repo
stringclasses
900 values
file
stringclasses
754 values
content
stringlengths
4
215k
https://github.com/rodneyosodo/variational-quantum-classifier-on-heartattack
rodneyosodo
# Basis encoding example with PennyLane import pandas as pd import pennylane as qml from pennylane import numpy as np # import the template from pennylane.templates.layers import StronglyEntanglingLayers from sklearn.utils import shuffle from sklearn.preprocessing import normalize from pennylane.templates.embeddings import AngleEmbedding, BasisEmbedding, AmplitudeEmbedding, DisplacementEmbedding from pennylane.init import strong_ent_layers_uniform np.random.seed(42) # Constants DATA_PATH = "../../Data/Processed/data.csv" num_qubits = 4 data = pd.read_csv(DATA_PATH) X, Y = data[['sex', 'cp', 'exang', 'oldpeak']].values, data['num'].values # normalize the data X = normalize(X) print(X[:5]) device = qml.device("default.qubit", wires=num_qubits) @qml.qnode(device) def circuit(data): for i in range(num_qubits): # Create superposition qml.Hadamard(wires=i) AngleEmbedding(features=data, wires=range(num_qubits), rotation="Y") return qml.expval(qml.PauliZ(0)), qml.expval(qml.PauliZ(1)), qml.expval(qml.PauliZ(2)), qml.expval(qml.PauliZ(3)) circuit(X[0])
https://github.com/rodneyosodo/variational-quantum-classifier-on-heartattack
rodneyosodo
from qiskit.ml.datasets import * from qiskit import QuantumCircuit from qiskit.aqua.components.optimizers import COBYLA, ADAM, SPSA, SLSQP, POWELL, L_BFGS_B, TNC, AQGD from qiskit.circuit.library import ZZFeatureMap, RealAmplitudes from qiskit.quantum_info import Statevector import numpy as np import pandas as pd import matplotlib.pyplot as plt from sklearn.preprocessing import normalize from sklearn.model_selection import train_test_split from sklearn.utils import shuffle import warnings warnings.filterwarnings("ignore") %matplotlib inline # constants n = 4 RANDOM_STATE = 42 LR = 1e-3 class_labels = ['yes', 'no'] def normalizeData(DATA_PATH = "../../Data/Processed/data.csv"): """ Normalizes the data """ # Reads the data data = pd.read_csv(DATA_PATH) data = shuffle(data, random_state=RANDOM_STATE) X, Y = data[['sex', 'cp', 'exang', 'oldpeak']].values, data['num'].values # normalize the data X = normalize(X) X_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size=0.3, random_state=RANDOM_STATE) return X_train, X_test, Y_train, Y_test X_train, X_test, Y_train, Y_test = normalizeData() sv = Statevector.from_label('0' * n) feature_map = ZZFeatureMap(n, reps=1) var_form = RealAmplitudes(n, reps=1) circuit = feature_map.combine(var_form) circuit.draw(output='mpl', filename="overallcircuit.png") def get_data_dict(params, x): parameters = {} for i, p in enumerate(feature_map.ordered_parameters): parameters[p] = x[i] for i, p in enumerate(var_form.ordered_parameters): parameters[p] = params[i] return parameters def assign_label(bit_string, class_labels): hamming_weight = sum([int(k) for k in list(bit_string)]) is_odd_parity = hamming_weight & 1 if is_odd_parity: return class_labels[1] else: return class_labels[0] def return_probabilities(counts, class_labels): shots = sum(counts.values()) result = {class_labels[0]: 0, class_labels[1]: 0} for key, item in counts.items(): label = assign_label(key, class_labels) result[label] += counts[key]/shots return result def classify(x_list, params, class_labels): qc_list = [] for x in x_list: circ_ = circuit.assign_parameters(get_data_dict(params, x)) qc = sv.evolve(circ_) qc_list += [qc] probs = [] for qc in qc_list: counts = qc.to_counts() prob = return_probabilities(counts, class_labels) probs += [prob] return probs def mse_cost(probs, expected_label): p = probs.get(expected_label) actual, pred = np.array(1), np.array(p) return np.square(np.subtract(actual,pred)).mean() cost_list = [] def cost_function(X, Y, class_labels, params, shots=100, print_value=False): # map training input to list of labels and list of samples cost = 0 training_labels = [] training_samples = [] for sample in X: training_samples += [sample] for label in Y: if label == 0: training_labels += [class_labels[0]] elif label == 1: training_labels += [class_labels[1]] probs = classify(training_samples, params, class_labels) # evaluate costs for all classified samples for i, prob in enumerate(probs): cost += mse_cost(prob, training_labels[i]) cost /= len(training_samples) # print resulting objective function if print_value: print('%.4f' % cost) # return objective value cost_list.append(cost) return cost cost_list = [] optimizer = ADAM(maxiter=100, ) # define objective function for training objective_function = lambda params: cost_function(X_train, Y_train, class_labels, params, print_value=True) # randomly initialize the parameters np.random.seed(RANDOM_STATE) init_params = 2*np.pi*np.random.rand(n*(1)*2) # train classifier opt_params, value, _ = optimizer.optimize(len(init_params), objective_function, initial_point=init_params) # print results print() print('opt_params:', opt_params) print('opt_value: ', value) fig = plt.figure() plt.plot(range(0,901,1), cost_list) plt.xlabel('Steps') plt.ylabel('Cost value') plt.title("ADAM Cost value against steps") plt.show() fig.savefig('../../Output/Figures/costvssteps.jpeg') def test_model(X, Y, class_labels, params): accuracy = 0 training_labels = [] training_samples = [] for sample in X: training_samples += [sample] probs = classify(training_samples, params, class_labels) for i, prob in enumerate(probs): if (prob.get('yes') >= prob.get('no')) and (Y_test[i] == 0): accuracy += 1 elif (prob.get('no') >= prob.get('yes')) and (Y_test[i] == 1): accuracy += 1 accuracy /= len(Y_test) print("Test accuracy: {}\n".format(accuracy)) test_model(X_test, Y_test, class_labels, opt_params)
https://github.com/rodneyosodo/variational-quantum-classifier-on-heartattack
rodneyosodo
import pandas as pd import numpy as np import matplotlib.pyplot as plt pd.set_option('display.max_rows', None) pd.set_option('display.max_columns', None) pd.set_option('display.width', None) pd.set_option('display.max_colwidth', -1) df = pd.read_csv("../../Data/Processed/costs.csv") df.columns = ['Spec', "CostValues"] df.head(1) df['Spec'] def check_lowest_cost(): """ Checks the configs which attained the lowest sost """ cost = [] lowest_cost = [] for i in range(df.shape[0]): data = [float(x.replace("[", "").replace("]", "")) for x in df['CostValues'][i].split(',')] lowest_cost.append(data[len(data) - 1]) lowest_cost = sorted(lowest_cost)[:80] # for m in lowest_cost: for i in range(df.shape[0]): data = df['CostValues'][i].split(',') data = [float(x.replace("[", "").replace("]", "")) for x in data] # if float(m) == float(data[len(data) - 1]): # print("{} : Cost: {}".format(df["Spec"][i], data[len(data) - 1])) cost.append(data[len(data) - 1]) print(len(cost)) df['exel'] = cost df.to_excel("n.xlsx") check_lowest_cost() # Plotting all for i in range(df.shape[0]): fig = plt.figure() data = df['CostValues'][i].split(',') data = [float(x.replace("[", "").replace("]", "")) for x in data] plt.plot(range(0, len(data), 1), data) plt.xlabel('Steps') plt.ylabel('Cost value') plt.title(str(df['Spec'][i])) plt.show() path = '../../Output/Figures/{}.jpeg'.format(df['Spec'][i].replace("(", "").replace("FeatureMap", "").replace(")", "").replace(",", "").replace("=", "").replace(" ", "").replace("_", "")) # fig.savefig(path) def plot_individual_optimizers(Optimizers): for opt in Optimizers: max_iter = 0 fig = plt.figure(figsize=(20,20)) for i in range(df.shape[0]): if df['Spec'][i].__contains__(opt): data = [float(x.replace("[", "").replace("]", "")) for x in df['CostValues'][i].split(',')] if len(data) >= max_iter: max_iter = len(data) for i in range(df.shape[0]): if df['Spec'][i].__contains__(opt): data = df['CostValues'][i].split(',') data = [float(x.replace("[", "").replace("]", "")) for x in data] if max_iter >= len(data): temp = np.full((max_iter, ), data[len(data) -1]) temp[:len(data)] = data data = temp plt.plot(range(0, len(data), 1), data, label='{}'.format(df['Spec'][i].replace("(", "").replace("FeatureMap", "").replace(")", "").replace(",", "").replace("=", "").replace(" ", "").replace("_", ""))) plt.xlabel('Steps') plt.ylabel('Cost value') plt.title(opt) plt.legend() plt.show() plot_individual_optimizers(["COBYLA", "SPSA", "ADAM"]) def plot_increase_vdepth(optimizer, featuremap, reps, checking): points = [] for i in range(df.shape[0]): title = str(df['Spec'][i]) if title.split("(")[0] == featuremap and str(title).__contains__(optimizer) and str(title).__contains__("reps={}".format(reps)): points.append(float(df['CostValues'][i].split(',')[-1].replace("[", "").replace("]", ""))) fig = plt.figure() plt.plot(range(0, len(points), 1), points) plt.xlabel(checking) plt.ylabel('Cost value') plt.title("{} {} featuremap {} fdepth".format(optimizer, featuremap, reps)) plt.show() def plot_increase_fdepth(optimizer, featuremap, fdepth, checking): points = [] for i in range(df.shape[0]): title = str(df['Spec'][i]) if title.split("(")[0] == featuremap and str(title).__contains__(optimizer) and str(title).__contains__("vdepth {}".format(fdepth)): points.append(float(df['CostValues'][i].split(',')[-1].replace("[", "").replace("]", ""))) fig = plt.figure() plt.plot(range(0, len(points), 1), points) plt.xlabel(checking) plt.ylabel('Cost value') plt.title("{} {} featuremap {} vdepth".format(optimizer, featuremap, fdepth)) plt.show() plot_increase_vdepth("SPSA", "PauliFeatureMap", 1, "fdepth") plot_increase_fdepth("SPSA", "PauliFeatureMap", 1, "vdepth") for opt in ['SPSA', "COBYLA", "ADAM"]: for fmap in ['ZZFeatureMap', 'ZFeatureMap', "PauliFeatureMap"]: for fdepth in [1, 2, 4]: plot_increase_vdepth(opt, fmap, fdepth, "vdepth") for opt in ['SPSA', "COBYLA", "ADAM"]: for fmap in ['ZZFeatureMap', 'ZFeatureMap', "PauliFeatureMap"]: for fdepth in [1, 3, 5]: plot_increase_fdepth(opt, fmap, fdepth, "fdepth")
https://github.com/rodneyosodo/variational-quantum-classifier-on-heartattack
rodneyosodo
from qiskit.ml.datasets import * from qiskit import QuantumCircuit from qiskit.aqua.components.optimizers import COBYLA, ADAM, SPSA, SLSQP, POWELL, L_BFGS_B, TNC, AQGD from qiskit.circuit.library import ZZFeatureMap, RealAmplitudes, ZFeatureMap, PauliFeatureMap from qiskit.quantum_info import Statevector import numpy as np import pandas as pd import matplotlib.pyplot as plt from sklearn.preprocessing import normalize from sklearn.model_selection import train_test_split from sklearn.utils import shuffle import warnings warnings.filterwarnings("ignore") %matplotlib inline # constants n = 4 RANDOM_STATE = 42 LR = 1e-3 class_labels = ['yes', 'no'] def normalizeData(DATA_PATH = "../../Data/Processed/data.csv"): """ Normalizes the data """ # Reads the data data = pd.read_csv(DATA_PATH) data = shuffle(data, random_state=RANDOM_STATE) X, Y = data[['sex', 'cp', 'exang', 'oldpeak']].values, data['num'].values # normalize the data X = normalize(X) X_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size=0.3, random_state=RANDOM_STATE) return X_train, X_test, Y_train, Y_test X_train, X_test, Y_train, Y_test = normalizeData() feature_map = ZZFeatureMap(n, reps=1) feature_map.draw(output="mpl", filename="../../Output/Figures/ZZFeaturemap.png") feature_map = ZFeatureMap(n, reps=1) feature_map.draw(output="mpl", filename="../../Output/Figures/ZFeaturemap.png") feature_map = PauliFeatureMap(n, reps=1) feature_map.draw(output="mpl", filename="../../Output/Figures/PauliFeaturemap.png") var_form = RealAmplitudes(n, reps=1) var_form.draw(output="mpl", filename="../../Output/Figures/RealAmplitudes.png") sv = Statevector.from_label('0' * n) feature_map = ZZFeatureMap(n, reps=1) feature_map.barrier() var_form = RealAmplitudes(n, reps=1) circuit = feature_map.combine(var_form) circuit.draw(output="mpl", filename="../../Output/Figures/overallcircuit.png") def get_data_dict(params, x): parameters = {} for i, p in enumerate(feature_map.ordered_parameters): parameters[p] = x[i] for i, p in enumerate(var_form.ordered_parameters): parameters[p] = params[i] return parameters def assign_label(bit_string, class_labels): hamming_weight = sum([int(k) for k in list(bit_string)]) is_odd_parity = hamming_weight & 1 if is_odd_parity: return class_labels[1] else: return class_labels[0] data = X_train[0] params = np.array([0.1, 1.2, 0.02, 0.1, 0.1, 1.2, 0.02, 0.1]) circ_ = circuit.assign_parameters(get_data_dict(params, data)) circ_.draw(plot_barriers=True, output="mpl", filename="../../Output/Figures/parameterisedcircuit.png")
