xopt) def rf(x): if isinstance(x, ParameterContainer): x = x.params return basef.f(dot(x, self._M)) self.f = rf def penalize(x, distance=5): ax = abs(x) tmp = clip(ax-distance, 0, ax.max()) return dot(tmp, tmp) #return sum([max(0, abs(xi) - distance) ** 2 for xi in x]) class SoftConstrainedFunction(FunctionEnvironment): """ Soft constraint handling through a penalization term. """ penalized = True def __init__(self, basef, distance=5, penalizationFactor=1.): FunctionEnvironment.__init__(self, basef.xdim, basef.xopt) self.desiredValue = basef.desiredValue self.toBeMinimized = basef.toBeMinimized if basef.penalized: # already OK self.f = basef.f else: if not self.toBeMinimized: penalizationFactor *= -1 def scf(x): if isinstance(x, ParameterContainer): x = x.params return basef.f(x) + penalize(x, distance) * penalizationFactor self.f = scf def generateDiags(alpha, dim, shuffled=False): diags = [power(alpha, i / (2 * dim - 2.)) for i in range(dim)] if shuffled: shuffle(diags) return diag(diags) class BBOBTransformationFunction(FunctionEnvironment): """ Reimplementation of the relatively complex set of function and variable transformations, and their non-trivial combinations from BBOB 2010. But in clean, reusable code. """ def __init__(self, basef, translate=True, rotate=False, conditioning=None, asymmetry=None, oscillate=False, penalized=0, desiredValue=1e-8, gnoise=None, unoise=None, cnoise=None, sparse=True, ): FunctionEnvironment.__init__(self, basef.xdim, basef.xopt) self._name = basef.__class__.__name__ self.desiredValue = desiredValue self.toBeMinimized = basef.toBeMinimized if self.xdim < 500: sparse = False if sparse: try: from scipy.sparse import csc_matrix except: sparse = False if translate: self.xopt = (rand(self.xdim) - 0.5) * 9.8 if conditioning: prefix = generateDiags(conditioning, self.xdim) if sparse: prefix = csc_matrix(prefix) if rotate: prefix = prefix * sparse_orth(self.xdim) if oscillate or not asymmetry: prefix = sparse_orth(self.xdim) * prefix else: if rotate: prefix = dot(prefix, dense_orth(self.xdim)) if oscillate or not asymmetry: prefix = dot(dense_orth(self.xdim), prefix) elif rotate and asymmetry and not oscillate: if sparse: prefix = sparse_orth(self.xdim) else: prefix = dense_orth(self.xdim) elif sparse: prefix = None else: prefix = eye(self.xdim) if penalized != 0: if self.penalized: penalized = 0 else: self.penalized = True # combine transformations if rotate: if sparse: r = sparse_orth(self.xdim) tmp1 = lambda x: ravel(x * r) else: r = dense_orth(self.xdim) tmp1 = lambda x: dot(x, r) else: tmp1 = lambda x: x if oscillate: tmp2 = lambda x: BBOBTransformationFunction.oscillatify(tmp1(x)) else: tmp2 = tmp1 if asymmetry is not None: tmp3 = lambda x: BBOBTransformationFunction.asymmetrify(tmp2(x), asymmetry) else: tmp3 = tmp2 # noise ntmp = None if gnoise: ntmp = lambda f: f * exp(gnoise * gauss(0, 1)) elif unoise: alpha = 0.49 * (1. / self.xdim) * unoise ntmp = lambda f: f * power(random(), unoise) * max(1, power(1e9 / (f + 1e-99), alpha * random())) elif cnoise: alpha, beta = cnoise ntmp = lambda f: f + alpha * max(0, 1000 * (random() < beta) * gauss(0, 1) / (abs(gauss(0, 1)) + 1e-199)) def noisetrans(f): if ntmp is None or f < 1e-8: return f else: return ntmp(f) + 1.01e-8 if sparse: if prefix is None: tmp4 = lambda x: tmp3(x - self.xopt) else: tmp4 = lambda x: ravel(prefix * tmp3(x - self.xopt)) else: tmp4 = lambda x: dot(prefix, tmp3(x - self.xopt)) self.f = lambda x: (noisetrans(basef.f(tmp4(x))) + penalized * penalize(x)) @staticmethod def asymmetrify(x, beta=0.2): dim = len(x) return x * (x<=0) + (x>0) * exp((1+beta*arange(dim)/(dim-1.)*sqrt(abs(x))) * log(abs(x)+1e-100)) #res = x.copy() #for i, xi in enumerate(x): # if xi > 0: # res[i] = power(xi, 1 + beta * i / (dim - 1.) * sqrt(xi)) #return res @staticmethod def _oscillatify(x): if isinstance(x, float): res = [x] else: res = x.copy() for i, xi in enumerate(res): if xi == 0: continue elif xi > 0: s = 1 c1 = 10 c2 = 7.9 else: s = 1 c1 = 5.5 c2 = 3.1 res[i] = s * exp(log(abs(xi)) + 0.049 * (sin(c1 * xi) + sin(c2 * xi))) if isinstance(x, float): return res[0] else: return res @staticmethod def oscillatify(x): return exp(log(abs(x)+1e-100) + (x>0) * 0.049 * (sin(10 * x) + sin(7.9 * x)) + (x<0) * 0.049 * (sin(5.5 * x) + sin(3.1 * x)))
bsd-3-clause
9341b8f303e3b07d2c5705c60e104644
33.557196
117
0.504485
3.90818
false
false
false
false
pybrain2/pybrain2
pybrain/rl/environments/mazes/polarmaze.py
25
1644
__author__ = 'Tom Schaul, tom@idsia.ch' from scipy import zeros from random import choice, random from .maze import Maze class PolarMaze(Maze): """ Mazes with the emphasis on Perseus: allow him to turn, go forward or backward. Thus there are 4 states per position. """ actions = 5 Stay = 0 Forward = 1 TurnAround = 2 TurnLeft = 3 TurnRight = 4 allActions = [Stay, Forward, TurnAround, TurnLeft, TurnRight] def reset(self): Maze.reset(self) self.perseusDir = choice(list(range(4))) def performAction(self, action): if self.stochAction > 0: if random() < self.stochAction: action = choice(list(range(len(PolarMaze.allActions)))) act = PolarMaze.allActions[action] self.bang = False if act == self.Forward: tmp = self._moveInDir(self.perseus, Maze.allActions[self.perseusDir]) if self.mazeTable[tmp] == False: self.perseus = tmp else: self.bang = True elif act == self.TurnLeft: self.perseusDir = (self.perseusDir + 1) % 4 elif act == self.TurnRight: self.perseusDir = (self.perseusDir - 1) % 4 elif act == self.TurnAround: self.perseusDir = (self.perseusDir + 2) % 4 def getSensors(self): obs = Maze.getSensors(self) res = zeros(4) res[:4 - self.perseusDir] = obs[self.perseusDir:] res[4 - self.perseusDir:] = obs[:self.perseusDir] return res def __str__(self): return Maze.__str__(self) + '(dir:' + str(self.perseusDir) + ')'
bsd-3-clause
f99abf1168db084ed16d4ed9465c3ba0
28.890909
86
0.579684
3.307847
false
false
false
false
pybrain2/pybrain2
pybrain/rl/environments/shipsteer/northwardtask.py
25
1592
__author__ = 'Martin Felder, felder@in.tum.de' from pybrain.rl.environments import EpisodicTask from .shipsteer import ShipSteeringEnvironment class GoNorthwardTask(EpisodicTask): """ The task of balancing some pole(s) on a cart """ def __init__(self, env=None, maxsteps=1000): """ :key env: (optional) an instance of a ShipSteeringEnvironment (or a subclass thereof) :key maxsteps: maximal number of steps (default: 1000) """ if env == None: env = ShipSteeringEnvironment(render=False) EpisodicTask.__init__(self, env) self.N = maxsteps self.t = 0 # scale sensors # [h, hdot, v] self.sensor_limits = [(-180.0, +180.0), (-180.0, +180.0), (-10.0, +40.0)] # actions: thrust, rudder self.actor_limits = [(-1.0, +2.0), (-90.0, +90.0)] # scale reward over episode, such that max. return = 100 self.rewardscale = 100. / maxsteps / self.sensor_limits[2][1] def reset(self): EpisodicTask.reset(self) self.t = 0 def performAction(self, action): self.t += 1 EpisodicTask.performAction(self, action) def isFinished(self): if self.t >= self.N: # maximal timesteps return True return False def getReward(self): if abs(self.env.getHeading()) < 5.: return self.env.getSpeed() * self.rewardscale else: return 0 def setMaxLength(self, n): self.N = n
bsd-3-clause
0453edd4d6296140cb3df1f36f268efd
29.615385
93
0.552764
3.537778
false
false
false
false
pybrain2/pybrain2
pybrain/structure/modules/gate.py
25
3451
# -*- coding: utf-8 -*- __author__ = 'Justin S Bayer, bayer.justin@googlemail.com' __version__ = '$Id$' from pybrain.structure.modules.module import Module from pybrain.structure.modules.neuronlayer import NeuronLayer from pybrain.tools.functions import sigmoid, sigmoidPrime class MultiplicationLayer(NeuronLayer): """Layer that implements pairwise multiplication.""" def __init__(self, dim, name=None): Module.__init__(self, 2 * dim, dim, name) self.setArgs(dim=dim, name=self.name) def _forwardImplementation(self, inbuf, outbuf): outbuf += inbuf[:self.outdim] * inbuf[self.outdim:] def _backwardImplementation(self, outerr, inerr, outbuf, inbuf): inerr[:self.outdim] += inbuf[self.outdim:] * outerr inerr[self.outdim:] += inbuf[:self.outdim] * outerr class GateLayer(NeuronLayer): """Layer that implements pairwise input multiplication, with one element of the pair being squashed. If a GateLayer of size n is created, it will have 2 * n inputs and n outputs. The i'th output is calculated as sigmoid(I_i) * I_(i + n) where I is the vector of inputs.""" def __init__(self, dim, name=None): Module.__init__(self, 2 * dim, dim, name) self.setArgs(dim=dim, name=self.name) def _forwardImplementation(self, inbuf, outbuf): outbuf += sigmoid(inbuf[:self.outdim]) * inbuf[self.outdim:] def _backwardImplementation(self, outerr, inerr, outbuf, inbuf): inerr[:self.outdim] += (sigmoidPrime(inbuf[:self.outdim]) * inbuf[self.outdim:] * outerr) inerr[self.outdim:] += (sigmoid(inbuf[:self.outdim]) * outerr) class DoubleGateLayer(NeuronLayer): """Layer that implements a continuous if-then-else. If a DoubleGateLayer of size n is created, it will have 2 * n inputs and 2 * n outputs. The i'th output is calculated as sigmoid(I_i) * I_(i + n) for i < n and as (1 - sigmoid(I_i) * I_(i + n) for i >= n where I is the vector of inputs.""" def __init__(self, dim, name=None): Module.__init__(self, 2 * dim, 2 * dim, name) self.setArgs(dim=dim, name=self.name) def _forwardImplementation(self, inbuf, outbuf): dim = self.indim // 2 outbuf[:dim] += sigmoid(inbuf[:dim]) * inbuf[dim:] outbuf[dim:] += (1 - sigmoid(inbuf[:dim])) * inbuf[dim:] def _backwardImplementation(self, outerr, inerr, outbuf, inbuf): dim = self.indim // 2 in0 = inbuf[:dim] in1 = inbuf[dim:] out0 = outerr[:dim] out1 = outerr[dim:] inerr[:dim] += sigmoidPrime(in0) * in1 * out0 inerr[dim:] += sigmoid(in0) * out0 inerr[:dim] -= sigmoidPrime(in0) * in1 * out1 inerr[dim:] += (1 - sigmoid(in0)) * out1 class SwitchLayer(NeuronLayer): """Layer that implements pairwise multiplication.""" #:TODO: Misleading docstring def __init__(self, dim, name=None): Module.__init__(self, dim, dim * 2, name) self.setArgs(dim=dim, name=self.name) def _forwardImplementation(self, inbuf, outbuf): outbuf[:self.indim] += sigmoid(inbuf) outbuf[self.indim:] += 1 - sigmoid(inbuf) def _backwardImplementation(self, outerr, inerr, outbuf, inbuf): inerr += sigmoidPrime(inbuf) * outerr[:self.indim] inerr -= sigmoidPrime(inbuf) * outerr[self.indim:]
bsd-3-clause
726faa132a61b1336806793bd8b44744
34.57732
80
0.616053
3.363548
false
false
false
false
pybrain2/pybrain2
examples/rl/valuebased/td.py
30
1972
#!/usr/bin/env python __author__ = 'Thomas Rueckstiess, ruecksti@in.tum.de' """ This example demonstrates how to use the discrete Temporal Difference Reinforcement Learning algorithms (SARSA, Q, Q(lambda)) in a classical fully observable MDP maze task. The goal point is the top right free field. """ from scipy import * #@UnusedWildImport import pylab from pybrain.rl.environments.mazes import Maze, MDPMazeTask from pybrain.rl.learners.valuebased import ActionValueTable from pybrain.rl.agents import LearningAgent from pybrain.rl.learners import Q, QLambda, SARSA #@UnusedImport from pybrain.rl.explorers import BoltzmannExplorer #@UnusedImport from pybrain.rl.experiments import Experiment # create the maze with walls (1) envmatrix = array([[1, 1, 1, 1, 1, 1, 1, 1, 1], [1, 0, 0, 1, 0, 0, 0, 0, 1], [1, 0, 0, 1, 0, 0, 1, 0, 1], [1, 0, 0, 1, 0, 0, 1, 0, 1], [1, 0, 0, 1, 0, 1, 1, 0, 1], [1, 0, 0, 0, 0, 0, 1, 0, 1], [1, 1, 1, 1, 1, 1, 1, 0, 1], [1, 0, 0, 0, 0, 0, 0, 0, 1], [1, 1, 1, 1, 1, 1, 1, 1, 1]]) env = Maze(envmatrix, (7, 7)) # create task task = MDPMazeTask(env) # create value table and initialize with ones table = ActionValueTable(81, 4) table.initialize(1.) # create agent with controller and learner - use SARSA(), Q() or QLambda() here learner = SARSA() # standard exploration is e-greedy, but a different type can be chosen as well # learner.explorer = BoltzmannExplorer() # create agent agent = LearningAgent(table, learner) # create experiment experiment = Experiment(task, agent) # prepare plotting pylab.gray() pylab.ion() for i in range(1000): # interact with the environment (here in batch mode) experiment.doInteractions(100) agent.learn() agent.reset() # and draw the table pylab.pcolor(table.params.reshape(81,4).max(1).reshape(9,9)) pylab.draw()
bsd-3-clause
8da42ab641707cc98db511c9f46127b5
29.338462
79
0.636917
3.029186
false
false
false
false
pybrain2/pybrain2
pybrain/rl/environments/ode/instances/johnnie.py
31
1429
__author__ = 'Frank Sehnke, sehnke@in.tum.de' from pybrain.rl.environments.ode import ODEEnvironment, sensors, actuators import imp from scipy import array class JohnnieEnvironment(ODEEnvironment): def __init__(self, renderer=True, realtime=False, ip="127.0.0.1", port="21590", buf='16384'): ODEEnvironment.__init__(self, renderer, realtime, ip, port, buf) # load model file self.loadXODE(imp.find_module('pybrain')[1] + "/rl/environments/ode/models/johnnie.xode") # standard sensors and actuators self.addSensor(sensors.JointSensor()) self.addSensor(sensors.JointVelocitySensor()) self.addActuator(actuators.JointActuator()) #set act- and obsLength, the min/max angles and the relative max touques of the joints self.actLen = self.indim self.obsLen = len(self.getSensors()) #ArmLeft, ArmRight, Hip, PevelLeft, PevelRight, TibiaLeft, TibiaRight, KneeLeft, KneeRight, FootLeft, FootRight self.tourqueList = array([0.2, 0.2, 0.2, 0.5, 0.5, 2.0, 2.0, 2.0, 2.0, 0.5, 0.5],) self.cHighList = array([1.0, 1.0, 0.5, 0.5, 0.5, 1.5, 1.5, 1.5, 1.5, 0.25, 0.25],) self.cLowList = array([-0.5, -0.5, -0.5, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, -0.25, -0.25],) self.stepsPerAction = 1 if __name__ == '__main__' : w = JohnnieEnvironment() while True: w.step() if w.stepCounter == 1000: w.reset()
bsd-3-clause
3445d223dbcfc69191fd7802e18d95a5
42.30303
119
0.627712
2.764023
false
false
false
false
pybrain2/pybrain2
pybrain/optimization/distributionbased/snes.py
25
4415
from __future__ import print_function __author__ = 'Tom Schaul, tom@idsia.ch' from pybrain.optimization.distributionbased.distributionbased import DistributionBasedOptimizer from scipy import dot, exp, log, sqrt, floor, ones, randn from pybrain.tools.rankingfunctions import HansenRanking class SNES(DistributionBasedOptimizer): """ Separable NES (diagonal). [As described in Schaul, Glasmachers and Schmidhuber (GECCO'11)] """ # parameters, which can be set but have a good (adapted) default value centerLearningRate = 1.0 covLearningRate = None batchSize = None uniformBaseline = True shapingFunction = HansenRanking() initVariance = 1. # fixed settings mustMaximize = True storeAllEvaluations = True storeAllEvaluated = True # for very long runs, we don't want to run out of memory clearStorage = False # minimal setting where to abort the search varianceCutoff = 1e-20 def _stoppingCriterion(self): if DistributionBasedOptimizer._stoppingCriterion(self): return True elif max(abs(self._sigmas)) < self.varianceCutoff: return True else: return False def _initLearningRate(self): """ Careful, robust default value. """ return 0.6 * (3 + log(self.numParameters)) / 3 / sqrt(self.numParameters) def _initBatchSize(self): """ as in CMA-ES """ return 4 + int(floor(3 * log(self.numParameters))) def _additionalInit(self): if self.covLearningRate is None: self.covLearningRate = self._initLearningRate() if self.batchSize is None: self.batchSize = self._initBatchSize() self._center = self._initEvaluable.copy() self._sigmas = ones(self.numParameters) * self.initVariance @property def _population(self): if self._wasUnwrapped: return [self._allEvaluated[i].params for i in self._pointers] else: return [self._allEvaluated[i] for i in self._pointers] @property def _currentEvaluations(self): fits = [self._allEvaluations[i] for i in self._pointers] if self._wasOpposed: fits = [-x for x in fits] return fits def _produceSample(self): return randn(self.numParameters) def _sample2base(self, sample): """ How does a sample look in the outside (base problem) coordinate system? """ return self._sigmas * sample + self._center def _base2sample(self, e): """ How does the point look in the present one reference coordinates? """ return (e - self._center) / self._sigmas def _produceSamples(self): """ Append batch size new samples and evaluate them. """ if self.clearStorage: self._allEvaluated = [] self._allEvaluations = [] tmp = [self._sample2base(self._produceSample()) for _ in range(self.batchSize)] list(map(self._oneEvaluation, tmp)) self._pointers = list(range(len(self._allEvaluated) - self.batchSize, len(self._allEvaluated))) def _learnStep(self): # produce samples self._produceSamples() samples = list(map(self._base2sample, self._population)) #compute utilities utilities = self.shapingFunction(self._currentEvaluations) utilities /= sum(utilities) # make the utilities sum to 1 if self.uniformBaseline: utilities -= 1. / self.batchSize # update center dCenter = dot(utilities, samples) self._center += self.centerLearningRate * self._sigmas * dCenter # update variances covGradient = dot(utilities, [s ** 2 - 1 for s in samples]) dA = 0.5 * self.covLearningRate * covGradient self._sigmas = self._sigmas * exp(dA) if __name__ == "__main__": from pybrain.rl.environments.functions.unimodal import ElliFunction print((SNES(ElliFunction(100), ones(100), verbose=True).learn()))
bsd-3-clause
c9fb8bab04dc4c8a15e77477a08978da
36.423729
123
0.579841
4.253372
false
false
false
false
pybrain2/pybrain2
examples/rl/environments/ode/johnnie_reinforce.py
30
2899
#!/usr/bin/env python ######################################################################### # Reinforcement Learning with PGPE on the Johnnie Environment # # The Johnnie robot is a body structure with 11 DoF . # Complex balancing tasks can be learned with this environment. # # Control/Actions: # The agent can control all 11 DOF of the robot. # # A wide variety of sensors are available for observation and reward: # - 11 angles of joints # - 11 angle velocitys of joints # - Number of foot parts that have contact to floor # - Height sensor in head for reward calculation # - Rotation sensor in 3 dimesnions # # Task available are: # - StandTask, agent has not to fall by himself # - Robust standing Task, agent has not to fall even then hit by reasonable random forces # - JumpTask, agent has to maximize the head-vertical position during the episode # # Requirements: pylab (for plotting only). If not available, comment the # last 3 lines out # Author: Frank Sehnke, sehnke@in.tum.de ######################################################################### __author__ = "Frank Sehnke" __version__ = '$Id$' from pybrain.tools.example_tools import ExTools from pybrain.rl.environments.ode import JohnnieEnvironment from pybrain.rl.environments.ode.tasks import StandingTask from pybrain.structure.modules.tanhlayer import TanhLayer from pybrain.tools.shortcuts import buildNetwork from pybrain.rl.agents import LearningAgent from pybrain.rl.learners import Reinforce from pybrain.rl.experiments import EpisodicExperiment hiddenUnits = 4 batch=2 #number of samples per learning step prnts=1 #number of learning steps after results are printed epis=5000000/batch/prnts #number of roleouts numbExp=10 #number of experiments et = ExTools(batch, prnts, kind = "learner")#tool for printing and plotting for runs in range(numbExp): # create environment #Options: Bool(OpenGL), Bool(Realtime simu. while client is connected), ServerIP(default:localhost), Port(default:21560) env = JohnnieEnvironment() # create task task = StandingTask(env) # create controller network net = buildNetwork(len(task.getObservation()), hiddenUnits, env.actLen, outclass=TanhLayer) # create agent with controller and learner (and its options) agent = LearningAgent(net, Reinforce()) et.agent = agent # create the experiment experiment = EpisodicExperiment(task, agent) #Do the experiment for updates in range(epis): for i in range(prnts): experiment.doEpisodes(batch) state, action, reward = agent.learner.dataset.getSequence(agent.learner.dataset.getNumSequences()-1) et.printResults(reward.sum(), runs, updates) et.addExps() et.showExps() #To view what the simulation is doing at the moment, go to pybrain/rl/environments/ode/ and start viewer.py (python-openGL musst be installed, see PyBrain documentation)
bsd-3-clause
7353ec26fb0ef560292f62974573c075
41.632353
169
0.714384
3.769831
false
false
false
false
pybrain2/pybrain2
pybrain/rl/environments/flexcube/environment.py
25
5905
__author__ = 'Frank Sehnke, sehnke@in.tum.de' from . import sensors import threading from pybrain.utilities import threaded from pybrain.tools.networking.udpconnection import UDPServer from pybrain.rl.environments.environment import Environment from scipy import ones, zeros, array, clip, arange, sqrt from time import sleep class FlexCubeEnvironment(Environment): def __init__(self, render=True, realtime=True, ip="127.0.0.1", port="21560"): # initialize base class self.render = render if self.render: self.updateDone = True self.updateLock = threading.Lock() self.server = UDPServer(ip, port) self.actLen = 12 self.mySensors = sensors.Sensors(["EdgesReal"]) self.dists = array([20.0, sqrt(2.0) * 20, sqrt(3.0) * 20]) self.gravVect = array([0.0, -100.0, 0.0]) self.centerOfGrav = zeros((1, 3), float) self.pos = ones((8, 3), float) self.vel = zeros((8, 3), float) self.SpringM = ones((8, 8), float) self.d = 60.0 self.dt = 0.02 self.startHight = 10.0 self.dumping = 0.4 self.fraktMin = 0.7 self.fraktMax = 1.3 self.minAkt = self.dists[0] * self.fraktMin self.maxAkt = self.dists[0] * self.fraktMax self.reset() self.count = 0 self.setEdges() self.act(array([20.0] * 12)) self.euler() self.realtime = realtime self.step = 0 def closeSocket(self): self.server.UDPInSock.close() sleep(10) def setEdges(self): self.edges = zeros((12, 2), int) count = 0 c1 = 0 for i in range(2): for j in range(2): for k in range(2): c2 = 0 for i2 in range(2): for j2 in range(2): for k2 in range(2): sum = abs(i - i2) + abs(j - j2) + abs(k - k2) if sum == 1 and i <= i2 and j <= j2 and k <= k2: self.edges[count] = [c1, c2] count += 1 c2 += 1 c1 += 1 def reset(self): self.action = ones((1, 12), float) * self.dists[0] for i in range(2): for j in range(2): for k in range(2): self.pos[i * 4 + j * 2 + k] = [i * self.dists[0] - self.dists[0] / 2.0, j * self.dists[0] - self.dists[0] / 2.0 + self.startHight, k * self.dists[0] - self.dists[0] / 2.0] self.vel = zeros((8, 3), float) idx0 = arange(8).repeat(8) idx1 = array(list(range(8)) * 8) self.difM = self.pos[idx0, :] - self.pos[idx1, :] #vectors from all points to all other points self.springM = sqrt((self.difM ** 2).sum(axis=1)).reshape(64, 1) self.distM = self.springM.copy() #distance matrix self.step = 0 self.mySensors.updateSensor(self.pos, self.vel, self.distM, self.centerOfGrav, self.step, self.action) if self.render: if self.server.clients > 0: # If there are clients send them reset signal self.server.send(["r", "r"]) def performAction(self, action): action = self.normAct(action) self.action = action.copy() self.act(action) self.euler() self.step += 1 if self.render: if self.updateDone: self.updateRenderer() if self.server.clients > 0 and self.realtime: sleep(0.02) def getSensors(self): self.mySensors.updateSensor(self.pos, self.vel, self.distM, self.centerOfGrav, self.step, self.action) return self.mySensors.getSensor()[:] def normAct(self, s): return clip(s, self.minAkt, self.maxAkt) def act(self, a): count = 0 for i in self.edges: self.springM[i[0] * 8 + i[1]] = a[count] self.springM[i[1] * 8 + i[0]] = a[count] count += 1 def euler(self): self.count += 1 #Inner Forces distM = self.distM.copy() disM = self.springM - distM #difference between wanted spring lengths and current ones disM = disM.reshape(64, 1) distM = distM + 0.0000000001 #hack to prevent divs by 0 #Forces to Velos #spring vectors normalized to 1 times the actual force from deformation vel = self.difM / distM vel *= disM * self.d * self.dt idx2 = arange(8) #TODO: arggggg!!!!! for i in range(8): self.vel[i] += vel[idx2 + i * 8, :].sum(axis=0) #Gravity self.vel += self.gravVect * self.dt #Dumping self.vel -= self.vel * self.dumping * self.dt #velos to positions self.pos += self.dt * self.vel #Collisions and friction for i in range(8): if self.pos[i][1] < 0.0: self.pos[i][1] = 0.0 self.vel[i] = self.vel[i] * [0.0, -1.0, 0.0] self.centerOfGrav = self.pos.sum(axis=0) / 8.0 #Distances of new state idx0 = arange(8).repeat(8) idx1 = array(list(range(8)) * 8) self.difM = self.pos[idx0, :] - self.pos[idx1, :] #vectors from all points to all other points self.distM = sqrt((self.difM ** 2).sum(axis=1)).reshape(64, 1) #distance matrix @threaded() def updateRenderer(self): self.updateDone = False if not self.updateLock.acquire(False): return # Listen for clients self.server.listen() if self.server.clients > 0: # If there are clients send them the new data self.server.send(repr([self.pos, self.centerOfGrav])) sleep(0.02) self.updateLock.release() self.updateDone = True
bsd-3-clause
043ce0edd103fb05fef300a679180d85
34.359281
191
0.52989
3.387837
false
false
false
false
pybrain2/pybrain2
pybrain/optimization/populationbased/coevolution/coevolution.py
25
10998
from __future__ import print_function __author__ = 'Tom Schaul, tom@idsia.ch' from scipy import argmax, array from random import sample, choice, shuffle from pybrain.utilities import fListToString, Named class Coevolution(Named): """ Population-based generational evolutionary algorithm with fitness being based (paritally) on a relative measure. """ # algorithm parameters populationSize = 50 selectionProportion = 0.5 elitism = False parentChildAverage = 1. # proportion of the child tournamentSize = None hallOfFameEvaluation = 0. # proportion of HoF evaluations in relative fitness useSharedSampling = False # an external absolute evaluator absEvaluator = None absEvalProportion = 0 # execution settings maxGenerations = None maxEvaluations = None verbose = False def __init__(self, relEvaluator, seeds, **args): """ :arg relevaluator: an anti-symmetric function that can evaluate 2 elements :arg seeds: a list of initial guesses """ # set parameters self.setArgs(**args) self.relEvaluator = relEvaluator if self.tournamentSize == None: self.tournamentSize = self.populationSize # initialize algorithm variables self.steps = 0 self.generation = 0 # the best host and the best parasite from each generation self.hallOfFame = [] # the relative fitnesses from each generation (of the selected individuals) self.hallOfFitnesses = [] # this dictionary stores all the results between 2 players (first one starting): # { (player1, player2): [games won, total games, cumulative score, list of scores] } self.allResults = {} # this dictionary stores the opponents a player has played against. self.allOpponents = {} # a list of all previous populations self.oldPops = [] # build initial populations self._initPopulation(seeds) def learn(self, maxSteps=None): """ Toplevel function, can be called iteratively. :return: best evaluable found in the last generation. """ if maxSteps != None: maxSteps += self.steps while True: if maxSteps != None and self.steps + self._stepsPerGeneration() > maxSteps: break if self.maxEvaluations != None and self.steps + self._stepsPerGeneration() > self.maxEvaluations: break if self.maxGenerations != None and self.generation >= self.maxGenerations: break self._oneGeneration() return self.hallOfFame[-1] def _oneGeneration(self): self.oldPops.append(self.pop) self.generation += 1 fitnesses = self._evaluatePopulation() # store best in hall of fame besti = argmax(array(fitnesses)) best = self.pop[besti] bestFits = sorted(fitnesses)[::-1][:self._numSelected()] self.hallOfFame.append(best) self.hallOfFitnesses.append(bestFits) if self.verbose: print(('Generation', self.generation)) print((' relat. fits:', fListToString(sorted(fitnesses), 4))) if len(best.params) < 20: print((' best params:', fListToString(best.params, 4))) self.pop = self._selectAndReproduce(self.pop, fitnesses) def _averageWithParents(self, pop, childportion): for i, p in enumerate(pop[:]): if p.parent != None: tmp = p.copy() tmp.parent = p.parent tmp._setParameters(p.params * childportion + p.parent.params * (1 - childportion)) pop[i] = tmp def _evaluatePopulation(self): hoFtournSize = min(self.generation, int(self.tournamentSize * self.hallOfFameEvaluation)) tournSize = self.tournamentSize - hoFtournSize if self.useSharedSampling: opponents = self._sharedSampling(tournSize, self.pop, self.oldPops[-1]) else: opponents = self.pop if len(opponents) < tournSize: tournSize = len(opponents) self._doTournament(self.pop, opponents, tournSize) if hoFtournSize > 0: hoF = list(set(self.hallOfFame)) self._doTournament(self.pop, hoF, hoFtournSize) fitnesses = [] for p in self.pop: fit = 0 for opp in opponents: fit += self._beats(p, opp) if hoFtournSize > 0: for opp in hoF: fit += self._beats(p, opp) if self.absEvalProportion > 0 and self.absEvaluator != None: fit = (1 - self.absEvalProportion) * fit + self.absEvalProportion * self.absEvaluator(p) fitnesses.append(fit) return fitnesses def _initPopulation(self, seeds): if self.parentChildAverage < 1: for s in seeds: s.parent = None self.pop = self._extendPopulation(seeds, self.populationSize) def _extendPopulation(self, seeds, size): """ build a population, with mutated copies from the provided seed pool until it has the desired size. """ res = seeds[:] for dummy in range(size - len(seeds)): chosen = choice(seeds) tmp = chosen.copy() tmp.mutate() if self.parentChildAverage < 1: tmp.parent = chosen res.append(tmp) return res def _selectAndReproduce(self, pop, fits): """ apply selection and reproduction to host population, according to their fitness.""" # combine population with their fitness, then sort, only by fitness s = list(zip(fits, pop)) shuffle(s) s.sort(key=lambda x:-x[0]) # select... selected = [x[1] for x in s[:self._numSelected()]] # ... and reproduce if self.elitism: newpop = self._extendPopulation(selected, self.populationSize) if self.parentChildAverage < 1: self._averageWithParents(newpop, self.parentChildAverage) else: newpop = self._extendPopulation(selected, self.populationSize + self._numSelected()) [self._numSelected():] if self.parentChildAverage < 1: self._averageWithParents(newpop[self._numSelected():], self.parentChildAverage) return newpop def _beats(self, h, p): """ determine the empirically observed score of p playing opp (starting or not). If they never played, assume 0. """ if (h, p) not in self.allResults: return 0 else: hpgames, hscore = self.allResults[(h, p)][1:3] phgames, pscore = self.allResults[(p, h)][1:3] return (hscore - pscore) / float(hpgames + phgames) def _doTournament(self, pop1, pop2, tournamentSize=None): """ Play a tournament. :key tournamentSize: If unspecified, play all-against-all """ # TODO: Preferably select high-performing opponents? for p in pop1: pop3 = pop2[:] while p in pop3: pop3.remove(p) if tournamentSize != None and tournamentSize < len(pop3): opps = sample(pop3, tournamentSize) else: opps = pop3 for opp in opps: self._relEval(p, opp) self._relEval(opp, p) def _globalScore(self, p): """ The average score over all evaluations for a player. """ if p not in self.allOpponents: return 0. scoresum, played = 0., 0 for opp in self.allOpponents[p]: scoresum += self.allResults[(p, opp)][2] played += self.allResults[(p, opp)][1] scoresum -= self.allResults[(opp, p)][2] played += self.allResults[(opp, p)][1] # slightly bias the global score in favor of players with more games (just for tie-breaking) played += 0.01 return scoresum / played def _sharedSampling(self, numSelect, selectFrom, relativeTo): """ Build a shared sampling set of opponents """ if numSelect < 1: return [] # determine the player of selectFrom with the most wins against players from relativeTo (and which ones) tmp = {} for p in selectFrom: beaten = [] for opp in relativeTo: if self._beats(p, opp) > 0: beaten.append(opp) tmp[p] = beaten beatlist = [(len(p_beaten[1]), self._globalScore(p_beaten[0]), p_beaten[0]) for p_beaten in list(tmp.items())] shuffle(beatlist) beatlist.sort(key=lambda x: x[:2]) best = beatlist[-1][2] unBeaten = list(set(relativeTo).difference(tmp[best])) otherSelect = selectFrom[:] otherSelect.remove(best) return [best] + self._sharedSampling(numSelect - 1, otherSelect, unBeaten) def _relEval(self, p, opp): """ a single relative evaluation (in one direction) with the involved bookkeeping.""" if p not in self.allOpponents: self.allOpponents[p] = [] self.allOpponents[p].append(opp) if (p, opp) not in self.allResults: self.allResults[(p, opp)] = [0, 0, 0., []] res = self.relEvaluator(p, opp) if res > 0: self.allResults[(p, opp)][0] += 1 self.allResults[(p, opp)][1] += 1 self.allResults[(p, opp)][2] += res self.allResults[(p, opp)][3].append(res) self.steps += 1 def __str__(self): s = 'Coevolution (' s += str(self._numSelected()) if self.elitism: s += '+' + str(self.populationSize - self._numSelected()) else: s += ',' + str(self.populationSize) s += ')' if self.parentChildAverage < 1: s += ' p_c_avg=' + str(self.parentChildAverage) return s def _numSelected(self): return int(self.populationSize * self.selectionProportion) def _stepsPerGeneration(self): res = self.populationSize * self.tournamentSize * 2 return res if __name__ == '__main__': # TODO: convert to unittest x = Coevolution(None, [None], populationSize=1) x.allResults[(1, 2)] = [1, 1, 1, []] x.allResults[(2, 1)] = [-1, 1, -1, []] x.allResults[(2, 5)] = [1, 1, 2, []] x.allResults[(5, 2)] = [-1, 1, -1, []] x.allResults[(2, 3)] = [1, 1, 3, []] x.allResults[(3, 2)] = [-1, 1, -1, []] x.allResults[(4, 3)] = [1, 1, 4, []] x.allResults[(3, 4)] = [-1, 1, -1, []] x.allOpponents[1] = [2] x.allOpponents[2] = [1, 5] x.allOpponents[3] = [2, 4] x.allOpponents[4] = [3] x.allOpponents[5] = [2] print((x._sharedSampling(4, [1, 2, 3, 4, 5], [1, 2, 3, 4, 6, 7, 8, 9]))) print(('should be', [4, 1, 2, 5]))
bsd-3-clause
e85a1b314884a30b5be883b80f46c875
36.535836
118
0.574104
3.670895
false
false
false
false
pybrain2/pybrain2
pybrain/rl/explorers/discrete/boltzmann.py
31
1506
__author__ = "Thomas Rueckstiess, ruecksti@in.tum.de" from scipy import array from pybrain.rl.explorers.discrete.discrete import DiscreteExplorer from pybrain.utilities import drawGibbs class BoltzmannExplorer(DiscreteExplorer): """ A discrete explorer, that executes the actions with probability that depends on their action values. The boltzmann explorer has a parameter tau (the temperature). for high tau, the actions are nearly equiprobable. for tau close to 0, this action selection becomes greedy. """ def __init__(self, tau = 2., decay = 0.9995): DiscreteExplorer.__init__(self) self.tau = tau self.decay = decay self._state = None def activate(self, state, action): """ The super class ignores the state and simply passes the action through the module. implement _forwardImplementation() in subclasses. """ self._state = state return DiscreteExplorer.activate(self, state, action) def _forwardImplementation(self, inbuf, outbuf): """ Draws a random number between 0 and 1. If the number is less than epsilon, a random action is chosen. If it is equal or larger than epsilon, the greedy action is returned. """ assert self.module values = self.module.getActionValues(self._state) action = drawGibbs(values, self.tau) self.tau *= self.decay outbuf[:] = array([action])
bsd-3-clause
a2ecfce27672db2d22e2d3a682816f38
34.023256
73
0.652722
4.126027
false
false
false
false
pybrain2/pybrain2
pybrain/rl/environments/mazes/tasks/shuttle.py
25
2692
__author__ = 'Tom Schaul, tom@idsia.ch' from scipy import array, zeros from random import random from .maze import MazeTask from pybrain.rl.environments.mazes import PolarMaze class ShuttleDocking(MazeTask): """ ####### #. *# ####### The spaceship needs to dock backwards into the goal station. """ actions = 3 observations = 5 discount = 0.95 mazeclass = PolarMaze finalReward = 10 bangPenalty = -3 initPos = [(1, 1)] topology = array([[1] * 7, [1, 0, 0, 0, 0, 0, 1], [1] * 7, ]) goal = (1, 5) Backup = 0 Forward = 1 TurnAround = 2 def reset(self): MazeTask.reset(self) self.env.perseusDir = 1 def getObservation(self): """ inold, seeold, black, seenew, innew """ res = zeros(5) if self.env.perseus == self.env.goal: res[4] = 1 elif self.env.perseus == self.env.initPos[0]: res[0] = 1 elif self.env.perseus[1] == 3: if random() > 0.7: res[self.env.perseusDir] = 1 else: res[(self.env.perseusDir + 2) % 4] = 1 else: res[(self.env.perseusDir + 2) % 4] = 1 return res def performAction(self, action): self.steps += 1 if action == self.TurnAround: self._turn() elif action == self.Forward: self._forward() else: # noisy backup r = random() if self.env.perseus[1] == 3: # in space if r < 0.1: self._turn() elif r < 0.9: self._backup() elif ((self.env.perseus[1] == 2 and self.env.perseusDir == 3) or (self.env.perseus[1] == 4 and self.env.perseusDir == 1)): # close to station, front to station if r < 0.3: self._turn() elif r < 0.6: self._backup() else: # close to station, back to station if r < 0.7: self._backup() def _backup(self): self.env.performAction(PolarMaze.TurnAround) self.env.performAction(PolarMaze.Forward) self.env.performAction(PolarMaze.TurnAround) def _turn(self): self.env.performAction(PolarMaze.TurnAround) def _forward(self): old = self.env.perseus self.env.performAction(PolarMaze.TurnAround) if self.env.perseus == self.env.goal or self.env.perseus == self.env.initPos[0]: self.env.perseus = old self.env.bang = True
bsd-3-clause
0befd08d824a682b4082ac983ba332cc
26.752577
88
0.497771
3.464607
false
false
false
false
pybrain2/pybrain2
pybrain/rl/experiments/tournament.py
25
4173
__author__ = 'Tom Schaul, tom@idsia.ch' from pybrain.rl.environments.twoplayergames.twoplayergame import TwoPlayerGame from pybrain.utilities import Named class Tournament(Named): """ the tournament class is a specific kind of experiment, that takes a pool of agents and has them compete against each other in a TwoPlayerGame. """ # do all moves need to be checked for legality? forcedLegality = False def __init__(self, env, agents): assert isinstance(env, TwoPlayerGame) self.startcolor = env.startcolor self.env = env self.agents = agents for a in agents: a.game = self.env self.reset() def reset(self): # a dictionnary attaching a list of outcomes to a player-couple-key self.results = {} self.rounds = 0 self.numGames = 0 def _produceAllPairs(self): """ produce a list of all pairs of agents (assuming ab <> ba)""" res = [] for a in self.agents: for b in self.agents: if a != b: res.append((a, b)) return res def _oneGame(self, p1, p2): """ play one game between two agents p1 and p2.""" self.numGames += 1 self.env.reset() players = (p1, p2) p1.color = self.startcolor p2.color = -p1.color p1.newEpisode() p2.newEpisode() i = 0 while not self.env.gameOver(): p = players[i] i = (i + 1) % 2 # alternate act = p.getAction() if self.forcedLegality: tries = 0 while not self.env.isLegal(*act): tries += 1 # CHECKME: maybe the legality check is too specific? act = p.getAction() if tries > 50: raise Exception('No legal move produced!') self.env.performAction(act) if players not in self.results: self.results[players] = [] wincolor = self.env.getWinner() if wincolor == p1.color: winner = p1 else: winner = p2 self.results[players].append(winner) def organize(self, repeat=1): """ have all agents play all others in all orders, and repeat. """ for dummy in range(repeat): self.rounds += 1 for p1, p2 in self._produceAllPairs(): self._oneGame(p1, p2) return self.results def eloScore(self, startingscore=1500, k=32): """ compute the elo score of all the agents, given the games played in the tournament. Also checking for potentially initial scores among the agents ('elo' variable). """ # initialize elos = {} for a in self.agents: if 'elo' in a.__dict__: elos[a] = a.elo else: elos[a] = startingscore # adjust ratings for i, a1 in enumerate(self.agents[:-1]): for a2 in self.agents[i + 1:]: # compute score (in favor of a1) s = 0 outcomes = self.results[(a1, a2)] + self.results[(a2, a1)] for r in outcomes: if r == a1: s += 1. elif r == self.env.DRAW: s += 0.5 # what score would have been estimated? est = len(outcomes) / (1. + 10 ** ((elos[a2] - elos[a1]) / 400.)) delta = k * (s - est) elos[a1] += delta elos[a2] -= delta for a, e in list(elos.items()): a.elo = e return elos def __str__(self): s = 'Tournament results (' + str(self.rounds) + ' rounds, ' + str(self.numGames) + ' games):\n' for p1, p2 in self._produceAllPairs(): wins = len([x for x in self.results[(p1, p2)] if x == p1]) losses = len([x for x in self.results[(p1, p2)] if x == p2]) s += ' ' * 3 + p1.name + ' won ' + str(wins) + ' times and lost ' + str(losses) + ' times against ' + p2.name + '\n' return s
bsd-3-clause
a46177cef8fa428e1e238d3eee1e32c0
34.666667
128
0.507788
3.790191
false
false
false
false
pybrain2/pybrain2
pybrain/tests/unittests/supervised/trainers/test_backprop.py
28
2025
""" >>> from pybrain.datasets.supervised import SupervisedDataSet >>> from pybrain.supervised.trainers import BackpropTrainer >>> from pybrain import FeedForwardNetwork >>> from pybrain.structure import LinearLayer, SigmoidLayer, FullConnection >>> from random import randrange >>> dataset = SupervisedDataSet(6, 2) >>> for i in range(1000): ... state = [randrange(0, 15), ... randrange(-70, 50), ... randrange(-70, 50), ... randrange(-70, 50), ... randrange(-70, 50), ... float(randrange(1, 5))/20.] ... action = [float(randrange(-1, 1))/10.0, ... randrange(0, 1)] ... dataset.addSample(state, action) >>> >>> net = FeedForwardNetwork() >>> >>> net.addInputModule(LinearLayer(6, name='in')) >>> net.addModule(SigmoidLayer(40, name='hidden_0')) >>> net.addModule(SigmoidLayer(16, name='hidden_1')) >>> net.addOutputModule(LinearLayer(2, name='out')) >>> >>> net.addConnection(FullConnection(net['in'], net['hidden_0'])) >>> net.addConnection(FullConnection(net['hidden_0'], net['hidden_1'])) >>> net.addConnection(FullConnection(net['hidden_1'], net['out'])) >>> >>> net.sortModules() >>> >>> trainer = BackpropTrainer(net, ... dataset=dataset, ... learningrate=0.01, ... lrdecay=1, ... momentum=0.5, ... verbose=False, ... weightdecay=0, ... batchlearning=False) >>> >>> trainingErrors, validationErrors = trainer.trainUntilConvergence( ... dataset=dataset, ... maxEpochs=10) """ __author__ = 'Steffen Kampmann, steffen.kampmann@gmail.com' from pybrain.tests import runModuleTestSuite if __name__ == "__main__": runModuleTestSuite(__import__('__main__'))
bsd-3-clause
f098086e1f8b2903964ea7dadb4b70b0
37.207547
79
0.521975
3.939689
false
true
false
false
pybrain2/pybrain2
examples/rl/environments/linear_fa/bicycle.py
26
14462
from __future__ import print_function """An attempt to implement Randlov and Alstrom (1998). They successfully use reinforcement learning to balance a bicycle, and to control it to drive to a specified goal location. Their work has been used since then by a few researchers as a benchmark problem. We only implement the balance task. This implementation differs at least slightly, since Randlov and Alstrom did not mention anything about how they annealed/decayed their learning rate, etc. As a result of differences, the results do not match those obtained by Randlov and Alstrom. """ __author__ = 'Chris Dembia, Bruce Cam, Johnny Israeli' from scipy import asarray from numpy import sin, cos, tan, sqrt, arcsin, arctan, sign, clip, argwhere from matplotlib import pyplot as plt import pybrain.rl.environments from pybrain.rl.environments.environment import Environment from pybrain.rl.learners.valuebased.linearfa import SARSALambda_LinFA from pybrain.rl.agents.linearfa import LinearFA_Agent from pybrain.rl.experiments import EpisodicExperiment from pybrain.utilities import one_to_n class BicycleEnvironment(Environment): """Randlov and Alstrom's bicycle model. This code matches nearly exactly some c code we found online for simulating Randlov and Alstrom's bicycle. The bicycle travels at a fixed speed. """ # For superclass. indim = 2 outdim = 10 # Environment parameters. time_step = 0.01 # Goal position and radius # Lagouakis (2002) uses angle to goal, not heading, as a state max_distance = 1000. # Acceleration on Earth's surface due to gravity (m/s^2): g = 9.82 # See the paper for a description of these quantities: # Distances (in meters): c = 0.66 dCM = 0.30 h = 0.94 L = 1.11 r = 0.34 # Masses (in kilograms): Mc = 15.0 Md = 1.7 Mp = 60.0 # Velocity of a bicycle (in meters per second), equal to 10 km/h: v = 10.0 * 1000.0 / 3600.0 # Derived constants. M = Mc + Mp # See Randlov's code. Idc = Md * r**2 Idv = 1.5 * Md * r**2 Idl = 0.5 * Md * r**2 Itot = 13.0 / 3.0 * Mc * h**2 + Mp * (h + dCM)**2 sigmad = v / r def __init__(self): Environment.__init__(self) self.reset() self.actions = [0.0, 0.0] self._save_wheel_contact_trajectories = False def performAction(self, actions): self.actions = actions self.step() def saveWheelContactTrajectories(self, opt): self._save_wheel_contact_trajectories = opt def step(self): # Unpack the state and actions. # ----------------------------- # Want to ignore the previous value of omegadd; it could only cause a # bug if we assign to it. (theta, thetad, omega, omegad, _, xf, yf, xb, yb, psi) = self.sensors (T, d) = self.actions # For recordkeeping. # ------------------ if self._save_wheel_contact_trajectories: self.xfhist.append(xf) self.yfhist.append(yf) self.xbhist.append(xb) self.ybhist.append(yb) # Intermediate time-dependent quantities. # --------------------------------------- # Avoid divide-by-zero, just as Randlov did. if theta == 0: rf = 1e8 rb = 1e8 rCM = 1e8 else: rf = self.L / np.abs(sin(theta)) rb = self.L / np.abs(tan(theta)) rCM = sqrt((self.L - self.c)**2 + self.L**2 / tan(theta)**2) phi = omega + np.arctan(d / self.h) # Equations of motion. # -------------------- # Second derivative of angular acceleration: omegadd = 1 / self.Itot * (self.M * self.h * self.g * sin(phi) - cos(phi) * (self.Idc * self.sigmad * thetad + sign(theta) * self.v**2 * ( self.Md * self.r * (1.0 / rf + 1.0 / rb) + self.M * self.h / rCM))) thetadd = (T - self.Idv * self.sigmad * omegad) / self.Idl # Integrate equations of motion using Euler's method. # --------------------------------------------------- # yt+1 = yt + yd * dt. # Must update omega based on PREVIOUS value of omegad. omegad += omegadd * self.time_step omega += omegad * self.time_step thetad += thetadd * self.time_step theta += thetad * self.time_step # Handlebars can't be turned more than 80 degrees. theta = np.clip(theta, -1.3963, 1.3963) # Wheel ('tyre') contact positions. # --------------------------------- # Front wheel contact position. front_temp = self.v * self.time_step / (2 * rf) # See Randlov's code. if front_temp > 1: front_temp = sign(psi + theta) * 0.5 * np.pi else: front_temp = sign(psi + theta) * arcsin(front_temp) xf += self.v * self.time_step * -sin(psi + theta + front_temp) yf += self.v * self.time_step * cos(psi + theta + front_temp) # Rear wheel. back_temp = self.v * self.time_step / (2 * rb) # See Randlov's code. if back_temp > 1: back_temp = np.sign(psi) * 0.5 * np.pi else: back_temp = np.sign(psi) * np.arcsin(back_temp) xb += self.v * self.time_step * -sin(psi + back_temp) yb += self.v * self.time_step * cos(psi + back_temp) # Preventing numerical drift. # --------------------------- # Copying what Randlov did. current_wheelbase = sqrt((xf - xb)**2 + (yf - yb)**2) if np.abs(current_wheelbase - self.L) > 0.01: relative_error = self.L / current_wheelbase - 1.0 xb += (xb - xf) * relative_error yb += (yb - yf) * relative_error # Update heading, psi. # -------------------- delta_y = yf - yb if (xf == xb) and delta_y < 0.0: psi = np.pi else: if delta_y > 0.0: psi = arctan((xb - xf) / delta_y) else: psi = sign(xb - xf) * 0.5 * np.pi - arctan(delta_y / (xb - xf)) self.sensors = np.array([theta, thetad, omega, omegad, omegadd, xf, yf, xb, yb, psi]) def reset(self): theta = 0 thetad = 0 omega = 0 omegad = 0 omegadd = 0 xf = 0 yf = self.L xb = 0 yb = 0 psi = np.arctan((xb - xf) / (yf - yb)) self.sensors = np.array([theta, thetad, omega, omegad, omegadd, xf, yf, xb, yb, psi]) self.xfhist = [] self.yfhist = [] self.xbhist = [] self.ybhist = [] def getSteer(self): return self.sensors[0] def getTilt(self): return self.sensors[2] def get_xfhist(self): return self.xfhist def get_yfhist(self): return self.yfhist def get_xbhist(self): return self.xbhist def get_ybhist(self): return self.ybhist def getSensors(self): return self.sensors class BalanceTask(pybrain.rl.environments.EpisodicTask): """The rider is to simply balance the bicycle while moving with the speed perscribed in the environment. This class uses a continuous 5 dimensional state space, and a discrete state space. This class is heavily guided by pybrain.rl.environments.cartpole.balancetask.BalanceTask. """ max_tilt = np.pi / 6. nactions = 9 def __init__(self, max_time=1000.0): super(BalanceTask, self).__init__(BicycleEnvironment()) self.max_time = max_time # Keep track of time in case we want to end episodes based on number of # time steps. self.t = 0 @property def indim(self): return 1 @property def outdim(self): return 5 def reset(self): super(BalanceTask, self).reset() self.t = 0 def performAction(self, action): """Incoming action is an int between 0 and 8. The action we provide to the environment consists of a torque T in {-2 N, 0, 2 N}, and a displacement d in {-.02 m, 0, 0.02 m}. """ self.t += 1 assert round(action[0]) == action[0] # -1 for action in {0, 1, 2}, 0 for action in {3, 4, 5}, 1 for # action in {6, 7, 8} torque_selector = np.floor(action[0] / 3.0) - 1.0 T = 2 * torque_selector # Random number in [-1, 1]: p = 2.0 * np.random.rand() - 1.0 # -1 for action in {0, 3, 6}, 0 for action in {1, 4, 7}, 1 for # action in {2, 5, 8} disp_selector = action[0] % 3 - 1.0 d = 0.02 * disp_selector + 0.02 * p super(BalanceTask, self).performAction([T, d]) def getObservation(self): (theta, thetad, omega, omegad, omegadd, xf, yf, xb, yb, psi) = self.env.getSensors() return self.env.getSensors()[0:5] def isFinished(self): # Criterion for ending an episode. From Randlov's paper: # "When the agent can balance for 1000 seconds, the task is considered # learned." if np.abs(self.env.getTilt()) > self.max_tilt: return True elapsed_time = self.env.time_step * self.t if elapsed_time > self.max_time: return True return False def getReward(self): # -1 reward for falling over; no reward otherwise. if np.abs(self.env.getTilt()) > self.max_tilt: return -1.0 return 0.0 class LinearFATileCoding3456BalanceTask(BalanceTask): """An attempt to exactly implement Randlov's function approximation. He discretized (tiled) the state space into 3456 bins. We use the same action space as in the superclass. """ # From Randlov, 1998: theta_bounds = np.array( [-0.5 * np.pi, -1.0, -0.2, 0, 0.2, 1.0, 0.5 * np.pi]) thetad_bounds = np.array( [-np.inf, -2.0, 0, 2.0, np.inf]) omega_bounds = np.array( [-BalanceTask.max_tilt, -0.15, -0.06, 0, 0.06, 0.15, BalanceTask.max_tilt]) omegad_bounds = np.array( [-np.inf, -0.5, -0.25, 0, 0.25, 0.5, np.inf]) omegadd_bounds = np.array( [-np.inf, -2.0, 0, 2.0, np.inf]) # http://stackoverflow.com/questions/3257619/numpy-interconversion-between-multidimensional-and-linear-indexing nbins_across_dims = [ len(theta_bounds) - 1, len(thetad_bounds) - 1, len(omega_bounds) - 1, len(omegad_bounds) - 1, len(omegadd_bounds) - 1] # This array, when dotted with the 5-dim state vector, gives a 'linear' # index between 0 and 3455. magic_array = np.cumprod([1] + nbins_across_dims)[:-1] @property def outdim(self): # Used when constructing LinearFALearner's. return 3456 def getBin(self, theta, thetad, omega, omegad, omegadd): bin_indices = [ np.digitize([theta], self.theta_bounds)[0] - 1, np.digitize([thetad], self.thetad_bounds)[0] - 1, np.digitize([omega], self.omega_bounds)[0] - 1, np.digitize([omegad], self.omegad_bounds)[0] - 1, np.digitize([omegadd], self.omegadd_bounds)[0] - 1, ] return np.dot(self.magic_array, bin_indices) def getBinIndices(self, linear_index): """Given a linear index (integer between 0 and outdim), returns the bin indices for each of the state dimensions. """ return linear_index / self.magic_array % self.nbins_across_dims def getObservation(self): (theta, thetad, omega, omegad, omegadd, xf, yf, xb, yb, psi) = self.env.getSensors() state = one_to_n(self.getBin(theta, thetad, omega, omegad, omegadd), self.outdim) return state class SARSALambda_LinFA_ReplacingTraces(SARSALambda_LinFA): """Randlov used replacing traces, but this doesn't exist in PyBrain's SARSALambda. """ def _updateEtraces(self, state, action, responsibility=1.): self._etraces *= self.rewardDiscount * self._lambda * responsibility # This assumes that state is an identity vector (like, from one_to_n). self._etraces[action] = clip(self._etraces[action] + state, -np.inf, 1.) # Set the trace for all other actions in this state to 0: action_bit = one_to_n(action, self.num_actions) for argstate in argwhere(state == 1) : self._etraces[argwhere(action_bit != 1), argstate] = 0. task = LinearFATileCoding3456BalanceTask() env = task.env # The learning is very sensitive to the learning rate decay. learner = SARSALambda_LinFA_ReplacingTraces(task.nactions, task.outdim, learningRateDecay=2000) learner._lambda = 0.95 task.discount = learner.rewardDiscount agent = LinearFA_Agent(learner) agent.logging = False exp = EpisodicExperiment(task, agent) performance_agent = LinearFA_Agent(learner) performance_agent.logging = False performance_agent.greedy = True performance_agent.learning = False env.saveWheelContactTrajectories(True) plt.ion() plt.figure(figsize=(8, 4)) ax1 = plt.subplot(1, 2, 1) ax2 = plt.subplot(1, 2, 2) def update_wheel_trajectories(): front_lines = ax2.plot(env.get_xfhist(), env.get_yfhist(), 'r') back_lines = ax2.plot(env.get_xbhist(), env.get_ybhist(), 'b') plt.axis('equal') perform_cumrewards = [] for irehearsal in range(7000): # Learn. # ------ r = exp.doEpisodes(1) # Discounted reward. cumreward = exp.task.getTotalReward() #print 'cumreward: %.4f; nsteps: %i; learningRate: %.4f' % ( # cumreward, len(r[0]), exp.agent.learner.learningRate) if irehearsal % 50 == 0: # Perform (no learning). # ---------------------- # Swap out the agent. exp.agent = performance_agent # Perform. r = exp.doEpisodes(1) perform_cumreward = task.getTotalReward() perform_cumrewards.append(perform_cumreward) print('PERFORMANCE: cumreward:', perform_cumreward, 'nsteps:', len(r[0])) # Swap back the learning agent. performance_agent.reset() exp.agent = agent ax1.cla() ax1.plot(perform_cumrewards, '.--') # Wheel trajectories. update_wheel_trajectories() plt.pause(0.001)
bsd-3-clause
576e50a290f8002c8a6f4c25357ec50d
32.554524
115
0.572673
3.325362
false
false
false
false
pybrain2/pybrain2
pybrain/tools/rankingfunctions.py
25
4779
""" Ranking functions that are used in Black-box optimization, or for selection. """ __author__ = 'Daan Wierstra and Tom Schaul' from pybrain.utilities import Named from random import randint from scipy import zeros, argmax, array, power, exp, sqrt, var, zeros_like, arange, mean, log def rankedFitness(R): """ produce a linear ranking of the fitnesses in R. (The highest rank is the best fitness)""" #l = sorted(list(enumerate(R)), cmp = lambda a,b: cmp(a[1],b[1])) #l = sorted(list(enumerate(l)), cmp = lambda a,b: cmp(a[1],b[1])) #return array(map(lambda (r, dummy): r, l)) res = zeros_like(R) l = list(zip(R, list(range(len(R))))) l.sort() for i, (_, j) in enumerate(l): res[j] = i return res def normalizedFitness(R): return array((R - mean(R)) / sqrt(var(R))).flatten() class RankingFunction(Named): """ Default: ranked and scaled to [0,1].""" def __init__(self, **args): self.setArgs(**args) n = self.__class__.__name__ for k, val in list(args.items()): n += '-' + str(k) + '=' + str(val) self.name = n def __call__(self, R): """ :key R: one-dimensional array containing fitnesses. """ res = rankedFitness(R) return res / float(max(res)) class TournamentSelection(RankingFunction): """ Standard evolution tournament selection, the returned array contains intergers for the samples that are selected indicating how often they are. """ tournamentSize = 2 def __call__(self, R): res = zeros(len(R)) for i in range(len(R)): l = [i] for dummy in range(self.tournamentSize - 1): randindex = i while randindex == i: randindex = randint(0, len(R) - 1) l.append(randindex) fits = [R[x] for x in l] res[argmax(fits)] += 1 return res class SmoothGiniRanking(RankingFunction): """ a smooth ranking function that gives more importance to examples with better fitness. Rescaled to be between 0 and 1""" gini = 0.1 linearComponent = 0. def __call__(self, R): def smoothup(x): """ produces a mapping from [0,1] to [0,1], with a specific gini coefficient. """ return power(x, 2 / self.gini - 1) ranks = rankedFitness(R) res = zeros(len(R)) for i in range(len(ranks)): res[i] = ranks[i] * self.linearComponent + smoothup(ranks[i] / float(len(R) - 1)) * (1 - self.linearComponent) res /= max(res) return res class ExponentialRanking(RankingFunction): """ Exponential transformation (with a temperature parameter) of the rank values. """ temperature = 10. def __call__(self, R): ranks = rankedFitness(R) ranks = ranks / (len(R) - 1.0) return exp(ranks * self.temperature) class HansenRanking(RankingFunction): """ Ranking, as used in CMA-ES """ def __call__(self, R): ranks = rankedFitness(R) return array([max(0., x) for x in log(len(R)/2.+1.0)-log(len(R)-array(ranks))]) class TopSelection(RankingFunction): """ Select the fraction of the best ranked fitnesses. """ topFraction = 0.1 def __call__(self, R): res = zeros(len(R)) ranks = rankedFitness(R) cutoff = len(R) * (1. - self.topFraction) for i in range(len(R)): if ranks[i] >= cutoff: res[i] = 1.0 else: res[i] = 0.0 return res class TopLinearRanking(TopSelection): """ Select the fraction of the best ranked fitnesses and scale them linearly between 0 and 1. """ topFraction = 0.2 def __call__(self, R): res = zeros(len(R)) ranks = rankedFitness(R) cutoff = len(R) * (1. - self.topFraction) for i in range(len(R)): if ranks[i] >= cutoff: res[i] = ranks[i] - cutoff else: res[i] = 0.0 res /= max(res) return res def getPossibleParameters(self, numberOfSamples): x = 1. / float(numberOfSamples) return arange(x * 2, 1 + x, x) def setParameter(self, p): self.topFraction = p class BilinearRanking(RankingFunction): """ Bi-linear transformation, rescaled. """ bilinearFactor = 20 def __call__(self, R): ranks = rankedFitness(R) res = zeros(len(R)) transitionpoint = 4 * len(ranks) / 5 for i in range(len(ranks)): if ranks[i] < transitionpoint: res[i] = ranks[i] else: res[i] = ranks[i] + (ranks[i] - transitionpoint) * self.bilinearFactor res /= max(res) return res
bsd-3-clause
7d60969e0fbdb90aee4feb62bc60abb6
28.319018
122
0.56267
3.458032
false
false
false
false
pybrain2/pybrain2
pybrain/auxiliary/pca.py
31
2607
# -*- coding: utf-8 -*- """Module that contains functionality for calculating the principal components of a dataset.""" __author__ = 'Justin S Bayer, bayerj@in.tum.de' from scipy import asmatrix, cov from scipy.linalg import inv, eig from numpy.random import standard_normal def reduceDim(data, dim, func='pca'): """Reduce the dimension of datapoints to dim via principal component analysis. A matrix of shape (n, d) specifies n points of dimension d. """ try: pcaFunc = globals()[func] except KeyError: raise ValueError('Unknown function to calc principal components') pc = pcaFunc(data, dim) return (pc * asmatrix(makeCentered(data)).T).T def makeCentered(data): """Move the mean of the data matrix into the origin. Rows are perceived as datapoints. """ return data - data.mean(axis=0) def pca(data, dim): """ Return the first dim principal components as colums of a matrix. Every row of the matrix resembles a point in the data space. """ assert dim <= data.shape[1], \ "dim must be less or equal than the original dimension" # We have to make a copy of the original data and substract the mean # of every entry data = makeCentered(data) cm = cov(data.T) # OPT only calculate the dim first eigenvectors here # The following calculation may seem a bit "weird" but also correct to me. # The eigenvectors with the dim highest eigenvalues have to be selected # We keep track of the indexes via enumerate to restore the right ordering # later. eigval, eigvec = eig(cm) eigval = [(val, ind) for ind, val in enumerate(eigval)] eigval.sort() eigval[:-dim] = [] # remove all but the highest dim elements # now we have to bring them back in the right order eig_indexes = [(ind, val) for val, ind in eigval] eig_indexes.sort(reverse=True) eig_indexes = [ind for ind, val in eig_indexes] return eigvec.take(eig_indexes, 1).T def pPca(data, dim): """Return a matrix which contains the first `dim` dimensions principal components of data. data is a matrix which's rows correspond to datapoints. Implementation of the 'probabilistic PCA' algorithm. """ num = data.shape[1] data = asmatrix(makeCentered(data)) # Pick a random reduction W = asmatrix(standard_normal((num, dim))) # Save for convergence check W_ = W[:] while True: E = inv(W.T * W) * W.T * data.T W, W_ = data.T * E.T * inv(E * E.T), W if abs(W - W_).max() < 0.001: break return W.T
bsd-3-clause
9954e705f7318676c77a5c6d0c82a80f
28.625
78
0.654776
3.729614
false
false
false
false
pybrain2/pybrain2
pybrain/supervised/trainers/backprop.py
2
11689
from __future__ import print_function from scipy import dot, argmax from random import shuffle from math import isnan from pybrain.supervised.trainers.trainer import Trainer from pybrain.utilities import fListToString from pybrain.auxiliary import GradientDescent from pybrain.tools.functions import abs_error __author__ = 'Daan Wierstra and Tom Schaul' class BackpropTrainer(Trainer): """Trainer that trains the parameters of a module according to a supervised dataset (potentially sequential) by backpropagating the errors (through time).""" def __init__(self, module, dataset=None, learningrate=0.01, lrdecay=1.0, momentum=0., verbose=False, batchlearning=False, weightdecay=0., errfun=None): """Create a BackpropTrainer to train the specified `module` on the specified `dataset`. The learning rate gives the ratio of which parameters are changed into the direction of the gradient. The learning rate decreases by `lrdecay`, which is used to to multiply the learning rate after each training step. The parameters are also adjusted with respect to `momentum`, which is the ratio by which the gradient of the last timestep is used. If `batchlearning` is set, the parameters are updated only at the end of each epoch. Default is False. `weightdecay` corresponds to the weightdecay rate, where 0 is no weight decay at all. Arguments: errfun (func): Function that takes 2 positional arguments, the target (true) and predicted (estimated) output vectors, and returns an estimate of the signed distance to the target (true) output. default = lambda targ, est: (targ - est)) """ Trainer.__init__(self, module) self.setData(dataset) self.verbose = verbose self.batchlearning = batchlearning self.weightdecay = weightdecay self.epoch = 0 self.totalepochs = 0 # set up gradient descender self.descent = GradientDescent() self.descent.alpha = learningrate self.descent.momentum = momentum self.descent.alphadecay = lrdecay self.descent.init(module.params) self.errfun = errfun or abs_error def train(self): """Train the associated module for one epoch.""" assert len(self.ds) > 0, "Dataset cannot be empty." self.module.resetDerivatives() errors = 0 ponderation = 0. shuffledSequences = [] for seq in self.ds._provideSequences(): shuffledSequences.append(seq) shuffle(shuffledSequences) for seq in shuffledSequences: e, p = self._calcDerivs(seq) errors += e ponderation += p if not self.batchlearning: gradient = (self.module.derivs - self.weightdecay * self.module.params) new = self.descent(gradient, errors) if new is not None: self.module.params[:] = new self.module.resetDerivatives() if self.verbose: print("Total error: {z: .12g}".format(z=errors / ponderation)) if self.batchlearning: self.module._setParameters(self.descent(self.module.derivs)) self.epoch += 1 self.totalepochs += 1 return errors / ponderation def _calcDerivs(self, seq): """Calculate error function and backpropagate output errors to yield the gradient.""" self.module.reset() for sample in seq: self.module.activate(sample[0]) error = 0 ponderation = 0. for offset, sample in reversed(list(enumerate(seq))): # need to make a distinction here between datasets containing # importance, and others target = sample[1] outerr = self.errfun(target, self.module.outputbuffer[offset]) if self.verbose > 1: print('output error: {}'.format(outerr)) if len(sample) > 2: importance = sample[2] error += 0.5 * dot(importance, outerr ** 2) ponderation += sum(importance) self.module.backActivate(outerr * importance) else: error += 0.5 * sum(outerr ** 2) ponderation += len(target) # FIXME: the next line keeps arac from producing NaNs. I don't # know why that is, but somehow the __str__ method of the # ndarray class fixes something, str(outerr) self.module.backActivate(outerr) if self.verbose > 1: print('total error so far: {}'.format(error)) if self.verbose > 1: print('TOTAL error: {}'.format(error)) return error, ponderation def _checkGradient(self, dataset=None, silent=False): """Numeric check of the computed gradient for debugging purposes.""" if dataset: self.setData(dataset) res = [] for seq in self.ds._provideSequences(): self.module.resetDerivatives() self._calcDerivs(seq) e = 1e-6 analyticalDerivs = self.module.derivs.copy() numericalDerivs = [] for p in range(self.module.paramdim): storedoldval = self.module.params[p] self.module.params[p] += e righterror, dummy = self._calcDerivs(seq) self.module.params[p] -= 2 * e lefterror, dummy = self._calcDerivs(seq) approxderiv = (righterror - lefterror) / (2 * e) self.module.params[p] = storedoldval numericalDerivs.append(approxderiv) r = list(zip(analyticalDerivs, numericalDerivs)) res.append(r) if not silent: print(r) return res def testOnData(self, dataset=None, verbose=False): """Compute the MSE of the module performance on the given dataset. If no dataset is supplied, the one passed upon Trainer initialization is used.""" if dataset is None: dataset = self.ds dataset.reset() if verbose: print('\nTesting on data:') errors = [] importances = [] ponderatedErrors = [] for seq in dataset._provideSequences(): self.module.reset() e, i = dataset._evaluateSequence(self.module.activate, seq, verbose) importances.append(i) errors.append(e) ponderatedErrors.append(e / i) if verbose: print(('All errors:', ponderatedErrors)) assert sum(importances) > 0 avgErr = sum(errors) / sum(importances) if verbose: print(('Average error:', avgErr)) print(('Max error:', max(ponderatedErrors), 'Median error:', sorted(ponderatedErrors)[len(errors) / 2])) return avgErr def testOnClassData(self, dataset=None, verbose=False, return_targets=False): """Return winner-takes-all classification output on a given dataset. If no dataset is given, the dataset passed during Trainer initialization is used. If return_targets is set, also return corresponding target classes. """ if dataset is None: dataset = self.ds dataset.reset() out = [] targ = [] for seq in dataset._provideSequences(): self.module.reset() for input, target in seq: res = self.module.activate(input) out.append(argmax(res)) targ.append(argmax(target)) if return_targets: return out, targ else: return out def trainUntilConvergence(self, dataset=None, maxEpochs=None, verbose=None, continueEpochs=10, validationProportion=0.25, trainingData=None, validationData=None, convergence_threshold=10): """Train the module on the dataset until it converges. Return the module with the parameters that gave the minimal validation error. If no dataset is given, the dataset passed during Trainer initialization is used. validationProportion is the ratio of the dataset that is used for the validation dataset. If the training and validation data is already set, the splitPropotion is ignored If maxEpochs is given, at most that many epochs are trained. Each time validation error hits a minimum, try for continueEpochs epochs to find a better one.""" epochs = 0 if dataset is None: dataset = self.ds if verbose is None: verbose = self.verbose if trainingData is None or validationData is None: # Split the dataset randomly: validationProportion of the samples # for validation. trainingData, validationData = ( dataset.splitWithProportion(1 - validationProportion)) if not (len(trainingData) > 0 and len(validationData)): raise ValueError("Provided dataset too small to be split into " "training and validation sets with proportion " + str(validationProportion)) self.ds = trainingData bestweights = self.module.params.copy() bestverr = self.testOnData(validationData) bestepoch = 0 self.trainingErrors = [] self.validationErrors = [bestverr] while True: trainingError = self.train() validationError = self.testOnData(validationData) if isnan(trainingError) or isnan(validationError): raise Exception("Training produced NaN results") self.trainingErrors.append(trainingError) self.validationErrors.append(validationError) if epochs == 0 or self.validationErrors[-1] < bestverr: # one update is always done bestverr = self.validationErrors[-1] bestweights = self.module.params.copy() bestepoch = epochs if maxEpochs is not None and epochs >= maxEpochs: self.module.params[:] = bestweights break epochs += 1 if len(self.validationErrors) >= continueEpochs * 2: # have the validation errors started going up again? # compare the average of the last few to the previous few old = self.validationErrors[-continueEpochs * 2:- continueEpochs] new = self.validationErrors[-continueEpochs:] if min(new) > max(old): self.module.params[:] = bestweights break lastnew = round(new[-1], convergence_threshold) if sum(round(y, convergence_threshold) - lastnew for y in new) == 0: self.module.params[:] = bestweights break self.ds = dataset if verbose: print(('train-errors:', fListToString(self.trainingErrors, 6))) print(('valid-errors:', fListToString(self.validationErrors, 6))) # slice off the inital bestverr return self.trainingErrors[:bestepoch], self.validationErrors[1:1 + bestepoch]
bsd-3-clause
bf82f5c85a687d8887518e79718cd24b
40.746429
86
0.583882
4.544712
false
false
false
false
pybrain2/pybrain2
pybrain/rl/environments/twoplayergames/tasks/capturetask.py
31
3766