https://github.com/rodneyosodo/variational-quantum-classifier-on-heartattack
rodneyosodo
import pandas as pd import numpy as np import matplotlib.pyplot as plt import seaborn as sns import pandas_profiling as pp from sklearn.datasets import load_wine %matplotlib inline # Constants DATA_PATH = "../../Data/Raw/data.csv" CLEAN_DATA_PATH = "../../Data/Processed/winedata.csv" raw_data = load_wine() features = pd.DataFrame(data=raw_data['data'],columns=raw_data['feature_names']) data = features data['target']=raw_data['target'] data['class']=data['target'].map(lambda ind: raw_data['target_names'][ind]) data.head() data.shape data.columns print(raw_data['DESCR']) data.info() data.describe() def check_unique(df): """" Checks the unique value in each column :param df: The dataframe """ for col in df.columns: unique = df[col].unique() print("Column: {} has {} unique values\n".format(col, unique)) check_unique(data) fig = plt.figure() sns.countplot(x="target", data=data, palette="gist_rainbow_r") plt.xlabel("Heart disease (0 = have, 1 = don't)") plt.title("Target distribution") plt.show() fig.savefig("../../Output/Figures/targetdist.png", ) data = data[data['class'] != 'class_2'] data.duplicated().sum() data.drop_duplicates(inplace=True) data.drop('target', axis=1).corrwith(data['target']).plot(kind='bar', grid=True, figsize=(12, 8), title="Correlation with target") pp.ProfileReport(df=data, dark_mode=True, explorative=True) data = data[['alcohol', 'flavanoids', 'color_intensity', 'proline', 'target']] data.to_csv(CLEAN_DATA_PATH, index=False)
https://github.com/rodneyosodo/variational-quantum-classifier-on-heartattack
rodneyosodo
from qiskit.ml.datasets import * from qiskit import QuantumCircuit from qiskit.aqua.components.optimizers import COBYLA, ADAM, SPSA, SLSQP, POWELL, L_BFGS_B, TNC, AQGD from qiskit.circuit.library import ZZFeatureMap, RealAmplitudes from qiskit.quantum_info import Statevector import numpy as np import pandas as pd import matplotlib.pyplot as plt from sklearn.preprocessing import MinMaxScaler from sklearn.model_selection import train_test_split from sklearn.utils import shuffle import warnings warnings.filterwarnings("ignore") %matplotlib inline # constants n = 4 RANDOM_STATE = 42 LR = 1e-3 class_labels = ['yes', 'no'] def normalizeData(DATA_PATH = "../../Data/Processed/winedata.csv"): """ Normalizes the data """ # Reads the data data = pd.read_csv(DATA_PATH) data = shuffle(data, random_state=RANDOM_STATE) X, Y = data[['alcohol', 'flavanoids', 'color_intensity', 'proline']].values, data['target'].values # normalize the data scaler = MinMaxScaler(feature_range=(-2 * np.pi, 2 * np.pi)) X = scaler.fit_transform(X) X_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size=0.3, random_state=RANDOM_STATE) return X_train, X_test, Y_train, Y_test X_train, X_test, Y_train, Y_test = normalizeData() sv = Statevector.from_label('0' * n) feature_map = ZZFeatureMap(n, reps=1) var_form = RealAmplitudes(n, reps=1) circuit = feature_map.combine(var_form) circuit.draw(output='mpl') def get_data_dict(params, x): parameters = {} for i, p in enumerate(feature_map.ordered_parameters): parameters[p] = x[i] for i, p in enumerate(var_form.ordered_parameters): parameters[p] = params[i] return parameters def assign_label(bit_string, class_labels): hamming_weight = sum([int(k) for k in list(bit_string)]) is_odd_parity = hamming_weight & 1 if is_odd_parity: return class_labels[1] else: return class_labels[0] def return_probabilities(counts, class_labels): shots = sum(counts.values()) result = {class_labels[0]: 0, class_labels[1]: 0} for key, item in counts.items(): label = assign_label(key, class_labels) result[label] += counts[key]/shots return result def classify(x_list, params, class_labels): qc_list = [] for x in x_list: circ_ = circuit.assign_parameters(get_data_dict(params, x)) qc = sv.evolve(circ_) qc_list += [qc] probs = [] for qc in qc_list: counts = qc.to_counts() prob = return_probabilities(counts, class_labels) probs += [prob] return probs def mse_cost(probs, expected_label): p = probs.get(expected_label) actual, pred = np.array(1), np.array(p) return np.square(np.subtract(actual,pred)).mean() cost_list = [] def cost_function(X, Y, class_labels, params, shots=100, print_value=False): # map training input to list of labels and list of samples cost = 0 training_labels = [] training_samples = [] for sample in X: training_samples += [sample] for label in Y: if label == 0: training_labels += [class_labels[0]] elif label == 1: training_labels += [class_labels[1]] probs = classify(training_samples, params, class_labels) # evaluate costs for all classified samples for i, prob in enumerate(probs): cost += mse_cost(prob, training_labels[i]) cost /= len(training_samples) # print resulting objective function if print_value: print('%.4f' % cost) # return objective value cost_list.append(cost) return cost cost_list = [] optimizer = SPSA(maxiter=100) # define objective function for training objective_function = lambda params: cost_function(X_train, Y_train, class_labels, params, print_value=True) # randomly initialize the parameters np.random.seed(RANDOM_STATE) init_params = 2*np.pi*np.random.rand(n*(1)*2) # train classifier opt_params, value, _ = optimizer.optimize(len(init_params), objective_function, initial_point=init_params) # print results print() print('opt_params:', opt_params) print('opt_value: ', value) fig = plt.figure() plt.plot(range(0,len(cost_list),1), cost_list) plt.xlabel('Steps') plt.ylabel('Cost value') plt.title("ADAM Cost value against steps") plt.show() def test_model(X, Y, class_labels, params): accuracy = 0 training_labels = [] training_samples = [] for sample in X: training_samples += [sample] probs = classify(training_samples, params, class_labels) for i, prob in enumerate(probs): if (prob.get('yes') >= prob.get('no')) and (Y_test[i] == 0): accuracy += 1 elif (prob.get('no') >= prob.get('yes')) and (Y_test[i] == 1): accuracy += 1 accuracy /= len(Y_test) print("Test accuracy: {}\n".format(accuracy)) test_model(X_test, Y_test, class_labels, opt_params)
https://github.com/rodneyosodo/variational-quantum-classifier-on-heartattack
rodneyosodo
import pandas as pd import numpy as np import matplotlib.pyplot as plt import seaborn as sns import pandas_profiling as pp from sklearn.datasets import load_wine, load_iris %matplotlib inline # Constants CLEAN_DATA_PATH = "../../Data/Processed/iris_csv.csv" raw_data = load_iris() features = pd.DataFrame(data=raw_data['data'],columns=raw_data['feature_names']) data = features data['target']=raw_data['target'] data['class']=data['target'].map(lambda ind: raw_data['target_names'][ind]) data.head() data.shape data.columns print(raw_data['DESCR']) data.info() data.describe() def check_unique(df): """" Checks the unique value in each column :param df: The dataframe """ for col in df.columns: unique = df[col].unique() print("Column: {} has {} unique values\n".format(col, unique)) check_unique(data) fig = plt.figure() sns.countplot(x="target", data=data, palette="gist_rainbow_r") plt.xlabel("Classes(0 = setosa, 1 = versicolor, 2 = verginica)") plt.title("Target distribution") plt.show() data = data[data['class'] != 'virginica'] data.duplicated().sum() data.drop_duplicates(inplace=True) data.drop('target', axis=1).corrwith(data['target']).plot(kind='bar', grid=True, figsize=(12, 8), title="Correlation with target") pp.ProfileReport(df=data, dark_mode=True, explorative=True) data data = data[['sepal length (cm)', 'sepal width (cm)', 'petal length (cm)', 'petal width (cm)', 'target']] data.to_csv(CLEAN_DATA_PATH, index=False)
https://github.com/rodneyosodo/variational-quantum-classifier-on-heartattack
rodneyosodo
from qiskit.ml.datasets import * from qiskit import QuantumCircuit from qiskit.aqua.components.optimizers import COBYLA, ADAM, SPSA, SLSQP, POWELL, L_BFGS_B, TNC, AQGD from qiskit.circuit.library import ZZFeatureMap, RealAmplitudes from qiskit.quantum_info import Statevector import numpy as np import pandas as pd import matplotlib.pyplot as plt from sklearn.preprocessing import MinMaxScaler from sklearn.model_selection import train_test_split from sklearn.utils import shuffle import warnings warnings.filterwarnings("ignore") %matplotlib inline # constants n = 4 RANDOM_STATE = 42 LR = 1e-3 class_labels = ['yes', 'no'] def normalizeData(DATA_PATH = "../../Data/Processed/iris_csv.csv"): """ Normalizes the data """ # Reads the data data = pd.read_csv(DATA_PATH) data = shuffle(data, random_state=RANDOM_STATE) X, Y = data[['sepal length (cm)', 'sepal width (cm)', 'petal length (cm)', 'petal width (cm)']].values, data['target'].values # normalize the data scaler = MinMaxScaler(feature_range=(-2 * np.pi, 2 * np.pi)) X = scaler.fit_transform(X) X_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size=0.3, random_state=RANDOM_STATE) return X_train, X_test, Y_train, Y_test X_train, X_test, Y_train, Y_test = normalizeData() sv = Statevector.from_label('0' * n) feature_map = ZZFeatureMap(n, reps=1) var_form = RealAmplitudes(n, reps=1) circuit = feature_map.combine(var_form) circuit.draw(output='mpl') def get_data_dict(params, x): parameters = {} for i, p in enumerate(feature_map.ordered_parameters): parameters[p] = x[i] for i, p in enumerate(var_form.ordered_parameters): parameters[p] = params[i] return parameters def assign_label(bit_string, class_labels): hamming_weight = sum([int(k) for k in list(bit_string)]) is_odd_parity = hamming_weight & 1 if is_odd_parity: return class_labels[1] else: return class_labels[0] def return_probabilities(counts, class_labels): shots = sum(counts.values()) result = {class_labels[0]: 0, class_labels[1]: 0} for key, item in counts.items(): label = assign_label(key, class_labels) result[label] += counts[key]/shots return result def classify(x_list, params, class_labels): qc_list = [] for x in x_list: circ_ = circuit.assign_parameters(get_data_dict(params, x)) qc = sv.evolve(circ_) qc_list += [qc] probs = [] for qc in qc_list: counts = qc.to_counts() prob = return_probabilities(counts, class_labels) probs += [prob] return probs def mse_cost(probs, expected_label): p = probs.get(expected_label) actual, pred = np.array(1), np.array(p) return np.square(np.subtract(actual,pred)).mean() cost_list = [] def cost_function(X, Y, class_labels, params, shots=100, print_value=False): # map training input to list of labels and list of samples cost = 0 training_labels = [] training_samples = [] for sample in X: training_samples += [sample] for label in Y: if label == 0: training_labels += [class_labels[0]] elif label == 1: training_labels += [class_labels[1]] probs = classify(training_samples, params, class_labels) # evaluate costs for all classified samples for i, prob in enumerate(probs): cost += mse_cost(prob, training_labels[i]) cost /= len(training_samples) # print resulting objective function if print_value: print('%.4f' % cost) # return objective value cost_list.append(cost) return cost cost_list = [] optimizer = SPSA(maxiter=100) # define objective function for training objective_function = lambda params: cost_function(X_train, Y_train, class_labels, params, print_value=True) # randomly initialize the parameters np.random.seed(RANDOM_STATE) init_params = 2*np.pi*np.random.rand(n*(1)*2) # train classifier opt_params, value, _ = optimizer.optimize(len(init_params), objective_function, initial_point=init_params) # print results print() print('opt_params:', opt_params) print('opt_value: ', value) fig = plt.figure() plt.plot(range(0,len(cost_list),1), cost_list) plt.xlabel('Steps') plt.ylabel('Cost value') plt.title("SPSA Cost value against steps") plt.show() def test_model(X, Y, class_labels, params): accuracy = 0 training_labels = [] training_samples = [] for sample in X: training_samples += [sample] probs = classify(training_samples, params, class_labels) for i, prob in enumerate(probs): if (prob.get('yes') >= prob.get('no')) and (Y_test[i] == 0): accuracy += 1 elif (prob.get('no') >= prob.get('yes')) and (Y_test[i] == 1): accuracy += 1 accuracy /= len(Y_test) print("Test accuracy: {}\n".format(accuracy)) test_model(X_test, Y_test, class_labels, opt_params)
https://github.com/rodneyosodo/variational-quantum-classifier-on-heartattack
rodneyosodo
import pandas as pd import numpy as np import matplotlib.pyplot as plt pd.set_option('display.max_rows', None) pd.set_option('display.max_columns', None) pd.set_option('display.width', None) pd.set_option('display.max_colwidth', -1) df = pd.read_csv("../../Data/Processed/winecosts.csv") df.columns = ['Spec', "CostValues"] df.head(1) df['Spec'] def check_lowest_cost(df): """ Checks the configs which attained the lowest sost """ cost = [] lowest_cost = [] for i in range(df.shape[0]): data = [float(x.replace("[", "").replace("]", "")) for x in df['CostValues'][i].split(',')] lowest_cost.append(data[len(data) - 1]) lowest_cost = sorted(lowest_cost)[:10] for m in lowest_cost: for i in range(df.shape[0]): data = df['CostValues'][i].split(',') data = [float(x.replace("[", "").replace("]", "")) for x in data] if float(m) == float(data[len(data) - 1]): print("{} : Cost: {}".format(df["Spec"][i], data[len(data) - 1])) cost.append(data[len(data) - 1]) #df['CostValues'] = [] #df.to_excel("n.xlsx") # print(df.head(80)) check_lowest_cost(df) # Plotting all for i in range(df.shape[0]): fig = plt.figure() data = df['CostValues'][i].split(',') data = [float(x.replace("[", "").replace("]", "")) for x in data] plt.plot(range(0, len(data), 1), data) plt.xlabel('Steps') plt.ylabel('Cost value') plt.title(str(df['Spec'][i])) plt.show() path = '../../Output/Figures/{}.jpeg'.format(df['Spec'][i].replace("(", "").replace("FeatureMap", "").replace(")", "").replace(",", "").replace("=", "").replace(" ", "").replace("_", "")) # fig.savefig(path) def plot_individual_optimizers(Optimizers): for opt in Optimizers: max_iter = 0 fig = plt.figure(figsize=(20,20)) for i in range(df.shape[0]): if df['Spec'][i].__contains__(opt): data = [float(x.replace("[", "").replace("]", "")) for x in df['CostValues'][i].split(',')] if len(data) >= max_iter: max_iter = len(data) for i in range(df.shape[0]): if df['Spec'][i].__contains__(opt): data = df['CostValues'][i].split(',') data = [float(x.replace("[", "").replace("]", "")) for x in data] if max_iter >= len(data): temp = np.full((max_iter, ), data[len(data) -1]) temp[:len(data)] = data data = temp plt.plot(range(0, len(data), 1), data, label='{}'.format(df['Spec'][i].replace("(", "").replace("FeatureMap", "").replace(")", "").replace(",", "").replace("=", "").replace(" ", "").replace("_", ""))) plt.xlabel('Steps') plt.ylabel('Cost value') plt.title(opt) plt.legend() plt.show() plot_individual_optimizers(["COBYLA", "SPSA", "ADAM"]) def plot_increase_vdepth(optimizer, featuremap, reps, checking): points = [] for i in range(df.shape[0]): title = str(df['Spec'][i]) if title.split("(")[0] == featuremap and str(title).__contains__(optimizer) and str(title).__contains__("reps={}".format(reps)): points.append(float(df['CostValues'][i].split(',')[-1].replace("[", "").replace("]", ""))) fig = plt.figure() plt.plot(range(0, len(points), 1), points) plt.xlabel(checking) plt.ylabel('Cost value') plt.title("{} {} featuremap {} fdepth".format(optimizer, featuremap, reps)) plt.show() def plot_increase_fdepth(optimizer, featuremap, fdepth, checking): points = [] for i in range(df.shape[0]): title = str(df['Spec'][i]) if title.split("(")[0] == featuremap and str(title).__contains__(optimizer) and str(title).__contains__("vdepth {}".format(fdepth)): points.append(float(df['CostValues'][i].split(',')[-1].replace("[", "").replace("]", ""))) fig = plt.figure() plt.plot(range(0, len(points), 1), points) plt.xlabel(checking) plt.ylabel('Cost value') plt.title("{} {} featuremap {} vdepth".format(optimizer, featuremap, fdepth)) plt.show() plot_increase_vdepth("SPSA", "PauliFeatureMap", 1, "fdepth") plot_increase_fdepth("SPSA", "PauliFeatureMap", 1, "vdepth") for opt in ['SPSA', "COBYLA", "ADAM"]: for fmap in ['ZZFeatureMap', 'ZFeatureMap', "PauliFeatureMap"]: for fdepth in [1, 2, 4]: plot_increase_vdepth(opt, fmap, fdepth, "vdepth") for opt in ['SPSA', "COBYLA", "ADAM"]: for fmap in ['ZZFeatureMap', 'ZFeatureMap', "PauliFeatureMap"]: for fdepth in [1, 3, 5]: plot_increase_fdepth(opt, fmap, fdepth, "fdepth")
https://github.com/rodneyosodo/variational-quantum-classifier-on-heartattack
rodneyosodo
import pandas as pd import numpy as np import matplotlib.pyplot as plt pd.set_option('display.max_rows', None) pd.set_option('display.max_columns', None) pd.set_option('display.width', None) pd.set_option('display.max_colwidth', -1) df = pd.read_csv("../../Data/Processed/iriscost.csv") df.columns = ['Spec', "CostValues"] df.head(1) df['Spec'] def check_lowest_cost(df): """ Checks the configs which attained the lowest sost """ cost = [] lowest_cost = [] for i in range(df.shape[0]): data = [float(x.replace("[", "").replace("]", "")) for x in df['CostValues'][i].split(',')] lowest_cost.append(data[len(data) - 1]) lowest_cost = sorted(lowest_cost)[:10] for m in lowest_cost: for i in range(df.shape[0]): data = df['CostValues'][i].split(',') data = [float(x.replace("[", "").replace("]", "")) for x in data] if float(m) == float(data[len(data) - 1]): print("{} : Cost: {}".format(df["Spec"][i], data[len(data) - 1])) cost.append(data[len(data) - 1]) #df['CostValues'] = [] #df.to_excel("n.xlsx") # print(df.head(80)) check_lowest_cost(df) # Plotting all for i in range(df.shape[0]): fig = plt.figure() data = df['CostValues'][i].split(',') data = [float(x.replace("[", "").replace("]", "")) for x in data] plt.plot(range(0, len(data), 1), data) plt.xlabel('Steps') plt.ylabel('Cost value') plt.title(str(df['Spec'][i])) plt.show() path = '../../Output/Figures/{}.jpeg'.format(df['Spec'][i].replace("(", "").replace("FeatureMap", "").replace(")", "").replace(",", "").replace("=", "").replace(" ", "").replace("_", "")) # fig.savefig(path) def plot_individual_optimizers(Optimizers): for opt in Optimizers: max_iter = 0 fig = plt.figure(figsize=(20,20)) for i in range(df.shape[0]): if df['Spec'][i].__contains__(opt): data = [float(x.replace("[", "").replace("]", "")) for x in df['CostValues'][i].split(',')] if len(data) >= max_iter: max_iter = len(data) for i in range(df.shape[0]): if df['Spec'][i].__contains__(opt): data = df['CostValues'][i].split(',') data = [float(x.replace("[", "").replace("]", "")) for x in data] if max_iter >= len(data): temp = np.full((max_iter, ), data[len(data) -1]) temp[:len(data)] = data data = temp plt.plot(range(0, len(data), 1), data, label='{}'.format(df['Spec'][i].replace("(", "").replace("FeatureMap", "").replace(")", "").replace(",", "").replace("=", "").replace(" ", "").replace("_", ""))) plt.xlabel('Steps') plt.ylabel('Cost value') plt.title(opt) plt.legend() plt.show() plot_individual_optimizers(["COBYLA", "SPSA", "ADAM"]) def plot_increase_vdepth(optimizer, featuremap, reps, checking): points = [] for i in range(df.shape[0]): title = str(df['Spec'][i]) if title.split("(")[0] == featuremap and str(title).__contains__(optimizer) and str(title).__contains__("reps={}".format(reps)): points.append(float(df['CostValues'][i].split(',')[-1].replace("[", "").replace("]", ""))) fig = plt.figure() plt.plot(range(0, len(points), 1), points) plt.xlabel(checking) plt.ylabel('Cost value') plt.title("{} {} featuremap {} fdepth".format(optimizer, featuremap, reps)) plt.show() def plot_increase_fdepth(optimizer, featuremap, fdepth, checking): points = [] for i in range(df.shape[0]): title = str(df['Spec'][i]) if title.split("(")[0] == featuremap and str(title).__contains__(optimizer) and str(title).__contains__("vdepth {}".format(fdepth)): points.append(float(df['CostValues'][i].split(',')[-1].replace("[", "").replace("]", ""))) fig = plt.figure() plt.plot(range(0, len(points), 1), points) plt.xlabel(checking) plt.ylabel('Cost value') plt.title("{} {} featuremap {} vdepth".format(optimizer, featuremap, fdepth)) plt.show() plot_increase_vdepth("SPSA", "PauliFeatureMap", 1, "fdepth") plot_increase_fdepth("SPSA", "PauliFeatureMap", 1, "vdepth") for opt in ['SPSA', "COBYLA", "ADAM"]: for fmap in ['ZZFeatureMap', 'ZFeatureMap', "PauliFeatureMap"]: for fdepth in [1, 2, 4]: plot_increase_vdepth(opt, fmap, fdepth, "vdepth") for opt in ['SPSA', "COBYLA", "ADAM"]: for fmap in ['ZZFeatureMap', 'ZFeatureMap', "PauliFeatureMap"]: for fdepth in [1, 3, 5]: plot_increase_fdepth(opt, fmap, fdepth, "fdepth")
https://github.com/rodneyosodo/variational-quantum-classifier-on-heartattack
rodneyosodo