__author__ = 'Tom Schaul, tom@idsia.ch' from pybrain.rl.environments.episodic import EpisodicTask from inspect import isclass from pybrain.utilities import Named from pybrain.rl.environments.twoplayergames import CaptureGame from pybrain.rl.environments.twoplayergames.capturegameplayers import RandomCapturePlayer, ModuleDecidingPlayer from pybrain.rl.environments.twoplayergames.capturegameplayers.captureplayer import CapturePlayer from pybrain.structure.modules.module import Module class CaptureGameTask(EpisodicTask, Named): """ The task of winning the maximal number of capture games against a fixed opponent. """ # first game, opponent is black opponentStart = True # on subsequent games, starting players are alternating alternateStarting = False # numerical reward value attributed to winning winnerReward = 1. # coefficient determining the importance of long vs. short games w.r. to winning/losing numMovesCoeff = 0. # average over some games for evaluations averageOverGames = 10 noisy = True def __init__(self, size, opponent = None, **args): EpisodicTask.__init__(self, CaptureGame(size)) self.setArgs(**args) if opponent == None: opponent = RandomCapturePlayer(self.env) elif isclass(opponent): # assume the agent can be initialized without arguments then. opponent = opponent(self.env) else: opponent.game = self.env if not self.opponentStart: opponent.color = CaptureGame.WHITE self.opponent = opponent self.maxmoves = self.env.size * self.env.size self.minmoves = 3 self.reset() def reset(self): self.switched = False EpisodicTask.reset(self) if self.opponent.color == CaptureGame.BLACK: # first move by opponent EpisodicTask.performAction(self, self.opponent.getAction()) def isFinished(self): res = self.env.gameOver() if res and self.alternateStarting and not self.switched: # alternate starting player self.opponent.color *= -1 self.switched = True return res def getReward(self): """ Final positive reward for winner, negative for loser. """ if self.isFinished(): win = (self.env.winner != self.opponent.color) moves = self.env.movesDone res = self.winnerReward - self.numMovesCoeff * (moves -self.minmoves)/(self.maxmoves-self.minmoves) if not win: res *= -1 if self.alternateStarting and self.switched: # opponent color has been inverted after the game! res *= -1 return res else: return 0 def performAction(self, action): EpisodicTask.performAction(self, action) if not self.isFinished(): EpisodicTask.performAction(self, self.opponent.getAction()) def f(self, x): """ If a module is given, wrap it into a ModuleDecidingAgent before evaluating it. Also, if applicable, average the result over multiple games. """ if isinstance(x, Module): agent = ModuleDecidingPlayer(x, self.env, greedySelection = True) elif isinstance(x, CapturePlayer): agent = x else: raise NotImplementedError('Missing implementation for '+x.__class__.__name__+' evaluation') res = 0 agent.game = self.env self.opponent.game = self.env for _ in range(self.averageOverGames): agent.color = -self.opponent.color x = EpisodicTask.f(self, agent) res += x return res / float(self.averageOverGames)
bsd-3-clause
4d04451032ffdb586dff51afd9bee288
35.563107
111
0.641795
4.071351
false
false
false
false
cobrateam/splinter
splinter/driver/zopetestbrowser.py
1
13544
# -*- coding: utf-8 -*- # Copyright 2012 splinter authors. All rights reserved. # Use of this source code is governed by a BSD-style # license that can be found in the LICENSE file. from __future__ import unicode_literals import mimetypes import re import time import warnings import lxml.html from lxml.cssselect import CSSSelector from zope.testbrowser.browser import Browser, ListControl from splinter.element_list import ElementList from splinter.exceptions import ElementDoesNotExist from splinter.driver import DriverAPI, ElementAPI from splinter.driver.element_present import ElementPresentMixIn from splinter.driver.find_links import FindLinks from splinter.driver.xpath_utils import _concat_xpath_from_str from splinter.cookie_manager import CookieManagerAPI class CookieManager(CookieManagerAPI): def add(self, cookie, **kwargs): for key, value in cookie.items(): kwargs['name'] = key kwargs['value'] = value if key not in self.driver.cookies: self.driver.cookies.create(**kwargs) else: self.driver.cookies.change(**kwargs) def delete(self, *cookies): if cookies: for cookie in cookies: try: del self.driver.cookies[cookie] except KeyError: pass else: warnings.warn( 'Deleting all cookies via CookieManager.delete() with no arguments ' 'has been deprecated. use CookieManager.delete_all().', FutureWarning, ) self.delete_all() def delete_all(self): self.driver.cookies.clearAll() def all(self, verbose=False): # NOQA: A003 cookies = {} for key, value in self.driver.cookies.items(): cookies[key] = value return cookies def __getitem__(self, item): return self.driver.cookies[item] def __contains__(self, key): return key in self.driver.cookies def __eq__(self, other_object): if isinstance(other_object, dict): return dict(self.driver.cookies) == other_object return False class ZopeTestBrowser(ElementPresentMixIn, DriverAPI): driver_name = "zope.testbrowser" def __init__(self, wait_time=2): self.wait_time = wait_time self._browser = Browser() self._cookie_manager = CookieManager(self._browser) self._last_urls = [] self.links = FindLinks(self) def __enter__(self): return self def __exit__(self, exc_type, exc_value, traceback): pass def visit(self, url): self._browser.open(url) def back(self): self._last_urls.insert(0, self.url) self._browser.goBack() def forward(self): try: self.visit(self._last_urls.pop()) except IndexError: pass def reload(self): self._browser.reload() def quit(self): # NOQA: A003 pass @property def htmltree(self): try: html = self.html.decode("utf-8") except AttributeError: html = self.html return lxml.html.fromstring(html) @property def title(self): return self._browser.title @property def html(self): return self._browser.contents @property def url(self): return self._browser.url def find_option_by_value(self, value): html = self.htmltree element = html.xpath('//option[@value="%s"]' % value)[0] control = self._browser.getControl(element.text) return ElementList( [ZopeTestBrowserOptionElement(control, self)], find_by="value", query=value ) def find_option_by_text(self, text): html = self.htmltree element = html.xpath('//option[normalize-space(text())="%s"]' % text)[0] control = self._browser.getControl(element.text) return ElementList( [ZopeTestBrowserOptionElement(control, self)], find_by="text", query=text ) def find_by_css(self, selector): xpath = CSSSelector(selector).path return self.find_by_xpath( xpath, original_find="css", original_query=selector ) def get_control(self, xpath_element): return xpath_element def find_by_xpath(self, xpath, original_find=None, original_query=None): html = self.htmltree elements = [] for xpath_element in html.xpath(xpath): if self._element_is_link(xpath_element): return self._find_links_by_xpath(xpath) elif self._element_is_control(xpath_element) and xpath_element.name: return self.find_by_name(xpath_element.name) else: elements.append(self.get_control(xpath_element)) find_by = original_find or "xpath" query = original_query or xpath return ElementList( [ZopeTestBrowserElement(element, self) for element in elements], find_by=find_by, query=query, ) def find_by_tag(self, tag): return self.find_by_xpath( "//%s" % tag, original_find="tag", original_query=tag ) def find_by_value(self, value): elem = self.find_by_xpath( '//*[@value="%s"]' % value, original_find="value", original_query=value ) if elem: return elem return self.find_by_xpath('//*[.="%s"]' % value) def find_by_text(self, text): xpath_str = _concat_xpath_from_str(text) return self.find_by_xpath( xpath_str, original_find="text", original_query=text, ) def find_by_id(self, id_value): return self.find_by_xpath( '//*[@id="%s"][1]' % id_value, original_find="id", original_query=id_value, ) def find_by_name(self, name): elements = [] index = 0 while True: try: control = self._browser.getControl(name=name, index=index) elements.append(control) index += 1 except LookupError: break except NotImplementedError: break return ElementList( [ZopeTestBrowserControlElement(element, self) for element in elements], find_by="name", query=name, ) def fill(self, name, value): self.find_by_name(name=name).first._control.value = value def fill_form(self, field_values, form_id=None, name=None, ignore_missing=False): form = self._browser if name or form_id: form = self._browser.getForm(name=name, id=form_id) for name, value in field_values.items(): try: control = form.getControl(name=name) if control.type == "checkbox": if value: control.value = control.options else: control.value = [] elif control.type == "radio": control.value = [ option for option in control.options if option == value ] elif control.type == "select": control.value = [value] else: control.value = value except NotImplementedError as e: if not ignore_missing: raise NotImplementedError(e) def choose(self, name, value): control = self._browser.getControl(name=name) control.value = [option for option in control.options if option == value] def check(self, name): control = self._browser.getControl(name=name) control.value = control.options def uncheck(self, name): control = self._browser.getControl(name=name) control.value = [] def attach_file(self, name, file_path): filename = file_path.split("/")[-1] control = self._browser.getControl(name=name) content_type, _ = mimetypes.guess_type(file_path) with open(file_path, 'rb') as f: control.add_file(f, content_type, filename) def _find_links_by_xpath(self, xpath): html = self.htmltree links = html.xpath(xpath) return ElementList( [ZopeTestBrowserLinkElement(link, self) for link in links], find_by="xpath", query=xpath, ) def select(self, name, value): self.find_by_name(name).first._control.value = [value] def is_text_present(self, text, wait_time=None): wait_time = wait_time or self.wait_time end_time = time.time() + wait_time while time.time() < end_time: if self._is_text_present(text): return True return False def _is_text_present(self, text): try: body = self.find_by_tag("body").first return text in body.text except ElementDoesNotExist: # This exception will be thrown if the body tag isn't present # This has occasionally been observed. Assume that the # page isn't fully loaded yet return False def is_text_not_present(self, text, wait_time=None): wait_time = wait_time or self.wait_time end_time = time.time() + wait_time while time.time() < end_time: if not self._is_text_present(text): return True return False def _element_is_link(self, element): return element.tag == "a" def _element_is_control(self, element): return hasattr(element, "type") @property def cookies(self): return self._cookie_manager re_extract_inner_html = re.compile(r"^<[^<>]+>(.*)</[^<>]+>$") class ZopeTestBrowserElement(ElementAPI): def __init__(self, element, parent): self._element = element self.parent = parent def __getitem__(self, attr): return self._element.attrib[attr] def find_by_css(self, selector): elements = self._element.cssselect(selector) return ElementList([self.__class__(element, self) for element in elements]) def find_by_xpath(self, selector): elements = self._element.xpath(selector) return ElementList([self.__class__(element, self) for element in elements]) def find_by_name(self, name): elements = self._element.cssselect('[name="%s"]' % name) return ElementList([self.__class__(element, self) for element in elements]) def find_by_tag(self, name): elements = self._element.cssselect(name) return ElementList([self.__class__(element, self) for element in elements]) def find_by_value(self, value): elements = self._element.cssselect('[value="%s"]' % value) return ElementList([self.__class__(element, self) for element in elements]) def find_by_text(self, text): # Add a period to the xpath to search only inside the parent. xpath_str = '.{}'.format(_concat_xpath_from_str(text)) return self.find_by_xpath(xpath_str) def find_by_id(self, id): # NOQA: A002 elements = self._element.cssselect("#%s" % id) return ElementList([self.__class__(element, self) for element in elements]) @property def value(self): return self._element.text_content() @property def text(self): return self.value @property def outer_html(self): return lxml.html.tostring(self._element, encoding="unicode").strip() @property def html(self): return re_extract_inner_html.match(self.outer_html).group(1) def has_class(self, class_name): return len(self._element.find_class(class_name)) > 0 class ZopeTestBrowserLinkElement(ZopeTestBrowserElement): def __init__(self, element, parent): super(ZopeTestBrowserLinkElement, self).__init__(element, parent) self._browser = parent._browser def __getitem__(self, attr): return super(ZopeTestBrowserLinkElement, self).__getitem__(attr) def click(self): return self._browser.open(self["href"]) class ZopeTestBrowserControlElement(ZopeTestBrowserElement): def __init__(self, control, parent): self._control = control self.parent = parent def __getitem__(self, attr): try: return getattr(self._control._control, attr) except AttributeError: return self._control._control.attrs[attr] @property def value(self): value = self._control.value if isinstance(self._control, ListControl) and len(value) == 1: return value[0] return value @property def checked(self): return bool(self._control.value) def click(self): return self._control.click() def fill(self, value): self._control.value = value def select(self, value): self._control.value = [value] class ZopeTestBrowserOptionElement(ZopeTestBrowserElement): def __init__(self, control, parent): self._control = control self.parent = parent def __getitem__(self, attr): return getattr(self._control, attr) @property def text(self): return self._control.labels[0] @property def value(self): return self._control.optionValue @property def selected(self): return self._control.selected
bsd-3-clause
bb54aa0fa4806c25d6923f5ebf2f999d
29.367713
87
0.588231
4.106731
false
false
false
false
ahmia/ahmia-site
ahmia/ahmia/management/commands/cleanup_db.py
1
1039
from datetime import timedelta from django.conf import settings from django.core.management import BaseCommand from django.utils import timezone from ... import utils from ...models import TorStats, I2PStats, SearchQuery, SearchResultsClick class Command(BaseCommand): def __init__(self): super(Command, self).__init__() self.days_to_keep = settings.USAGE_STATS_DAYS def handle(self, *args, **options): self.cleanup_stats_etc() def cleanup_stats_etc(self): # *Stats tables oldest_day_to_keep = utils.timezone_today() - timedelta(days=self.days_to_keep) TorStats.objects.filter(day__lt=oldest_day_to_keep).delete() I2PStats.objects.filter(day__lt=oldest_day_to_keep).delete() # SearchQueries and Clicks oldest_datetime_to_keep = timezone.now() - timedelta(days=self.days_to_keep) SearchQuery.objects.filter(updated__lt=oldest_datetime_to_keep).delete() SearchResultsClick.objects.filter(updated__lt=oldest_datetime_to_keep).delete()
bsd-3-clause
7d226a562d0786c9ab39efa0be075728
36.107143
87
0.698749
3.632867
false
false
false
false
django-helpdesk/django-helpdesk
helpdesk/views/public.py
1
9647
""" django-helpdesk - A Django powered ticket tracker for small enterprise. (c) Copyright 2008 Jutda. All Rights Reserved. See LICENSE for details. views/public.py - All public facing views, eg non-staff (no authentication required) views. """ from django.conf import settings from django.core.exceptions import ImproperlyConfigured, ObjectDoesNotExist, PermissionDenied from django.http import HttpResponseRedirect from django.shortcuts import render from django.urls import reverse from django.utils.translation import gettext as _ from django.views.decorators.clickjacking import xframe_options_exempt from django.views.decorators.csrf import csrf_exempt from django.views.generic.base import TemplateView from django.views.generic.edit import FormView from helpdesk import settings as helpdesk_settings from helpdesk.decorators import is_helpdesk_staff, protect_view from helpdesk.lib import text_is_spam from helpdesk.models import Queue, Ticket, UserSettings from helpdesk.user import huser_from_request import helpdesk.views.abstract_views as abstract_views import helpdesk.views.staff as staff from importlib import import_module import logging from urllib.parse import quote logger = logging.getLogger(__name__) def create_ticket(request, *args, **kwargs): if is_helpdesk_staff(request.user): return staff.CreateTicketView.as_view()(request, *args, **kwargs) else: return CreateTicketView.as_view()(request, *args, **kwargs) class BaseCreateTicketView(abstract_views.AbstractCreateTicketMixin, FormView): def get_form_class(self): try: the_module, the_form_class = helpdesk_settings.HELPDESK_PUBLIC_TICKET_FORM_CLASS.rsplit( ".", 1) the_module = import_module(the_module) the_form_class = getattr(the_module, the_form_class) except Exception as e: raise ImproperlyConfigured( f"Invalid custom form class {helpdesk_settings.HELPDESK_PUBLIC_TICKET_FORM_CLASS}" ) from e return the_form_class def dispatch(self, *args, **kwargs): request = self.request if not request.user.is_authenticated and helpdesk_settings.HELPDESK_REDIRECT_TO_LOGIN_BY_DEFAULT: return HttpResponseRedirect(reverse('login')) if is_helpdesk_staff(request.user) or \ (request.user.is_authenticated and helpdesk_settings.HELPDESK_ALLOW_NON_STAFF_TICKET_UPDATE): try: if request.user.usersettings_helpdesk.login_view_ticketlist: return HttpResponseRedirect(reverse('helpdesk:list')) else: return HttpResponseRedirect(reverse('helpdesk:dashboard')) except UserSettings.DoesNotExist: return HttpResponseRedirect(reverse('helpdesk:dashboard')) return super().dispatch(*args, **kwargs) def get_initial(self): initial_data = super().get_initial() # add pre-defined data for public ticket if hasattr(settings, 'HELPDESK_PUBLIC_TICKET_QUEUE'): # get the requested queue; return an error if queue not found try: initial_data['queue'] = Queue.objects.get( slug=settings.HELPDESK_PUBLIC_TICKET_QUEUE, allow_public_submission=True ).id except Queue.DoesNotExist as e: logger.fatal( "Public queue '%s' is configured as default but can't be found", settings.HELPDESK_PUBLIC_TICKET_QUEUE ) raise ImproperlyConfigured( "Wrong public queue configuration") from e if hasattr(settings, 'HELPDESK_PUBLIC_TICKET_PRIORITY'): initial_data['priority'] = settings.HELPDESK_PUBLIC_TICKET_PRIORITY if hasattr(settings, 'HELPDESK_PUBLIC_TICKET_DUE_DATE'): initial_data['due_date'] = settings.HELPDESK_PUBLIC_TICKET_DUE_DATE return initial_data def get_form_kwargs(self, *args, **kwargs): kwargs = super().get_form_kwargs(*args, **kwargs) if '_hide_fields_' in self.request.GET: kwargs['hidden_fields'] = self.request.GET.get( '_hide_fields_', '').split(',') kwargs['readonly_fields'] = self.request.GET.get( '_readonly_fields_', '').split(',') return kwargs def form_valid(self, form): request = self.request if text_is_spam(form.cleaned_data['body'], request): # This submission is spam. Let's not save it. return render(request, template_name='helpdesk/public_spam.html') else: ticket = form.save( user=self.request.user if self.request.user.is_authenticated else None) try: return HttpResponseRedirect('%s?ticket=%s&email=%s&key=%s' % ( reverse('helpdesk:public_view'), ticket.ticket_for_url, quote(ticket.submitter_email), ticket.secret_key) ) except ValueError: # if someone enters a non-int string for the ticket return HttpResponseRedirect(reverse('helpdesk:home')) class CreateTicketIframeView(BaseCreateTicketView): template_name = 'helpdesk/public_create_ticket_iframe.html' @csrf_exempt @xframe_options_exempt def dispatch(self, *args, **kwargs): return super().dispatch(*args, **kwargs) def form_valid(self, form): if super().form_valid(form).status_code == 302: return HttpResponseRedirect(reverse('helpdesk:success_iframe')) class SuccessIframeView(TemplateView): template_name = 'helpdesk/success_iframe.html' @xframe_options_exempt def dispatch(self, *args, **kwargs): return super().dispatch(*args, **kwargs) class CreateTicketView(BaseCreateTicketView): template_name = 'helpdesk/public_create_ticket.html' def get_form(self, form_class=None): form = super().get_form(form_class) # Add the CSS error class to the form in order to better see them in # the page form.error_css_class = 'text-danger' return form class Homepage(CreateTicketView): template_name = 'helpdesk/public_homepage.html' def get_context_data(self, **kwargs): context = super().get_context_data(**kwargs) context['kb_categories'] = huser_from_request( self.request).get_allowed_kb_categories() return context def search_for_ticket(request, error_message=None): if hasattr(settings, 'HELPDESK_VIEW_A_TICKET_PUBLIC') and settings.HELPDESK_VIEW_A_TICKET_PUBLIC: email = request.GET.get('email', None) return render(request, 'helpdesk/public_view_form.html', { 'ticket': False, 'email': email, 'error_message': error_message, 'helpdesk_settings': helpdesk_settings, }) else: raise PermissionDenied( "Public viewing of tickets without a secret key is forbidden.") @protect_view def view_ticket(request): ticket_req = request.GET.get('ticket', None) email = request.GET.get('email', None) key = request.GET.get('key', '') if not (ticket_req and email): if ticket_req is None and email is None: return search_for_ticket(request) else: return search_for_ticket(request, _('Missing ticket ID or e-mail address. Please try again.')) queue, ticket_id = Ticket.queue_and_id_from_query(ticket_req) try: if hasattr(settings, 'HELPDESK_VIEW_A_TICKET_PUBLIC') and settings.HELPDESK_VIEW_A_TICKET_PUBLIC: ticket = Ticket.objects.get( id=ticket_id, submitter_email__iexact=email) else: ticket = Ticket.objects.get( id=ticket_id, submitter_email__iexact=email, secret_key__iexact=key) except (ObjectDoesNotExist, ValueError): return search_for_ticket(request, _('Invalid ticket ID or e-mail address. Please try again.')) if is_helpdesk_staff(request.user): redirect_url = reverse('helpdesk:view', args=[ticket_id]) if 'close' in request.GET: redirect_url += '?close' return HttpResponseRedirect(redirect_url) if 'close' in request.GET and ticket.status == Ticket.RESOLVED_STATUS: from helpdesk.views.staff import update_ticket # Trick the update_ticket() view into thinking it's being called with # a valid POST. request.POST = { 'new_status': Ticket.CLOSED_STATUS, 'public': 1, 'title': ticket.title, 'comment': _('Submitter accepted resolution and closed ticket'), } if ticket.assigned_to: request.POST['owner'] = ticket.assigned_to.id request.GET = {} return update_ticket(request, ticket_id, public=True) # redirect user back to this ticket if possible. redirect_url = '' if helpdesk_settings.HELPDESK_NAVIGATION_ENABLED: redirect_url = reverse('helpdesk:view', args=[ticket_id]) return render(request, 'helpdesk/public_view_ticket.html', { 'key': key, 'mail': email, 'ticket': ticket, 'helpdesk_settings': helpdesk_settings, 'next': redirect_url, }) def change_language(request): return_to = '' if 'return_to' in request.GET: return_to = request.GET['return_to'] return render(request, 'helpdesk/public_change_language.html', {'next': return_to})
bsd-3-clause
849b25adf0750d28d1f976f435c4f1c6
37.899194
106
0.639577
4.058477
false
false
false
false
django-helpdesk/django-helpdesk
demo/demodesk/config/settings.py
1
7279
""" Django settings for django-helpdesk demodesk project. For more information on this file, see https://docs.djangoproject.com/en/1.11/topics/settings/ For the full list of settings and their values, see https://docs.djangoproject.com/en/1.11/ref/settings/ """ import os # Build paths inside the project like this: os.path.join(BASE_DIR, ...) BASE_DIR = os.path.dirname(os.path.dirname(os.path.abspath(__file__))) # Quick-start development settings - unsuitable for production # See https://docs.djangoproject.com/en/1.11/howto/deployment/checklist/ # SECURITY WARNING: keep the secret key used in production secret! SECRET_KEY = '_crkn1+fnzu5$vns_-d+^ayiq%z4k*s!!ag0!mfy36(y!vrazd' # SECURITY WARNING: don't run with debug turned on in production! DEBUG = True ALLOWED_HOSTS = [] # SECURITY WARNING: you probably want to configure your server # to use HTTPS with secure cookies, then you'd want to set # the following settings: # #SECURE_PROXY_SSL_HEADER = ('HTTP_X_FORWARDED_PROTO', 'https') #SESSION_COOKIE_SECURE = True #CSRF_COOKIE_SECURE = True # # We leave them commented out here because most likely for # an internal demo you don't need such security, but please # remember when setting up your own development / production server! # Application definition INSTALLED_APPS = [ 'django.contrib.admin', 'django.contrib.auth', 'django.contrib.contenttypes', 'django.contrib.sessions', 'django.contrib.messages', 'django.contrib.staticfiles', 'django.contrib.sites', 'django.contrib.humanize', 'bootstrap4form', 'account', # Required by pinax-teams 'pinax.invitations', # required by pinax-teams 'pinax.teams', # team support 'reversion', # required by pinax-teams 'helpdesk', # This is us! 'rest_framework', # required for the API ] MIDDLEWARE = [ 'django.middleware.security.SecurityMiddleware', 'django.contrib.sessions.middleware.SessionMiddleware', 'django.middleware.common.CommonMiddleware', 'django.middleware.csrf.CsrfViewMiddleware', 'django.contrib.auth.middleware.AuthenticationMiddleware', 'django.contrib.messages.middleware.MessageMiddleware', 'django.middleware.clickjacking.XFrameOptionsMiddleware', ] ROOT_URLCONF = 'demo.demodesk.config.urls' TEMPLATES = [ { 'BACKEND': 'django.template.backends.django.DjangoTemplates', 'DIRS': [], 'APP_DIRS': True, 'OPTIONS': { 'debug': True, 'context_processors': [ 'django.template.context_processors.debug', 'django.template.context_processors.request', 'django.contrib.auth.context_processors.auth', 'django.contrib.messages.context_processors.messages', ], }, }, ] WSGI_APPLICATION = 'demo.demodesk.config.wsgi.application' # django-helpdesk configuration settings # You can override django-helpdesk's defaults by redefining them here. # To see what settings are available, see the docs/configuration.rst # file for more information. # Some common settings are below. HELPDESK_DEFAULT_SETTINGS = { 'use_email_as_submitter': True, 'email_on_ticket_assign': True, 'email_on_ticket_change': True, 'login_view_ticketlist': True, 'email_on_ticket_apichange': True, 'preset_replies': True, 'tickets_per_page': 25 } # Should the public web portal be enabled? HELPDESK_PUBLIC_ENABLED = True HELPDESK_VIEW_A_TICKET_PUBLIC = True HELPDESK_SUBMIT_A_TICKET_PUBLIC = True # Should the Knowledgebase be enabled? HELPDESK_KB_ENABLED = True HELPDESK_TICKETS_TIMELINE_ENABLED = True # Allow users to change their passwords HELPDESK_SHOW_CHANGE_PASSWORD = True # Activate the API HELPDESK_ACTIVATE_API_ENDPOINT = True # Instead of showing the public web portal first, # we can instead redirect users straight to the login page. HELPDESK_REDIRECT_TO_LOGIN_BY_DEFAULT = False LOGIN_URL = 'helpdesk:login' LOGIN_REDIRECT_URL = 'helpdesk:home' # Database # - by default, we use SQLite3 for the demo, but you can also # configure MySQL or PostgreSQL, see the docs for more: # https://docs.djangoproject.com/en/1.11/ref/settings/#databases DATABASES = { 'default': { 'ENGINE': 'django.db.backends.sqlite3', 'NAME': os.path.join(BASE_DIR, 'db.sqlite3'), } } # Sites # - this allows hosting of more than one site from a single server, # in practice you can probably just leave this default if you only # host a single site, but read more in the docs: # https://docs.djangoproject.com/en/1.11/ref/contrib/sites/ SITE_ID = 1 # Sessions # https://docs.djangoproject.com/en/1.11/topics/http/sessions SESSION_COOKIE_AGE = 86400 # = 1 day # For better default security, set these cookie flags, but # these are likely to cause problems when testing locally #CSRF_COOKIE_SECURE = True #SESSION_COOKIE_SECURE = True #CSRF_COOKIE_HTTPONLY = True #SESSION_COOKIE_HTTPONLY = True # Password validation # https://docs.djangoproject.com/en/1.11/ref/settings/#auth-password-validators AUTH_PASSWORD_VALIDATORS = [ { 'NAME': 'django.contrib.auth.password_validation.UserAttributeSimilarityValidator', }, { 'NAME': 'django.contrib.auth.password_validation.MinimumLengthValidator', }, { 'NAME': 'django.contrib.auth.password_validation.CommonPasswordValidator', }, { 'NAME': 'django.contrib.auth.password_validation.NumericPasswordValidator', }, ] # Email # https://docs.djangoproject.com/en/1.11/topics/email/ # This demo uses the console backend, which simply prints emails to the console # rather than actually sending them out. DEFAULT_FROM_EMAIL = 'helpdesk@example.com' SERVER_EMAIL = 'helpdesk@example.com' EMAIL_BACKEND = 'django.core.mail.backends.console.EmailBackend' # If you want to test sending real emails, uncomment and modify the following: #EMAIL_BACKEND = 'django.core.mail.backends.smtp.EmailBackend' #EMAIL_HOST = 'smtp.example.com' #EMAIL_PORT = '25' # Internationalization # https://docs.djangoproject.com/en/1.11/topics/i18n/ # By default, django-helpdesk uses en, but other languages are also available. # The most complete translations are: es-MX, ru, zh-Hans # Contribute to our translations via Transifex if you can! # See CONTRIBUTING.rst for more info. LANGUAGE_CODE = 'en-US' TIME_ZONE = 'UTC' USE_I18N = True USE_L10N = True USE_TZ = True # Static files (CSS, JavaScript, Images) # https://docs.djangoproject.com/en/1.11/howto/static-files/ STATIC_URL = '/static/' # static root needs to be defined in order to use collectstatic STATIC_ROOT = os.path.join(BASE_DIR, 'static') # MEDIA_ROOT is where media uploads are stored. # We set this to a directory to host file attachments created # with tickets. MEDIA_URL = '/media/' MEDIA_ROOT = os.path.join(BASE_DIR, 'media') # Fixtures # https://docs.djangoproject.com/en/1.11/ref/settings/#std:setting-FIXTURE_DIRS # - This is only necessary to make the demo project work, not needed for # your own projects unless you make your own fixtures FIXTURE_DIRS = [os.path.join(BASE_DIR, 'fixtures')] # for Django 3.2+, set default for autofields: DEFAULT_AUTO_FIELD = 'django.db.models.AutoField' try: from .local_settings import * except ImportError: pass
bsd-3-clause
d65ee46c15fef37afe5e5f72cd924b16
28.831967
91
0.716994
3.441608
false
false
false
false
lepture/flask-oauthlib
example/douban.py
16
1583
from flask import Flask, redirect, url_for, session, request, jsonify from flask_oauthlib.client import OAuth app = Flask(__name__) app.debug = True app.secret_key = 'development' oauth = OAuth(app) douban = oauth.remote_app( 'douban', consumer_key='0cfc3c5d9f873b1826f4b518de95b148', consumer_secret='3e209e4f9ecf6a4a', base_url='https://api.douban.com/', request_token_url=None, request_token_params={'scope': 'douban_basic_common,shuo_basic_r'}, access_token_url='https://www.douban.com/service/auth2/token', authorize_url='https://www.douban.com/service/auth2/auth', access_token_method='POST', ) @app.route('/') def index(): if 'douban_token' in session: resp = douban.get('shuo/v2/statuses/home_timeline') return jsonify(status=resp.status, data=resp.data) return redirect(url_for('login')) @app.route('/login') def login(): return douban.authorize(callback=url_for('authorized', _external=True)) @app.route('/logout') def logout(): session.pop('douban_token', None) return redirect(url_for('index')) @app.route('/login/authorized') def authorized(): resp = douban.authorized_response() if resp is None: return 'Access denied: reason=%s error=%s' % ( request.args['error_reason'], request.args['error_description'] ) session['douban_token'] = (resp['access_token'], '') return redirect(url_for('index')) @douban.tokengetter def get_douban_oauth_token(): return session.get('douban_token') if __name__ == '__main__': app.run()
bsd-3-clause
ee74c82d8c27bee1c6aae0b4cc3f74fc
25.383333
75
0.660139
3.122288
false
false
false
false
lepture/flask-oauthlib
flask_oauthlib/contrib/apps.py
6
7551
""" flask_oauthlib.contrib.apps ~~~~~~~~~~~~~~~~~~~~~~~~~~~ The bundle of remote app factories for famous third platforms. Usage:: from flask import Flask from flask_oauthlib.client import OAuth from flask_oauthlib.contrib.apps import github app = Flask(__name__) oauth = OAuth(app) github.register_to(oauth, scope=['user:email']) github.register_to(oauth, name='github2') Of course, it requires consumer keys in your config:: GITHUB_CONSUMER_KEY = '' GITHUB_CONSUMER_SECRET = '' GITHUB2_CONSUMER_KEY = '' GITHUB2_CONSUMER_SECRET = '' Some apps with OAuth 1.0a such as Twitter could not accept the ``scope`` argument. Contributed by: tonyseek """ import copy from oauthlib.common import unicode_type, bytes_type __all__ = ['douban', 'dropbox', 'facebook', 'github', 'google', 'linkedin', 'twitter', 'weibo'] class RemoteAppFactory(object): """The factory to create remote app and bind it to given extension. :param default_name: the default name which be used for registering. :param kwargs: the pre-defined kwargs. :param docstring: the docstring of factory. """ def __init__(self, default_name, kwargs, docstring=''): assert 'name' not in kwargs assert 'register' not in kwargs self.default_name = default_name self.kwargs = kwargs self._kwargs_processor = None self.__doc__ = docstring.lstrip() def register_to(self, oauth, name=None, **kwargs): """Creates a remote app and registers it.""" kwargs = self._process_kwargs( name=(name or self.default_name), **kwargs) return oauth.remote_app(**kwargs) def create(self, oauth, **kwargs): """Creates a remote app only.""" kwargs = self._process_kwargs( name=self.default_name, register=False, **kwargs) return oauth.remote_app(**kwargs) def kwargs_processor(self, fn): """Sets a function to process kwargs before creating any app.""" self._kwargs_processor = fn return fn def _process_kwargs(self, **kwargs): final_kwargs = copy.deepcopy(self.kwargs) # merges with pre-defined kwargs final_kwargs.update(copy.deepcopy(kwargs)) # use name as app key final_kwargs.setdefault('app_key', final_kwargs['name'].upper()) # processes by pre-defined function if self._kwargs_processor is not None: final_kwargs = self._kwargs_processor(**final_kwargs) return final_kwargs def make_scope_processor(default_scope): def processor(**kwargs): # request_token_params scope = kwargs.pop('scope', [default_scope]) # default scope if not isinstance(scope, (unicode_type, bytes_type)): scope = ','.join(scope) # allows list-style scope request_token_params = kwargs.setdefault('request_token_params', {}) request_token_params.setdefault('scope', scope) # doesn't override return kwargs return processor douban = RemoteAppFactory('douban', { 'base_url': 'https://api.douban.com/v2/', 'request_token_url': None, 'access_token_url': 'https://www.douban.com/service/auth2/token', 'authorize_url': 'https://www.douban.com/service/auth2/auth', 'access_token_method': 'POST', }, """ The OAuth app for douban.com API. :param scope: optional. default: ``['douban_basic_common']``. see also: http://developers.douban.com/wiki/?title=oauth2 """) douban.kwargs_processor(make_scope_processor('douban_basic_common')) dropbox = RemoteAppFactory('dropbox', { 'base_url': 'https://www.dropbox.com/1/', 'request_token_url': None, 'access_token_url': 'https://api.dropbox.com/1/oauth2/token', 'authorize_url': 'https://www.dropbox.com/1/oauth2/authorize', 'access_token_method': 'POST', 'request_token_params': {}, }, """The OAuth app for Dropbox API.""") facebook = RemoteAppFactory('facebook', { 'request_token_params': {'scope': 'email'}, 'base_url': 'https://graph.facebook.com', 'request_token_url': None, 'access_token_url': '/oauth/access_token', 'authorize_url': 'https://www.facebook.com/dialog/oauth', }, """ The OAuth app for Facebook API. :param scope: optional. default: ``['email']``. """) facebook.kwargs_processor(make_scope_processor('email')) github = RemoteAppFactory('github', { 'base_url': 'https://api.github.com/', 'request_token_url': None, 'access_token_method': 'POST', 'access_token_url': 'https://github.com/login/oauth/access_token', 'authorize_url': 'https://github.com/login/oauth/authorize', }, """ The OAuth app for GitHub API. :param scope: optional. default: ``['user:email']``. """) github.kwargs_processor(make_scope_processor('user:email')) google = RemoteAppFactory('google', { 'base_url': 'https://www.googleapis.com/oauth2/v1/', 'request_token_url': None, 'access_token_method': 'POST', 'access_token_url': 'https://accounts.google.com/o/oauth2/token', 'authorize_url': 'https://accounts.google.com/o/oauth2/auth', }, """ The OAuth app for Google API. :param scope: optional. default: ``['email']``. """) google.kwargs_processor(make_scope_processor( 'email')) twitter = RemoteAppFactory('twitter', { 'base_url': 'https://api.twitter.com/1.1/', 'request_token_url': 'https://api.twitter.com/oauth/request_token', 'access_token_url': 'https://api.twitter.com/oauth/access_token', 'authorize_url': 'https://api.twitter.com/oauth/authenticate', }, """The OAuth app for Twitter API.""") weibo = RemoteAppFactory('weibo', { 'base_url': 'https://api.weibo.com/2/', 'authorize_url': 'https://api.weibo.com/oauth2/authorize', 'request_token_url': None, 'access_token_method': 'POST', 'access_token_url': 'https://api.weibo.com/oauth2/access_token', # since weibo's response is a shit, we need to force parse the content 'content_type': 'application/json', }, """ The OAuth app for weibo.com API. :param scope: optional. default: ``['email']`` """) weibo.kwargs_processor(make_scope_processor('email')) def change_weibo_header(uri, headers, body): """Since weibo is a rubbish server, it does not follow the standard, we need to change the authorization header for it.""" auth = headers.get('Authorization') if auth: auth = auth.replace('Bearer', 'OAuth2') headers['Authorization'] = auth return uri, headers, body weibo.pre_request = change_weibo_header linkedin = RemoteAppFactory('linkedin', { 'request_token_params': {'state': 'RandomString'}, 'base_url': 'https://api.linkedin.com/v1/', 'request_token_url': None, 'access_token_method': 'POST', 'access_token_url': 'https://www.linkedin.com/uas/oauth2/accessToken', 'authorize_url': 'https://www.linkedin.com/uas/oauth2/authorization', }, """ The OAuth app for LinkedIn API. :param scope: optional. default: ``['r_basicprofile']`` """) linkedin.kwargs_processor(make_scope_processor('r_basicprofile')) def change_linkedin_query(uri, headers, body): auth = headers.pop('Authorization') headers['x-li-format'] = 'json' if auth: auth = auth.replace('Bearer', '').strip() if '?' in uri: uri += '&oauth2_access_token=' + auth else: uri += '?oauth2_access_token=' + auth return uri, headers, body linkedin.pre_request = change_linkedin_query