#!/usr/bin/env python # coding: utf-8 from qiskit import QuantumCircuit from qiskit.aqua.components.optimizers import COBYLA, ADAM, SPSA from qiskit.circuit.library import ZZFeatureMap, RealAmplitudes, ZFeatureMap, PauliFeatureMap from qiskit.quantum_info import Statevector import numpy as np import pandas as pd from sklearn.preprocessing import MinMaxScaler from sklearn.model_selection import train_test_split from sklearn.utils import shuffle import csv import warnings warnings.filterwarnings("ignore") class Benchmark: """ Benchmarking different optimizers, featuremaps and depth of variational circuits """ def __init__(self, optimizer, variational_depth, feature_map, X_train, X_test, Y_train, Y_test): """ Initial function :param optimizer: The optimizer to benchmark :param variational_depth: The depth of the variational circuit :param feature_map: The featuremap that encodes data :param X_train: The x data for training :param X_test: The x data for testing :param Y_train: The y data for training :param Y_test: The y data for testing """ self.optimizer = optimizer self.variational_depth = variational_depth self.feature_map = feature_map self.no_qubit = 4 self.random_state = 42 self.class_labels = ['yes', 'no'] self.circuit = None self.var_form = RealAmplitudes(self.no_qubit, reps=self.variational_depth) self.sv = Statevector.from_label('0' * self.no_qubit) self.X_train, self.X_test, self.Y_train, self.Y_test = X_train, X_test, Y_train, Y_test self.cost_list = [] def prepare_circuit(self): """ Prepares the circuit. Combines an encoding circuit, feature map, to a variational circuit, RealAmplitudes :return: """ self.circuit = self.feature_map.combine(self.var_form) # circuit.draw(output='mpl') def get_data_dict(self, params, x): """ Assign the params to the variational circuit and the data to the featuremap :param params: Parameter for training the variational circuit :param x: The data :return parameters: """ parameters = {} for i, p in enumerate(self.feature_map.ordered_parameters): parameters[p] = x[i] for i, p in enumerate(self.var_form.ordered_parameters): parameters[p] = params[i] return parameters def assign_label(self, bit_string): """ Based on the output from measurements assign no if it odd parity and yes if it is even parity :param bit_string: The bit string eg 00100 :return class_label: Yes or No """ hamming_weight = sum([int(k) for k in list(bit_string)]) is_odd_parity = hamming_weight & 1 if is_odd_parity: return self.class_labels[1] else: return self.class_labels[0] def return_probabilities(self, counts): """ Calculates the probabilities of the class label after assigning the label from the bit string measured as output :type counts: dict :param counts: The counts from the measurement of the quantum circuit :return result: The probability of each class """ shots = sum(counts.values()) result = {self.class_labels[0]: 0, self.class_labels[1]: 0} for key, item in counts.items(): label = self.assign_label(key) result[label] += counts[key] / shots return result def classify(self, x_list, params): """ Assigns the x and params to the quantum circuit the runs a measurement to return the probabilities of each class :type params: List :type x_list: List :param x_list: The x data :param params: Parameters for optimizing the variational circuit :return probs: The probabilities """ qc_list = [] for x in x_list: circ_ = self.circuit.assign_parameters(self.get_data_dict(params, x)) qc = self.sv.evolve(circ_) qc_list += [qc] probs = [] for qc in qc_list: counts = qc.to_counts() prob = self.return_probabilities(counts) probs += [prob] return probs @staticmethod def mse_cost(probs, expected_label): """ Calculates the mean squared error from the expected values and calculated values :type expected_label: List :type probs: List :param probs: The expected values :param expected_label: The real values :return mse: The mean squared error """ p = probs.get(expected_label) actual, pred = np.array(1), np.array(p) mse = np.square(np.subtract(actual, pred)).mean() return mse def cost_function(self, X, Y, params, print_value=False): """ This is the cost function and returns cost for optimization :type print_value: Boolean :type params: List :type Y: List :type X: List :param X: The x data :param Y: The label :param params: The parameters :param print_value: If you want values to be printed :return cost: """ # map training input to list of labels and list of samples cost = 0 training_labels = [] training_samples = [] for sample in X: training_samples += [sample] for label in Y: if label == 0: training_labels += [self.class_labels[0]] elif label == 1: training_labels += [self.class_labels[1]] probs = self.classify(training_samples, params) # evaluate costs for all classified samples for i, prob in enumerate(probs): cost += self.mse_cost(prob, training_labels[i]) cost /= len(training_samples) # print resulting objective function if print_value: print('%.4f' % cost) # return objective value self.cost_list.append(cost) return cost def test_model(self, X, Y, params): """ Test the model based on x test and y test :type params: List :type Y: List :type X: List :param X: The x test set :param Y: The y test set :param params: The parameters :return: """ accuracy = 0 training_samples = [] for sample in X: training_samples += [sample] probs = self.classify(training_samples, params) for i, prob in enumerate(probs): if (prob.get('yes') >= prob.get('no')) and (Y[i] == 0): accuracy += 1 elif (prob.get('no') >= prob.get('yes')) and (Y[i] == 1): accuracy += 1 accuracy /= len(Y) print("Test accuracy: {}".format(accuracy)) def run(self): """ Runs the whole code 1. Prepares the circuit 2. define the objective function 3. Initialize the paramters 4. Optimize the paramters by training the classifier :return: """ self.prepare_circuit() # define objective function for training objective_function = lambda params: self.cost_function(self.X_train, self.Y_train, params, print_value=False) # randomly initialize the parameters np.random.seed(self.random_state) init_params = 2 * np.pi * np.random.rand(self.no_qubit * self.variational_depth * 2) # train classifier opt_params, value, _ = self.optimizer.optimize(len(init_params), objective_function, initial_point=init_params) # print results # print() # print('opt_params:', opt_params) # print('opt_value: ', value) self.test_model(self.X_test, self.Y_test, opt_params) def get_cost_list(self): """ Return the cost list :return cost list: """ return self.cost_list def normalize_data(dataPath="../../Data/Processed/heartdata.csv"): """ Normalizes the data :return X_train, X_test, Y_train, Y_test: """ # Reads the data data = pd.read_csv(dataPath) data = shuffle(data, random_state=42) X, Y = data[['sex', 'cp', 'exang', 'oldpeak']].values, data['num'].values # normalize the data scaler = MinMaxScaler(feature_range=(-2 * np.pi, 2 * np.pi)) X = scaler.fit_transform(X) X_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size=0.3, random_state=42) return X_train, X_test, Y_train, Y_test def main(): data = {} feature_maps = ['ZZFeatureMap(4, reps=1)', 'ZZFeatureMap(4, reps=2)', 'ZZFeatureMap(4, reps=4)', 'ZFeatureMap(4, reps=1)', 'ZFeatureMap(4, reps=2)', 'ZFeatureMap(4, reps=4)', 'PauliFeatureMap(4, reps=1)', 'PauliFeatureMap(4, reps=2)', 'PauliFeatureMap(4, reps=4)'] optimizers = ["COBYLA(maxiter=50)", "SPSA(max_trials=50)", "ADAM(maxiter=50)"] x_train, x_test, y_train, y_test = normalize_data() for fe in feature_maps: for i in [1, 3, 5]: for opt in optimizers: print("FE: {}\tDepth: {}\tOpt: {}".format(fe, i, opt)) test_benchmark = Benchmark(optimizer=eval(opt), variational_depth=i, feature_map=eval(fe), X_train=x_train, X_test=x_test, Y_train=y_train, Y_test=y_test) test_benchmark.run() data_list = "{} {} vdepth {}".format(fe, opt, i) data[data_list] = test_benchmark.get_cost_list() w = csv.writer(open("../../Data/Processed/heartcosts.csv", "w")) for key, val in data.items(): w.writerow([key, val]) if __name__ == "__main__": main()
https://github.com/MuhammadMiqdadKhan/Quantum-Teleportation-Using-Qiskit-and-Real-Quantum-Computer-Tutorial
MuhammadMiqdadKhan
# Do the necessary imports import numpy as np from qiskit import QuantumCircuit, QuantumRegister, ClassicalRegister, execute, BasicAer, IBMQ from qiskit.visualization import plot_histogram, plot_bloch_multivector qr = QuantumRegister(3) # Protocol uses 3 qubits crz = ClassicalRegister(1) # and 2 classical bits crx = ClassicalRegister(1) # in 2 different registers teleportation_circuit = QuantumCircuit(qr, crz, crx) def create_bell_pair(qc, a, b): """Creates a bell pair in qc using qubits a & b""" qc.h(a) # Put qubit a into state |+> qc.cx(a,b) # CNOT with a as control and b as target # In our case, Eve entangles qubits q1 and q2 # Let's apply this to our circuit: create_bell_pair(teleportation_circuit, 1, 2) # And view the circuit so far: teleportation_circuit.draw() def alice_gates(qc, psi, a): qc.cx(psi, a) qc.h(psi) # Let's apply this to our circuit: teleportation_circuit.barrier() # Use barrier to separate steps alice_gates(teleportation_circuit, 0, 1) teleportation_circuit.draw() def measure_and_send(qc, a, b): """Measures qubits a & b and 'sends' the results to Bob""" qc.barrier() qc.measure(a,0) qc.measure(b,1) measure_and_send(teleportation_circuit, 0 ,1) teleportation_circuit.draw() # This function takes a QuantumCircuit (qc), integer (qubit) # and ClassicalRegisters (crz & crx) to decide which gates to apply def bob_gates(qc, qubit, crz, crx): # Here we use c_if to control our gates with a classical # bit instead of a qubit qc.z(qubit).c_if(crz, 1) # Apply gates if the registers qc.x(qubit).c_if(crx, 1) # are in the state '1' # Let's apply this to our circuit: teleportation_circuit.barrier() # Use barrier to separate steps bob_gates(teleportation_circuit, 2, crz, crx) teleportation_circuit.draw() from qiskit_textbook.tools import random_state, vector2latex # Create random 1-qubit state psi = random_state(1) # Display it nicely vector2latex(psi, pretext="|\\psi\\rangle =") # Show it on a Bloch sphere plot_bloch_multivector(psi) from qiskit.extensions import Initialize init_gate = Initialize(psi) qr = QuantumRegister(3) # Protocol uses 3 qubits crz = ClassicalRegister(1) # and 2 classical registers crx = ClassicalRegister(1) qc = QuantumCircuit(qr, crz, crx) # First, let's initialise Alice's q0 qc.append(init_gate, [0]) qc.barrier() # Now begins the teleportation protocol create_bell_pair(qc, 1, 2) qc.barrier() # Send q1 to Alice and q2 to Bob alice_gates(qc, 0, 1) # Alice then sends her classical bits to Bob measure_and_send(qc, 0, 1) # Bob decodes qubits bob_gates(qc, 2, crz, crx) qc.draw() backend = BasicAer.get_backend('statevector_simulator') out_vector = execute(qc, backend).result().get_statevector() plot_bloch_multivector(out_vector) inverse_init_gate = init_gate.gates_to_uncompute() qc.append(inverse_init_gate, [2]) qc.draw() # Need to add a new ClassicalRegister # to see the result cr_result = ClassicalRegister(1) qc.add_register(cr_result) qc.measure(2,2) qc.draw() backend = BasicAer.get_backend('qasm_simulator') counts = execute(qc, backend, shots=1024).result().get_counts() plot_histogram(counts) def bob_gates(qc, a, b, c): qc.cz(a, c) qc.cx(b, c) qc = QuantumCircuit(3,1) # First, let's initialise Alice's q0 qc.append(init_gate, [0]) qc.barrier() # Now begins the teleportation protocol create_bell_pair(qc, 1, 2) qc.barrier() # Send q1 to Alice and q2 to Bob alice_gates(qc, 0, 1) qc.barrier() # Alice sends classical bits to Bob bob_gates(qc, 0, 1, 2) # We undo the initialisation process qc.append(inverse_init_gate, [2]) # See the results, we only care about the state of qubit 2 qc.measure(2,0) # View the results: qc.draw() # First, see what devices we are allowed to use by loading our saved accounts IBMQ.load_account() provider = IBMQ.get_provider(hub='ibm-q') provider.backends() # get the least-busy backend at IBM and run the quantum circuit there from qiskit.providers.ibmq import least_busy backend = least_busy(provider.backends(filters=lambda b: b.configuration().n_qubits >= 3 and not b.configuration().simulator and b.status().operational==True)) job_exp = execute(qc, backend=backend, shots=8192) # Get the results and display them exp_result = job_exp.result() exp_measurement_result = exp_result.get_counts(qc) print(exp_measurement_result) plot_histogram(exp_measurement_result) error_rate_percent = sum([exp_measurement_result[result] for result in exp_measurement_result.keys() if result[0]=='1']) \ * 100./ sum(list(exp_measurement_result.values())) print("The experimental error rate : ", error_rate_percent, "%") import qiskit qiskit.__qiskit_version__
https://github.com/qiskit-community/qiskit-alt
qiskit-community
import subprocess def python_commands(commands): try: result = subprocess.run( ['python', '-c', commands], check=True, capture_output=True, encoding='utf8' ) except subprocess.CalledProcessError as err: return err return result def basic_inits(): all_coms = [] for _compile in ("False", "True"): for calljulia in ("pyjulia", "juliacall"): for depot in ("False", "True"): args = f"compile={_compile}, calljulia='{calljulia}', depot={depot}" coms = f"import qiskit_alt; qiskit_alt.project.ensure_init({args})" all_coms.append(coms) if calljulia == "pyjulia": other_calljulia = "juliacall" else: other_calljulia = "pyjulia" args = f"compile={_compile}, calljulia='{other_calljulia}', depot={depot}" coms = f"import qiskit_alt; qiskit_alt.project.ensure_init({args}); qiskit_alt.project.clean_all()" all_coms.append(coms) return all_coms # def basic_inits(): # all_coms = ["import sys", "import os", "import sdsdff", "import sdsdff", "import shutil"] # return all_coms def run_tests(all_commands=None, verbose=False): num_passed = 0 if all_commands is None: all_commands = basic_inits() for commands in all_commands: print(f"running '{commands}'") result = python_commands(commands) if isinstance(result, subprocess.CalledProcessError): print(f"**** Commands '{commands}' failed with error code {result}") print(result.stderr) else: num_passed += 1 if verbose: print(result) msg = f"{num_passed} of {len(all_commands)} installation tests passed" if num_passed < len(all_commands): print("**** " + msg) else: print(msg) if __name__ == '__main__': run_tests()
https://github.com/qiskit-community/qiskit-alt
qiskit-community
# Benchmark qiskit_alt constructing Fermionic operators from pyscf integrals. import qiskit_alt qiskit_alt.project.ensure_init() import timeit def make_setup_code(basis, geometry): return f""" import qiskit_alt h2_geometry = [['H', [0., 0., 0.]], ['H', [0., 0., 0.7414]]] h2o_geometry = [['O', [0., 0., 0.]], ['H', [0.757, 0.586, 0.]], ['H', [-0.757, 0.586, 0.]]] from qiskit_alt.electronic_structure import fermionic_hamiltonian fermionic_hamiltonian({geometry}, {basis}) #qiskit_alt.fermionic_hamiltonian({geometry}, {basis}) """ def run_one_basis(basis, geometry, num_repetitions): setup_code = make_setup_code(basis, geometry) bench_code = f"fermionic_hamiltonian({geometry}, {basis})" time = timeit.timeit(stmt=bench_code, setup=setup_code, number=num_repetitions) t = 1000 * time / num_repetitions print(f"geometry={geometry}, basis={basis} {t:0.2f}", "ms") return t def run_benchmarks(): alt_times = [] for basis, geometry, num_repetitions in (("'sto3g'", "h2_geometry", 10), ("'631g'", "h2_geometry", 10), ("'631++g'", "h2_geometry", 10), ("'sto3g'", "h2o_geometry", 10), ("'631g'", "h2o_geometry", 5)): t = run_one_basis(basis, geometry, num_repetitions) alt_times.append(t) return alt_times if __name__ == '__main__': alt_times = run_benchmarks()
https://github.com/qiskit-community/qiskit-alt
qiskit-community
# Benchmark qiskit_alt transforming an operator from the computational- to the Pauli basis. import qiskit_alt qiskit_alt.project.ensure_init() import timeit Main = qiskit_alt.project.julia.Main def make_setup_code(nqubits): return f""" import qiskit_alt from qiskit_alt.pauli_operators import PauliSum_to_SparsePauliOp QuantumOps = qiskit_alt.project.simple_import("QuantumOps") import numpy as np Main = qiskit_alt.project.julia.Main m = np.random.rand(2**{nqubits}, 2**{nqubits}) """ def run_one(nqubits, num_repetitions): setup_code = make_setup_code(nqubits) if qiskit_alt.project._calljulia_name == 'juliacall': bench_code = "PauliSum_to_SparsePauliOp(QuantumOps.PauliSum(Main.convert(Main.Matrix, m)))" else: bench_code = "PauliSum_to_SparsePauliOp(QuantumOps.PauliSum(m))" time = timeit.timeit(stmt=bench_code, setup=setup_code, number=num_repetitions) t = 1000 * time / num_repetitions print(f"nqubits={nqubits}, {t:0.2f}", "ms") return t def run_benchmarks(): qk_alt_times = [] for nqubits, num_repetitions in ((2, 50), (3, 50), (4, 10), (5, 10), (6, 10), (7, 10), (8, 3)): t = run_one(nqubits, num_repetitions) qk_alt_times.append(t) return qk_alt_times if __name__ == '__main__': qk_alt_times = run_benchmarks()
https://github.com/qiskit-community/qiskit-alt
qiskit-community
# Benchmark qiskit_alt peforming the Jordan-Wigner transform on a Fermi operator. import qiskit_alt qiskit_alt.project.ensure_init() import timeit def make_setup_code(basis, geometry): return f""" import qiskit_alt.electronic_structure h2_geometry = [['H', [0., 0., 0.]], ['H', [0., 0., 0.7414]]] h2o_geometry = [['O', [0., 0., 0.]], ['H', [0.757, 0.586, 0.]], ['H', [-0.757, 0.586, 0.]]] basis = {basis} fermi_op = qiskit_alt.electronic_structure.fermionic_hamiltonian({geometry}, basis) qiskit_alt.electronic_structure.jordan_wigner(fermi_op); """ def run_one_basis(basis, geometry, num_repetitions): setup_code = make_setup_code(basis, geometry) bench_code = "qiskit_alt.electronic_structure.jordan_wigner(fermi_op)" time = timeit.timeit(stmt=bench_code, setup=setup_code, number=num_repetitions) t = 1000 * time / num_repetitions print(f"geometry={geometry}, basis={basis} {t:0.2f}", "ms") return t def run_benchmarks(): alt_times = [] for basis, geometry, num_repetitions in (("'sto3g'", "h2_geometry", 10), ("'631g'", "h2_geometry", 10), ("'631++g'", "h2_geometry", 10), ("'sto3g'", "h2o_geometry", 10), ("'631g'", "h2o_geometry", 5)): t = run_one_basis(basis, geometry, num_repetitions) alt_times.append(t) return alt_times if __name__ == '__main__': alt_times = run_benchmarks()
https://github.com/qiskit-community/qiskit-alt
qiskit-community
# Benchmark qiskit_alt creating a SparsePauliOp from a list of strings. import sys import qiskit_alt qiskit_alt.project.ensure_init() import random from timeit import timeit Main = qiskit_alt.project.julia.Main QuantumOps = qiskit_alt.project.simple_import("QuantumOps") from qiskit_alt.pauli_operators import PauliSum_to_SparsePauliOp random.seed(123) def rand_label(k, n): return ["".join(random.choices("IXYZ", k=k)) for _ in range(n)] def run_benchmarks(): qkalt_times = [] for k in (10, 100): for n in (10, 100, 1000, 5000, 10_000, 100_000): label = rand_label(k, n) if qiskit_alt.project._calljulia_name == 'juliacall': label = Main.pyconvert_list(Main.String, label) PauliSum_to_SparsePauliOp(QuantumOps.PauliSum(label)) number = 20 t = timeit(lambda: PauliSum_to_SparsePauliOp(QuantumOps.PauliSum(label)), number=number) t = t * 1000 / number qkalt_times.append(t) print(f'k={k}, n={n}, {t} ms') return qkalt_times if __name__ == '__main__': qkalt_times = run_benchmarks()
https://github.com/qiskit-community/qiskit-alt
qiskit-community
import subprocess import shutil import os # Include these lines if we run all files in one process import qiskit_alt qiskit_alt.project.ensure_init() bench_scripts = [ "fermionic_alt_time.py", "fermionic_nature_time.py", "from_matrix_alt.py", "from_matrix_quantum_info.py", "jordan_wigner_alt_time.py", "jordan_wigner_nature_time.py", "pauli_from_list_alt.py", "pauli_from_list_qinfo.py", ] _python = shutil.which("python") ## Run each benchmark script in a separate process def run_bench(fname): dirname = os.path.dirname(os.path.abspath(__file__)) full = os.path.join(dirname, fname) res = subprocess.run( [_python, full], check=True, capture_output=True, encoding='utf8' ).stdout print(res) def exec_full_dir(fname): dirname = os.path.dirname(os.path.abspath(__file__)) filepath = os.path.join(dirname, fname) exec_full(filepath) def exec_full(filepath): global_namespace = { "__file__": filepath, "__name__": "__main__", } with open(filepath, 'rb') as file: exec(compile(file.read(), filepath, 'exec'), global_namespace) for fname in bench_scripts: print(fname) exec_full_dir(fname) print()
https://github.com/qiskit-community/qiskit-alt
qiskit-community
import subprocess import shutil import os # Include these lines if we run all files in one process import qiskit_alt qiskit_alt.project.ensure_init() bench_scripts = [ "fermionic_alt_time.py", "from_matrix_alt.py", "jordan_wigner_alt_time.py", "pauli_from_list_alt.py" ] _python = shutil.which("python") ## Run each benchmark script in a separate process def run_bench(fname): dirname = os.path.dirname(os.path.abspath(__file__)) full = os.path.join(dirname, fname) res = subprocess.run( [_python, full], check=True, capture_output=True, encoding='utf8' ).stdout print(res) def exec_full_dir(fname): dirname = os.path.dirname(os.path.abspath(__file__)) filepath = os.path.join(dirname, fname) exec_full(filepath) def exec_full(filepath): global_namespace = { "__file__": filepath, "__name__": "__main__", } with open(filepath, 'rb') as file: exec(compile(file.read(), filepath, 'exec'), global_namespace) def run_all(): for fname in bench_scripts: print(fname) exec_full_dir(fname) print() if __name__ == '__main__': run_all()
https://github.com/qiskit-community/qiskit-alt
qiskit-community
import qiskit_alt qiskit_alt.project.ensure_init() # The following line would also initialize qiskit_alt behind the scenes, but we have made it explicity above. # import qiskit_alt.electronic_structure from qiskit.providers import aer # make sure this is available from qiskit_nature.drivers import UnitsType, Molecule from qiskit_nature.drivers.second_quantization import ( ElectronicStructureDriverType, ElectronicStructureMoleculeDriver, ) from qiskit_nature.problems.second_quantization import ElectronicStructureProblem from qiskit_nature.converters.second_quantization import QubitConverter from qiskit_nature.mappers.second_quantization import JordanWignerMapper from qiskit.algorithms import VQE from qiskit.circuit.library import TwoLocal from qiskit.utils import QuantumInstance from qiskit import Aer from qiskit.algorithms.optimizers import COBYLA import qiskit_alt, qiskit_alt.electronic_structure # already done above, but this does not hurt from qiskit_alt.pauli_operators import PauliSum_to_SparsePauliOp from qiskit.opflow import PauliSumOp geometry = [["H", [0.0, 0.0, 0.0]], ["H", [0.0, 0.0, 0.735]]] basis="sto3g" molecule = Molecule( geometry=geometry, charge=0, multiplicity=1 ) driver = ElectronicStructureMoleculeDriver( molecule, basis=basis, driver_type=ElectronicStructureDriverType.PYSCF ) es_problem = ElectronicStructureProblem(driver) second_q_op= es_problem.second_q_ops() qubit_converter = QubitConverter(mapper=JordanWignerMapper()) qubit_op = qubit_converter.convert(second_q_op[0]) # Print the underlying data stored backing qubit_op qubit_op.primitive.simplify() # Compute the Fermionic operator of the molecule fermi_op = qiskit_alt.electronic_structure.fermionic_hamiltonian(geometry, basis) # Convert the Fermionic operator to a Pauli operator using the Jordan-Wigner transform pauli_op = qiskit_alt.electronic_structure.jordan_wigner(fermi_op) type(pauli_op) pauli_op.simplify() (pauli_op - qubit_op.primitive).simplify() fermi_op[1] # In qiskit_alt, the Fermi operator carries the identity term nuclear_energy_term = fermi_op[1].coeff nuclear_energy_term # Define the ansatz ansatz = TwoLocal( rotation_blocks=["h", "rx"], entanglement_blocks="cz", entanglement="full", reps=2, parameter_prefix="y", ) # Initialize the COBYLA optimizer optimizer = COBYLA(maxiter=500) # Select the backend for the quantum_instance backend = QuantumInstance(Aer.get_backend('statevector_simulator'), seed_simulator=100, seed_transpiler=100) # Run VQE algorithm vqe = VQE(ansatz = ansatz, optimizer = optimizer, quantum_instance = backend) # Compute the ground-state energy of the molecule nature_result = vqe.compute_minimum_eigenvalue(operator=qubit_op) nature_energy = nature_result.eigenvalue.real + nuclear_energy_term # Include constant term that is normally thrown away print("The ground-state energy of the Hydrogen molecule is {} Hatree".format(round(nature_energy,3))) # Convert the Pauli operator into a sum of Pauli operators, the required input format pauli_sum_op = PauliSumOp(pauli_op) # Compute the ground-state energy of the molecule alt_result = vqe.compute_minimum_eigenvalue(operator=pauli_sum_op) print("The ground-state energy of the Hydrogen molecule is {} Hatree".format(round(alt_result.eigenvalue.real,3)))