bsd-3-clause
87efb75a528885aace6e72467a7008dc
31.973799
76
0.639518
3.550071
false
false
false
false
lepture/flask-oauthlib
flask_oauthlib/client.py
1
24833
# coding: utf-8 """ flask_oauthlib.client ~~~~~~~~~~~~~~~~~~~~~ Implemnts OAuth1 and OAuth2 support for Flask. :copyright: (c) 2013 - 2014 by Hsiaoming Yang. """ import logging import oauthlib.oauth1 import oauthlib.oauth2 from copy import copy from functools import wraps from oauthlib.common import to_unicode, PY3, add_params_to_uri from flask import request, redirect, json, session, current_app from werkzeug.urls import url_quote, url_decode, url_encode from werkzeug.http import parse_options_header from werkzeug.utils import cached_property from .utils import to_bytes try: from urlparse import urljoin import urllib2 as http except ImportError: from urllib import request as http from urllib.parse import urljoin log = logging.getLogger('flask_oauthlib') if PY3: string_types = (str,) else: string_types = (str, unicode) __all__ = ('OAuth', 'OAuthRemoteApp', 'OAuthResponse', 'OAuthException') class OAuth(object): """Registry for remote applications. :param app: the app instance of Flask Create an instance with Flask:: oauth = OAuth(app) """ state_key = 'oauthlib.client' def __init__(self, app=None): self.remote_apps = {} self.app = app if app: self.init_app(app) def init_app(self, app): """Init app with Flask instance. You can also pass the instance of Flask later:: oauth = OAuth() oauth.init_app(app) """ self.app = app app.extensions = getattr(app, 'extensions', {}) app.extensions[self.state_key] = self def remote_app(self, name, register=True, **kwargs): """Registers a new remote application. :param name: the name of the remote application :param register: whether the remote app will be registered Find more parameters from :class:`OAuthRemoteApp`. """ remote = OAuthRemoteApp(self, name, **kwargs) if register: assert name not in self.remote_apps self.remote_apps[name] = remote return remote def __getattr__(self, key): try: return object.__getattribute__(self, key) except AttributeError: app = self.remote_apps.get(key) if app: return app raise AttributeError('No such app: %s' % key) _etree = None def get_etree(): global _etree if _etree is not None: return _etree try: from lxml import etree as _etree except ImportError: try: from xml.etree import cElementTree as _etree except ImportError: try: from xml.etree import ElementTree as _etree except ImportError: raise TypeError('lxml or etree not found') return _etree def parse_response(resp, content, strict=False, content_type=None): """Parse the response returned by :meth:`OAuthRemoteApp.http_request`. :param resp: response of http_request :param content: content of the response :param strict: strict mode for form urlencoded content :param content_type: assign a content type manually """ if not content_type: content_type = resp.headers.get('content-type', 'application/json') ct, options = parse_options_header(content_type) if ct in ('application/json', 'text/javascript'): if not content: return {} return json.loads(content) if ct in ('application/xml', 'text/xml'): return get_etree().fromstring(content) if ct != 'application/x-www-form-urlencoded' and strict: return content charset = options.get('charset', 'utf-8') return url_decode(content, charset=charset).to_dict() def prepare_request(uri, headers=None, data=None, method=None): """Make request parameters right.""" if headers is None: headers = {} if data and not method: method = 'POST' elif not method: method = 'GET' if method == 'GET' and data: uri = add_params_to_uri(uri, data) data = None return uri, headers, data, method def encode_request_data(data, format): if format is None: return data, None if format == 'json': return json.dumps(data or {}), 'application/json' if format == 'urlencoded': return url_encode(data or {}), 'application/x-www-form-urlencoded' raise TypeError('Unknown format %r' % format) class OAuthResponse(object): def __init__(self, resp, content, content_type=None): self._resp = resp self.raw_data = content self.data = parse_response( resp, content, strict=True, content_type=content_type, ) @property def status(self): """The status code of the response.""" return self._resp.code class OAuthException(RuntimeError): def __init__(self, message, type=None, data=None): self.message = message self.type = type self.data = data def __str__(self): if PY3: return self.message return self.message.encode('utf-8') def __unicode__(self): return self.message class OAuthRemoteApp(object): """Represents a remote application. :param oauth: the associated :class:`OAuth` object :param name: the name of the remote application :param base_url: the base url for every request :param request_token_url: the url for requesting new tokens :param access_token_url: the url for token exchange :param authorize_url: the url for authorization :param consumer_key: the application specific consumer key :param consumer_secret: the application specific consumer secret :param request_token_params: an optional dictionary of parameters to forward to the request token url or authorize url depending on oauth version :param request_token_method: the HTTP method that should be used for the access_token_url. Default is ``GET`` :param access_token_params: an optional dictionary of parameters to forward to the access token url :param access_token_method: the HTTP method that should be used for the access_token_url. Default is ``GET`` :param access_token_headers: additonal headers that should be used for the access_token_url. :param content_type: force to parse the content with this content_type, usually used when the server didn't return the right content type. .. versionadded:: 0.3.0 :param app_key: lazy load configuration from Flask app config with this app key """ def __init__( self, oauth, name, base_url=None, request_token_url=None, access_token_url=None, authorize_url=None, consumer_key=None, consumer_secret=None, rsa_key=None, signature_method=None, request_token_params=None, request_token_method=None, access_token_params=None, access_token_method=None, access_token_headers=None, content_type=None, app_key=None, encoding='utf-8', ): self.oauth = oauth self.name = name self._base_url = base_url self._request_token_url = request_token_url self._access_token_url = access_token_url self._authorize_url = authorize_url self._consumer_key = consumer_key self._consumer_secret = consumer_secret self._rsa_key = rsa_key self._signature_method = signature_method self._request_token_params = request_token_params self._request_token_method = request_token_method self._access_token_params = access_token_params self._access_token_method = access_token_method self._access_token_headers = access_token_headers or {} self._content_type = content_type self._tokengetter = None self.app_key = app_key self.encoding = encoding # Check for required authentication information. # Skip this check if app_key is specified, since the information is # specified in the Flask config, instead. if not app_key: if signature_method == oauthlib.oauth1.SIGNATURE_RSA: # check for consumer_key and rsa_key if not consumer_key or not rsa_key: raise TypeError( "OAuthRemoteApp with RSA authentication requires " "consumer key and rsa key" ) else: # check for consumer_key and consumer_secret if not consumer_key or not consumer_secret: raise TypeError( "OAuthRemoteApp requires consumer key and secret" ) @cached_property def base_url(self): return self._get_property('base_url') @cached_property def request_token_url(self): return self._get_property('request_token_url', None) @cached_property def access_token_url(self): return self._get_property('access_token_url') @cached_property def authorize_url(self): return self._get_property('authorize_url') @cached_property def consumer_key(self): return self._get_property('consumer_key') @cached_property def consumer_secret(self): return self._get_property('consumer_secret') @cached_property def rsa_key(self): return self._get_property('rsa_key') @cached_property def signature_method(self): return self._get_property('signature_method') @cached_property def request_token_params(self): return self._get_property('request_token_params', {}) @cached_property def request_token_method(self): return self._get_property('request_token_method', 'GET') @cached_property def access_token_params(self): return self._get_property('access_token_params', {}) @cached_property def access_token_method(self): return self._get_property('access_token_method', 'POST') @cached_property def content_type(self): return self._get_property('content_type', None) def _get_property(self, key, default=False): attr = getattr(self, '_%s' % key) if attr is not None: return attr if not self.app_key: if default is not False: return default return attr app = self.oauth.app or current_app if self.app_key in app.config: # works with dict config config = app.config[self.app_key] if default is not False: return config.get(key, default) return config[key] # works with plain text config config_key = "%s_%s" % (self.app_key, key.upper()) if default is not False: return app.config.get(config_key, default) return app.config[config_key] def get_oauth1_client_params(self, token): params = copy(self.request_token_params) or {} if token and isinstance(token, (tuple, list)): params["resource_owner_key"] = token[0] params["resource_owner_secret"] = token[1] # Set params for SIGNATURE_RSA if self.signature_method == oauthlib.oauth1.SIGNATURE_RSA: params["signature_method"] = self.signature_method params["rsa_key"] = self.rsa_key return params def make_client(self, token=None): # request_token_url is for oauth1 if self.request_token_url: # get params for client params = self.get_oauth1_client_params(token) client = oauthlib.oauth1.Client( client_key=self.consumer_key, client_secret=self.consumer_secret, **params ) else: if token: if isinstance(token, (tuple, list)): token = {'access_token': token[0]} elif isinstance(token, string_types): token = {'access_token': token} client = oauthlib.oauth2.WebApplicationClient( self.consumer_key, token=token ) return client @staticmethod def http_request(uri, headers=None, data=None, method=None): uri, headers, data, method = prepare_request( uri, headers, data, method ) log.debug('Request %r with %r method' % (uri, method)) req = http.Request(uri, headers=headers, data=data) req.get_method = lambda: method.upper() try: resp = http.urlopen(req) content = resp.read() resp.close() return resp, content except http.HTTPError as resp: content = resp.read() resp.close() return resp, content def get(self, *args, **kwargs): """Sends a ``GET`` request. Accepts the same parameters as :meth:`request`. """ kwargs['method'] = 'GET' return self.request(*args, **kwargs) def post(self, *args, **kwargs): """Sends a ``POST`` request. Accepts the same parameters as :meth:`request`. """ kwargs['method'] = 'POST' return self.request(*args, **kwargs) def put(self, *args, **kwargs): """Sends a ``PUT`` request. Accepts the same parameters as :meth:`request`. """ kwargs['method'] = 'PUT' return self.request(*args, **kwargs) def delete(self, *args, **kwargs): """Sends a ``DELETE`` request. Accepts the same parameters as :meth:`request`. """ kwargs['method'] = 'DELETE' return self.request(*args, **kwargs) def patch(self, *args, **kwargs): """Sends a ``PATCH`` request. Accepts the same parameters as :meth:`post`. """ kwargs['method'] = 'PATCH' return self.request(*args, **kwargs) def request(self, url, data=None, headers=None, format='urlencoded', method='GET', content_type=None, token=None): """ Sends a request to the remote server with OAuth tokens attached. :param data: the data to be sent to the server. :param headers: an optional dictionary of headers. :param format: the format for the `data`. Can be `urlencoded` for URL encoded data or `json` for JSON. :param method: the HTTP request method to use. :param content_type: an optional content type. If a content type is provided, the data is passed as it, and the `format` is ignored. :param token: an optional token to pass, if it is None, token will be generated by tokengetter. """ headers = dict(headers or {}) if token is None: token = self.get_request_token() client = self.make_client(token) url = self.expand_url(url) if method == 'GET': assert format == 'urlencoded' if data: url = add_params_to_uri(url, data) data = None else: if content_type is None: data, content_type = encode_request_data(data, format) if content_type is not None: headers['Content-Type'] = content_type if self.request_token_url: # oauth1 uri, headers, body = client.sign( url, http_method=method, body=data, headers=headers ) else: # oauth2 uri, headers, body = client.add_token( url, http_method=method, body=data, headers=headers ) if hasattr(self, 'pre_request'): # This is designed for some rubbish services like weibo. # Since they don't follow the standards, we need to # change the uri, headers, or body. uri, headers, body = self.pre_request(uri, headers, body) if body: data = to_bytes(body, self.encoding) else: data = None resp, content = self.http_request( uri, headers, data=to_bytes(body, self.encoding), method=method ) return OAuthResponse(resp, content, self.content_type) def authorize(self, callback=None, state=None, **kwargs): """ Returns a redirect response to the remote authorization URL with the signed callback given. :param callback: a redirect url for the callback :param state: an optional value to embed in the OAuth request. Use this if you want to pass around application state (e.g. CSRF tokens). :param kwargs: add optional key/value pairs to the query string """ params = dict(self.request_token_params) or {} params.update(**kwargs) if self.request_token_url: token = self.generate_request_token(callback)[0] url = '%s?oauth_token=%s' % ( self.expand_url(self.authorize_url), url_quote(token) ) if params: url += '&' + url_encode(params) else: assert callback is not None, 'Callback is required for OAuth2' client = self.make_client() if 'scope' in params: scope = params.pop('scope') else: scope = None if isinstance(scope, str): # oauthlib need unicode scope = _encode(scope, self.encoding) if 'state' in params: if not state: state = params.pop('state') else: # remove state in params params.pop('state') if callable(state): # state can be function for generate a random string state = state() session['%s_oauthredir' % self.name] = callback url = client.prepare_request_uri( self.expand_url(self.authorize_url), redirect_uri=callback, scope=scope, state=state, **params ) return redirect(url) def tokengetter(self, f): """ Register a function as token getter. """ self._tokengetter = f return f def expand_url(self, url): return urljoin(self.base_url, url) def generate_request_token(self, callback=None): # for oauth1 only if callback is not None: callback = urljoin(request.url, callback) client = self.make_client() client.callback_uri = _encode(callback, self.encoding) realm = self.request_token_params.get('realm') realms = self.request_token_params.get('realms') if not realm and realms: realm = ' '.join(realms) uri, headers, _ = client.sign( self.expand_url(self.request_token_url), http_method=self.request_token_method, realm=realm, ) log.debug('Generate request token header %r', headers) resp, content = self.http_request( uri, headers, method=self.request_token_method, ) data = parse_response(resp, content) if not data: raise OAuthException( 'Invalid token response from %s' % self.name, type='token_generation_failed' ) if resp.code not in (200, 201): message = 'Failed to generate request token' if 'oauth_problem' in data: message += ' (%s)' % data['oauth_problem'] raise OAuthException( message, type='token_generation_failed', data=data, ) tup = (data['oauth_token'], data['oauth_token_secret']) session['%s_oauthtok' % self.name] = tup return tup def get_request_token(self): assert self._tokengetter is not None, 'missing tokengetter' rv = self._tokengetter() if rv is None: raise OAuthException('No token available', type='token_missing') return rv def handle_oauth1_response(self, args): """Handles an oauth1 authorization response.""" client = self.make_client() client.verifier = args.get('oauth_verifier') tup = session.get('%s_oauthtok' % self.name) if not tup: raise OAuthException( 'Token not found, maybe you disabled cookie', type='token_not_found' ) client.resource_owner_key = tup[0] client.resource_owner_secret = tup[1] uri, headers, data = client.sign( self.expand_url(self.access_token_url), _encode(self.access_token_method) ) headers.update(self._access_token_headers) resp, content = self.http_request( uri, headers, to_bytes(data, self.encoding), method=self.access_token_method ) data = parse_response(resp, content) if resp.code not in (200, 201): raise OAuthException( 'Invalid response from %s' % self.name, type='invalid_response', data=data ) return data def handle_oauth2_response(self, args): """Handles an oauth2 authorization response.""" client = self.make_client() remote_args = { 'code': args.get('code'), 'client_secret': self.consumer_secret, 'redirect_uri': session.get('%s_oauthredir' % self.name) } log.debug('Prepare oauth2 remote args %r', remote_args) remote_args.update(self.access_token_params) headers = copy(self._access_token_headers) if self.access_token_method == 'POST': headers.update({'Content-Type': 'application/x-www-form-urlencoded'}) body = client.prepare_request_body(**remote_args) resp, content = self.http_request( self.expand_url(self.access_token_url), headers=headers, data=to_bytes(body, self.encoding), method=self.access_token_method, ) elif self.access_token_method == 'GET': qs = client.prepare_request_body(**remote_args) url = self.expand_url(self.access_token_url) url += ('?' in url and '&' or '?') + qs resp, content = self.http_request( url, headers=headers, method=self.access_token_method, ) else: raise OAuthException( 'Unsupported access_token_method: %s' % self.access_token_method ) data = parse_response(resp, content, content_type=self.content_type) if resp.code not in (200, 201): raise OAuthException( 'Invalid response from %s' % self.name, type='invalid_response', data=data ) return data def handle_unknown_response(self): """Handles a unknown authorization response.""" return None def authorized_response(self, args=None): """Handles authorization response smartly.""" if args is None: args = request.args if 'oauth_verifier' in args: data = self.handle_oauth1_response(args) elif 'code' in args: data = self.handle_oauth2_response(args) else: data = self.handle_unknown_response() # free request token session.pop('%s_oauthtok' % self.name, None) session.pop('%s_oauthredir' % self.name, None) return data def authorized_handler(self, f): """Handles an OAuth callback. .. versionchanged:: 0.7 @authorized_handler is deprecated in favor of authorized_response. """ @wraps(f) def decorated(*args, **kwargs): log.warn( '@authorized_handler is deprecated in favor of ' 'authorized_response' ) data = self.authorized_response() return f(*((data,) + args), **kwargs) return decorated def _encode(text, encoding='utf-8'): if encoding: return to_unicode(text, encoding) return text
bsd-3-clause
8de9cec9d8f243135c251d91208bd3b6
32.694708
81
0.571337
4.33537
false
false
false
false
lepture/flask-oauthlib
example/google.py
5
1864
""" google example ~~~~~~~~~~~~~~ This example is contributed by Bruno Rocha GitHub: https://github.com/rochacbruno """ from flask import Flask, redirect, url_for, session, request, jsonify from flask_oauthlib.client import OAuth app = Flask(__name__) app.config['GOOGLE_ID'] = "cloud.google.com/console and get your ID" app.config['GOOGLE_SECRET'] = "cloud.google.com/console and get the secret" app.debug = True app.secret_key = 'development' oauth = OAuth(app) google = oauth.remote_app( 'google', consumer_key=app.config.get('GOOGLE_ID'), consumer_secret=app.config.get('GOOGLE_SECRET'), request_token_params={ 'scope': 'email' }, base_url='https://www.googleapis.com/oauth2/v1/', request_token_url=None, access_token_method='POST', access_token_url='https://accounts.google.com/o/oauth2/token', authorize_url='https://accounts.google.com/o/oauth2/auth', ) @app.route('/') def index(): if 'google_token' in session: me = google.get('userinfo') return jsonify({"data": me.data}) return redirect(url_for('login')) @app.route('/login') def login(): return google.authorize(callback=url_for('authorized', _external=True)) @app.route('/logout') def logout(): session.pop('google_token', None) return redirect(url_for('index')) @app.route('/login/authorized') def authorized(): resp = google.authorized_response() if resp is None: return 'Access denied: reason=%s error=%s' % ( request.args['error_reason'], request.args['error_description'] ) session['google_token'] = (resp['access_token'], '') me = google.get('userinfo') return jsonify({"data": me.data}) @google.tokengetter def get_google_oauth_token(): return session.get('google_token') if __name__ == '__main__': app.run()
bsd-3-clause
b5578256bcf2c7ac691fe8be0bcdc127
24.534247
75
0.642167
3.445471
false
true
false
false
lepture/flask-oauthlib
example/linkedin.py
16
2007
from flask import Flask, redirect, url_for, session, request, jsonify from flask_oauthlib.client import OAuth app = Flask(__name__) app.debug = True app.secret_key = 'development' oauth = OAuth(app) linkedin = oauth.remote_app( 'linkedin', consumer_key='k8fhkgkkqzub', consumer_secret='ZZtLETQOQYNDjMrz', request_token_params={ 'scope': 'r_basicprofile', 'state': 'RandomString', }, base_url='https://api.linkedin.com/v1/', request_token_url=None, access_token_method='POST', access_token_url='https://www.linkedin.com/uas/oauth2/accessToken', authorize_url='https://www.linkedin.com/uas/oauth2/authorization', ) @app.route('/') def index(): if 'linkedin_token' in session: me = linkedin.get('people/~') return jsonify(me.data) return redirect(url_for('login')) @app.route('/login') def login(): return linkedin.authorize(callback=url_for('authorized', _external=True)) @app.route('/logout') def logout(): session.pop('linkedin_token', None) return redirect(url_for('index')) @app.route('/login/authorized') def authorized(): resp = linkedin.authorized_response() if resp is None: return 'Access denied: reason=%s error=%s' % ( request.args['error_reason'], request.args['error_description'] ) session['linkedin_token'] = (resp['access_token'], '') me = linkedin.get('people/~') return jsonify(me.data) @linkedin.tokengetter def get_linkedin_oauth_token(): return session.get('linkedin_token') def change_linkedin_query(uri, headers, body): auth = headers.pop('Authorization') headers['x-li-format'] = 'json' if auth: auth = auth.replace('Bearer', '').strip() if '?' in uri: uri += '&oauth2_access_token=' + auth else: uri += '?oauth2_access_token=' + auth return uri, headers, body linkedin.pre_request = change_linkedin_query if __name__ == '__main__': app.run()
bsd-3-clause
440e9e905ecca293298c2ec0b747074e
24.730769
77
0.63129
3.378788
false
false
false
false
ufal/neuralmonkey
neuralmonkey/runners/ctc_debug_runner.py
1
1826
from typing import Dict, List import numpy as np import tensorflow as tf from typeguard import check_argument_types from neuralmonkey.runners.base_runner import BaseRunner from neuralmonkey.decoders.ctc_decoder import CTCDecoder from neuralmonkey.decorators import tensor class CTCDebugRunner(BaseRunner[CTCDecoder]): """A runner that print out raw CTC output including the blank symbols.""" # pylint: disable=too-few-public-methods # Pylint issue here: https://github.com/PyCQA/pylint/issues/2607 class Executable(BaseRunner.Executable["CTCDebugRunner"]): def collect_results(self, results: List[Dict]) -> None: vocabulary = self.executor.decoder.vocabulary if len(results) != 1: raise RuntimeError("CTCDebugRunners do not support ensembles.") logits = results[0]["logits"] argmaxes = np.argmax(logits, axis=2).T decoded_batch = [] for indices in argmaxes: decoded_instance = [] for index in indices: if index == len(vocabulary): symbol = "<BLANK>" else: symbol = vocabulary.index_to_word[index] decoded_instance.append(symbol) decoded_batch.append(decoded_instance) self.set_runner_result(outputs=decoded_batch, losses=[]) # pylint: enable=too-few-public-methods def __init__(self, output_series: str, decoder: CTCDecoder) -> None: check_argument_types() super().__init__(output_series, decoder) @tensor def fetches(self) -> Dict[str, tf.Tensor]: return {"logits": self.decoder.logits} @property def loss_names(self) -> List[str]: return []
bsd-3-clause
2933ee68d837f15c4ee464ed58aee44b
32.814815
79
0.608434
4.296471
false
false
false
false
ufal/neuralmonkey
neuralmonkey/processors/speech.py
2
2025
from typing import Callable import numpy as np from python_speech_features import mfcc, fbank, logfbank, ssc, delta from neuralmonkey.readers.audio_reader import Audio # pylint: disable=invalid-name def SpeechFeaturesPreprocessor(feature_type: str = "mfcc", delta_order: int = 0, delta_window: int = 2, **kwargs) -> Callable: """Calculate speech features. First, the given type of features (e.g. MFCC) is computed using a window of length `winlen` and step `winstep`; for additional keyword arguments (specific to each feature type), see http://python-speech-features.readthedocs.io/. Then, delta features up to `delta_order` are added. By default, 13 MFCCs per frame are computed. To add delta and delta-delta features (resulting in 39 coefficients per frame), set `delta_order=2`. Arguments: feature_type: mfcc, fbank, logfbank or ssc (default is mfcc) delta_order: maximum order of the delta features (default is 0) delta_window: window size for delta features (default is 2) **kwargs: keyword arguments for the appropriate function from python_speech_features Returns: A numpy array of shape [num_frames, num_features]. """ if feature_type not in FEATURE_TYPES: raise ValueError( "Unknown speech feature type '{}'".format(feature_type)) def preprocess(audio: Audio) -> np.ndarray: features = [FEATURE_TYPES[feature_type]( audio.data, samplerate=audio.rate, **kwargs)] for _ in range(delta_order): features.append(delta(features[-1], delta_window)) return np.concatenate(features, axis=1) return preprocess def _fbank(*args, **kwargs) -> np.ndarray: feat, _ = fbank(*args, **kwargs) return feat FEATURE_TYPES = {"mfcc": mfcc, "fbank": _fbank, "logfbank": logfbank, "ssc": ssc}
bsd-3-clause
61a824d082b9fdd1586ed54073eb50f7
32.75
77
0.627654
4.05
false
false
false
false
ufal/neuralmonkey
neuralmonkey/attention/base_attention.py
1
8051
"""Decoding functions using multiple attentions for RNN decoders. See http://arxiv.org/abs/1606.07481 The attention mechanisms used in Neural Monkey are inherited from the ``BaseAttention`` class defined in this module. The attention function can be viewed as a soft lookup over an associative memory. The *query* vector is used to compute a similarity score of the *keys* of the associative memory and the resulting scores are used as weights in a weighted sum of the *values* associated with the keys. We call the (unnormalized) similarity scores *energies*, we call *attention distribution* the energies after (softmax) normalization, and we call the resulting weighted sum of states a *context vector*. Note that it is possible (and true in most cases) that the attention keys are equal to the values. In case of self-attention, even queries are from the same set of vectors. To abstract over different flavors of attention mechanism, we conceptualize the procedure as follows: Each attention object has the ``attention`` function which operates on the query tensor. The attention function receives the query tensor (the decoder state) and optionally the previous state of the decoder, and computes the context vector. The function also receives a *loop state*, which is used to store data in an autoregressive loop that generates a sequence. The attention uses the loop state to store to store attention distributions and context vectors in time. This structure is called ``AttentionLoopState``. To be able to initialize the loop state, each attention object that uses this feature defines the ``initial_loop_state`` function with empty tensors. Since there can be many *modes* in which the decoder that uses the attention operates, the attention objects have the ``finalize_loop`` method, which takes the last attention loop state and the name of the mode (a string) and processes this data to be available in the ``histories`` dictionary. The single and most used example of two *modes* are the *train* and *runtime* modes of the autoregressive decoder. """ from typing import Dict, Optional, Any, Tuple, Union import tensorflow as tf from neuralmonkey.attention.namedtuples import AttentionLoopState from neuralmonkey.model.model_part import ModelPart from neuralmonkey.model.parameterized import InitializerSpecs from neuralmonkey.model.stateful import TemporalStateful, SpatialStateful # pylint: disable=invalid-name Attendable = Union[TemporalStateful, SpatialStateful] # pylint: enable=invalid-name def empty_attention_loop_state( batch_size: Union[int, tf.Tensor], length: Union[int, tf.Tensor], dimension: Union[int, tf.Tensor]) -> AttentionLoopState: """Create an empty attention loop state. The attention loop state is a technical object for storing the attention distributions and the context vectors in time. It is used with the ``tf.while_loop`` dynamic implementation of decoders. Arguments: batch_size: The size of the batch. length: The number of encoder states (keys). dimension: The dimension of the context vector Returns: This function returns an empty attention loop state which means there are two empty Tensors one for attention distributions in time, and one for the attention context vectors in time. """ return AttentionLoopState( contexts=tf.zeros(shape=[0, batch_size, dimension], name="contexts"), weights=tf.zeros(shape=[0, batch_size, length], name="distributions")) def get_attention_states(encoder: Attendable) -> tf.Tensor: """Return the temporal or spatial states of an encoder. Arguments: encoder: The encoder with the states to attend. Returns: Either a 3D or a 4D tensor, depending on whether the encoder is temporal (e.g. recurrent encoder) or spatial (e.g. a CNN encoder). The first two dimensions are (batch, time). """ if isinstance(encoder, TemporalStateful): return encoder.temporal_states if isinstance(encoder, SpatialStateful): shape = encoder.spatial_states.get_shape().as_list() return tf.reshape(encoder.spatial_states, [-1, shape[1] * shape[2], shape[3]]) raise TypeError("Unknown encoder type") def get_attention_mask(encoder: Attendable) -> Optional[tf.Tensor]: """Return the temporal or spatial mask of an encoder. Arguments: encoder: The encoder to get the mask from. Returns: Either a 2D or a 3D tensor, depending on whether the encoder is temporal (e.g. recurrent encoder) or spatial (e.g. a CNN encoder). """ if isinstance(encoder, TemporalStateful): if encoder.temporal_mask is None: raise ValueError("The encoder temporal mask should not be none") return encoder.temporal_mask if isinstance(encoder, SpatialStateful): if encoder.spatial_mask is None: return None shape = encoder.spatial_states.get_shape().as_list() return tf.reshape(encoder.spatial_mask, [-1, shape[1] * shape[2]]) raise TypeError("Unknown encoder type") class BaseAttention(ModelPart): """The abstract class for the attenion mechanism flavors.""" def __init__(self, name: str, reuse: ModelPart = None, save_checkpoint: str = None, load_checkpoint: str = None, initializers: InitializerSpecs = None) -> None: """Create a new ``BaseAttention`` object.""" ModelPart.__init__( self, name, reuse, save_checkpoint, load_checkpoint, initializers) self.query_state_size = None # type: tf.Tensor self._histories = {} # type: Dict[str, tf.Tensor] @property def histories(self) -> Dict[str, tf.Tensor]: """Return the attention histories dictionary. Use this property after it has been populated. Returns: The attention histories dictionary. """ return self._histories def attention(self, query: tf.Tensor, decoder_prev_state: tf.Tensor, decoder_input: tf.Tensor, loop_state: Any) -> Tuple[tf.Tensor, Any]: """Get context vector for a given query.""" raise NotImplementedError("Abstract method") def initial_loop_state(self) -> Any: """Get initial loop state for the attention object. Returns: The newly created initial loop state object. """ raise NotImplementedError("Abstract method") def finalize_loop(self, key: str, last_loop_state: Any) -> None: """Store the attention histories from loop state under a given key. Arguments: key: The key to the histories dictionary to store the data in. last_loop_state: The loop state object from the last state of the decoding loop. """ raise NotImplementedError("Abstract method") @property def context_vector_size(self) -> int: """Return the static size of the context vector. Returns: An integer specifying the context vector dimension. """ raise NotImplementedError("Abstract property") def visualize_attention(self, key: str, max_outputs: int = 16) -> None: """Include the attention histories under a given key into a summary. Arguments: key: The key to the attention histories dictionary. max_outputs: Maximum number of images to save. """ if key not in self.histories: raise KeyError( "Key {} not among attention histories".format(key)) alignments = tf.expand_dims( tf.transpose(self.histories[key], perm=[1, 2, 0]), -1) summary_name = "{}.{}".format(self.name, key) tf.summary.image( summary_name, alignments, collections=["summary_att_plots"], max_outputs=max_outputs)
bsd-3-clause
f15e8306ec4adf5e1b15dbd18f787c89
38.273171
79
0.680164
4.397051
false
false
false
false
ufal/neuralmonkey
neuralmonkey/processors/helpers.py
1
1266
from typing import Any, Callable, Generator, List from random import randint def preprocess_char_based(sentence: List[str]) -> List[str]: return list(" ".join(sentence)) def preprocess_add_noise(sentence: List[str]) -> List[str]: sent = sentence[:] length = len(sentence) if length > 1: for _ in range(length // 2): swap = randint(0, length - 2) sent[swap] = sent[swap + 1] sent[swap + 1] = sent[swap] return sent # TODO refactor post-processors to work on sentence level def postprocess_char_based(sentences: List[List[str]]) -> List[List[str]]: result = [] for sentence in sentences: joined = "".join(sentence) tokenized = joined.split(" ") result.append(tokenized) return result def untruecase( sentences: List[List[str]]) -> Generator[List[str], None, None]: for sentence in sentences: if sentence: yield [sentence[0].capitalize()] + sentence[1:] else: yield [] def pipeline(processors: List[Callable]) -> Callable: """Concatenate processors.""" def process(data: Any) -> Any: for processor in processors: data = processor(data) return data return process
bsd-3-clause
7272700bbac13b5da5c6a041f172d878
24.836735
74
0.605845
4.044728
false
false
false
false
ufal/neuralmonkey
neuralmonkey/runners/regression_runner.py
1
1912
from typing import Dict, List, Callable import numpy as np import tensorflow as tf from typeguard import check_argument_types from neuralmonkey.decoders.sequence_regressor import SequenceRegressor from neuralmonkey.decorators import tensor from neuralmonkey.runners.base_runner import BaseRunner # pylint: disable=invalid-name Postprocessor = Callable[[List[float]], List[float]] # pylint: enable=invalid-name class RegressionRunner(BaseRunner[SequenceRegressor]): """A runnner that takes the predictions of a sequence regressor.""" # pylint: disable=too-few-public-methods # Pylint issue here: https://github.com/PyCQA/pylint/issues/2607 class Executable(BaseRunner.Executable["RegressionRunner"]): def collect_results(self, results: List[Dict]) -> None: predictions_sum = np.zeros_like(results[0]["prediction"]) mse_loss = 0. for sess_result in results: if "mse" in sess_result: mse_loss += sess_result["mse"] predictions_sum += sess_result["prediction"] predictions = (predictions_sum / len(results)).tolist() if self.executor.postprocess is not None: predictions = self.executor.postprocess(predictions) self.set_runner_result(outputs=predictions, losses=[mse_loss]) # pylint: enable=too-few-public-methods def __init__(self, output_series: str, decoder: SequenceRegressor, postprocess: Postprocessor = None) -> None: check_argument_types() super().__init__(output_series, decoder) self.postprocess = postprocess @tensor def fetches(self) -> Dict[str, tf.Tensor]: return {"prediction": self.decoder.predictions, "mse": self.decoder.cost} @property def loss_names(self) -> List[str]: return ["mse"]
bsd-3-clause
0d807aff5d86e00f44146fb93df8d5c7
33.142857
74
0.650105
4.248889
false
false
false
false
ufal/neuralmonkey
neuralmonkey/model/model_part.py
1
4048
"""Basic functionality of all model parts.""" from abc import ABCMeta from typing import MutableSet, Set, List, Tuple, Iterable from neuralmonkey.model.parameterized import Parameterized, InitializerSpecs from neuralmonkey.model.feedable import Feedable class GenericModelPart(metaclass=ABCMeta): """Base class for Neural Monkey model parts. Neural Monkey dynamically decides which model parts are in use when using a specific trainer or a runner. Each trainer/runner holds a reference to a top-level model part, which is then responsible for collecting references to all `Parameterized` and `Feedable` objects that contribute to the computation of its Tensors. This behavior is implemented using the `get_dependencies` method, which is called recursively on all instances of `GenericModelPart` class that are references from within a model part. Apart from the `get_dependencies` method, this class also provides the `dependencies` property which store the names of the Python class attributes that are regarded as potential dependents of the `GenericModelPart` object. These dependents are automatically checked for type and when they are instances of the `GenericModelPart` class, results of their `get_dependencies` are united and returned as dependencies of the current object. """ @property def dependencies(self) -> List[str]: """Return a list of attribute names regarded as dependents.""" return ["encoder", "parent_decoder", "input_sequence", "attentions", "encoders"] def __get_deps( self, attr: str, feedables: MutableSet[Feedable], parameterizeds: MutableSet[Parameterized]) -> None: attr_val = getattr(self, attr, None) if attr_val is None: return deps = [] # type: List[GenericModelPart] if isinstance(attr_val, GenericModelPart): deps = [attr_val] elif isinstance(attr_val, Iterable): deps = [a for a in attr_val if isinstance(a, GenericModelPart)] for dep in deps: feeds, params = dep.get_dependencies() feedables |= feeds parameterizeds |= params def get_dependencies(self) -> Tuple[Set[Feedable], Set[Parameterized]]: """Collect all dependents of this object recursively. The dependents are collected using the `dependencies` property. Each stores a potential dependent object. If the object exsits and is an instance of `GenericModelPart`, dependents are collected recursively by calling its `get_dependencies` method. If the object itself is instance of `Feedable` or `Parameterized` class, it is added among the respective sets returned. Returns: A `Tuple` of `Set`s of `Feedable` and `Parameterized` objects. """ feedables = set() # type: Set[Feedable] parameterizeds = set() # type: Set[Parameterized] if isinstance(self, Feedable): feedables |= {self} if isinstance(self, Parameterized): parameterizeds |= {self} for attr in self.dependencies: self.__get_deps(attr, feedables, parameterizeds) return feedables, parameterizeds class ModelPart(Parameterized, GenericModelPart, Feedable): """Base class of all parametric feedable model parts. Serves as a syntactic sugar for labeling `Feedable`, `Parameterized`, and `GenericModelPart` objects. """ def __init__(self, name: str, reuse: "ModelPart" = None, save_checkpoint: str = None, load_checkpoint: str = None, initializers: InitializerSpecs = None) -> None: Parameterized.__init__(self, name, reuse, save_checkpoint, load_checkpoint, initializers) GenericModelPart.__init__(self) with self.use_scope(): Feedable.__init__(self)
bsd-3-clause
6c2f15c766ae24fd8fe8dd6114a0abf3
38.300971
79
0.656374
4.482835
false
false
false
false
ufal/neuralmonkey
neuralmonkey/attention/transformer_cross_layer.py
1
9898
"""Input combination strategies for multi-source Transformer decoder.""" # TODO add citation when URL becomes available from typing import Callable, List import tensorflow as tf from neuralmonkey.attention.scaled_dot_product import attention from neuralmonkey.tf_utils import layer_norm # pylint: disable=too-many-arguments def single( queries: tf.Tensor, states: tf.Tensor, mask: tf.Tensor, n_heads: int, attention_dropout_callback: Callable[[tf.Tensor], tf.Tensor], dropout_callback: Callable[[tf.Tensor], tf.Tensor], normalize: bool = True, use_dropout: bool = True, residual: bool = True, use_att_transform_bias: bool = False): """Run attention on a single encoder. Arguments: queries: The input for the attention. states: The encoder states (keys & values). mask: The temporal mask of the encoder. n_heads: Number of attention heads to use. attention_dropout_callback: Dropout function to apply in attention. dropout_callback: Dropout function to apply on the attention output. normalize: If True, run layer normalization on the queries. use_dropout: If True, perform dropout on the attention output. residual: If True, sum the context vector with the input queries. use_att_transform_bias: If True, enable bias in the attention head projections (for all queries, keys and values). Returns: A Tensor that contains the context vector. """ # Layer normalization normalized_queries = layer_norm(queries) if normalize else queries # Attend to the encoder # TODO handle attention histories encoder_context, _ = attention( queries=normalized_queries, keys=states, values=states, keys_mask=mask, num_heads=n_heads, dropout_callback=attention_dropout_callback, use_bias=use_att_transform_bias) # Apply dropout if use_dropout: encoder_context = dropout_callback(encoder_context) # Add residual connections if residual: encoder_context += queries return encoder_context # pylint: enable=too-many-arguments def serial(queries: tf.Tensor, encoder_states: List[tf.Tensor], encoder_masks: List[tf.Tensor], heads: List[int], attention_dropout_callbacks: List[Callable[[tf.Tensor], tf.Tensor]], dropout_callback: Callable[[tf.Tensor], tf.Tensor]) -> tf.Tensor: """Run attention with serial input combination. The procedure is as follows: 1. repeat for every encoder: - lnorm + attend + dropout + add residual 2. update queries between layers Arguments: queries: The input for the attention. encoder_states: The states of each encoder. encoder_masks: The temporal mask of each encoder. heads: Number of attention heads to use for each encoder. attention_dropout_callbacks: Dropout functions to apply in attention over each encoder. dropout_callback: The dropout function to apply on the outputs of each sub-attention. Returns: A Tensor that contains the context vector. """ context = queries for i, (states, mask, n_heads, attn_drop_cb) in enumerate(zip( encoder_states, encoder_masks, heads, attention_dropout_callbacks)): with tf.variable_scope("enc_{}".format(i)): context = single(context, states, mask, n_heads, attention_dropout_callback=attn_drop_cb, dropout_callback=dropout_callback) return context def parallel( queries: tf.Tensor, encoder_states: List[tf.Tensor], encoder_masks: List[tf.Tensor], heads: List[int], attention_dropout_callbacks: List[Callable[[tf.Tensor], tf.Tensor]], dropout_callback: Callable[[tf.Tensor], tf.Tensor]) -> tf.Tensor: """Run attention with parallel input combination. The procedure is as follows: 1. normalize queries, 2. attend and dropout independently for every encoder, 3. sum up the results 4. add residual and return Arguments: queries: The input for the attention. encoder_states: The states of each encoder. encoder_masks: The temporal mask of each encoder. heads: Number of attention heads to use for each encoder. attention_dropout_callbacks: Dropout functions to apply in attention over each encoder. dropout_callback: The dropout function to apply on the outputs of each sub-attention. Returns: A Tensor that contains the context vector. """ normalized_queries = layer_norm(queries) contexts = [] for i, (states, mask, n_heads, attn_drop_cb) in enumerate(zip( encoder_states, encoder_masks, heads, attention_dropout_callbacks)): with tf.variable_scope("enc_{}".format(i)): contexts.append( single(normalized_queries, states, mask, n_heads, attention_dropout_callback=attn_drop_cb, dropout_callback=dropout_callback, normalize=False, residual=False)) return sum(contexts) + queries # pylint: disable=too-many-locals def hierarchical( queries: tf.Tensor, encoder_states: List[tf.Tensor], encoder_masks: List[tf.Tensor], heads: List[int], heads_hier: int, attention_dropout_callbacks: List[Callable[[tf.Tensor], tf.Tensor]], dropout_callback: Callable[[tf.Tensor], tf.Tensor]) -> tf.Tensor: """Run attention with hierarchical input combination. The procedure is as follows: 1. normalize queries 2. attend to every encoder 3. attend to the resulting context vectors (reuse normalized queries) 4. apply dropout, add residual connection and return Arguments: queries: The input for the attention. encoder_states: The states of each encoder. encoder_masks: The temporal mask of each encoder. heads: Number of attention heads to use for each encoder. heads_hier: Number of attention heads to use in the second attention. attention_dropout_callbacks: Dropout functions to apply in attention over each encoder. dropout_callback: The dropout function to apply in the second attention and over the outputs of each sub-attention. Returns: A Tensor that contains the context vector. """ normalized_queries = layer_norm(queries) contexts = [] batch = tf.shape(queries)[0] time_q = tf.shape(queries)[1] dimension = tf.shape(queries)[2] for i, (states, mask, n_heads, attn_drop_cb) in enumerate(zip( encoder_states, encoder_masks, heads, attention_dropout_callbacks)): with tf.variable_scope("enc_{}".format(i)): contexts.append( single(normalized_queries, states, mask, n_heads, attention_dropout_callback=attn_drop_cb, dropout_callback=dropout_callback, normalize=False, residual=False)) # context is of shape [batch, time(q), channels(v)], # stack to [batch, time(q), n_encoders, channels(v)] # reshape to [batch x time(q), n_encoders, channels(v)] stacked_contexts = tf.reshape( tf.stack(contexts, axis=2), [batch * time_q, len(encoder_states), dimension]) # hierarchical mask: ones of shape [batch x time(q), n_encoders] hier_mask = tf.ones([batch * time_q, len(encoder_states)]) # reshape queries to [batch x time(q), 1, channels(v)] reshaped_queries = tf.reshape( normalized_queries, [batch * time_q, 1, dimension]) # returned shape [batch x time(q), 1, channels(v)] with tf.variable_scope("enc_hier"): # NOTE as attention dropout keep probability, we use the # dropout_keep_prob value instead of attention_dropout_keep_prob. encoder_context_stacked_batch = single( reshaped_queries, stacked_contexts, hier_mask, heads_hier, attention_dropout_callback=dropout_callback, dropout_callback=lambda x: x, normalize=False, use_dropout=False, residual=False) # reshape back to [batch, time(q), channels(v)] encoder_context = tf.reshape( encoder_context_stacked_batch, [batch, time_q, dimension]) encoder_context = dropout_callback(encoder_context) return encoder_context + queries # pylint: enable=too-many-locals def flat(queries: tf.Tensor, encoder_states: List[tf.Tensor], encoder_masks: List[tf.Tensor], heads: int, attention_dropout_callback: Callable[[tf.Tensor], tf.Tensor], dropout_callback: Callable[[tf.Tensor], tf.Tensor]) -> tf.Tensor: """Run attention with flat input combination. The procedure is as follows: 1. concatenate the states and mask along the time axis 2. run attention over the concatenation Arguments: queries: The input for the attention. encoder_states: The states of each encoder. encoder_masks: The temporal mask of each encoder. heads: Number of attention heads to use for each encoder. attention_dropout_callbacks: Dropout functions to apply in attention over each encoder. dropout_callback: The dropout function to apply on the output of the attention. Returns: A Tensor that contains the context vector. """ concat_states = tf.concat(encoder_states, 1) concat_mask = tf.concat(encoder_masks, 1) return single(queries, concat_states, concat_mask, heads, attention_dropout_callback, dropout_callback)
bsd-3-clause
bdbe0c40fbb7b2e738b3338beaf87f48
36.634981
79
0.646797
4.352682
false
false
false
false
ufal/neuralmonkey
lib/subword_nmt/bpe_toy.py
3
1687
#!/usr/bin/env python # -*- coding: utf-8 -*- # Author: Rico Sennrich """Use byte pair encoding (BPE) to learn a variable-length encoding of the vocabulary in a text. Unlike the original BPE, it does not compress the plain text, but can be used to reduce the vocabulary of a text to a configurable number of symbols, with only a small increase in the number of tokens. This is an (inefficient) toy implementation that shows the algorithm. For processing large datasets, indexing and incremental updates can be used to speed up the implementation (see learn_bpe.py). Reference: Rico Sennrich, Barry Haddow and Alexandra Birch (2016). Neural Machine Translation of Rare Words with Subword Units. Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics (ACL 2016). Berlin, Germany. """ import re import sys import collections def get_stats(vocab): pairs = collections.defaultdict(int) for word, freq in vocab.items(): symbols = word.split() for i in range(len(symbols)-1): pairs[symbols[i],symbols[i+1]] += freq return pairs def merge_vocab(pair, v_in): v_out = {} bigram_pattern = re.escape(' '.join(pair)) p = re.compile(r'(?<!\S)' + bigram_pattern + r'(?!\S)') for word in v_in: w_out = p.sub(''.join(pair), word) v_out[w_out] = v_in[word] return v_out vocab = {'l o w </w>' : 5, 'l o w e r </w>' : 2, 'n e w e s t </w>' : 6, 'w i d e s t </w>' : 3} num_merges = 15 for i in range(num_merges): pairs = get_stats(vocab) best = max(pairs, key=pairs.get) if pairs[best] < 2: sys.stderr.write('no pair has frequency > 1. Stopping\n') break vocab = merge_vocab(best, vocab) print(best)
bsd-3-clause
18f9ea72a60b33c71b80fb0204f53c6a
34.145833
116
0.681091
3.153271
false
false
false
false
ufal/neuralmonkey
neuralmonkey/decoders/autoregressive.py
1
22666
# pylint: disable=too-many-lines """Abstract class for autoregressive decoding. Either for the recurrent decoder, or for the transformer decoder. The autoregressive decoder uses the while loop to get the outputs. Descendants should only specify the initial state and the while loop body. """ from typing import NamedTuple, Callable, Optional, Any, List, Dict, Tuple import tensorflow as tf from neuralmonkey.dataset import Dataset from neuralmonkey.decorators import tensor from neuralmonkey.model.feedable import FeedDict from neuralmonkey.model.parameterized import InitializerSpecs from neuralmonkey.model.model_part import ModelPart from neuralmonkey.logging import warn from neuralmonkey.model.sequence import EmbeddedSequence from neuralmonkey.nn.utils import dropout from neuralmonkey.tf_utils import ( append_tensor, get_variable, get_state_shape_invariants) from neuralmonkey.vocabulary import ( Vocabulary, pad_batch, sentence_mask, UNK_TOKEN_INDEX, START_TOKEN_INDEX, END_TOKEN_INDEX, PAD_TOKEN_INDEX) class LoopState(NamedTuple( "LoopState", [("histories", Any), ("constants", Any), ("feedables", Any)])): """The loop state object. The LoopState is a structure that works with the tf.while_loop function the decoder loop state stores all the information that is not invariant for the decoder run. Attributes: histories: A set of tensors that grow in time as the decoder proceeds. constants: A set of independent tensors that do not change during the entire decoder run. feedables: A set of tensors used as the input of a single decoder step. """ class DecoderHistories(NamedTuple( "DecoderHistories", [("logits", tf.Tensor), ("output_states", tf.Tensor), ("output_symbols", tf.Tensor), ("output_mask", tf.Tensor), ("other", Any)])): """The values collected during the run of an autoregressive decoder. This should only record decoding history and the decoding should not be dependent on these values. Attributes defined here (and in the `other`) substructure should always be time-major (e.g., shape(time, batch, ...)). Attributes: logits: A tensor of shape ``(time, batch, vocabulary)`` which contains the unnormalized output scores of words in a vocabulary. output_states: A tensor of shape ``(time, batch, state_size)``. The states of the decoder before the final output (logit) projection. output_symbols: An int tensor of shape ``(time, batch)``. Stores the generated symbols. (Either an argmax-ed value from the logits, or a target token, during training.) output_mask: A float tensor of zeros and ones of shape ``(time, batch)``. Keeps track of valid positions in the decoded data. other: A structure related to a specific AutoregressiveDecoder implementation. """ class DecoderConstants(NamedTuple( "DecoderConstants", [("train_inputs", Optional[tf.Tensor])])): """The constants used by an autoregressive decoder. Attributes: train_inputs: During training, this is populated by the target token ids. """ class DecoderFeedables(NamedTuple( "DecoderFeedables", [("step", tf.Tensor), ("finished", tf.Tensor), ("embedded_input", tf.Tensor), ("other", Any)])): """The input of a single step of an autoregressive decoder. The decoder should be able to generate an output symbol only using the information contained in this structure. Attributes defined here (and in the `other`) substructure should always be batch-major (e.g., shape(batch, ...)). Attributes: step: A scalar int tensor, stores the number of the current time step. finished: A boolean tensor of shape ``(batch)``, which says whether the decoding of a sentence in the batch is finished or not. (E.g. whether the end token has already been generated.) embedded_input: A ``batch``-sized tensor with embedded inputs to the decoder. During inference, this contains the previously generated tokens. During training, this contains the reference tokens. other: A structure related to a specific AutoregressiveDecoder implementation. """ # pylint: disable=too-many-public-methods,too-many-instance-attributes class AutoregressiveDecoder(ModelPart): # pylint: disable=too-many-arguments,too-many-locals def __init__(self, name: str, vocabulary: Vocabulary, data_id: str, max_output_len: int, dropout_keep_prob: float = 1.0, embedding_size: int = None, embeddings_source: EmbeddedSequence = None, tie_embeddings: bool = False, label_smoothing: float = None, supress_unk: bool = False, reuse: ModelPart = None, save_checkpoint: str = None, load_checkpoint: str = None, initializers: InitializerSpecs = None) -> None: """Initialize parameters common for all autoregressive decoders. Arguments: name: Name of the decoder. Should be unique accross all Neural Monkey objects. vocabulary: Target vocabulary. data_id: Target data series. max_output_len: Maximum length of an output sequence. reuse: Reuse the variables from the model part. dropout_keep_prob: Probability of keeping a value during dropout. embedding_size: Size of embedding vectors for target words. embeddings_source: Embedded sequence to take embeddings from. tie_embeddings: Use decoder.embedding_matrix also in place of the output decoding matrix. label_smoothing: Label smoothing parameter. supress_unk: If true, decoder will not produce symbols for unknown tokens. """ ModelPart.__init__(self, name, reuse, save_checkpoint, load_checkpoint, initializers) self.vocabulary = vocabulary self.data_id = data_id self.max_output_len = max_output_len self.dropout_keep_prob = dropout_keep_prob self._embedding_size = embedding_size self.embeddings_source = embeddings_source self.label_smoothing = label_smoothing self.tie_embeddings = tie_embeddings self.supress_unk = supress_unk self.encoder_states = lambda: [] # type: Callable[[], List[tf.Tensor]] self.encoder_masks = lambda: [] # type: Callable[[], List[tf.Tensor]] # Check the values of the parameters (max_output_len, ...) if self.max_output_len <= 0: raise ValueError( "Maximum sequence length must be a positive integer.") if self._embedding_size is not None and self._embedding_size <= 0: raise ValueError("Embedding size must be a positive integer.") if self.dropout_keep_prob < 0.0 or self.dropout_keep_prob > 1.0: raise ValueError("Dropout keep probability must be a real number " "in the interval [0,1].") # pylint: enable=too-many-arguments,too-many-locals @property def embedding_size(self) -> int: if self.embeddings_source is None: if self._embedding_size is None: raise ValueError( "You must specify either embedding size or the embedded " "sequence from which to reuse the embeddings (e.g. set " "'embedding_size' or 'embeddings_source' parameter)") return self._embedding_size if self.embeddings_source is not None: if self._embedding_size is not None: warn("Overriding the embedding_size parameter with the " "size of the reused embeddings from the encoder.") return self.embeddings_source.embedding_matrix.get_shape()[1].value @tensor def go_symbols(self) -> tf.Tensor: return tf.fill([self.batch_size], tf.constant(START_TOKEN_INDEX, dtype=tf.int64)) @property def input_types(self) -> Dict[str, tf.DType]: return {self.data_id: tf.string} @property def input_shapes(self) -> Dict[str, tf.TensorShape]: return {self.data_id: tf.TensorShape([None, None])} @tensor def train_tokens(self) -> tf.Tensor: return self.dataset[self.data_id] @tensor def train_inputs(self) -> tf.Tensor: return tf.transpose( self.vocabulary.strings_to_indices(self.train_tokens)) @tensor def train_mask(self) -> tf.Tensor: return sentence_mask(self.train_inputs) @tensor def decoding_w(self) -> tf.Variable: if (self.tie_embeddings and self.embedding_size != self.output_dimension): raise ValueError( "`embedding_size must be equal to the output_projection " "size when using the `tie_embeddings` option") with tf.name_scope("output_projection"): if self.tie_embeddings: return tf.transpose(self.embedding_matrix) return get_variable( "state_to_word_W", [self.output_dimension, len(self.vocabulary)], initializer=tf.random_uniform_initializer(-0.5, 0.5)) @tensor def decoding_b(self) -> Optional[tf.Variable]: if self.tie_embeddings: return tf.zeros(len(self.vocabulary)) with tf.name_scope("output_projection"): return get_variable( "state_to_word_b", [len(self.vocabulary)], initializer=tf.zeros_initializer()) @tensor def embedding_matrix(self) -> tf.Variable: """Variables and operations for embedding of input words. If we are reusing word embeddings, this function takes the embedding matrix from the first encoder """ if self.embeddings_source is not None: return self.embeddings_source.embedding_matrix assert self.embedding_size is not None return get_variable( name="word_embeddings", shape=[len(self.vocabulary), self.embedding_size]) def embed_input_symbols(self, input_symbols: tf.Tensor) -> tf.Tensor: embedded_input = tf.nn.embedding_lookup( self.embedding_matrix, input_symbols) return dropout(embedded_input, self.dropout_keep_prob, self.train_mode) @tensor def train_loop_result(self) -> LoopState: return self.decoding_loop(train_mode=True) @tensor def train_logits(self) -> tf.Tensor: train_result = LoopState(*self.train_loop_result) return train_result.histories.logits @tensor def train_output_states(self) -> tf.Tensor: train_result = LoopState(*self.train_loop_result) return train_result.histories.output_states @tensor def train_logprobs(self) -> tf.Tensor: return tf.nn.log_softmax(self.train_logits) @tensor def train_xents(self) -> tf.Tensor: train_targets = tf.transpose(self.train_inputs) softmax_function = None if self.label_smoothing: softmax_function = ( lambda labels, logits: tf.losses.softmax_cross_entropy( tf.one_hot(labels, len(self.vocabulary)), logits, label_smoothing=self.label_smoothing)) # Return losses of shape (batch, time). Losses on invalid positions # are zero. return tf.contrib.seq2seq.sequence_loss( tf.transpose(self.train_logits, perm=[1, 0, 2]), train_targets, tf.transpose(self.train_mask), average_across_batch=False, average_across_timesteps=False, softmax_loss_function=softmax_function) @tensor def train_loss(self) -> tf.Tensor: # Cross entropy mean over all words in the batch # (could also be done as a mean over sentences) return tf.reduce_sum(self.train_xents) / tf.reduce_sum(self.train_mask) @property def cost(self) -> tf.Tensor: return self.train_loss @tensor def runtime_loop_result(self) -> LoopState: return self.decoding_loop(train_mode=False) @tensor def runtime_logits(self) -> tf.Tensor: runtime_result = LoopState(*self.runtime_loop_result) return runtime_result.histories.logits @tensor def runtime_output_states(self) -> tf.Tensor: runtime_result = LoopState(*self.runtime_loop_result) return runtime_result.histories.output_states @tensor def runtime_mask(self) -> tf.Tensor: runtime_result = LoopState(*self.runtime_loop_result) return runtime_result.histories.output_mask @tensor def decoded(self) -> tf.Tensor: # We disable generating of <pad> tokens at index 0 # (self.runtime_logits[:, :, 1:]). This shifts the indices # of the decoded tokens (therefore, we add +1 to the decoded # output indices). # self.runtime_logits is of size [batch, sentence_len, vocabulary_size] return tf.argmax(self.runtime_logits[:, :, 1:], -1) + 1 @tensor def runtime_xents(self) -> tf.Tensor: train_targets = tf.transpose(self.train_inputs) batch_major_logits = tf.transpose(self.runtime_logits, [1, 0, 2]) min_time = tf.minimum(tf.shape(train_targets)[1], tf.shape(batch_major_logits)[1]) # NOTE if done properly, there should be padding of the shorter # sequence instead of cropping to the length of the shorter one return tf.contrib.seq2seq.sequence_loss( logits=batch_major_logits[:, :min_time], targets=train_targets[:, :min_time], weights=tf.transpose(self.train_mask)[:, :min_time], average_across_batch=False, average_across_timesteps=False) @tensor def runtime_loss(self) -> tf.Tensor: return (tf.reduce_sum(self.runtime_xents) / tf.reduce_sum(tf.to_float(self.runtime_mask))) @tensor def runtime_logprobs(self) -> tf.Tensor: return tf.nn.log_softmax(self.runtime_logits) @property def output_dimension(self) -> int: raise NotImplementedError("Abstract property") def get_initial_feedables(self) -> DecoderFeedables: return DecoderFeedables( step=tf.constant(0, tf.int32), finished=tf.zeros([self.batch_size], dtype=tf.bool), embedded_input=self.embed_input_symbols(self.go_symbols), other=None) def get_initial_histories(self) -> DecoderHistories: output_states = tf.zeros( shape=[0, self.batch_size, self.embedding_size], dtype=tf.float32, name="hist_output_states") output_mask = tf.zeros( shape=[0, self.batch_size], dtype=tf.bool, name="hist_output_mask") output_symbols = tf.zeros( shape=[0, self.batch_size], dtype=tf.int64, name="hist_output_symbols") logits = tf.zeros( shape=[0, self.batch_size, len(self.vocabulary)], dtype=tf.float32, name="hist_logits") return DecoderHistories( logits=logits, output_states=output_states, output_mask=output_mask, output_symbols=output_symbols, other=None) def get_initial_constants(self) -> DecoderConstants: return DecoderConstants(train_inputs=self.train_inputs) def get_initial_loop_state(self) -> LoopState: return LoopState( feedables=self.get_initial_feedables(), histories=self.get_initial_histories(), constants=self.get_initial_constants()) def loop_continue_criterion(self, *args) -> tf.Tensor: """Decide whether to break out of the while loop. Arguments: loop_state: ``LoopState`` instance (see the docs for this module). Represents current decoder loop state. """ loop_state = LoopState(*args) finished = loop_state.feedables.finished not_all_done = tf.logical_not(tf.reduce_all(finished)) before_max_len = tf.less(loop_state.feedables.step, self.max_output_len) return tf.logical_and(not_all_done, before_max_len) def next_state(self, loop_state: LoopState) -> Tuple[tf.Tensor, Any, Any]: raise NotImplementedError("Abstract method.") def get_body(self, train_mode: bool, sample: bool = False, temperature: float = 1.) -> Callable: """Return the while loop body function.""" def is_finished(finished: tf.Tensor, symbols: tf.Tensor) -> tf.Tensor: has_just_finished = tf.equal(symbols, END_TOKEN_INDEX) return tf.logical_or(finished, has_just_finished) def state_to_logits(state: tf.Tensor) -> tf.Tensor: logits = tf.matmul(state, self.decoding_w) logits += self.decoding_b if self.supress_unk: unk_mask = tf.one_hot( UNK_TOKEN_INDEX, depth=len(self.vocabulary), on_value=-1e9) logits += unk_mask return logits def logits_to_symbols(logits: tf.Tensor, loop_state: LoopState) -> tf.Tensor: step = loop_state.feedables.step if sample: next_symbols = tf.squeeze( tf.multinomial(logits, num_samples=1), axis=1) elif train_mode: next_symbols = loop_state.constants.train_inputs[step] else: next_symbols = tf.argmax(logits, axis=1) int_unfinished_mask = tf.to_int64( tf.logical_not(loop_state.feedables.finished)) # Note this works only when PAD_TOKEN_INDEX is 0. Otherwise # this have to be rewritten assert PAD_TOKEN_INDEX == 0 next_symbols = next_symbols * int_unfinished_mask return next_symbols def body(*args) -> LoopState: loop_state = LoopState(*args) feedables = loop_state.feedables histories = loop_state.histories with tf.variable_scope(self._variable_scope, reuse=tf.AUTO_REUSE): output_state, dec_other, hist_other = self.next_state( loop_state) logits = state_to_logits(output_state) logits /= temperature next_symbols = logits_to_symbols(logits, loop_state) finished = is_finished(feedables.finished, next_symbols) next_feedables = DecoderFeedables( step=feedables.step + 1, finished=finished, embedded_input=self.embed_input_symbols(next_symbols), other=dec_other) next_histories = DecoderHistories( logits=append_tensor(histories.logits, logits), output_states=append_tensor( histories.output_states, output_state), output_symbols=append_tensor( histories.output_symbols, next_symbols), output_mask=append_tensor( histories.output_mask, tf.logical_not(finished)), other=hist_other) return LoopState( feedables=next_feedables, histories=next_histories, constants=loop_state.constants) return body def finalize_loop(self, final_loop_state: LoopState, train_mode: bool) -> None: """Execute post-while loop operations. Arguments: final_loop_state: Decoder loop state at the end of the decoding loop. train_mode: Boolean flag, telling whether this is a training run. """ def decoding_loop(self, train_mode: bool, sample: bool = False, temperature: float = 1) -> LoopState: """Run the decoding while loop. Calls get_initial_loop_state and constructs tf.while_loop with the continuation criterion returned from loop_continue_criterion, and body function returned from get_body. After finishing the tf.while_loop, it calls finalize_loop to further postprocess the final decoder loop state (usually by stacking Tensors containing decoding histories). Arguments: train_mode: Boolean flag, telling whether this is a training run. sample: Boolean flag, telling whether we should sample the output symbols from the output distribution instead of using argmax or gold data. temperature: float value specifying the softmax temperature """ initial_loop_state = self.get_initial_loop_state() with tf.control_dependencies([self.decoding_w, self.decoding_b]): final_loop_state = tf.while_loop( self.loop_continue_criterion, self.get_body(train_mode, sample, temperature), initial_loop_state, shape_invariants=tf.contrib.framework.nest.map_structure( get_state_shape_invariants, initial_loop_state)) self.finalize_loop(final_loop_state, train_mode) return final_loop_state def feed_dict(self, dataset: Dataset, train: bool = False) -> FeedDict: """Populate the feed dictionary for the decoder object. Arguments: dataset: The dataset to use for the decoder. train: Boolean flag, telling whether this is a training run. """ fd = ModelPart.feed_dict(self, dataset, train) sentences = dataset.maybe_get_series(self.data_id) if sentences is None and train: raise ValueError("When training, you must feed " "reference sentences") if sentences is not None: fd[self.train_tokens] = pad_batch( list(sentences), self.max_output_len, add_start_symbol=False, add_end_symbol=True) return fd
bsd-3-clause