https://github.com/qiskit-community/qiskit-alt
qiskit-community
import qiskit_alt qiskit_alt.project.ensure_init(calljulia="pyjulia", compile=False, depot=True) julia = qiskit_alt.project.julia julia.Main.zeros(10) from qiskit_nature.drivers import UnitsType, Molecule from qiskit_nature.drivers.second_quantization import ElectronicStructureDriverType, ElectronicStructureMoleculeDriver # Specify the geometry of the H_2 molecule geometry = [['H', [0., 0., 0.]], ['H', [0., 0., 0.735]]] basis = 'sto3g' molecule = Molecule(geometry=geometry, charge=0, multiplicity=1) driver = ElectronicStructureMoleculeDriver(molecule, basis=basis, driver_type=ElectronicStructureDriverType.PYSCF) from qiskit_nature.problems.second_quantization import ElectronicStructureProblem from qiskit_nature.converters.second_quantization import QubitConverter from qiskit_nature.mappers.second_quantization import JordanWignerMapper es_problem = ElectronicStructureProblem(driver) second_q_op = es_problem.second_q_ops() fermionic_hamiltonian = second_q_op[0] qubit_converter = QubitConverter(mapper=JordanWignerMapper()) nature_qubit_op = qubit_converter.convert(fermionic_hamiltonian) nature_qubit_op.primitive import qiskit_alt.electronic_structure fermi_op = qiskit_alt.electronic_structure.fermionic_hamiltonian(geometry, basis) pauli_op = qiskit_alt.electronic_structure.jordan_wigner(fermi_op) pauli_op.simplify() # The Julia Pauli operators use a different sorting convention; we sort again for comparison. %run ../bench/jordan_wigner_nature_time.py nature_times %run ../bench/jordan_wigner_alt_time.py alt_times [t_nature / t_qk_alt for t_nature, t_qk_alt in zip(nature_times, alt_times)] %run ../bench/fermionic_nature_time.py nature_times %run ../bench/fermionic_alt_time.py alt_times [t_nature / t_qk_alt for t_nature, t_qk_alt in zip(nature_times, alt_times)] from qiskit_alt.pauli_operators import QuantumOps, PauliSum_to_SparsePauliOp import numpy as np m = np.random.rand(2**3, 2**3) # 3-qubit operator pauli_sum = QuantumOps.PauliSum(m) # This is a wrapped Julia object PauliSum_to_SparsePauliOp(pauli_sum) # Convert to qiskit.quantum_info.SparsePauliOp %run ../bench/from_matrix_quantum_info.py %run ../bench/from_matrix_alt.py [t_pyqk / t_qk_alt for t_pyqk, t_qk_alt in zip(pyqk_times, qk_alt_times)] %run ../bench/pauli_from_list_qinfo.py %run ../bench/pauli_from_list_alt.py [x / y for x,y in zip(quantum_info_times, qkalt_times)] %load_ext julia.magic %julia using Random: randstring %julia pauli_strings = [randstring("IXYZ", 10) for _ in 1:1000] None; %julia import Pkg; Pkg.add("BenchmarkTools") %julia using BenchmarkTools: @btime %julia @btime QuantumOps.PauliSum($pauli_strings) None; %julia pauli_sum = QuantumOps.PauliSum(pauli_strings); %julia println(length(pauli_sum)) %julia println(pauli_sum[1]) 6.9 * 2.29 / .343 # Ratio of time to construct PauliSum via qiskit_alt to time in pure Julia import qiskit.tools.jupyter d = qiskit.__qiskit_version__._version_dict d['qiskit_alt'] = qiskit_alt.__version__ %qiskit_version_table
https://github.com/qiskit-community/qiskit-alt
qiskit-community
import qiskit_alt qiskit_alt.project.ensure_init(calljulia="juliacall", compile=False) julia = qiskit_alt.project.julia Main = julia.Main julia.Main.zeros(3) type(julia.Main.zeros(3)) from qiskit_nature.drivers import UnitsType, Molecule from qiskit_nature.drivers.second_quantization import ElectronicStructureDriverType, ElectronicStructureMoleculeDriver # Specify the geometry of the H_2 molecule geometry = [['H', [0., 0., 0.]], ['H', [0., 0., 0.735]]] basis = 'sto3g' molecule = Molecule(geometry=geometry, charge=0, multiplicity=1) driver = ElectronicStructureMoleculeDriver(molecule, basis=basis, driver_type=ElectronicStructureDriverType.PYSCF) from qiskit_nature.problems.second_quantization import ElectronicStructureProblem from qiskit_nature.converters.second_quantization import QubitConverter from qiskit_nature.mappers.second_quantization import JordanWignerMapper es_problem = ElectronicStructureProblem(driver) second_q_op = es_problem.second_q_ops() fermionic_hamiltonian = second_q_op[0] qubit_converter = QubitConverter(mapper=JordanWignerMapper()) nature_qubit_op = qubit_converter.convert(fermionic_hamiltonian) nature_qubit_op.primitive import qiskit_alt.electronic_structure fermi_op = qiskit_alt.electronic_structure.fermionic_hamiltonian(geometry, basis) pauli_op = qiskit_alt.electronic_structure.jordan_wigner(fermi_op) pauli_op.simplify() # The Julia Pauli operators use a different sorting convention; we sort again for comparison. %run ../bench/jordan_wigner_nature_time.py nature_times %run ../bench/jordan_wigner_alt_time.py alt_times [t_nature / t_qk_alt for t_nature, t_qk_alt in zip(nature_times, alt_times)] %run ../bench/fermionic_nature_time.py nature_times %run ../bench/fermionic_alt_time.py alt_times [t_nature / t_qk_alt for t_nature, t_qk_alt in zip(nature_times, alt_times)] from qiskit_alt.pauli_operators import QuantumOps, PauliSum_to_SparsePauliOp import numpy as np m = np.random.rand(2**3, 2**3) # 3-qubit operator m = Main.convert(Main.Matrix, m) # Convert PythonCall.PyArray to a native Julia type pauli_sum = QuantumOps.PauliSum(m) # This is a wrapped Julia object PauliSum_to_SparsePauliOp(pauli_sum) # Convert to qiskit.quantum_info.SparsePauliOp %run ../bench/from_matrix_quantum_info.py %run ../bench/from_matrix_alt.py [t_pyqk / t_qk_alt for t_pyqk, t_qk_alt in zip(pyqk_times, qk_alt_times)] %run ../bench/pauli_from_list_qinfo.py %run ../bench/pauli_from_list_alt.py [x / y for x,y in zip(quantum_info_times, qkalt_times)] import qiskit.tools.jupyter d = qiskit.__qiskit_version__._version_dict d['qiskit_alt'] = qiskit_alt.__version__ %qiskit_version_table # %load_ext julia.magic # %julia using Random: randstring #%julia pauli_strings = [randstring("IXYZ", 10) for _ in 1:1000] # None; # %julia import Pkg; Pkg.add("BenchmarkTools") #%julia using BenchmarkTools: @btime #%julia @btime QuantumOps.PauliSum($pauli_strings) #None; #%julia pauli_sum = QuantumOps.PauliSum(pauli_strings); #%julia println(length(pauli_sum)) #%julia println(pauli_sum[1]) #6.9 * 2.29 / .343 # Ratio of time to construct PauliSum via qiskit_alt to time in pure Julia
https://github.com/qiskit-community/qiskit-alt
qiskit-community
from qiskit.algorithms import VQE from qiskit.algorithms.optimizers import COBYLA from qiskit.opflow import X, Y, Z, I, PauliSumOp from qiskit.circuit.library import TwoLocal from qiskit import Aer import numpy as np import qiskit_alt, qiskit_alt.electronic_structure from qiskit_alt.pauli_operators import PauliSum_to_SparsePauliOp from qiskit.opflow import PauliSumOp # Describe the Hydrogen molecule by fixing the location of each nucleus geometry = [['H', [0., 0., 0.]], ['H', [0., 0., 0.739]]] basis = 'sto3g' # Compute the Fermionic operator of the molecule fermi_op = qiskit_alt.electronic_structure.fermionic_hamiltonian(geometry, basis) # Convert the Fermionic operator to a Pauli operator using the Jordan-Wigner transform pauli_op = qiskit_alt.electronic_structure.jordan_wigner(fermi_op); # Convert the Pauli operator into a sum of Pauli operators # input to the VQE algorithm to compute the minimum eigenvalue pauli_sum_op = PauliSumOp(pauli_op) print(pauli_sum_op) # Set up the ansatz of type TwoLOcal ansatz = TwoLocal(num_qubits=4, rotation_blocks=['ry'], entanglement_blocks='cx', reps=1, entanglement='linear', skip_final_rotation_layer= False) # Initialize the COBYLA optimizer optimizer = COBYLA(maxiter=500) # Select the backend for the quantum_instance backend = Aer.get_backend('aer_simulator_statevector') # Run VQE algorithm vqe = VQE(ansatz = ansatz, optimizer = optimizer, quantum_instance = backend) # Compute the ground-state energy of the molecule result = vqe.compute_minimum_eigenvalue(operator=pauli_sum_op) print("The ground-state energy of the Hydrogen molecule is {} Hatree".format(round(result.eigenvalue.real,3)))
https://github.com/qiskit-community/qiskit-alt
qiskit-community
import qiskit_alt geometry = [['H', [0., 0., 0.]], ['H', [0., 0., 0.7414]]] #basis = 'sto3g' basis = '631++g' fermi_op = qiskit_alt.fermionic_hamiltonian(geometry, basis) pauli_op = qiskit_alt.jordan_wigner(fermi_op) #basis = '631g' #basis = 'dzvp' # Too big #basis = 'dzp' #basis = 'dzvp2'
https://github.com/qiskit-community/qiskit-alt
qiskit-community
import qiskit_alt def do_jw_problem(): # qiskit_alt.Main.eval('include("examples/jw_example.jl")') qiskit_alt.Main.eval('include("jw_example.jl")') pauli_op = qiskit_alt.Main.eval("pauli_op") spop_jl = qiskit_alt.QiskitQuantumInfo.SparsePauliOp(pauli_op) spop = qiskit_alt.jlSparsePauliOp(spop_jl) return spop
https://github.com/qiskit-community/qiskit-alt
qiskit-community
import qiskit_alt #geometry = [['H', [0., 0., 0.]], ['H', [0., 0., 0.7414]]] geometry = [['O', [0., 0., 0.]], ['H', [0.757, 0.586, 0.]], ['H', [-0.757, 0.586, 0.]]] #basis = 'sto3g' #basis = '631g' basis = 'dzvp2' pauli_op = qiskit_alt.jordan_wigner(geometry, basis)
https://github.com/qiskit-community/qiskit-alt
qiskit-community
import pytest import qiskit_alt project = qiskit_alt.project project.ensure_init() # calljulia="pyjulia" import qiskit_alt.electronic_structure Main = project.julia.Main def test_always_passes(): assert True def test_interface_lib(): assert qiskit_alt.project.julia.__name__ == 'julia' def test_import_QuantumOps(): project.simple_import("QuantumOps") assert True def test_import_ElectronicStructure(): project.simple_import("ElectronicStructure") assert True def test_import_QiskitQuantumInfo(): project.simple_import("QiskitQuantumInfo") assert True @pytest.fixture def conv_geometry(): geometry = [['H', [0., 0., 0.]], ['H', [0., 0., 0.7414]]] # geometry = [['O', [0., 0., 0.]], # ['H', [0.757, 0.586, 0.]], # ['H', [-0.757, 0.586, 0.]]] return qiskit_alt.electronic_structure.Geometry(geometry) def test_Geometry_length(conv_geometry): assert Main.length(conv_geometry) == 2 def test_Geometry_atom(conv_geometry): atom = Main.getindex(conv_geometry, 1) assert atom.coords == (0.0, 0.0, 0.0) def test_fermionic_hamiltonian(conv_geometry): fermi_op = Main.fermionic_hamiltonian(conv_geometry, "sto3g") assert Main.length(fermi_op) == 25
https://github.com/qiskit-community/qiskit-alt
qiskit-community
import pytest import qiskit_alt project = qiskit_alt.project project.ensure_init(calljulia="juliacall") def test_always_passes(): assert True def test_interface_lib(): assert qiskit_alt.project.julia.__name__ == 'juliacall' def test_Main(): Main = qiskit_alt.project.julia.Main assert Main.sind(90) == 1.0
https://github.com/BryceFuller/qiskit_camp
BryceFuller
# Python import logging from copy import deepcopy from functools import partial from contextlib import contextmanager import time from multideterminant_prep import PrepareMultiDeterminantState as pmds # Local # External import pandas as pd import numpy as np import xarray as xr import xyzpy as xy from qiskit import QuantumCircuit, QuantumRegister, ClassicalRegister, execute, Aer, tools from qiskit.aqua.components.optimizers import SPSA, SLSQP def build_model_from_parameters(circuit, theta): layers, num_qubits, rotations = theta.shape assert rotations == 3, "Expected a 3-sized vector for each theta." # Combine all the qubits in the circuit into one flat list all_qubits = [(reg, qubit) for register in circuit.qregs for reg, qubit in register] assert len(all_qubits) == num_qubits, \ "There are {} qubits but expected same as parameter's shape ({})!".format(len(all_qubits), num_qubits) for l in range(layers): # Add layer of (pseudo)random single qubit rotations for n, (reg, qubit_idx) in enumerate(all_qubits): circuit.u3(theta[l][n][0], theta[l][n][1], theta[l][n][2], reg[qubit_idx]) # Pattern with 2-qubit gates # change the step size in range to affect density of cx layer for n in range(0, num_qubits, 2): if (n + 1) + (l % 2) < num_qubits: # Out-of-bounds check reg_1, qubit_idx_1 = all_qubits[n + (l % 2)] reg_2, qubit_idx_2 = all_qubits[(n + 1) + (l % 2)] circuit.cx(reg_1[qubit_idx_1], reg_2[qubit_idx_2]) def build_compression_model(registers, model_parameters): """ Given a set of input registers, builds a parametrised model of random gate operations which seek to approximate some other (equally-sized) target circuit. Returns: QuantumCircuit: a circuit consisting of random operations on the given registers. """ model_circuit = QuantumCircuit(*registers) # Reshape model parameters so they make sense n_layers = int(len(model_parameters) / 3 / model_circuit.width()) model_parameters = np.array(model_parameters).reshape((n_layers, model_circuit.width(), 3)) # TODO: Store some internal parametrization that can be optimised over. # TODO: Create random set of gate operations based on parametrization. build_model_from_parameters(model_circuit, model_parameters) return model_circuit def swap_test_with_compression_model(target_circuit, model_parameters): """ Given a circuit, builds a parametrised sub-circuit that runs in parallel with and approximately models (with compression) the original circuit. Return: QuantumCircuit: new combined circuit consisting of the original circuit and the model circuit. """ # We assume all quantum registers of the original circuit are to be # modelled. new_registers = [reg.__class__(reg.size, '{}_model'.format(reg.name)) for reg in target_circuit.qregs] # TODO: Build the model's compression circuit here. model_circuit = build_compression_model(new_registers, model_parameters) # Append the two circuits together. top_circuit = target_circuit + model_circuit # Synchronisation barrier just so we know where the original circuits ends. top_circuit.barrier() # Performs a swap test between the compression model and original circuit. # ------------------------------------------------------------------------ # Firstly, create an ancillary in superposition to store the swap test result. swap_tester = QuantumRegister(1, 'swap_tester') top_circuit.add_register(swap_tester) top_circuit.h(swap_tester) # Next, we perform controlled SWAPs using the swap tester. for orig_reg, model_reg in zip(target_circuit.qregs, model_circuit.qregs): for i in range(orig_reg.size): top_circuit.cswap(swap_tester[0], orig_reg[i], model_reg[i]) # Finally, we re-interfere the branches and measure the swap tester. top_circuit.h(swap_tester) top_circuit.x(swap_tester) # Make it so test measurement is P(model == target). swap_test_result = ClassicalRegister(1, 'swap_test_result') top_circuit.add_register(swap_test_result) top_circuit.measure(swap_tester, swap_test_result) return top_circuit def compute_approximation_fidelity(target_circuit, backend='qasm_simulator', n_shots=1000, results_fidelity_list=[], model_parameters=[]): """ Performs a set of runs on the target circuit and an approximation model to compute the approximation's fidelity. Returns: float: a measure of fidelity in [0, 1], where 1 represents the maximum achievable approximation of the target circuit. """ final_circuit = swap_test_with_compression_model(target_circuit, model_parameters) # Execute the SWAP test circuit simulator = Aer.get_backend(backend) job = execute(final_circuit, backend=simulator, shots=n_shots) # Return a measurement of fidelity result_counts = job.result().get_counts(final_circuit) fidelity = result_counts.get('1', n_shots) / n_shots results_fidelity_list.append(fidelity) logging.critical("Fidelity: {}".format(fidelity)) return fidelity def cross_validate_qnn_depth(target_circuit, n_shots, n_iters, n_layers, run=0): """ Runs a single cross-validation experiment with the given parameters. Returns: xr.Dataset: an xarray Dataset consisting of any number of DataArrays of data regarding the results of this experiment. This Dataset will be merged with all other experiment datasets, so in theory every experiment should return a fairly consistent set of data. """ # TODO: Compile the circuit to determine a measure of optimised circuit depth # tools.compiler.compile(target_circuit, backend) # logging.critical("Circuit depth (compiled): {}".format(compiled_depth)) # Configuration n_params = target_circuit.width() * n_layers * 3 # Build variable bounds variable_bounds_single = (0., 2*np.pi) variable_bounds = [variable_bounds_single] * n_params initial_point = np.random.uniform(low=variable_bounds_single[0], high=variable_bounds_single[1], size=(n_params,)).tolist() # logging.critical("Initial point: {}".format(initial_point)) # Store resulting information results_fidelity_list = [] # Partially define the objective function maximisation_function = partial(compute_approximation_fidelity, target_circuit, "qasm_simulator", n_shots, results_fidelity_list) minimization_function = lambda params: -maximisation_function(params) # Call the optimiser optimizer = SPSA(max_trials=n_iters, save_steps=1) result = optimizer.optimize(n_params, minimization_function, variable_bounds=variable_bounds, initial_point=initial_point) last_params, last_score, _ = result logging.critical("FINAL SCORE: {}".format(-last_score)) # Ignore the first set of fidelities (calibration) and very last one (equal to last_score) results_fidelity_list = results_fidelity_list[-((n_iters * 2) + 1):-1] # TODO calculate compiled depth (for simulator this is always 2*l) # Output results return xr.Dataset({ "fidelity": xr.DataArray(np.array(results_fidelity_list).reshape((n_iters, 2)), coords={"iteration": range(n_iters), "plusminus": range(2)}, dims=["iteration", "plusminus"]), "last_theta": xr.DataArray(np.array(last_params).reshape((n_layers, target_circuit.width(), 3)), coords={"layer": range(n_layers), "qubit": range(target_circuit.width()), "angle": ["theta", "phi", "lambda"]}, dims=["layer", "qubit", "angle"]), "uncompiled_target_depth": xr.DataArray(target_circuit.depth()), # "compiled_target_depth": xr.DataArray(compiled_depth), "uncompiled_model_depth": xr.DataArray(2 * n_layers), }) @contextmanager def experiment_crop(fn, experiment_name): """ Defines how to run a crop of experiments (i.e. all the experiments in the grid) in parallel and store results in a file. """ experiment_runner = xy.Runner(fn, var_names=None) timestr = time.strftime("%Y%m%d-%H%M%S") filename = "experiment_results/{}_{}.h5".format(experiment_name, timestr) logging.critical("Running experiments: {}".format(filename)) experiment_harvester = xy.Harvester(experiment_runner, data_name=filename) experiment = experiment_harvester.Crop() yield experiment experiment.grow_missing(parallel=True) results = experiment.reap(wait=True, overwrite=True) logging.critical(results) logging.critical("Saved experiments: {}".format(filename)) if __name__ == "__main__": """ Example which defines a circuit and then runs a set of experiments, defined by the grid_search parameters, across a parallel set of processes. """ logging.critical("Creating the circuit...") in_strings = ["0101", "0011"] in_weights = [4, 7] target_circuit = pmds(in_weights, in_strings, mode='noancilla') logging.critical("Circuit depth (uncompiled): {}".format(target_circuit.depth())) logging.critical("Running the experiments...") with experiment_crop(cross_validate_qnn_depth, "experiments") as experiment: grid_search = { 'n_shots': [1000], 'n_iters': [200], 'n_layers': [2, 3, 4, 5], 'run': range(50), } constants = { 'target_circuit': target_circuit } experiment.sow_combos(grid_search, constants=constants)
https://github.com/BryceFuller/qiskit_camp
BryceFuller
#!/usr/bin/env python # coding: utf-8 # Python import math # Local # External import numpy as np from qiskit import QuantumCircuit, QuantumRegister import qiskit.aqua def PrepareMultiDeterminantState(weights, determinants, normalize=True, mode='basic'): ''' Constructs multideterminant state : a linear superposition of many fermionic determinants. Appropriate for us as an input state for VQE. Input: np.array of REAL weights, and list of determinants. Each determinant is a bitstring of length n. Output: a QuantumCircuit object. If mode == basic, acting on 2n qubits (n logical, n ancillas). Otherwise, acting on n+1 qubits. The depth of the circuit goes like O(n L) where L = len(weights), in basic mode, and exponentially in n when only 1 ancilla. This is an implementation of the circuit in https://arxiv.org/abs/1809.05523 ''' # number of determinants in the state L = len(weights) if L != len(determinants): raise Exception('# weights != # determinants') # Check normalization norm = np.linalg.norm(weights) # print('Norm is: {}'.format(norm)) if normalize: weights /= norm elif abs(norm-1) > 10**-5: raise Exception('Weights do not produce a normalized vector.') # TODO: check that weights are normalized # number of orbitals that determinants can be made of n = len(determinants[0]) # the first n qubits will be used for orbitals, the next as the controlled rotation ancilla, # the last n as ancillas will be used for the n-Toffoli qubits = QuantumRegister(n, 'q') cqub = QuantumRegister(1, 'c') if mode == 'noancilla': # no ancillas registers = [qubits, cqub] else: ancillas = QuantumRegister(n-1, 'a') registers = [qubits, cqub, ancillas] # initialize circ = QuantumCircuit(*registers) for i in range(n): if determinants[0][i] == 1: circ.x(qubits[i]) circ.x(cqub) b = 1 # the beta a = 0 # iterate over all determinants that must be in the final state for step in range(L-1): # choose first qubit on which D_l and D_{l+1} differ old_det = determinants[step] new_det = determinants[step+1] a = weights[step] # Find first bit of difference different_qubit = 0 while different_qubit <= n: if old_det[different_qubit] != new_det[different_qubit]: break else: different_qubit += 1 if different_qubit == n: raise Exception('Determinants {} and {} are the same.'.format(old_det, new_det)) # Compute the rotation angle # Equation is cos(g) beta_l = alpha_l angle = math.acos(a/b) # print('Step {} angle is {}'.format(step, angle)) # beta_{l+1} is b = b * math.sin(angle) if old_det[different_qubit] == 1: b = -b # print('Flipped beta sign.') # want a controlled-Y rotation, but can do controlled-Z, so map Y basis to Z basis. # 1) first map Y to X (with U1(-pi/2) gate) # 2) X to Z, with Hadamard # 3) apply z-rotation # 4) undo this # --- # circ.u1(-math.pi/2, qubits[different_qubit]) # circ.h(qubits[different_qubit]) # circ.crz(angle, cqub, qubits[different_qubit]) # circ.h(qubits[different_qubit]) # circ.u1(-math.pi/2, qubits[different_qubit]) circ.cu3(2 * angle, 0, 0, cqub, qubits[different_qubit]) # Now must do an n-qubit Toffoli to change the |1> to |0> # but controlled on the |D_l> bitstring flip_all(circ, qubits, old_det) if mode == 'noancilla': # no ancillas circ.mct(qubits, cqub[0], None, mode='noancilla') else: circ.mct(qubits, cqub[0], ancillas, mode='basic') # nToffoliAncillas(circ, qubits, ancillas, cqub) flip_all(circ, qubits, old_det) # undo # if step > 0: # break # If b not same sign as weight, flip sign # if b * weights[step+1]: # circ.z(cqub) # b = -b # Now continue flipping the rest of the bits for i in range(different_qubit+1, n): if old_det[i] != new_det[i]: circ.cx(cqub, qubits[i]) circ.barrier() # Finally check that the sign of the last weight is correct # and set the ancilla to zero. if b * weights[-1] < 0: # must flip sign circ.z(cqub) # Remove the |1> flip_all(circ, qubits, new_det) if mode == 'noancilla': # no ancillas circ.mct(qubits, cqub[0], None, mode='noancilla') else: circ.mct(qubits, cqub[0], ancillas, mode='basic') # nToffoliAncillas(circ, qubits, ancillas, cqub) flip_all(circ, qubits, new_det) # undo return circ def flip_all(circ, tgt, bits): ''' Flips the qubits of tgt where the bit in the bitstring is 0. ''' for i in range(len(bits)): if bits[i] == 0: # flip circ.x(tgt[i]) def nToffoliAncillas(circ, ctrl, anc, tgt): ''' Returns a circuit that implements an n-qubit Toffoli using n ancillas. ''' n = len(ctrl) # compute circ.ccx(ctrl[0], ctrl[1], anc[0]) for i in range(2, n): circ.ccx(ctrl[i], anc[i-2], anc[i-1]) # copy circ.cx(anc[n-2], tgt[0]) # uncompute for i in range(n-1, 1, -1): circ.ccx(ctrl[i], anc[i-2], anc[i-1]) circ.ccx(ctrl[0], ctrl[1], anc[0]) return circ