5c204f2ea76de930d1c8c54e8e276e30
37.811644
79
0.61215
4.269354
false
false
false
false
ufal/neuralmonkey
neuralmonkey/evaluators/beer.py
1
2482
import tempfile import subprocess from typing import List from typeguard import check_argument_types from neuralmonkey.logging import log from neuralmonkey.evaluators.evaluator import Evaluator class BeerWrapper(Evaluator[List[str]]): """Wrapper for BEER scorer. Paper: http://aclweb.org/anthology/D14-1025 Code: https://github.com/stanojevic/beer """ def __init__(self, wrapper: str, name: str = "BEER", encoding: str = "utf-8") -> None: """Initialize the BEER wrapper. Args: name: Name of the evaluator. wrapper: Path to the BEER's executable. encoding: Data encoding. """ check_argument_types() super().__init__(name) self.wrapper = wrapper self.encoding = encoding def serialize_to_bytes(self, sentences: List[List[str]]) -> bytes: joined = [" ".join(r) for r in sentences] string = "\n".join(joined) + "\n" return string.encode(self.encoding) def score_batch(self, hypotheses: List[List[str]], references: List[List[str]]) -> float: ref_bytes = self.serialize_to_bytes(references) hyp_bytes = self.serialize_to_bytes(hypotheses) with tempfile.NamedTemporaryFile() as reffile, \ tempfile.NamedTemporaryFile() as hypfile: reffile.write(ref_bytes) reffile.flush() hypfile.write(hyp_bytes) hypfile.flush() args = [self.wrapper, "-r", reffile.name, "-s", hypfile.name] output_proc = subprocess.run(args, stderr=subprocess.PIPE, stdout=subprocess.PIPE) proc_stdout = output_proc.stdout.decode("utf-8") # type: ignore lines = proc_stdout.splitlines() if not lines: return 0.0 try: beer_score = float(lines[0].split()[-1]) return beer_score except IndexError: log("Error: Malformed output from BEER wrapper:", color="red") log(proc_stdout, color="red") log("=======", color="red") return 0.0 except ValueError: log("Value error - beer '{}' is not a number.".format( lines[0]), color="red") return 0.0
bsd-3-clause
5984f38326c6b692c25fa9cd129e2c29
30.820513
78
0.532635
4.339161
false
false
false
false
ufal/neuralmonkey
neuralmonkey/trainers/self_critical_objective.py
1
8044
"""Training objective for self-critical learning. Self-critic learning is a modification of the REINFORCE algorithm that uses the reward of the train-time decoder output as a baseline in the update step. For more details see: https://arxiv.org/pdf/1612.00563.pdf """ from typing import Callable, Iterable, Tuple, Optional from itertools import takewhile from collections import Counter import numpy as np import tensorflow as tf from typeguard import check_argument_types from neuralmonkey.decoders.decoder import Decoder from neuralmonkey.decorators import tensor from neuralmonkey.trainers.generic_trainer import Objective from neuralmonkey.vocabulary import END_TOKEN_INDEX # pylint: disable=invalid-name RewardFunction = Callable[[np.ndarray, np.ndarray], np.ndarray] # pylint: enable=invalid-name class SelfCriticalObjective(Objective[Decoder]): def __init__(self, decoder: Decoder, reward_function: RewardFunction, weight: float = None) -> None: """Self-critical objective. Args: decoder: A recurrent decoder. reward_function: A reward function computing score in Python. weight: Mixing weight for a trainer. Returns: Objective object to be used in generic trainer. """ check_argument_types() name = "{}_self_critical".format(decoder.name) super().__init__(name, decoder) self.reward_function = reward_function self._weight = weight @tensor def weight(self) -> Optional[tf.Tensor]: if self._weight is None: return None return tf.constant(self._weight) @tensor def loss(self) -> tf.Tensor: # decoded, shape (time, batch) train_decoded = tf.argmax(self.decoder.train_logits, axis=2) runtime_decoded = tf.argmax(self.decoder.runtime_logits, axis=2) reference = self.decoder.train_inputs # rewards, shape (batch) train_reward = tf.py_func( self.reward_function, [reference, train_decoded], tf.float32) runtime_reward = tf.py_func( self.reward_function, [reference, runtime_decoded], tf.float32) tf.summary.scalar( "train_{}/{}".format(self.decoder.data_id, self.reward_function.__name__), tf.reduce_mean(runtime_reward), collections=["summary_train"]) # REINFORCE score: shape (time, batch, vocab) score_by_word = reinforce_score( runtime_reward, train_reward, runtime_decoded, self.decoder.runtime_logits) float_mask = tf.to_float(self.decoder.runtime_mask) masked_score_by_word = score_by_word * float_mask # sum the matrix (dot product of rows, sum over time, and over batch) # pylint: disable=invalid-unary-operand-type loss = -tf.reduce_sum(masked_score_by_word) / tf.reduce_sum(float_mask) # pylint: enable=invalid-unary-operand-type tf.summary.scalar( "train_{}/self_critical_cost".format(self.decoder.data_id), loss, collections=["summary_train"]) return loss def reinforce_score(reward: tf.Tensor, baseline: tf.Tensor, decoded: tf.Tensor, logits: tf.Tensor) -> tf.Tensor: """Cost function whose derivative is the REINFORCE equation. This implements the primitive function to the central equation of the REINFORCE algorithm that estimates the gradients of the loss with respect to decoder logits. It uses the fact that the second term of the product (the difference of the word distribution and one hot vector of the decoded word) is a derivative of negative log likelihood of the decoded word. The reward function and the baseline are however treated as a constant, so they influence the derivate only multiplicatively. """ # shape (1, batch, 1) reward_diff = tf.expand_dims(reward - baseline, 0) # runtime probabilities, shape (time, batch, vocab) decoded_neg_likelihood = tf.nn.sparse_softmax_cross_entropy_with_logits( labels=decoded, logits=logits) # REINFORCE gradient, shape (time, batch, vocab) score = tf.stop_gradient(reward_diff) * decoded_neg_likelihood return score def sentence_bleu(references: np.ndarray, hypotheses: np.ndarray) -> np.ndarray: """Compute index-based sentence-level BLEU score. Computes sentence level BLEU on indices outputed by the decoder, i.e. whatever the decoder uses as a unit is used a token in the BLEU computation, ignoring the tokens may be sub-word units. """ bleu_scores = [] for ref, hyp in zip(np.transpose(references), np.transpose(hypotheses)): matched_counts = [] hyp_n_grams_counts = [] for n in range(1, 5): matched, total, _ = _count_matching_n_grams(ref, hyp, n) if n > 1: matched += 1 total += 1 matched_counts.append(matched) hyp_n_grams_counts.append(total) if hyp_n_grams_counts[0] == 0: bleu_scores.append(0.) else: precision = ( np.prod(matched_counts) / np.prod(hyp_n_grams_counts)) ** .25 ref_len = sum(1 for _ in takewhile(lambda i: i != END_TOKEN_INDEX, ref)) brevity_penalty = np.min([ 1., np.exp(1 - ref_len / hyp_n_grams_counts[0])]) bleu_scores.append(brevity_penalty * precision) assert all(0 <= s <= 1 for s in bleu_scores) return np.array(bleu_scores, dtype=np.float32) def sentence_gleu(references: np.ndarray, hypotheses: np.ndarray) -> np.ndarray: """Compute index-based GLEU score. GLEU score is a sentence-level metric used in Google's Neural MT as a reward in reinforcement learning (https://arxiv.org/abs/1609.08144). It is a minimum of precision and recall on 1- to 4-grams. It operates over the indices emitted by the decoder which are not necessarily tokens (could be characters or subword units). """ gleu_scores = [] for ref, hyp in zip(np.transpose(references), np.transpose(hypotheses)): matched_counts = [] hyp_n_grams_counts = [] ref_n_grams_counts = [] for n in range(1, 5): matched, total_hyp, total_ref = _count_matching_n_grams( ref, hyp, n) matched_counts.append(matched) hyp_n_grams_counts.append(total_hyp) ref_n_grams_counts.append(total_ref) precision = np.sum(matched_counts) / np.sum(hyp_n_grams_counts) recall = np.sum(matched_counts) / np.sum(ref_n_grams_counts) assert 0. <= precision <= 1.0 assert 0. <= recall <= 1.0 gleu_scores.append(min(precision, recall)) return np.array(gleu_scores, dtype=np.float32) def _count_matching_n_grams(ref: np.ndarray, hyp: np.ndarray, n: int) -> Tuple[int, int, int]: ref_counts = Counter() # type: Counter total_ref_n_grams = 0 for n_gram in _get_n_grams(ref, n): ref_counts[str(n_gram)] += 1 total_ref_n_grams += 1 matched_n_grams = 0 total_hyp_n_grams = 0 hyp_n_grams = _get_n_grams(hyp, n) for n_gram in hyp_n_grams: n_gram_s = str(n_gram) if ref_counts[n_gram_s] > 0: matched_n_grams += 1 ref_counts[n_gram_s] -= 1 total_hyp_n_grams += 1 assert matched_n_grams <= total_hyp_n_grams assert matched_n_grams <= total_ref_n_grams return matched_n_grams, total_hyp_n_grams, total_ref_n_grams def _get_n_grams(indices: np.ndarray, order: int) -> Iterable[np.ndarray]: all_n_grams = [indices[i:i + order] for i in range(len(indices) - order + 1)] return takewhile(lambda g: g[-1] != END_TOKEN_INDEX, all_n_grams)
bsd-3-clause
4c7221fa0ad05c93d6e1866875894899
33.822511
79
0.620338
3.783631
false
false
false
false
ufal/neuralmonkey
neuralmonkey/runners/runner.py
1
3169
from typing import Dict, List, Callable, Union import numpy as np import tensorflow as tf from typeguard import check_argument_types from neuralmonkey.runners.base_runner import BaseRunner, NextExecute from neuralmonkey.decoders.autoregressive import AutoregressiveDecoder from neuralmonkey.decoders.classifier import Classifier from neuralmonkey.decorators import tensor # pylint: disable=invalid-name SupportedDecoder = Union[AutoregressiveDecoder, Classifier] Postprocessor = Callable[[List[List[str]]], List[List[str]]] # pylint: enable=invalid-name class GreedyRunner(BaseRunner[SupportedDecoder]): class Executable(BaseRunner.Executable["GreedyRunner"]): def next_to_execute(self) -> NextExecute: """Get the tensors and additional feed dicts for execution.""" fetches = self.executor.fetches if not self.summaries: fetches["image_summaries"] = None if not self.compute_losses: fetches["train_xent"] = tf.zeros([]) fetches["runtime_xent"] = tf.zeros([]) return fetches, [] def collect_results(self, results: List[Dict]) -> None: train_loss = 0. runtime_loss = 0. summed_logprobs = [-np.inf for _ in range( results[0]["decoded_logprobs"].shape[0])] for sess_result in results: train_loss += sess_result["train_xent"] runtime_loss += sess_result["runtime_xent"] for i, logprob in enumerate(sess_result["decoded_logprobs"]): summed_logprobs[i] = np.logaddexp( summed_logprobs[i], logprob) argmaxes = [np.argmax(l, axis=1) for l in summed_logprobs] decoded_tokens = self.executor.vocabulary.vectors_to_sentences( argmaxes) if self.executor.postprocess is not None: decoded_tokens = self.executor.postprocess(decoded_tokens) summaries = None if "image_summaries" in results[0]: summaries = [results[0]["image_summaries"]] self.set_runner_result( outputs=decoded_tokens, losses=[train_loss, runtime_loss], summaries=summaries) def __init__(self, output_series: str, decoder: SupportedDecoder, postprocess: Postprocessor = None) -> None: check_argument_types() super().__init__(output_series, decoder) self.postprocess = postprocess self.vocabulary = self.decoder.vocabulary @tensor def fetches(self) -> Dict[str, tf.Tensor]: fetches = {"decoded_logprobs": self.decoder.runtime_logprobs, "train_xent": self.decoder.train_loss, "runtime_xent": self.decoder.runtime_loss} att_plot_summaries = tf.get_collection("summary_att_plots") if att_plot_summaries: fetches["image_summaries"] = tf.summary.merge(att_plot_summaries) return fetches @property def loss_names(self) -> List[str]: return ["train_xent", "runtime_xent"]
bsd-3-clause
ffe5a67c40dd1556f7e821448283390c
34.211111
77
0.611549
4.253691
false
false
false
false
ufal/neuralmonkey
lib/subword_nmt/segment-char-ngrams.py
3
2377
#!/usr/bin/env python # -*- coding: utf-8 -*- # Author: Rico Sennrich from __future__ import unicode_literals, division import sys import codecs import argparse # hack for python2/3 compatibility from io import open argparse.open = open # python 2/3 compatibility if sys.version_info < (3, 0): sys.stderr = codecs.getwriter('UTF-8')(sys.stderr) sys.stdout = codecs.getwriter('UTF-8')(sys.stdout) sys.stdin = codecs.getreader('UTF-8')(sys.stdin) def create_parser(): parser = argparse.ArgumentParser( formatter_class=argparse.RawDescriptionHelpFormatter, description="segment rare words into character n-grams") parser.add_argument( '--input', '-i', type=argparse.FileType('r'), default=sys.stdin, metavar='PATH', help="Input file (default: standard input).") parser.add_argument( '--vocab', type=argparse.FileType('r'), metavar='PATH', required=True, help="Vocabulary file.") parser.add_argument( '--shortlist', type=int, metavar='INT', default=0, help="do not segment INT most frequent words in vocabulary (default: '%(default)s')).") parser.add_argument( '-n', type=int, metavar='INT', default=2, help="segment rare words into character n-grams of size INT (default: '%(default)s')).") parser.add_argument( '--output', '-o', type=argparse.FileType('w'), default=sys.stdout, metavar='PATH', help="Output file (default: standard output)") parser.add_argument( '--separator', '-s', type=str, default='@@', metavar='STR', help="Separator between non-final subword units (default: '%(default)s'))") return parser if __name__ == '__main__': parser = create_parser() args = parser.parse_args() vocab = [line.split()[0] for line in args.vocab if len(line.split()) == 2] vocab = dict((y,x) for (x,y) in enumerate(vocab)) for line in args.input: for word in line.split(): if word not in vocab or vocab[word] > args.shortlist: i = 0 while i*args.n < len(word): args.output.write(word[i*args.n:i*args.n+args.n]) i += 1 if i*args.n < len(word): args.output.write(args.separator) args.output.write(' ') else: args.output.write(word + ' ') args.output.write('\n')
bsd-3-clause
d64c519fbba7bf78a48e718f32b45bab
32.478873
96
0.611275
3.656923
false
false
false
false
ufal/neuralmonkey
neuralmonkey/trainers/delayed_update_trainer.py
1
9072
# pylint: disable=unused-import from typing import Dict, List, Tuple, Optional # pylint: enable=unused-import import tensorflow as tf from typeguard import check_argument_types from neuralmonkey.decorators import tensor from neuralmonkey.runners.base_runner import GraphExecutor, NextExecute from neuralmonkey.trainers.generic_trainer import (GenericTrainer, Objective, Gradients) class DelayedUpdateTrainer(GenericTrainer): class Executable(GraphExecutor.Executable["DelayedUpdateTrainer"]): def __init__(self, executor: "DelayedUpdateTrainer", compute_losses: bool, summaries: bool, num_sessions: int) -> None: assert compute_losses if num_sessions != 1: raise ValueError( "Trainer only supports execution in a single session") super().__init__(executor, compute_losses, summaries, num_sessions) self.state = 0 self.res_sums = [] # type: List[tf.Summary] self.res_losses = None # type: Optional[List[float]] self.res_batch = None # type: Optional[int] def next_to_execute(self) -> NextExecute: if self.state == 0: # ACCUMULATING fetches = {"accumulators": self.executor.accumulate_ops, "counter": self.executor.cumulator_counter, "batch_size": self.executor.batch_size, "losses": self.executor.objective_values} elif self.state == 1: # UPDATING fetches = { "train_op": self.executor.train_op, "_update_ops": tf.get_collection(tf.GraphKeys.UPDATE_OPS)} if self.summaries: fetches.update(self.executor.summaries) else: # RESETTING fetches = {"resets": self.executor.reset_ops} return fetches, [] def collect_results(self, results: List[Dict]) -> None: assert len(results) == 1 result = results[0] if self.state == 0: # ACCUMULATING self.res_losses = result["losses"] self.res_batch = result["batch_size"] # Are we updating? counter = result["counter"] if counter == self.executor.batches_per_update: self.state = 1 return elif self.state == 1: if self.summaries: self.res_sums = [result["scalar_summaries"], result["histogram_summaries"]] self.state = 2 return assert self.res_losses is not None assert self.res_batch is not None objective_names = [obj.name for obj in self.executor.objectives] objective_names += ["L1", "L2"] losses = dict(zip(objective_names, self.res_losses)) self.set_result({}, losses, self.res_batch, self.res_sums) # pylint: disable=too-many-arguments def __init__(self, batches_per_update: int, objectives: List[Objective], l1_weight: float = 0.0, l2_weight: float = 0.0, clip_norm: float = None, optimizer: tf.train.Optimizer = None, var_scopes: List[str] = None, var_collection: str = None) -> None: check_argument_types() GenericTrainer.__init__(self, objectives, l1_weight, l2_weight, clip_norm, optimizer, var_scopes, var_collection) self.batches_per_update = batches_per_update # pylint: enable=too-many-arguments @tensor def existing_grads_and_vars(self) -> Tuple[ List[tf.Tensor], List[tf.Variable]]: orig_grads = super().raw_gradients # pylint: disable=not-an-iterable # Pylint does not understand @tensor annotations transposed = tuple(zip( *[(grad, var) for grad, var in orig_grads if grad is not None])) # pylint: enable=not-an-iterable return list(transposed[0]), list(transposed[1]) @tensor def gradient_buffers(self) -> List[tf.Variable]: # pylint: disable=unpacking-non-sequence existing_gradients, _ = self.existing_grads_and_vars # pylint: enable=unpacking-non-sequence with tf.variable_scope("gradient_buffer"): return [tf.Variable(initial_value=tf.zeros_like(grad), trainable=False) for grad in existing_gradients] @tensor def objective_buffers(self) -> List[tf.Variable]: with tf.variable_scope("loss_buffers"): return [tf.Variable(0.0, trainable=False) for _ in self.objectives] # pylint: disable=no-self-use @tensor def diff_buffer(self) -> tf.Variable: return tf.Variable(0.0, trainable=False) @tensor def cumulator_counter(self) -> tf.Variable: return tf.Variable(0, trainable=False, name="cumulator_counter") # pylint: enable=no-self-use @tensor def accumulate_ops(self) -> List[tf.Operation]: # pylint: disable=unpacking-non-sequence existing_gradients, _ = self.existing_grads_and_vars # pylint: enable=unpacking-non-sequence # pylint: disable=not-an-iterable # Pylint does not understand @tensor annotations accumulate_ops = [ tf.assign_add(gradbuf, grad) for gradbuf, grad in zip( self.gradient_buffers, existing_gradients)] accumulate_ops.extend( tf.assign_add(objbuf, obj.loss) for objbuf, obj in zip(self.objective_buffers, self.objectives)) # pylint: enable=not-an-iterable accumulate_ops.append( tf.assign_add(self.diff_buffer, self.differentiable_loss_sum)) accumulate_ops.append( tf.assign_add(self.cumulator_counter, 1)) return accumulate_ops @tensor def reset_ops(self) -> List[tf.Operation]: # pylint: disable=not-an-iterable # Pylint does not understand @tensor annotations reset_ops = [tf.assign(gradbuf, tf.zeros_like(gradbuf)) for gradbuf in self.gradient_buffers] reset_ops.extend( tf.assign(objbuf, 0.0) for objbuf in self.objective_buffers) # pylint: enable=not-an-iterable reset_ops.append(tf.assign(self.diff_buffer, 0.0)) reset_ops.append(tf.assign(self.cumulator_counter, 0)) return reset_ops @tensor def raw_gradients(self) -> Gradients: """Return averaged gradients over buffers.""" # pylint: disable=not-an-iterable # Pylint does not understand @tensor annotations averaged_grads = [grad / tf.to_float(self.cumulator_counter) for grad in self.gradient_buffers] # pylint: enable=not-an-iterable tf.summary.scalar( "train_opt_cost", self.diff_buffer / tf.to_float(self.cumulator_counter), collections=["summary_train"]) # log all objectives for obj, objbuf in zip(self.objectives, self.objective_buffers): tf.summary.scalar( obj.name, objbuf / tf.to_float(self.cumulator_counter), collections=["summary_train"]) # now, zip averaged grads with associated vars to a Gradients struct. # pylint: disable=unpacking-non-sequence _, existing_vars = self.existing_grads_and_vars # pylint: enable=unpacking-non-sequence return list(zip(averaged_grads, existing_vars)) @tensor def summaries(self) -> Dict[str, tf.Tensor]: # pylint: disable=protected-access if isinstance(self.optimizer._lr, tf.Tensor): tf.summary.scalar("learning_rate", self.optimizer._lr, collections=["summary_train"]) # pylint: enable=protected-access # pylint: disable=unpacking-non-sequence l1_norm, l2_norm = self.regularization_losses # pylint: enable=unpacking-non-sequence tf.summary.scalar("train_l1", l1_norm, collections=["summary_train"]) tf.summary.scalar("train_l2", l2_norm, collections=["summary_train"]) # pylint: disable=not-an-iterable # Pylint does not understand @tensor annotations for grad, var in self.gradients: if grad is not None: summary_name = "gr_{}".format(var.name) tf.summary.histogram( summary_name, grad, collections=["summary_gradients"]) # pylint: enable=not-an-iterable return { "scalar_summaries": tf.summary.merge( tf.get_collection("summary_train")), "histogram_summaries": tf.summary.merge( tf.get_collection("summary_gradients"))}
bsd-3-clause
4df8a04e1c83750dfe09805d23f4167f
37.769231
79
0.578924
4.182573
false
false
false
false
ufal/neuralmonkey
neuralmonkey/writers/plain_text_writer.py
1
1974
from typing import Iterator, List, Any, Callable from neuralmonkey.logging import log from neuralmonkey.readers.plain_text_reader import ALNUM_CHARSET # pylint: disable=invalid-name # Writer: function that gets file and the data Writer = Callable[[str, Any], None] # pylint: enable=invalid-name def t2t_detokenize(data: Iterator[List[str]]) -> Iterator[str]: """Detokenize text tokenized by t2t_tokenized_text_reader. Method is inspired by tensor2tensor tokenizer.decode method: https://github.com/tensorflow/tensor2tensor/blob/v1.5.5/tensor2tensor/data_generators/tokenizer.py """ for sentence in data: is_alnum = [t[0] in ALNUM_CHARSET for t in sentence] ret = [] for i, token in enumerate(sentence): if i > 0 and is_alnum[i - 1] and is_alnum[i]: ret.append(" ") ret.append(token) yield "".join(ret) def text_writer(encoding: str = "utf-8") -> Writer: def writer(path: str, data: Iterator) -> None: with open(path, "w", encoding=encoding) as f_out: for sentence in data: f_out.write(str(sentence) + "\n") log("Result saved as plain text in '{}'".format(path)) return writer def tokenized_text_writer(encoding: str = "utf-8") -> Writer: """Get a writer that is reversed to the tokenized_text_reader.""" def writer(path: str, data: Iterator[List[str]]) -> None: wrt = text_writer(encoding) wrt(path, (" ".join(s) for s in data)) return writer def t2t_tokenized_text_writer(encoding: str = "utf-8") -> Writer: """Get a writer that is reversed to the t2t_tokenized_text_reader.""" def writer(path: str, data: Iterator[List[str]]) -> None: wrt = text_writer(encoding) wrt(path, t2t_detokenize(data)) return writer # pylint: disable=invalid-name UtfPlainTextWriter = tokenized_text_writer() T2TWriter = t2t_tokenized_text_writer() # pylint: enable=invalid-name
bsd-3-clause
79323426254959c8ca5ff25fa1d2af6e
31.9
102
0.653495
3.433043
false
false
false
false
ufal/neuralmonkey
neuralmonkey/encoders/transformer.py
1
12857
"""Implementation of the encoder of the Transformer model. Described in Vaswani et al. (2017), arxiv.org/abs/1706.03762 """ from typing import List import math import tensorflow as tf from typeguard import check_argument_types from neuralmonkey.attention.base_attention import ( Attendable, get_attention_states, get_attention_mask) from neuralmonkey.decorators import tensor from neuralmonkey.attention.scaled_dot_product import attention from neuralmonkey.model.parameterized import InitializerSpecs from neuralmonkey.model.model_part import ModelPart from neuralmonkey.model.stateful import (TemporalStateful, TemporalStatefulWithOutput) from neuralmonkey.nn.utils import dropout from neuralmonkey.tf_utils import get_variable, layer_norm def position_signal(dimension: int, length: tf.Tensor) -> tf.Tensor: # Code simplified and copied from github.com/tensorflow/tensor2tensor # TODO write this down on a piece of paper and understand the code and # compare it to the paper positions = tf.to_float(tf.range(length)) num_timescales = dimension // 2 # see: github.com/tensorflow/tensor2tensor/blob/v1.5.5/tensor2tensor/ # layers/common_attention.py#L425 log_timescale_increment = math.log(1.0e4) / (num_timescales - 1) inv_timescales = tf.exp(tf.range(num_timescales, dtype=tf.float32) * -log_timescale_increment) scaled_time = tf.expand_dims(positions, 1) * tf.expand_dims( inv_timescales, 0) signal = tf.concat([tf.sin(scaled_time), tf.cos(scaled_time)], axis=1) signal = tf.pad(signal, [[0, 0], [0, tf.mod(dimension, 2)]]) signal = tf.reshape(signal, [1, length, dimension]) return signal class TransformerLayer(TemporalStateful): def __init__(self, states: tf.Tensor, mask: tf.Tensor) -> None: self._states = states self._mask = mask @property def temporal_states(self) -> tf.Tensor: return self._states @property def temporal_mask(self) -> tf.Tensor: return self._mask # pylint: disable=too-many-instance-attributes class TransformerEncoder(ModelPart, TemporalStatefulWithOutput): # pylint: disable=too-many-arguments,too-many-locals def __init__(self, name: str, input_sequence: TemporalStateful, ff_hidden_size: int, depth: int, n_heads: int, dropout_keep_prob: float = 1.0, attention_dropout_keep_prob: float = 1.0, target_space_id: int = None, use_att_transform_bias: bool = False, use_positional_encoding: bool = True, input_for_cross_attention: Attendable = None, n_cross_att_heads: int = None, reuse: ModelPart = None, save_checkpoint: str = None, load_checkpoint: str = None, initializers: InitializerSpecs = None) -> None: """Create an encoder of the Transformer model. Described in Vaswani et al. (2017), arxiv.org/abs/1706.03762 Arguments: input_sequence: Embedded input sequence. name: Name of the decoder. Should be unique accross all Neural Monkey objects. reuse: Reuse the model variables. dropout_keep_prob: Probability of keeping a value during dropout. target_space_id: Specifies the modality of the target space. use_att_transform_bias: Add bias when transforming qkv vectors for attention. use_positional_encoding: If True, position encoding signal is added to the input. Keyword arguments: ff_hidden_size: Size of the feedforward sublayers. n_heads: Number of the self-attention heads. depth: Number of sublayers. attention_dropout_keep_prob: Probability of keeping a value during dropout on the attention output. input_for_cross_attention: An attendable model part that is attended using cross-attention on every layer of the decoder, analogically to how encoder is attended in the decoder. n_cross_att_heads: Number of heads used in the cross-attention. """ check_argument_types() ModelPart.__init__(self, name, reuse, save_checkpoint, load_checkpoint, initializers) self.input_sequence = input_sequence self.ff_hidden_size = ff_hidden_size self.depth = depth self.n_heads = n_heads self.dropout_keep_prob = dropout_keep_prob self.attention_dropout_keep_prob = attention_dropout_keep_prob self.target_space_id = target_space_id self.use_att_transform_bias = use_att_transform_bias self.use_positional_encoding = use_positional_encoding self.input_for_cross_attention = input_for_cross_attention self.n_cross_att_heads = n_cross_att_heads if self.depth <= 0: raise ValueError("Depth must be a positive integer.") if self.ff_hidden_size <= 0: raise ValueError("Feed forward hidden size must be a " "positive integer.") if self.dropout_keep_prob <= 0.0 or self.dropout_keep_prob > 1.0: raise ValueError("Dropout keep prob must be inside (0,1].") if (self.attention_dropout_keep_prob <= 0.0 or self.attention_dropout_keep_prob > 1.0): raise ValueError("Dropout keep prob for attn must be in (0,1].") if self.target_space_id is not None and (self.target_space_id >= 32 or self.target_space_id < 0): raise ValueError( "If provided, the target space ID should be between 0 and 31. " "Was: {}".format(self.target_space_id)) if (input_for_cross_attention is None) != (n_cross_att_heads is None): raise ValueError( "Either both input_for_cross_attention and n_cross_att_heads " "must be provided or none of them.") self._variable_scope.set_initializer(tf.variance_scaling_initializer( mode="fan_avg", distribution="uniform")) # pylint: enable=too-many-arguments,too-many-locals @property def model_dimension(self) -> int: dim = self.input_sequence.dimension if self.input_for_cross_attention is not None: cross_att_dim = get_attention_states( self.input_for_cross_attention).get_shape()[-1].value if cross_att_dim != dim: raise ValueError( "The input for cross-attention must be of the same " "dimension as the model, was {}.".format(cross_att_dim)) return dim @tensor def output(self) -> tf.Tensor: return tf.reduce_sum(self.temporal_states, axis=1) @tensor def modality_matrix(self) -> tf.Tensor: """Create an embedding matrix for varyining target modalities. Used to embed different target space modalities in the tensor2tensor models (e.g. during the zero-shot translation). """ emb_size = self.input_sequence.temporal_states.shape.as_list()[-1] return get_variable( name="target_modality_embedding_matrix", shape=[32, emb_size], dtype=tf.float32, initializer=tf.variance_scaling_initializer( mode="fan_avg", distribution="uniform")) @tensor def target_modality_embedding(self) -> tf.Tensor: """Gather correct embedding of the target space modality. See TransformerEncoder.modality_matrix for more information. """ return tf.gather(self.modality_matrix, tf.constant(self.target_space_id)) @tensor def encoder_inputs(self) -> tf.Tensor: inputs = self.input_sequence.temporal_states if self.target_space_id is not None: inputs += tf.reshape(self.target_modality_embedding, [1, 1, -1]) length = tf.shape(inputs)[1] if self.use_positional_encoding: inputs += position_signal(self.model_dimension, length) return dropout(inputs, self.dropout_keep_prob, self.train_mode) def self_attention_sublayer( self, prev_layer: TransformerLayer) -> tf.Tensor: """Create the encoder self-attention sublayer.""" # Layer normalization normalized_states = layer_norm(prev_layer.temporal_states) # Run self-attention self_context, _ = attention( queries=normalized_states, keys=normalized_states, values=normalized_states, keys_mask=prev_layer.temporal_mask, num_heads=self.n_heads, dropout_callback=lambda x: dropout( x, self.attention_dropout_keep_prob, self.train_mode), use_bias=self.use_att_transform_bias) # Apply dropout self_context = dropout( self_context, self.dropout_keep_prob, self.train_mode) # Add residual connections return self_context + prev_layer.temporal_states def cross_attention_sublayer(self, queries: tf.Tensor) -> tf.Tensor: assert self.cross_attention_sublayer is not None assert self.n_cross_att_heads is not None assert self.input_for_cross_attention is not None encoder_att_states = get_attention_states( self.input_for_cross_attention) encoder_att_mask = get_attention_mask(self.input_for_cross_attention) # Layer normalization normalized_queries = layer_norm(queries) encoder_context, _ = attention( queries=normalized_queries, keys=encoder_att_states, values=encoder_att_states, keys_mask=encoder_att_mask, num_heads=self.n_cross_att_heads, dropout_callback=lambda x: dropout( x, self.attention_dropout_keep_prob, self.train_mode), use_bias=self.use_att_transform_bias) # Apply dropout encoder_context = dropout( encoder_context, self.dropout_keep_prob, self.train_mode) # Add residual connections return encoder_context + queries def feedforward_sublayer(self, layer_input: tf.Tensor) -> tf.Tensor: """Create the feed-forward network sublayer.""" # Layer normalization normalized_input = layer_norm(layer_input) # Feed-forward network hidden layer + ReLU ff_hidden = tf.layers.dense( normalized_input, self.ff_hidden_size, activation=tf.nn.relu, name="hidden_state") # Apply dropout on hidden layer activations ff_hidden = dropout(ff_hidden, self.dropout_keep_prob, self.train_mode) # Feed-forward output projection ff_output = tf.layers.dense( ff_hidden, self.model_dimension, name="output") # Apply dropout on feed-forward output projection ff_output = dropout(ff_output, self.dropout_keep_prob, self.train_mode) # Add residual connections return ff_output + layer_input def layer(self, level: int) -> TransformerLayer: # Recursive implementation. Outputs of the zeroth layer # are normalized inputs. if level == 0: return TransformerLayer(self.encoder_inputs, self.temporal_mask) # Compute the outputs of the previous layer prev_layer = self.layer(level - 1) with tf.variable_scope("layer_{}".format(level - 1)): with tf.variable_scope("self_attention"): self_context = self.self_attention_sublayer(prev_layer) if self.input_for_cross_attention is not None: with tf.variable_scope("cross_attention"): self_context = self.cross_attention_sublayer(self_context) with tf.variable_scope("feedforward"): output_states = self.feedforward_sublayer(self_context) # Layer normalization on the encoder outputs if self.depth == level: output_states = layer_norm(output_states) return TransformerLayer(states=output_states, mask=self.temporal_mask) @tensor def temporal_states(self) -> tf.Tensor: return self.layer(self.depth).temporal_states @tensor def temporal_mask(self) -> tf.Tensor: return self.input_sequence.temporal_mask @property def dependencies(self) -> List[str]: deps = super().dependencies if self.input_for_cross_attention is not None: return deps + ["input_for_cross_attention"] return deps
bsd-3-clause
10628950ed3fdbc2e29225c2c6805a6a
37.960606
79
0.623085
4.123477
false
false
false
false
ufal/neuralmonkey
scripts/caffe_image_features.py
1
4420
import argparse import sys import os os.environ['GLOG_minloglevel'] = '4' sys.path.append("caffe/python") import caffe import numpy as np import skimage def crop_image(x, target_height=227, target_width=227): image = skimage.img_as_float(skimage.io.imread(x)).astype(np.float32) if len(image.shape) == 2: image = np.tile(image[:,:,None], 3) elif len(image.shape) == 4: image = image[:,:,:,0] height, width, rgb = image.shape if width == height: resized_image = skimage.transform.resize(image, (target_height,target_width)) elif height < width: resized_image = skimage.transform.resize(image, (int(width * float(target_height)/height), target_width)) cropping_length = int((resized_image.shape[1] - target_height) / 2) resized_image = resized_image[:,cropping_length:resized_image.shape[1] - cropping_length] else: resized_image = skimage.transform.resize(image, (target_height, int(height * float(target_width) / width))) cropping_length = int((resized_image.shape[0] - target_width) / 2) resized_image = resized_image[cropping_length:resized_image.shape[0] - cropping_length,:] return skimage.transform.resize(resized_image, (target_height, target_width)) class CNN: def __init__(self, deploy, model, mean, batch_size=10, width=227, height=227): self.deploy = deploy self.model = model self.mean = mean self.batch_size = batch_size self.net, self.transformer = self.get_net() self.net.blobs['data'].reshape(self.batch_size, 3, height, width) self.width = width self.height = height def get_net(self): #caffe.set_mode_cpu() net = caffe.Net(self.deploy, self.model, caffe.TEST) transformer = caffe.io.Transformer({'data':net.blobs['data'].data.shape}) transformer.set_transpose('data', (2,0,1)) transformer.set_mean('data', np.load(self.mean).mean(1).mean(1)) transformer.set_raw_scale('data', 255) transformer.set_channel_swap('data', (2,1,0)) return net, transformer def get_features(self, image_list, layers='fc7', layer_sizes=[4096]): iter_until = len(image_list) + self.batch_size all_feats = np.zeros([len(image_list)] + layer_sizes, dtype=np.float32) for start, end in zip(list(range(0, iter_until, self.batch_size)), \ list(range(self.batch_size, iter_until, self.batch_size))): image_batch_file = image_list[start:end] image_batch = np.array([crop_image(x, target_width=self.width, target_height=self.height) for x in image_batch_file]) caffe_in = np.zeros(np.array(image_batch.shape)[[0,3,1,2]], dtype=np.float32) for idx, in_ in enumerate(image_batch): caffe_in[idx] = self.transformer.preprocess('data', in_) out = self.net.forward_all(blobs=[layers], **{'data':caffe_in}) feats = out[layers] all_feats[start:end] = feats return all_feats def shape(string): return [int(s) for s in string.split("x")] if __name__ == "__main__": parser = argparse.ArgumentParser(description="Image feature extraction") parser.add_argument("--model-prototxt", type=str, required=True) parser.add_argument("--model-parameters", type=str, required=True) parser.add_argument("--img-mean", type=str, required=True) parser.add_argument("--feature-layer", type=str, required=True) parser.add_argument("--image-directory", type=str, required=True) parser.add_argument("--image-list", type=argparse.FileType('r'), required=True) parser.add_argument("--output-file", type=argparse.FileType('wb'), required=True) parser.add_argument("--img-shape", type=shape, required=True) parser.add_argument("--output-shape", type=shape, required=True) args = parser.parse_args() cnn = CNN(deploy=args.model_prototxt, model=args.model_parameters, mean=args.img_mean, batch_size=10, width=args.img_shape[0], height=args.img_shape[1]) path_list = [os.path.join(args.image_directory, f.rstrip()) for f in args.image_list] features_shape = [args.output_shape[2]] + args.output_shape[:2] features = cnn.get_features(path_list, layers=args.feature_layer, layer_sizes=features_shape) np.save(args.output_file, features.transpose((0, 2, 3, 1)))
bsd-3-clause
53d1a8cec8541cbc6a43296e708b369c
41.095238
129
0.647285
3.335849
false
false
false
false
ufal/neuralmonkey
scripts/postedit_reconstruct_data.py
2
2259
#!/usr/bin/env python3 """ This a script that takes the result of automatic postediting encoded as a sequence of <keep>, <delete> and insert operations and applies them on the original text being post-edited. The inverse script to this one is 'postedit_prepare_data.py'. """ import argparse from neuralmonkey.processors.german import GermanPreprocessor from neuralmonkey.processors.german import GermanPostprocessor from postedit_prepare_data import load_tokenized # TODO make reconstruct a postprocessor def reconstruct(source, edits): index = 0 target = [] for edit in edits: if edit == '<keep>': if index < len(source): target.append(source[index]) index += 1 elif edit == '<delete>': index += 1 else: target.append(edit) # we may have created a shorter sequence of edit ops due to the # decoder limitations -> now copy the rest of source if index < len(source): target.extend(source[index:]) return target def main(): parser = argparse.ArgumentParser( description="Convert postediting target data to sequence of edits") parser.add_argument("--edits", type=argparse.FileType('r'), required=True) parser.add_argument("--translated-sentences", type=argparse.FileType('r'), required=True) parser.add_argument("--target-german", type=bool, default=False) args = parser.parse_args() postprocess = lambda x: x preprocess = None # type: GermanPreprocessor if args.target_german: # pylint: disable=redefined-variable-type postprocess = GermanPostprocessor() preprocess = GermanPreprocessor() trans_sentences = load_tokenized( args.translated_sentences, preprocess=preprocess) edit_sequences = load_tokenized(args.edits, preprocess=None) for trans, edits in zip(trans_sentences, edit_sequences): target = reconstruct(trans, edits) # TODO refactor this (change postprocessor api) print(" ".join(postprocess([target])[0])) if __name__ == '__main__': # edits = ['<keep>', 'ahoj', '<delete>', 'proc?'] # source = ['Karle', 'co', 'kdy'] # print reconstruct(source, edits) main()
bsd-3-clause
7c005d1b00019bc6044ef2ea54cc6ee4
29.945205
78
0.656042
4.06295
false
false
false
false
ledatelescope/bifrost
python/bifrost/map.py
1
7421
# Copyright (c) 2016-2022, The Bifrost Authors. All rights reserved. # # Redistribution and use in source and binary forms, with or without # modification, are permitted provided that the following conditions # are met: # * Redistributions of source code must retain the above copyright # notice, this list of conditions and the following disclaimer. # * Redistributions in binary form must reproduce the above copyright # notice, this list of conditions and the following disclaimer in the # documentation and/or other materials provided with the distribution. # * Neither the name of The Bifrost Authors nor the names of its # contributors may be used to endorse or promote products derived # from this software without specific prior written permission. # # THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY # EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE # IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR # PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR # CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, # EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, # PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR # PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY # OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT # (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. # Python2 compatibility from __future__ import absolute_import import sys if sys.version_info > (3,): long = int from bifrost.libbifrost import _bf, _check, _array from bifrost.ndarray import asarray import numpy as np import ctypes import glob import os from bifrost.libbifrost_generated import BF_MAP_KERNEL_DISK_CACHE from bifrost import telemetry telemetry.track_module() def _is_literal(x): return isinstance(x, (int, long, float, complex)) def _convert_to_array(arg): if _is_literal(arg): arr = np.array(arg) if isinstance(arg, (int, long)) and -(1 << 31) <= arg < (1 << 31): arr = arr.astype(np.int32) # TODO: Any way to decide when these should be double-precision? elif isinstance(arg, float): arr = arr.astype(np.float32) elif isinstance(arg, complex): arr = arr.astype(np.complex64) arr.flags['WRITEABLE'] = False arg = arr return asarray(arg) def map(func_string, data, axis_names=None, shape=None, func_name=None, extra_code=None, block_shape=None, block_axes=None): """Apply a function to a set of ndarrays. Args: func_string (str): The function to apply to the arrays, as a string (see below for examples). data (dict): Map of string names to ndarrays or scalars. axis_names (list): List of string names by which each axis is referenced in func_string. shape: The shape of the computation. If None, the broadcast shape of all data arrays is used. func_name (str): Name of the function, for debugging purposes. extra_code (str): Additional code to be included at global scope. block_shape: The 2D shape of the thread block (y,x) with which the kernel is launched. This is a performance tuning parameter. If NULL, a heuristic is used to select the block shape. Changes to this parameter do _not_ require re-compilation of the kernel. block_axes: List of axis indices (or names) specifying the 2 computation axes to which the thread block (y,x) is mapped. This is a performance tuning parameter. If NULL, a heuristic is used to select the block axes. Values may be negative for reverse indexing. Changes to this parameter _do_ require re-compilation of the kernel. Note: Only GPU computation is currently supported. Examples:: # Add two arrays together bf.map("c = a + b", {'c': c, 'a': a, 'b': b}) # Compute outer product of two arrays bf.map("c(i,j) = a(i) * b(j)", {'c': c, 'a': a, 'b': b}, axis_names=('i','j')) # Split the components of a complex array bf.map("a = c.real; b = c.imag", {'c': c, 'a': a, 'b': b}) # Raise an array to a scalar power bf.map("c = pow(a, p)", {'c': c, 'a': a, 'p': 2.0}) # Slice an array with a scalar index bf.map("c(i) = a(i,k)", {'c': c, 'a': a, 'k': 7}, ['i'], shape=c.shape) """ try: func_string = func_string.encode() if func_name is not None: func_name = func_name.encode() if extra_code is not None: extra_code = extra_code.encode() except AttributeError: # Python2 catch pass narg = len(data) ndim = len(shape) if shape is not None else 0 arg_arrays = [] args = [] arg_names = [] if block_axes is not None: # Allow referencing axes by name block_axes = [axis_names.index(bax) if isinstance(bax, str) else bax for bax in block_axes] if block_axes is not None and len(block_axes) != 2: raise ValueError("block_axes must contain exactly 2 entries") if block_shape is not None and len(block_shape) != 2: raise ValueError("block_shape must contain exactly 2 entries") for key, arg in data.items(): arg = _convert_to_array(arg) # Note: We must keep a reference to each array lest they be garbage # collected before their corresponding BFarray is used. arg_arrays.append(arg) args.append(arg.as_BFarray()) arg_names.append(key) _check(_bf.bfMap(ndim, _array(shape, dtype=ctypes.c_long), _array(axis_names), narg, _array(args), _array(arg_names), func_name, func_string, extra_code, _array(block_shape), _array(block_axes))) def list_map_cache(): output = "Cache enabled: %s" % ('yes' if BF_MAP_KERNEL_DISK_CACHE else 'no') if BF_MAP_KERNEL_DISK_CACHE: cache_path = os.path.join(os.path.expanduser('~'), '.bifrost', _bf.BF_MAP_KERNEL_DISK_CACHE_SUBDIR) try: with open(os.path.join(cache_path, _bf.BF_MAP_KERNEL_DISK_CACHE_VERSION_FILE), 'r') as fh: version = fh.read() mapcache, runtime, driver = version.split(None, 2) mapcache = int(mapcache, 10) mapcache = "%i.%i" % (mapcache//1000, (mapcache//10) % 1000) runtime = int(runtime, 10) runtime = "%i.%i" % (runtime//1000, (runtime//10) % 1000) driver = int(driver, 10) driver = "%i.%i" % (driver//1000, (driver//10) % 1000) entries = glob.glob(os.path.join(cache_path, '*.inf')) output += "\nCache version: %s (map cache) %s (runtime), %s (driver)" % (mapcache, runtime, driver) output += "\nCache entries: %i" % len(entries) except OSError: pass print(output) def clear_map_cache(): _check(_bf.bfMapClearCache())
bsd-3-clause
0023894c9677eae8af4baa997e0903db
41.164773
111
0.612047
3.883307
false
false
false
false
ledatelescope/bifrost
python/bifrost/addon/leda/make_header.py
1
13543
# Copyright (c) 2016-2021, The Bifrost Authors. All rights reserved. # # Redistribution and use in source and binary forms, with or without # modification, are permitted provided that the following conditions # are met: # * Redistributions of source code must retain the above copyright # notice, this list of conditions and the following disclaimer. # * Redistributions in binary form must reproduce the above copyright # notice, this list of conditions and the following disclaimer in the # documentation and/or other materials provided with the distribution. # * Neither the name of The Bifrost Authors nor the names of its # contributors may be used to endorse or promote products derived # from this software without specific prior written permission. # # THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY # EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE # IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR # PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR # CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, # EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, # PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR # PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY # OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT # (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. """ make_header.py ======= Modified by Hugh Garsden from Danny Price's dada.py and pipeline.py Makes header.txt files that is used by corr2uvfit and DuCT. """ from __future__ import print_function, division import numpy as np import os, sys, ephem, datetime from dateutil import tz from bifrost import telemetry telemetry.track_module() class DadaReader(object): """ Dada file reader for raw LEDA correlator data. Reads the header of a dada file Parameters ---------- filename: str name of dada file to open n_int: int number of integrations to read. If None, will only read header inspectOnly: bool If inspectOnly, will only read header and will not unpack data. """ DEFAULT_HEADER_SIZE = 4096 def __init__(self, filename, warnings, file_size): self.filename = filename self.warnings = warnings self.file_size = file_size # Externally supplied #print(filename, warnings, file_size) self.generate_info() def generate_info(self): """ Parse dada header and form useful quantities. Calculate everything that can be calculated based on what's in the header. For the rest, call them UNKNOWN. """ f = open(self.filename, 'rb') headerstr = f.read(self.DEFAULT_HEADER_SIZE) f.close() header = {} for line in headerstr.split('\n'): try: key, value = line.split() except ValueError: break key = key.strip() value = value.strip() header[key] = value self.source = header["SOURCE"] self.mode = header['MODE'] if "UTC_START" in header: self.datestamp = header['UTC_START'] else: self.datestamp = "UNKNOWN" if "CFREQ" in header: self.c_freq_mhz = float(header['CFREQ']) else: self.c_freq_mhz = "UNKNOWN" if "BW" in header: self.bandwidth_mhz = float(header['BW']) else: self.bandwidth_mhz = "UNKNOWN" if "NCHAN" in header: self.n_chans = int(header["NCHAN"]) else: self.n_chans = "UNKNOWN" if "DATA_ORDER" in header: self.data_order = header["DATA_ORDER"] else: self.data_order = "UNKNOWN" have_size = True # If we can settle on a file size for the zipped files. # Calculate number of integrations within this file # File may not be complete, hence file_size_dsk is read too. # However this is now complicated by zipping files. I am # trying to be clever to figure the size. - HG if self.filename[-8:] == ".dadazip": # Will not unzip to get actual size. Must be specified somehow. if self.file_size: # We are given the complete file size which overrides everything else. data_size_dsk = int(self.file_size)-self.DEFAULT_HEADER_SIZE data_size_hdr = data_size_dsk elif "FILE_SIZE" in header: # Hope that this is right data_size_dsk = int(header["FILE_SIZE"]) # these data sizes don't include header data_size_hdr = data_size_dsk else: # Failure if self.warnings: print("WARNING: File is zipped and FILE_SIZE is not in header and file_size not supplied. ") have_size = False data_size_hdr = data_size_dsk = 0 else: # File not zipped. Can get true complete file size data_size_dsk = os.path.getsize(self.filename)-self.DEFAULT_HEADER_SIZE if "FILE_SIZE" in header: data_size_hdr = int(header["FILE_SIZE"]) else: data_size_hdr = data_size_dsk if data_size_hdr != data_size_dsk: if self.warnings: print("WARNING: Data size in file doesn't match actual size. Using actual size.") data_size = data_size_dsk # Settle on this as the size of the data self.file_size = data_size+self.DEFAULT_HEADER_SIZE # Try to be clever and generate values that can be generated, while leaving # undefined values as UNKNOWN. if "BYTES_PER_AVG" in header: bpa = int(header["BYTES_PER_AVG"]) if "BYTES_PER_AVG" in header and have_size: if data_size % bpa != 0: if self.warnings: print("WARNING: BYTES_PER_AVG does not result in an integral number of scans") if "DATA_ORDER" in header and self.data_order == 'TIME_SUBSET_CHAN_TRIANGULAR_POL_POL_COMPLEX': if self.warnings: print('DATA_ORDER is TIME_SUBSET_CHAN_TRIANGULAR_POL_POL_COMPLEX, resetting BYTES_PER_AVG to',(109*32896*2*2+9*109*1270*2*2)*8,"(fixed)") bpa = (109*32896*2*2+9*109*1270*2*2)*8 if data_size % bpa != 0 and self.warnings: print("WARNING: BYTES_PER_AVG still doesn't give integral number of scans") self.n_int = float(data_size) / bpa else: self.n_int = "UNKNOWN" if "TSAMP" in header and "NAVG" in header: # Calculate integration time per accumulation tsamp = float(header["TSAMP"]) * 1e-6 # Sampling time per channel, in microseconds navg = int(header["NAVG"]) # Number of averages per integration int_tim = tsamp * navg # Integration time is tsamp * navg self.t_int = int_tim if "OBS_OFFSET" in header and "BYTES_PER_AVG" in header: # Calculate the time offset since the observation started byte_offset = int(header["OBS_OFFSET"]) num_int_since_obs_start = byte_offset // bpa time_offset_since_obs_start = num_int_since_obs_start * int_tim self.t_offset = time_offset_since_obs_start else: self.t_offset = "UNKNOWN" else: self.t_int = "UNKNOWN" self.t_offset = "UNKNOWN" class DadaTimes(object): """ Handle the generation of true times and RA/DEC for the observation in the DADA file. Use pyephem for the tricky stuff. Includes the new calculation of RA/DEC in terms of long/lat rather than just using long/lat. """ def time_at_timezone(self, dt, zone): from_zone = tz.gettz('UTC') to_zone = tz.gettz(zone) # Tell the datetime object that it's in UTC time zone since # datetime objects are 'naive' by default dt = dt.replace(tzinfo=from_zone) # Convert time zone return dt.astimezone(to_zone) def __init__(self, header): ovro = ephem.Observer() (ovro.lat, ovro.lon, ovro.elev) = ('37.23978', '-118.281667', 1184.120) if header.datestamp == "UNKNOWN" or header.t_offset == "UNKNOWN": self.lst = "UNKNOWN" self.date_str = "UNKNOWN" self.time_str = "UNKNOWN" self.localtime_str = "UNKNOWN" self.lst_str = "UNKNOWN" self.dec_str = "UNKNOWN" return # Calculate times including LST dt = datetime.datetime.strptime(header.datestamp, "%Y-%m-%d-%H:%M:%S")+datetime.timedelta(seconds=header.t_offset) ovro.date = dt self.lst = ovro.sidereal_time() localt = self.time_at_timezone(dt, "America/Los_Angeles") self.date_str = "%04d%02d%02d"%(dt.year,dt.month,dt.day) self.time_str = "%02d%02d%02d"%(dt.hour,dt.minute,dt.second) self.localtime_str = "%02d%02d%02d"%(localt.hour,localt.minute,localt.second) ra, dec = ovro.radec_of(0, np.pi/2) self.lst_str = str(float(ra) / 2 / np.pi * 24) self.dec_str = str(float(repr(dec))*180/np.pi) #print("UTC START: %s"%dada_file.datestamp) #print("TIME OFFSET: %s"%datetime.timedelta(seconds=dada_file.t_offset)) #print("NEW START: (%s, %s)"%(date_str, time_str)) def make_header(filename, write=True, warn=True, size=None): """ Create useful/necessary information about an observation. Used by other programs like corr2uvfits and DuCT. filename: DADA file, can be zipped warn: print warnings write: write a header.txt files size: specify a true file size in case of zipped file """ # Get information from the DADA file dada_file = DadaReader(filename, warn, size) dada_times = DadaTimes(dada_file) # Fill and either dump or return header. Slight differences depending on which. header_params = {'N_CHANS' : dada_file.n_chans, 'N_SCANS' : dada_file.n_int, 'INT_TIME' : dada_file.t_int, 'FREQCENT' : dada_file.c_freq_mhz, 'BANDWIDTH' : dada_file.bandwidth_mhz, 'RA_HRS' : dada_times.lst_str, 'DEC_DEGS' : dada_times.dec_str, 'DATE' : dada_times.date_str, 'TIME' : dada_times.time_str, 'LOCALTIME' : dada_times.localtime_str, 'LST' : dada_times.lst_str, 'DATA_ORDER' : dada_file.data_order, 'FILE_SIZE' : dada_file.file_size, 'MODE' : dada_file.mode, 'TIME_OFFSET': dada_file.t_offset, 'SOURCE' : dada_file.source } if header_params["N_SCANS"] == "UNKNOWN": n_scans = "UNKNOWN" else: n_scans = str(int(header_params['N_SCANS'])) if write: # This format is used by corr2uvfits and DuCT for transforming a DADA file. output = open("header.txt","w") output.write("# Generated by make_header.py\n\n") output.write("FIELDNAME Zenith\n") output.write("N_SCANS "+n_scans+"\n") output.write("N_INPUTS 512\n") output.write("N_CHANS "+str(header_params['N_CHANS'])+" # number of channels in spectrum\n") output.write("CORRTYPE B # correlation type to use. 'C'(cross), 'B'(both), or 'A'(auto)\n") output.write("INT_TIME "+str(header_params['INT_TIME'])+" # integration time of scan in seconds\n") output.write("FREQCENT "+str(header_params['FREQCENT'])+" # observing center freq in MHz\n") output.write("BANDWIDTH "+str(header_params['BANDWIDTH'])+" # total bandwidth in MHz\n") output.write("# To phase to the zenith, these must be the HA, RA and Dec of the zenith.\n") output.write("HA_HRS 0.000000 # the RA of the desired phase centre (hours)\n") output.write("RA_HRS "+header_params['RA_HRS']+" # the RA of the desired phase centre (hours)\n") output.write("DEC_DEGS "+str(header_params['DEC_DEGS'])+" # the DEC of the desired phase centre (degs)\n") output.write("DATE "+header_params['DATE']+" # YYYYMMDD\n") output.write("TIME "+header_params['TIME']+" # HHMMSS\n") output.write("LOCALTIME "+str(dada_times.localtime_str)+"\n") output.write("LST "+str(dada_times.lst)+"\n") output.write("INVERT_FREQ 0 # 1 if the freq decreases with channel number\n") output.write("CONJUGATE 1 # conjugate the raw data to fix sign convention problem if necessary\n") output.write("GEOM_CORRECT 0\n") output.close() return header_params # If this function is called from other scripts (e.g. plot scripts) it can supply useful information if __name__ == "__main__": if len(sys.argv) == 2: make_header(sys.argv[1]) elif len(sys.argv) == 3: make_header(sys.argv[1],size=sys.argv[2]) else: print("Expecting file name and optionally file size")
bsd-3-clause
3a070df3706248d3d106a0e8a7edac6c
43.549342
153
0.594994
3.697243
false
false
false
false
ledatelescope/bifrost
python/bifrost/blocks/convert_visibilities.py
1
9672
# Copyright (c) 2016-2021, The Bifrost Authors. All rights reserved. # # Redistribution and use in source and binary forms, with or without # modification, are permitted provided that the following conditions # are met: # * Redistributions of source code must retain the above copyright # notice, this list of conditions and the following disclaimer. # * Redistributions in binary form must reproduce the above copyright # notice, this list of conditions and the following disclaimer in the # documentation and/or other materials provided with the distribution. # * Neither the name of The Bifrost Authors nor the names of its # contributors may be used to endorse or promote products derived # from this software without specific prior written permission. # # THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY # EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE # IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR # PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR # CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, # EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, # PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR # PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY # OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT # (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. from __future__ import absolute_import from bifrost.map import map as bf_map from bifrost.pipeline import TransformBlock from bifrost.DataType import DataType from copy import deepcopy from math import sqrt from bifrost import telemetry telemetry.track_module() class ConvertVisibilitiesBlock(TransformBlock): def __init__(self, iring, fmt, *args, **kwargs): super(ConvertVisibilitiesBlock, self).__init__(iring, *args, **kwargs) self.ofmt = fmt def define_valid_input_spaces(self): return ('cuda',) def on_sequence(self, iseq): ihdr = iseq.header itensor = ihdr['_tensor'] ilabels = itensor['labels'] assert(ilabels[0] == 'time') ohdr = deepcopy(ihdr) otensor = ohdr['_tensor'] if ilabels[1:] == ['freq', 'station_i', 'pol_i', 'station_j', 'pol_j']: nchan, nstand, npol, nstand_j, npol_j = itensor['shape'][1:] assert(nstand_j == nstand) assert( npol_j == npol) self.ifmt = 'matrix' if self.ofmt == 'matrix': ohdr['matrix_fill_mode'] = 'hermitian' elif self.ofmt == 'storage': nbaseline = nstand*(nstand+1)//2 del ohdr['matrix_fill_mode'] otensor['labels'] = ['time', 'baseline', 'freq', 'stokes'] otensor['shape'] = [-1, nbaseline, nchan, npol*npol] time_units, freq_units, stand_units, pol_units, _, _ = itensor['units'] otensor['units'] = [time_units, None, freq_units, ('I', 'Q', 'U', 'V')] else: raise NotImplementedError("Unsupported conversion from " + self.ifmt + " to " + self.ofmt) elif ilabels[1:] == ['baseline', 'freq', 'stokes']: nbaseline, nchan, nstokes = itensor['shape'][1:] assert(nstokes == 1 or nstokes == 4) npol = 1 if nstokes == 1 else 2 nstand = int(sqrt(8 * nbaseline + 1) - 1) // 2 time_units, baseline_units, freq_units, stokes_units, = itensor['units'] pol_units = ('X', 'Y') # TODO: Support L/R (using additional metadata?) self.ifmt = 'storage' if self.ofmt == 'matrix': otensor['labels'] = ['time', 'freq', 'station_i', 'pol_i', 'station_j', 'pol_j'] otensor['shape'] = [-1, nchan, nstand, npol, nstand, npol] otensor['units'] = [time_units, freq_units, None, pol_units, None, pol_units] else: raise NotImplementedError("Cannot convert input from %s to %s" % (ilabels, self.ofmt)) return ohdr def on_data(self, ispan, ospan): idata = ispan.data odata = ospan.data itype = DataType(idata.dtype) otype = DataType(odata.dtype) if self.ifmt == 'matrix' and self.ofmt == 'matrix': # Make a full-matrix copy of the lower-only input matrix # odata[t,c,i,p,j,q] = idata[t,c,i,p,j,q] (lower filled only) shape_nopols = list(idata.shape) del shape_nopols[5] del shape_nopols[3] idata = idata.view(itype.as_vector(2)) odata = odata.view(otype.as_vector(2)) bf_map( ''' bool in_lower_triangle = (i > j); if( in_lower_triangle ) { odata(t,c,i,0,j,0) = idata(t,c,i,0,j,0); odata(t,c,i,1,j,0) = idata(t,c,i,1,j,0); } else { auto x = idata(t,c,j,0,i,0); auto y = idata(t,c,j,1,i,0); auto x1 = x[1]; x[0] = x[0].conj(); x[1] = y[0].conj(); if( i != j ) { y[0] = x1.conj(); } y[1] = y[1].conj(); odata(t,c,i,0,j,0) = x; odata(t,c,i,1,j,0) = y; } ''', shape=shape_nopols, axis_names=['t', 'c', 'i', 'j'], data={'idata': idata, 'odata': odata}) elif self.ifmt == 'matrix' and self.ofmt == 'storage': assert(idata.shape[2] <= 2048) idata = idata.view(itype.as_vector(2)) odata = odata.view(otype.as_vector(4)) # TODO: Support L/R as well as X/Y pols bf_map(''' // TODO: This only works up to 2048 in single-precision #define project_triangular(i, j) ((i)*((i)+1)/2 + (j)) int i = int((sqrt(8.f*(b)+1)-1)/2); int j = b - project_triangular(i, 0); auto x = idata(t,c,i,0,j,0); auto y = idata(t,c,i,1,j,0); if( i == j ) { x[1] = y[0].conj(); } idata_type::value_type eye(0, 1); auto I = (x[0] + y[1]); auto Q = (x[0] - y[1]); auto U = (x[1] + y[0]); auto V = (x[1] - y[0]) * eye; odata(t,b,c,0) = odata_type(I,Q,U,V); ''', shape=odata.shape[:-1], axis_names=['t', 'b', 'c'], data={'idata': idata, 'odata': odata}, block_shape=[64,8]) # TODO: Tune this #elif self.ifmt == 'matrix' and self.ofmt == 'triangular': elif self.ifmt == 'storage' and self.ofmt == 'matrix': oshape_nopols = list(odata.shape) del oshape_nopols[5] del oshape_nopols[3] idata = idata.view(itype.as_vector(4)) odata = odata.view(otype.as_vector(2)) bf_map(''' bool in_upper_triangle = (i < j); auto b = in_upper_triangle ? j*(j+1)/2 + i : i*(i+1)/2 + j; auto IQUV = idata(t,b,c,0); auto I = IQUV[0], Q = IQUV[1], U = IQUV[2], V = IQUV[3]; idata_type::value_type eye(0, 1); auto xx = 0.5f*(I + Q); auto xy = 0.5f*(U - V*eye); auto yx = 0.5f*(U + V*eye); auto yy = 0.5f*(I - Q); if( i == j ) { xy = yx.conj(); } if( in_upper_triangle ) { auto tmp_xy = xy; xx = xx.conj(); xy = yx.conj(); yx = tmp_xy.conj(); yy = yy.conj(); } odata(t,c,i,0,j,0) = odata_type(xx, xy); odata(t,c,i,1,j,0) = odata_type(yx, yy); ''', shape=oshape_nopols, axis_names=['t', 'c', 'i', 'j'], data={'idata': idata, 'odata': odata}, block_shape=[64,8]) # TODO: Tune this else: raise NotImplementedError def convert_visibilities(iring, fmt, *args, **kwargs): """Convert visibility data to a new format. Supported values of 'fmt' are: matrix, storage Args: iring (Ring or Block): Input data source. fmt (str): The desired output format: matrix, storage. *args: Arguments to ``bifrost.pipeline.TransformBlock``. **kwargs: Keyword Arguments to ``bifrost.pipeline.TransformBlock``. **Tensor semantics**:: Input: ['time', 'freq', 'station_i', 'pol_i', 'station_j', 'pol_j'], dtype = any complex, space = CUDA fmt = 'matrix' (produces a fully-filled matrix from a lower-filled one) Output: ['time', 'freq', 'station_i', 'pol_i', 'station_j', 'pol_j'], dtype = any complex, space = CUDA fmt = 'storage' (suitable for common on-disk data formats such as UVFITS, FITS-IDI, MS etc.) Output: ['time', 'baseline', 'freq', 'stokes'], dtype = any complex, space = CUDA Input: ['time', 'baseline', 'freq', 'stokes'], dtype = any complex, space = CUDA fmt = 'matrix' (fully-filled matrix suitable for linear algebra operations) Output: ['time', 'freq', 'station_i', 'pol_i', 'station_j', 'pol_j'], dtype = any complex, space = CUDA Returns: ConvertVisibilitiesBlock: A new block instance. """ return ConvertVisibilitiesBlock(iring, fmt, *args, **kwargs)
bsd-3-clause
ede8708991dbfdc068d82e8fe33153d7
44.838863
111
0.531638
3.475386
false
false
false
false
ledatelescope/bifrost
python/bifrost/psrdada.py
1
9835
# Copyright (c) 2016-2021, The Bifrost Authors. All rights reserved. # # Redistribution and use in source and binary forms, with or without # modification, are permitted provided that the following conditions # are met: # * Redistributions of source code must retain the above copyright # notice, this list of conditions and the following disclaimer. # * Redistributions in binary form must reproduce the above copyright # notice, this list of conditions and the following disclaimer in the # documentation and/or other materials provided with the distribution. # * Neither the name of The Bifrost Authors nor the names of its # contributors may be used to endorse or promote products derived # from this software without specific prior written permission. # # THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY # EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE # IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR # PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR # CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, # EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, # PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR # PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY # OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT # (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. """ This provides an interface for reading/writing PSRDADA ring buffers. bifrost.libpsrdada_generated is generated at build time using ctypesgen.py PSRDADA must be built as a shared library to use this. This can be accomplished by adding the following lines to psrdada/configure.in: #AC_DISABLE_SHARED LT_INIT lib_LTLIBRARIES = libpsrdada.la libtest_la_LDFLAGS = -version-info 0:0:0 """ from __future__ import absolute_import, print_function import bifrost.libpsrdada_generated as _dada import numpy as np from bifrost.ndarray import _address_as_buffer from bifrost.libbifrost import EndOfDataStop import ctypes from bifrost import telemetry telemetry.track_module() def get_pointer_value(ptr): return ctypes.c_void_p.from_buffer(ptr).value class MultiLog(object): count = 0 def __init__(self, name=None): if name is None: name = "MultiLog%i" % MultiLog.count MultiLog.count += 1 self.obj = _dada.multilog_open(name, '\0') def __del__(self): _dada.multilog_close(self.obj) class IpcBufBlock(object): def __init__(self, buf, mutable=False): self.buf = buf self.ptr, self.nbyte, self.block_id = self.buf.open() self.nbyte_commit = self.nbyte self.ptr = get_pointer_value(self.ptr) if self.ptr is not None: self.data = np.ndarray( shape=(self.nbyte,), buffer=_address_as_buffer(self.ptr, self.nbyte), dtype=np.uint8) self.data.flags['WRITEABLE'] = mutable def __del__(self): self.close() def commit(self, nbyte=None): if nbyte is None: nbyte = self.nbyte self.nbyte_commit = nbyte def close(self): if self.ptr is not None: self.buf.close(self.nbyte_commit) self.ptr = None def enable_eod(self): #print('>ipcbuf_enable_eod') if _dada.ipcbuf_enable_eod(self.buf.buf) < 0: raise IOError("Failed to enable EOD flag") def size_bytes(self): return self.nbyte def __enter__(self): return self def __exit__(self, type, value, tb): self.close() class IpcBaseBuf(object): def __init__(self, ipcbuf, mutable=False): self.buf = ipcbuf self.mutable = mutable def size_bytes(self): return _dada.ipcbuf_get_bufsz(self.buf) def eod(self): #print('>ipcbuf_eod') return bool(_dada.ipcbuf_eod(self.buf)) def reset(self): #print('>ipcbuf_reset') if _dada.ipcbuf_reset(self.buf) < 0: raise IOError("Failed to reset buffer") def __iter__(self): return self def __next__(self): block = IpcBufBlock(self, self.mutable) if block.nbyte > 0: return block else: del block self.reset() raise EndOfDataStop('IpcBufBlock empty') def next(self): return self.__next__() def open(self): raise NotImplementedError() def close(self, nbyte): raise NotImplementedError() class IpcBaseIO(IpcBaseBuf): def __init__(self, ipcio, mutable=False): ipcbuf = ctypes.cast(ipcio, ctypes.POINTER(_dada.ipcbuf_t)) super(IpcBaseIO, self).__init__(ipcbuf, mutable) self.io = ipcio def stop(self): #print('>ipcio_stop') if _dada.ipcio_stop(self.io) < 0: raise IOError("Failed to write EOD marker to block") class IpcReadHeaderBuf(IpcBaseBuf): def __init__(self, ipcbuf): super(IpcReadHeaderBuf, self).__init__(ipcbuf) def open(self): nbyte = ctypes.c_uint64() #print('>ipcbuf_get_next_read') ptr = _dada.ipcbuf_get_next_read(self.buf, nbyte) nbyte = nbyte.value block_id = 0 return ptr, nbyte, block_id def close(self, nbyte): #print('>ipcbuf_mark_cleared') if _dada.ipcbuf_mark_cleared(self.buf) < 0: raise IOError("Failed to mark block as cleared") class IpcWriteHeaderBuf(IpcBaseBuf): def __init__(self, ipcbuf): super(IpcWriteHeaderBuf, self).__init__(ipcbuf, mutable=True) def open(self): nbyte = self.size_bytes() #print('>ipcbuf_get_next_write') ptr = _dada.ipcbuf_get_next_write(self.buf) block_id = 0 return ptr, nbyte, block_id def close(self, nbyte): #print('>ipcbuf_mark_filled') if _dada.ipcbuf_mark_filled(self.buf, nbyte) < 0: raise IOError("Failed to mark block as filled") class IpcReadDataBuf(IpcBaseIO): def __init__(self, ipcio): super(IpcReadDataBuf, self).__init__(ipcio) def open(self): nbyte = ctypes.c_uint64() block_id = ctypes.c_uint64() #print('>ipcio_open_block_read') ptr = _dada.ipcio_open_block_read(self.io, nbyte, block_id) nbyte = nbyte.value block_id = block_id.value #print('block_id =', block_id) return ptr, nbyte, block_id def close(self, nbyte): #print('>ipcio_close_block_read(nbyte=%i)' % nbyte) if _dada.ipcio_close_block_read(self.io, nbyte) < 0: raise IOError("Failed to close block for reading") class IpcWriteDataBuf(IpcBaseIO): def __init__(self, ipcio): super(IpcWriteDataBuf, self).__init__(ipcio, mutable=True) self.nbyte_commit = 0 # Default to committing nothing def open(self): nbyte = self.size_bytes() block_id = ctypes.c_uint64() #print('>ipcio_open_block_write') ptr = _dada.ipcio_open_block_write(self.io, block_id) block_id = block_id.value #print('block_id =', block_id) return ptr, nbyte, block_id def close(self, nbyte): #print('>ipcio_close_block_write(nbyte=%i)' % nbyte) if _dada.ipcio_close_block_write(self.io, nbyte) < 0: raise IOError("Failed to close block for writing") class Hdu(object): def __init__(self): self._dada = _dada self.log = MultiLog() self.hdu = _dada.dada_hdu_create(self.log.obj) self.connected = False def __del__(self): self.disconnect() _dada.dada_hdu_destroy(self.hdu) def _connect(self, buffer_key=0xDADA): self.buffer_key = buffer_key _dada.dada_hdu_set_key(self.hdu, self.buffer_key) if _dada.dada_hdu_connect(self.hdu) < 0: raise IOError("Could not connect to buffer '%x'" % self.buffer_key) def _disconnect(self): if _dada.dada_hdu_disconnect(self.hdu) < 0: raise IOError("Could not disconnect from buffer '%x'" % self.buffer_key) def _lock(self, mode): self.mode = mode if mode == 'read': if _dada.dada_hdu_lock_read(self.hdu) < 0: raise IOError("Could not lock buffer '%x' for reading" % self.buffer_key) else: if _dada.dada_hdu_lock_write(self.hdu) < 0: raise IOError("Could not lock buffer '%x' for writing" % self.buffer_key) def _unlock(self): if self.mode == 'read': if _dada.dada_hdu_unlock_read(self.hdu) < 0: raise IOError("Could not unlock buffer '%x' for reading" % self.buffer_key) else: if _dada.dada_hdu_unlock_write(self.hdu) < 0: raise IOError("Could not unlock buffer '%x' for writing" % self.buffer_key) def relock(self): self._unlock() self._lock(self.mode) def open_HACK(self): if _dada.ipcio_open(self.data_block.io, 'w') < 0: raise IOError("ipcio_open failed") def connect_read(self, buffer_key=0xDADA): self._connect(buffer_key) self._lock('read') self.header_block = IpcReadHeaderBuf(self.hdu.contents.header_block) self.data_block = IpcReadDataBuf(self.hdu.contents.data_block) self.connected = True def connect_write(self, buffer_key=0xDADA): self._connect(buffer_key) self._lock('write') self.header_block = IpcWriteHeaderBuf(self.hdu.contents.header_block) self.data_block = IpcWriteDataBuf(self.hdu.contents.data_block) self.connected = True def disconnect(self): if self.connected: self._unlock() self._disconnect() self.connected = False
bsd-3-clause
68d36c1c8878d30d112c3c53f4827197
37.268482
91
0.629181
3.449667
false
false
false
false
ledatelescope/bifrost
python/bifrost/blocks/serialize.py
1
11508
# Copyright (c) 2016-2020, The Bifrost Authors. All rights reserved. # # Redistribution and use in source and binary forms, with or without # modification, are permitted provided that the following conditions # are met: # * Redistributions of source code must retain the above copyright # notice, this list of conditions and the following disclaimer. # * Redistributions in binary form must reproduce the above copyright # notice, this list of conditions and the following disclaimer in the # documentation and/or other materials provided with the distribution. # * Neither the name of The Bifrost Authors nor the names of its # contributors may be used to endorse or promote products derived # from this software without specific prior written permission. # # THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY # EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE # IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR # PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR # CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, # EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, # PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR # PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY # OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT # (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. # Python2 compatibility from __future__ import absolute_import, print_function import sys if sys.version_info < (3,): range = xrange from bifrost.pipeline import SinkBlock, SourceBlock import os import warnings try: import simplejson as json except ImportError: warnings.warn("Install simplejson for better performance", RuntimeWarning) import json import glob from functools import reduce from bifrost import telemetry telemetry.track_module() def _parse_bifrost_filename(fname): inds = fname[fname.find('.bf.') + 4:].split('.')[:-1] inds = [int(i) for i in inds] frame0, ringlet_inds = inds[0], inds[1:] return frame0, ringlet_inds class BifrostReader(object): def __init__(self, basename): assert(basename.endswith('.bf')) hdr_filename = basename + '.json' with open(hdr_filename, 'r') as hdr_file: self.header = json.load(hdr_file) data_filenames = glob.glob(basename + '.*.dat') inds = [_parse_bifrost_filename(fname) for fname in data_filenames] frame0s, ringlet_inds = zip(*inds) nringlets = [max(r) + 1 for r in zip(*ringlet_inds)] # TODO: Support multiple ringlet axes (needed in SerializeBlock too) assert(len(nringlets) <= 1) self.nringlet = nringlets[0] if len(nringlets) else 0 if self.nringlet > 0: ringlet_inds = [inds[0] for inds in ringlet_inds] self.ringlet_files = [] for ringlet in range(self.nringlet): ringlet_filenames = [f for f, r in zip(data_filenames, ringlet_inds) if r == ringlet] ringlet_filenames.sort() ringlet_files = [open(f, 'rb') for f in ringlet_filenames] self.ringlet_files.append(ringlet_files) self.nfile = len(self.ringlet_files[0]) if not all([len(files) == self.nfile for files in self.ringlet_files]): raise IOError("Number of files in each ringlet does not match") else: data_filenames.sort() self.files = [open(f, 'rb') for f in data_filenames] self.nfile = len(self.files) self.cur_file = 0 def __enter__(self): return self def __exit__(self, type, value, tb): if self.nringlet > 0: for ringlet in self.ringlet_files: for f in ringlet: f.close() else: for f in self.files: f.close() def readinto(self, buf, frame_nbyte): if self.cur_file == self.nfile: return 0 nframe_read = 0 if self.nringlet > 0: # First dimension of buf is ringlets bufs = buf nbyte_reads = [ringlet_file[self.cur_file].readinto(buf) for ringlet_file, buf in zip(self.ringlet_files, bufs)] nbyte_read = min(nbyte_reads) else: nbyte_read = self.files[self.cur_file].readinto(buf) if nbyte_read % frame_nbyte != 0: raise IOError("Unexpected end of file") nframe_read += nbyte_read // frame_nbyte while nbyte_read < buf[0].nbytes: self.cur_file += 1 if self.cur_file == self.nfile: break if self.nringlet > 0: nbyte_reads = [ringlet_file[self.cur_file].readinto(buf) for ringlet_file, buf in zip(self.ringlet_files, bufs)] nbyte_read = min(nbyte_reads) else: nbyte_read = self.files[self.cur_file].readinto(buf) if nbyte_read % frame_nbyte != 0: raise IOError("Unexpected end of file") nframe_read += nbyte_read // frame_nbyte return nframe_read class DeserializeBlock(SourceBlock): def __init__(self, filenames, gulp_nframe, *args, **kwargs): super(DeserializeBlock, self).__init__(filenames, gulp_nframe, *args, **kwargs) def create_reader(self, sourcename): return BifrostReader(sourcename) def on_sequence(self, ireader, sourcename): return [ireader.header] def on_data(self, reader, ospans): ospan = ospans[0] return [reader.readinto(ospan.data, ospan.frame_nbyte)] def deserialize(filenames, gulp_nframe, *args, **kwargs): """Deserializes a data stream from a set of files using a simple data format Sequence headers are read as JSON files, and sequence data are read directly as binary from separate files. The actual header and data files must have the following general form:: # Header <filename>.json # Single-ringlet data <filename>.<frame_offset>.dat # Multi-ringlet data <filename>.<frame_offset>.<ringlet>.dat See also: ``serialize`` Args: filenames (list): List of input filenames (each ending with '.bf') gulp_nframe (int): No. frames to read at a time. *args: Arguments to ``bifrost.pipeline.SourceBlock``. **kwargs: Keyword Arguments to ``bifrost.pipeline.SourceBlock``. **Tensor semantics**:: Input: One data file per sequence Output: [frame, ...], dtype = any, space = SYSTEM Input: One data file per ringlet Output: [ringlet, frame, ...], dtype = any, space = SYSTEM Returns: DeserializeBlock: A new block instance. """ return DeserializeBlock(filenames, gulp_nframe, *args, **kwargs) # **TODO: Write a DeserializeBlock that does the inverse of this class SerializeBlock(SinkBlock): def __init__(self, iring, path, max_file_size=None, *args, **kwargs): super(SerializeBlock, self).__init__(iring, *args, **kwargs) if path is None: path = '' self.path = path if max_file_size is None: max_file_size = 1024**3 self.max_file_size = max_file_size def _close_data_files(self): if hasattr(self, 'ofiles'): for ofile in self.ofiles: ofile.close() def _open_new_data_files(self, frame_offset): self._close_data_files() self.bytes_written = 0 if self.frame_axis == 0: # No ringlets, we can write all data to one file filenames = [self.basename + '.bf.%012i.dat' % frame_offset] elif self.frame_axis == 1: # Ringlets, we must write each to a separate file ndigit = len(str(self.nringlet-1)) filenames = [self.basename + ('.bf.%012i.%0'+str(ndigit)+'i.dat') % (frame_offset, i) for i in range(self.nringlet)] else: # TODO: Need to deal with separating multiple ringlet axes # E.g., separate each ringlet dim with a dot # Will have to lift/project the indices raise NotImplementedError("Multiple ringlet axes not supported") # Open data files self.ofiles = [open(fname, 'wb') for fname in filenames] def on_sequence(self, iseq): hdr = iseq.header tensor = hdr['_tensor'] if hdr['name'] != '': self.basename = hdr['name'] else: self.basename = '%020i' % hdr['time_tag'] if self.path != '': # TODO: May need more flexibility in path handling # E.g., may want to keep subdirs from original name self.basename = os.path.basename(self.basename) self.basename = os.path.join(self.path, self.basename) # Write sequence header file with open(self.basename + '.bf.json', 'w') as hdr_file: hdr_file.write(json.dumps(hdr, indent=4, sort_keys=True)) shape = tensor['shape'] self.frame_axis = shape.index(-1) self.nringlet = reduce(lambda a, b: a * b, shape[:self.frame_axis], 1) self._open_new_data_files(frame_offset=0) def on_sequence_end(self, iseq): self._close_data_files() def on_data(self, ispan): if self.nringlet == 1: bytes_to_write = ispan.data.nbytes else: bytes_to_write = ispan.data[0].nbytes # Check if file size limit has been reached if self.bytes_written + bytes_to_write > self.max_file_size: self._open_new_data_files(ispan.frame_offset) self.bytes_written += bytes_to_write # Write data to file(s) if self.nringlet == 1: ispan.data.tofile(self.ofiles[0]) else: for r in range(self.nringlet): ispan.data[r].tofile(self.ofiles[r]) def serialize(iring, path=None, max_file_size=None, *args, **kwargs): """Serializes a data stream to a set of files using a simple data format Sequence headers are written as JSON files, and sequence data are written directly as binary to separate files. Filenames begin with the sequence name if present, or the time tag if not. The general form is:: # Header <name_or_time_tag>.bf.json # Single-ringlet data <name_or_time_tag>.bf.<frame_offset>.dat # Multi-ringlet data <name_or_time_tag>.bf.<frame_offset>.<ringlet>.dat Args: iring (Ring or Block): Input data source.o path (str): Path specifying where to write output files. max_file_size (int): Max no. bytes to write to a single file. If set to -1, no limit is applied. *args: Arguments to ``bifrost.pipeline.SinkBlock``. **kwargs: Keyword Arguments to ``bifrost.pipeline.SinkBlock``. **Tensor semantics**:: Input: [frame, ...], dtype = any, space = SYSTEM Output: One data file per sequence Input: [ringlet, frame, ...], dtype = any, space = SYSTEM Output: One data file per ringlet Returns: SerializeBlock: A new block instance. """ return SerializeBlock(iring, path, max_file_size, *args, **kwargs)
bsd-3-clause
cbd626e3315ab52ba518f65e2196d0af
40.1
87
0.614095
3.774352
false
false
false
false
ledatelescope/bifrost
test/test_ndarray.py
1
7630
# Copyright (c) 2016-2022, The Bifrost Authors. All rights reserved. # # Redistribution and use in source and binary forms, with or without # modification, are permitted provided that the following conditions # are met: # * Redistributions of source code must retain the above copyright # notice, this list of conditions and the following disclaimer. # * Redistributions in binary form must reproduce the above copyright # notice, this list of conditions and the following disclaimer in the # documentation and/or other materials provided with the distribution. # * Neither the name of The Bifrost Authors nor the names of its # contributors may be used to endorse or promote products derived # from this software without specific prior written permission. # # THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY # EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE # IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR # PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR # CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, # EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, # PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR # PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY # OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT # (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. import unittest import numpy as np import bifrost as bf import ctypes from bifrost.libbifrost_generated import BF_CUDA_ENABLED class NDArrayTest(unittest.TestCase): def setUp(self): self.known_vals = [[0,1],[2,3],[4,5]] self.known_array = np.array(self.known_vals, dtype=np.float32) def test_construct(self): a = bf.ndarray(self.known_vals, dtype='f32') np.testing.assert_equal(a, self.known_array) def test_assign(self): b = bf.ndarray(shape=(3,2), dtype='f32') b[...] = self.known_array np.testing.assert_equal(b, self.known_array) @unittest.skipUnless(BF_CUDA_ENABLED, "requires GPU support") def test_space_copy(self): c = bf.ndarray(self.known_vals, dtype='f32') c = c.copy(space='cuda').copy(space='cuda_host').copy(space='system') np.testing.assert_equal(c, self.known_array) def run_contiguous_copy(self, space='system'): a = np.random.rand(2,3,4,5) a = a.astype(np.float64) b = a.transpose(0,3,2,1).copy() c = bf.zeros(a.shape, dtype=a.dtype, space='system') c[...] = a c = c.copy(space=space) d = c.transpose(0,3,2,1).copy(space='system') # Use ctypes to directly access the memory b_data = ctypes.cast(b.ctypes.data, ctypes.POINTER(ctypes.c_double)) b_data = np.array([b_data[i] for i in range(b.size)]) d_data = ctypes.cast(d.ctypes.data, ctypes.POINTER(ctypes.c_double)) d_data = np.array([d_data[i] for i in range(d.size)]) np.testing.assert_equal(d_data, b_data) def test_contiguous_copy(self): self.run_contiguous_copy() @unittest.skipUnless(BF_CUDA_ENABLED, "requires GPU support") def test_space_contiguous_copy(self): self.run_contiguous_copy(space='cuda') def run_slice_copy(self, space='system'): a = np.random.rand(2,3,4,5) a = a.astype(np.float64) b = a[:,1:,:,:].copy() c = bf.zeros(a.shape, dtype=a.dtype, space='system') c[...] = a c = c.copy(space=space) d = c[:,1:,:,:].copy(space='system') # Use ctypes to directly access the memory b_data = ctypes.cast(b.ctypes.data, ctypes.POINTER(ctypes.c_double)) b_data = np.array([b_data[i] for i in range(b.size)]) d_data = ctypes.cast(d.ctypes.data, ctypes.POINTER(ctypes.c_double)) d_data = np.array([d_data[i] for i in range(d.size)]) np.testing.assert_equal(d_data, b_data) def test_slice_copy(self): self.run_slice_copy() @unittest.skipUnless(BF_CUDA_ENABLED, "requires GPU support") def test_space_slice_copy(self): self.run_slice_copy(space='cuda') def run_contiguous_slice_copy(self, space='system'): a = np.random.rand(2,3,4,5) a = a.astype(np.float64) b = a.transpose(0,3,2,1)[:,1:,:,:].copy() c = bf.zeros(a.shape, dtype=a.dtype, space='system') c[...] = a c = c.copy(space=space) d = c.transpose(0,3,2,1)[:,1:,:,:].copy(space='system') # Use ctypes to directly access the memory b_data = ctypes.cast(b.ctypes.data, ctypes.POINTER(ctypes.c_double)) b_data = np.array([b_data[i] for i in range(b.size)]) d_data = ctypes.cast(d.ctypes.data, ctypes.POINTER(ctypes.c_double)) d_data = np.array([d_data[i] for i in range(d.size)]) np.testing.assert_equal(d_data, b_data) def test_contiguous_slice_copy(self): self.run_contiguous_slice_copy() @unittest.skipUnless(BF_CUDA_ENABLED, "requires GPU support") def test_space_contiguous_slice_copy(self): self.run_contiguous_slice_copy(space='cuda') def test_view(self): d = bf.ndarray(self.known_vals, dtype='f32') d = d.view(dtype='cf32') np.testing.assert_equal(d, np.array([[0 + 1j], [2 + 3j], [4 + 5j]])) @unittest.skipUnless(BF_CUDA_ENABLED, "requires GPU support") def test_str(self): e = bf.ndarray(self.known_vals, dtype='f32', space='cuda') self.assertEqual(str(e), str(self.known_array)) @unittest.skipUnless(BF_CUDA_ENABLED, "requires GPU support") def test_repr(self): f = bf.ndarray(self.known_vals, dtype='f32', space='cuda') repr_f = repr(f) # Note: This chops off the class name repr_f = repr_f[repr_f.find('('):] repr_k = repr(self.known_array) repr_k = repr_k[repr_k.find('('):] # Remove whitespace (for some reason the indentation differs) repr_f = repr_f.replace(' ', '') repr_k = repr_k.replace(' ', '') self.assertEqual(repr_f, repr_k) @unittest.skipUnless(BF_CUDA_ENABLED, "requires GPU support") def test_zeros_like(self): g = bf.ndarray(self.known_vals, dtype='f32', space='cuda') g = bf.zeros_like(g) g = g.copy('system') known = np.zeros_like(self.known_array) np.testing.assert_equal(g, known) @unittest.skipUnless(BF_CUDA_ENABLED, "requires GPU support") def test_getitem(self): g = bf.ndarray(self.known_vals, space='cuda') np.testing.assert_equal(g[0].copy('system'), self.known_array[0]) np.testing.assert_equal(g[(0,)].copy('system'), self.known_array[(0,)]) np.testing.assert_equal(int(g[0,0]), self.known_array[0,0]) np.testing.assert_equal(g[:1,1:].copy('system'), self.known_array[:1,1:]) @unittest.skipUnless(BF_CUDA_ENABLED, "requires GPU support") def test_setitem(self): g = bf.zeros_like(self.known_vals, space='cuda') g[...] = self.known_vals np.testing.assert_equal(g.copy('system'), self.known_vals) g[:1,1:] = [[999]] np.testing.assert_equal(g.copy('system'), np.array([[0,999],[2,3],[4,5]])) g[0,0] = 888 np.testing.assert_equal(g.copy('system'), np.array([[888,999],[2,3],[4,5]])) g[0] = [99,88] np.testing.assert_equal(g.copy('system'), np.array([[99,88],[2,3],[4,5]])) g[:,1] = [77,66,55] np.testing.assert_equal(g.copy('system'), np.array([[99,77],[2,66],[4,55]]))
bsd-3-clause
852389bf49de40e4d6bd17f5e2a56e48
48.545455
84
0.634731
3.264869
false
true
false
false
ledatelescope/bifrost
testbench/your_first_block.py
1
3507
#!/usr/bin/env python # Copyright (c) 2017-2020, The Bifrost Authors. All rights reserved. # # Redistribution and use in source and binary forms, with or without # modification, are permitted provided that the following conditions # are met: # * Redistributions of source code must retain the above copyright # notice, this list of conditions and the following disclaimer. # * Redistributions in binary form must reproduce the above copyright # notice, this list of conditions and the following disclaimer in the # documentation and/or other materials provided with the distribution. # * Neither the name of The Bifrost Authors nor the names of its # contributors may be used to endorse or promote products derived # from this software without specific prior written permission. # # THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY # EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE # IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR # PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR # CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, # EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, # PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR # PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY # OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT # (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. """ # your_first_block.py This testbench initializes a simple bifrost pipeline that reads from a binary file, and then writes the data to an output file. """ # Python2 compatibility from __future__ import print_function import os import numpy as np import bifrost.pipeline as bfp from bifrost.blocks import BinaryFileReadBlock, BinaryFileWriteBlock import glob from datetime import datetime from copy import deepcopy from pprint import pprint class UselessAddBlock(bfp.TransformBlock): def __init__(self, iring, n_to_add, *args, **kwargs): super(UselessAddBlock, self).__init__(iring, *args, **kwargs) self.n_to_add = n_to_add def on_sequence(self, iseq): ohdr = deepcopy(iseq.header) ohdr["name"] += "_with_added_value" return ohdr def on_data(self, ispan, ospan): in_nframe = ispan.nframe out_nframe = in_nframe idata = ispan.data + self.n_to_add odata = ospan.data odata[...] = idata return out_nframe class PrintStuffBlock(bfp.SinkBlock): def __init__(self, iring, *args, **kwargs): super(PrintStuffBlock, self).__init__(iring, *args, **kwargs) self.n_iter = 0 def on_sequence(self, iseq): print("[%s]" % datetime.now()) print(iseq.name) pprint(iseq.header) self.n_iter = 0 def on_data(self, ispan): now = datetime.now() if self.n_iter % 100 == 0: print("[%s] %s" % (now, ispan.data)) self.n_iter += 1 if __name__ == "__main__": # Setup pipeline filenames = sorted(glob.glob('testdata/sin_data*.bin')) b_read = BinaryFileReadBlock(filenames, 32768, 1, 'f32') b_add = UselessAddBlock(b_read, n_to_add=100) b_print = PrintStuffBlock(b_read) b_print2 = PrintStuffBlock(b_add) # Run pipeline pipeline = bfp.get_default_pipeline() print(pipeline.dot_graph()) pipeline.run()
bsd-3-clause
da8cfbf465817a1e4c6ce7e6c371587a
34.424242
83
0.692615
3.742796
false
false
false
false
ledatelescope/bifrost
python/bifrost/views/basic_views.py
1
8870
# Copyright (c) 2016-2021, The Bifrost Authors. All rights reserved. # # Redistribution and use in source and binary forms, with or without # modification, are permitted provided that the following conditions # are met: # * Redistributions of source code must retain the above copyright # notice, this list of conditions and the following disclaimer. # * Redistributions in binary form must reproduce the above copyright # notice, this list of conditions and the following disclaimer in the # documentation and/or other materials provided with the distribution. # * Neither the name of The Bifrost Authors nor the names of its # contributors may be used to endorse or promote products derived # from this software without specific prior written permission. # # THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS ``AS IS'' AND ANY # EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE # IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR # PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR # CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, # EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, # PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR # PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY # OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT # (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. from __future__ import absolute_import, division from bifrost.pipeline import block_view from bifrost.DataType import DataType from bifrost.units import convert_units from numpy import isclose from bifrost import telemetry telemetry.track_module() def custom(block, hdr_transform): """An alias to `bifrost.pipeline.block_view` """ return block_view(block, hdr_transform) def rename_axis(block, old, new): def header_transform(hdr, old=old, new=new): axis = hdr['_tensor']['labels'].index(old) hdr['_tensor']['labels'][axis] = new return hdr return block_view(block, header_transform) def reinterpret_axis(block, axis, label, scale=None, units=None): """ Manually reinterpret the scale and/or units on an axis """ def header_transform(hdr, axis=axis, label=label, scale=scale, units=units): tensor = hdr['_tensor'] if isinstance(axis, str): axis = tensor['labels'].index(axis) if label is not None: tensor['labels'][axis] = label if scale is not None: tensor['scales'][axis] = scale if units is not None: tensor['units'][axis] = units return hdr return block_view(block, header_transform) def reverse_scale(block, axis): """ Manually reverse the scale factor on a given axis""" def header_transform(hdr, axis=axis): tensor = hdr['_tensor'] if isinstance(axis, str): axis = tensor['labels'].index(axis) tensor['scales'][axis][1] *= -1 return hdr return block_view(block, header_transform) def add_axis(block, axis, label=None, scale=None, units=None): """Add an extra dimension to the frame at position 'axis' E.g., if the shape is [-1, 3, 2], then selecting axis=1 would change the shape to be [-1, 1, 3, 2]. Axis may be negative, or a string corresponding to an existing axis label, in which case the new axis is inserted after the referenced axis. """ def header_transform(hdr, axis=axis, label=label, scale=scale, units=units): tensor = hdr['_tensor'] if isinstance(axis, str): axis = tensor['labels'].index(axis) + 1 if axis < 0: axis += len(tensor['shape']) + 1 tensor['shape'].insert(axis, 1) if 'labels' in tensor: tensor['labels'].insert(axis, label) if 'scales' in tensor: tensor['scales'].insert(axis, scale) if 'units' in tensor: tensor['units'].insert(axis, units) return hdr return block_view(block, header_transform) def delete_axis(block, axis): """Remove a unitary dimension from the frame E.g., if the shape is [-1, 1, 3, 2], then selecting axis=1 would change the shape to be [-1, 3, 2]. Axis may be negative, or a string corresponding to an existing axis label. """ def header_transform(hdr, axis=axis): tensor = hdr['_tensor'] specified_axis = axis if isinstance(axis, str): specified_axis = "'%s'" % specified_axis axis = tensor['labels'].index(axis) if axis < 0: axis += len(tensor['shape']) + 1 if tensor['shape'][axis] != 1: raise ValueError("Cannot delete non-unitary axis %s with shape %i" % (specified_axis, tensor['shape'][axis])) del tensor['shape'][axis] if 'labels' in tensor: del tensor['labels'][axis] if 'scales' in tensor: del tensor['scales'][axis] if 'units' in tensor: del tensor['units'][axis] return hdr return block_view(block, header_transform) def astype(block, dtype): def header_transform(hdr, new_dtype=dtype): tensor = hdr['_tensor'] old_dtype = tensor['dtype'] old_itemsize = DataType(old_dtype).itemsize new_itemsize = DataType(new_dtype).itemsize old_axissize = old_itemsize * tensor['shape'][-1] if old_axissize % new_itemsize: raise ValueError("New type not compatible with data shape") tensor['shape'][-1] = old_axissize // new_itemsize tensor['dtype'] = dtype return hdr return block_view(block, header_transform) def split_axis(block, axis, n, label=None): # Set function attributes to enable capture in nested function (closure) def header_transform(hdr, axis=axis, n=n, label=label): tensor = hdr['_tensor'] if isinstance(axis, str): axis = tensor['labels'].index(axis) shape = tensor['shape'] if shape[axis] == -1: # Axis is frame axis # TODO: Should assert even division here instead? # ***TODO: Why does pipeline deadlock when this doesn't divide? hdr['gulp_nframe'] = (hdr['gulp_nframe'] - 1) // n + 1 else: # Axis is not frame axis if shape[axis] % n: raise ValueError("Split does not evenly divide axis (%i // %i)" % (tensor['shape'][axis], n)) shape[axis] //= n shape.insert(axis + 1, n) if 'units' in tensor: tensor['units'].insert(axis + 1, tensor['units'][axis]) if 'labels' in tensor: if label is None: label = tensor['labels'][axis] + "_split" tensor['labels'].insert(axis + 1, label) if 'scales' in tensor: tensor['scales'].insert(axis + 1, [0, tensor['scales'][axis][1]]) tensor['scales'][axis][1] *= n return hdr return block_view(block, header_transform) def merge_axes(block, axis1, axis2, label=None): def header_transform(hdr, axis1=axis1, axis2=axis2, label=label): tensor = hdr['_tensor'] if isinstance(axis1, str): axis1 = tensor['labels'].index(axis1) if isinstance(axis2, str): axis2 = tensor['labels'].index(axis2) axis1, axis2 = sorted([axis1, axis2]) if axis2 != axis1 + 1: raise ValueError("Merge axes must be adjacent") n = tensor['shape'][axis2] if n == -1: # Axis2 is frame axis raise ValueError("Second merge axis cannot be frame axis") elif tensor['shape'][axis1] == -1: # Axis1 is frame axis hdr['gulp_nframe'] *= n else: # Neither axis is frame axis tensor['shape'][axis1] *= n del tensor['shape'][axis2] if 'scales' in tensor and 'units' in tensor: scale1 = tensor['scales'][axis1][1] scale2 = tensor['scales'][axis2][1] units1 = tensor['units'][axis1] units2 = tensor['units'][axis2] scale2 = convert_units(scale2, units2, units1) if not isclose(scale1, n * scale2): raise ValueError("Scales of merge axes do not line up: " "%f != %f" % (scale1, n * scale2)) tensor['scales'][axis1][1] = scale2 del tensor['scales'][axis2] del tensor['units'][axis2] if 'labels' in tensor: if label is not None: tensor['labels'][axis1] = label del tensor['labels'][axis2] return hdr return block_view(block, header_transform)
bsd-3-clause
dcbd0ba286d51edb6225615217ce2b6b
40.448598
81
0.609808
4.059497
false
false
false
false
ledatelescope/bifrost
docs/source/conf.py
1
5080
# -*- coding: utf-8 -*- # # bifrost documentation build configuration file, created by # sphinx-quickstart on Thu Apr 27 09:41:02 2017. # # This file is execfile()d with the current directory set to its # containing dir. # # Note that not all possible configuration values are present in this # autogenerated file. # # All configuration values have a default; values that are commented out # serve to show the default. # If extensions (or modules to document with autodoc) are in another directory, # add these directories to sys.path here. If the directory is relative to the # documentation root, use os.path.abspath to make it absolute, like shown here. # # import os # import sys # sys.path.insert(0, os.path.abspath('.')) # -- General configuration ------------------------------------------------ # If your documentation needs a minimal Sphinx version, state it here. # # needs_sphinx = '1.0' # Add any Sphinx extension module names here, as strings. They can be # extensions coming with Sphinx (named 'sphinx.ext.*') or your custom # ones. extensions = ['sphinx.ext.autodoc', 'sphinx.ext.doctest', 'sphinx.ext.todo', 'sphinx.ext.coverage', 'sphinx.ext.mathjax', 'sphinx.ext.ifconfig', 'sphinx.ext.viewcode', 'sphinx.ext.napoleon', 'breathe'] # Add any paths that contain templates here, relative to this directory. templates_path = ['_templates'] # The suffix(es) of source filenames. # You can specify multiple suffix as a list of string: # # source_suffix = ['.rst', '.md'] source_suffix = '.rst' # The master toctree document. master_doc = 'index' # General information about the project. project = u'bifrost' copyright = u'2017, ledatelescope' author = u'ledatelescope' # The version info for the project you're documenting, acts as replacement for # |version| and |release|, also used in various other places throughout the # built documents. # # The short X.Y version. version = u'0.6' # The full version, including alpha/beta/rc tags. release = u'0.6' # The language for content autogenerated by Sphinx. Refer to documentation # for a list of supported languages. # # This is also used if you do content translation via gettext catalogs. # Usually you set "language" from the command line for these cases. language = None # List of patterns, relative to source directory, that match files and # directories to ignore when looking for source files. # This patterns also effect to html_static_path and html_extra_path exclude_patterns = [] # The name of the Pygments (syntax highlighting) style to use. pygments_style = 'sphinx' # If true, `todo` and `todoList` produce output, else they produce nothing. todo_include_todos = True # -- Options for HTML output ---------------------------------------------- # The theme to use for HTML and HTML Help pages. See the documentation for # a list of builtin themes. # html_theme = 'alabaster' # Theme options are theme-specific and customize the look and feel of a theme # further. For a list of options available for each theme, see the # documentation. # # html_theme_options = {} # Add any paths that contain custom static files (such as style sheets) here, # relative to this directory. They are copied after the builtin static files, # so a file named "default.css" will overwrite the builtin "default.css". html_static_path = ['_static'] # -- Options for HTMLHelp output ------------------------------------------ # Output file base name for HTML help builder. htmlhelp_basename = 'bifrostdoc' # -- Options for LaTeX output --------------------------------------------- latex_elements = { # The paper size ('letterpaper' or 'a4paper'). # # 'papersize': 'letterpaper', # The font size ('10pt', '11pt' or '12pt'). # # 'pointsize': '10pt', # Additional stuff for the LaTeX preamble. # # 'preamble': '', # Latex figure (float) alignment # # 'figure_align': 'htbp', } # Grouping the document tree into LaTeX files. List of tuples # (source start file, target name, title, # author, documentclass [howto, manual, or own class]). latex_documents = [ (master_doc, 'bifrost.tex', u'bifrost Documentation', u'ledatelescope', 'manual'), ] # -- Options for manual page output --------------------------------------- # One entry per manual page. List of tuples # (source start file, name, description, authors, manual section). man_pages = [ (master_doc, 'bifrost', u'bifrost Documentation', [author], 1) ] # -- Options for Texinfo output ------------------------------------------- # Grouping the document tree into Texinfo files. List of tuples # (source start file, target name, title, author, # dir menu entry, description, category) texinfo_documents = [ (master_doc, 'bifrost', u'bifrost Documentation', author, 'bifrost', 'A stream processing framework for high-throughput applications.', 'Miscellaneous'), ] # -- Set up breathe ------------------------------------------------------- breathe_projects = {"bifrost": "../doxygen/xml"} breathe_default_project = "bifrost"
bsd-3-clause
c0a1cf71c182e9bc2246ba5ed68f1afd
29.787879
90
0.666339
3.816679
false
true
false
false