https://github.com/BryceFuller/qiskit_camp
BryceFuller
%matplotlib inline import xarray as xr import seaborn as sns import matplotlib.pyplot as plt get_ipython().run_line_magic('pylab', 'inline') #sns.set_style("darkgrid") pylab.rcParams['figure.figsize'] = (16, 8) pylab.rcParams.update({'font.size': 10}) df = xr.open_dataset("../experiment_results/experiments_20190301-052407.h5").squeeze() df sns.lineplot(x='iteration', y='fidelity', hue='n_layers', data=df.fidelity.sel(n_shots=1000).to_dataframe().reset_index()) plt.legend(df.n_layers.values) plt.title("Fidelity improvement with optimiser iterations. (Two determinants with four qubits)") df.fidelity.max(dim="iteration").mean(dim='run')#.mean(dim='n_layers') sns.boxenplot(orient='h', x='fidelity', y='n_layers', data=df.fidelity.max(dim='iteration').squeeze().to_dataframe().reset_index()) df.fidelity.max(dim='iteration').squeeze()
https://github.com/BryceFuller/qiskit_camp
BryceFuller
%matplotlib inline import sys sys.path.append("..") import compress from qiskit import QuantumRegister, ClassicalRegister, QuantumCircuit, execute, Aer import numpy as np from functools import partial from qiskit.aqua.components.optimizers import SPSA from multideterminant_prep import PrepareMultiDeterminantState as pmds import seaborn as sns import matplotlib.pyplot as plt get_ipython().run_line_magic('pylab', 'inline') pylab.rcParams['figure.figsize'] = (12, 6) pylab.rcParams.update({'font.size': 10}) in_strings = ["101", "011", "110"] in_weights = [4, 7, 2] circuit = pmds(in_weights, in_strings) circuit.draw(scale=0.5, style={'fold':100}, output='mpl') # Configuration n_shots = 1000 n_trials = 300 n_layers = 3 n_params = circuit.width() * n_layers * 3 # Build variable bounds variable_bounds_single = (0., 2*np.pi) variable_bounds = [variable_bounds_single] * n_params initial_point = np.random.uniform(low=variable_bounds_single[0], high=variable_bounds_single[1], size=(n_params,)).tolist() circuit.draw(scale=0.5, style={'fold':100}, output='mpl') print("Circuit has a depth of: {}".format(circuit.depth())) final_circuit = compress.swap_test_with_compression_model(circuit, initial_point) final_circuit.draw(scale=0.5, style={'fold':100}, output='mpl') results = compress.cross_validate_qnn_depth(circuit, n_shots, n_trials, n_layers, run=0) results sns.lineplot(x='iteration', y='fidelity', data=results.fidelity.to_dataframe().reset_index()) plt.title("Fidelity improvement with optimiser iterations")
https://github.com/BryceFuller/qiskit_camp
BryceFuller
# Python import logging # Local import compress # External import pytest import numpy as np from qiskit import QuantumRegister, QuantumCircuit def test_compression_model(): # Create a simple circuit that applies one simple flip. q0 = QuantumRegister(1, 'q0') circuit = QuantumCircuit(q0) fidelity = compress.compute_approximation_fidelity(circuit) assert fidelity == 1.0, "Fidelity should be 100% for the same circuit!" circuit.h(q0) fidelity = compress.compute_approximation_fidelity(circuit) assert np.isclose(fidelity, 0.75, rtol=1e-1) # XXX: Very loose test :3 # Optimise a random compression circuit and ensure it converges to fidelity=1.0 results = compress.cross_validate_qnn_depth(circuit, n_shots=100, n_iters=30, n_layers=3) assert np.isclose(results.fidelity.isel(iteration=-1).max(dim='plusminus').squeeze(), 1.0, rtol=1e-1)
https://github.com/HayleySummer/Qiskit_Hackathon_Europe
HayleySummer
import numpy from PIL import Image from IPython.display import Image as dispImage #Generate initial image data = numpy.zeros((8, 8, 3), dtype=numpy.uint8) white = [255, 255, 255] grey = [128, 128, 128] black = [0,0,0] boardRows = 8 boardColumns = 8 for i in range(boardRows): for j in range(boardColumns): if (i == 3 and j == 3) or (i == 4 and j == 4): data[i, j] = white elif (i == 3 and j == 4) or (i == 4 and j == 3): data[i, j] = black else: data[i, j] = grey image = Image.fromarray(data) image.show() image = image.save("prueba.jpg") dispImage("prueba.jpg", width=100, height=100) #Path to the file 15x10_w.txt with open('/content/drive/MyDrive/Colab Notebooks/Hackathon/Board_To_Image/Data/Training Set/15x10_w.txt') as f: lines = f.readlines() f.close() auxParams = lines[0].split("\n") params = auxParams[0].split(" ") print(params) #0 5 4 ==> #First 0 == black - 1 == white #Second parameter column index (from 0 to 7) #Third parameter row index (from 0 to 7) #Print first movement #FROM THIS POINTS, THE FIRST PARAMETER "i" IS THE ROW AND THE SECOND "j" THE COLUMN if params[0] == "0": color = black else: color = white i = int(params[2]) j = int(params[1]) data[i,j] = color print(i,j) #Print all movements (WITHOUT FLIPPLIG SPOTS) """ for line in lines: auxParams = line.split("\n") params = auxParams[0].split(" ") if params[0] == "0": color = black else: color = white i = int(params[2]) #rows j = int(params[1]) #colummns data[i,j] = color print(data) """ #Save image with movement image = Image.fromarray(data) image.show() image = image.save("prueba.jpg") dispImage("prueba.jpg", width=100, height=100)
https://github.com/HayleySummer/Qiskit_Hackathon_Europe
HayleySummer
import numpy as np # Importing standard Qiskit libraries from qiskit import QuantumCircuit, transpile, Aer, IBMQ, ClassicalRegister, QuantumRegister, execute from qiskit.tools.jupyter import * from qiskit.visualization import * from ibm_quantum_widgets import * # import basic plot tools from qiskit.tools.visualization import plot_histogram # Loading your IBM Q account(s) provider = IBMQ.load_account() # Board A. Only valid move is first Tile # 011 # 111 # 111 boardA = [ 0, 1, 1, 1, 1, 1, 1, 1, 1] # Board B. Has Seven Valid moves #000 #010 #010 boardB = [ 0, 0, 0, 0, 1, 0, 0, 1, 0] # Board C. Has 5 Valid moves # 000 # 000 # 010 boardC = [ 0, 0, 0, 0, 0, 0, 0, 1, 0] # Board D. Has 2 Valid moves # 001 # 111 # 111 boardD = [ 0, 0, 1, 1, 1, 1, 1, 1, 1] # Params : circuit, BoardState, Auxillary Qubits, oracle def othello_3x3_oracle(qc, boardQ, auxillary, oracle): #Compute the search for the valid moves # Start with contrived solution, Use a board with only 1 free space. The only valid answer is of the form 100 000 000 #qc.cx(auxillary[0],boardQ[0]) qc.cx(auxillary,boardQ) qc.mct(boardQ, oracle) ### If the board is solved flip the output bit. #Phase kickback because output_qubit is in superposition #Uncompute the search for the valid moves #qc.cx(auxillary[0],boardQ[0]) qc.cx(auxillary,boardQ) return def diffuser(nqubits): qcDiff = QuantumCircuit(nqubits) # Apply transformation |s> -> |00..0> (H-gates) for qubit in range(nqubits): qcDiff.h(qubit) # Apply transformation |00..0> -> |11..1> (X-gates) for qubit in range(nqubits): qcDiff.x(qubit) # Do multi-controlled-Z gate qcDiff.h(nqubits-1) qcDiff.mct(list(range(nqubits-1)), nqubits-1) # multi-controlled-toffoli qcDiff.h(nqubits-1) # Apply transformation |11..1> -> |00..0> for qubit in range(nqubits): qcDiff.x(qubit) # Apply transformation |00..0> -> |s> for qubit in range(nqubits): qcDiff.h(qubit) # We will return the diffuser as a gate U_s = qcDiff.to_gate() U_s.name = "$U_s$" return U_s def othelloMoveSearch ( boardRepresentation ): # Create a 9 Qubit Register for the Quantum represntation of the board. boardQ = QuantumRegister(9) # Create a 9 Qubit Register for Auxillary Qubits for Grover Search auxillary = QuantumRegister(9) # Create a single Qubit Register for the Grover Search Oracle oracle = QuantumRegister(1) # Create a 2 Qubit Quantum Register [Not used yet] qr = QuantumRegister(2) # Create a 9 Bit Classical Register for the output cr = ClassicalRegister(9) # Create a quantum circuit containing the registers above qc = QuantumCircuit(boardQ, auxillary, oracle, qr, cr) # Put the Auxillary Qubits into SuperPosition of the Results. qc.h(auxillary[:]) # Prepare the Oracle Qubit qc.x(oracle[0]) qc.h(oracle[0]) # Set up the Quantum Representation of the board to match the current state of the board for x in range(len( boardRepresentation )): if boardRepresentation[x] > 0 : qc.x(boardQ[x]) # Call the Grover Algorithm and Diffuser 16 times [Number of iterations varies according to search] for i in range (16) : othello_3x3_oracle(qc, boardQ, auxillary, oracle) qc.append(diffuser(9), [9,10,11,12,13,14,15,16,17]) # Measure the status of the Quantum Board to verify the setup has worked. qc.measure(auxillary,cr) # Change the endian qc = qc.reverse_bits() #print ("Diagnostic Log 001") return qc # Create the circuit and print it on screen qc1 = othelloMoveSearch( boardA ) qc1.draw(output="mpl") #backend = BasicAer.get_backend('qasm_simulator') backend = Aer.get_backend('qasm_simulator') shots = 1024 results = execute(qc1, backend=backend, shots=shots).result() answer = results.get_counts() plot_histogram(answer, number_to_keep=6)
https://github.com/jhlee29/quantum-meets-hangul
jhlee29
from PIL import Image, ImageOps import os, glob import numpy as np from sklearn import model_selection #cross_validation # from keras.utils import np_utils # General imports import os import gzip import numpy as np import matplotlib.pyplot as plt from pylab import cm import warnings warnings.filterwarnings("ignore") # scikit-learn imports from sklearn import datasets from sklearn.model_selection import train_test_split from sklearn.preprocessing import StandardScaler, MinMaxScaler from sklearn.decomposition import PCA from sklearn.svm import SVC from sklearn.metrics import accuracy_score # Qiskit imports from qiskit import Aer, execute from qiskit.circuit import QuantumCircuit, Parameter, ParameterVector from qiskit.circuit.library import PauliFeatureMap, ZFeatureMap, ZZFeatureMap from qiskit.circuit.library import TwoLocal, NLocal, RealAmplitudes, EfficientSU2 from qiskit.circuit.library import HGate, RXGate, RYGate, RZGate, CXGate, CRXGate, CRZGate from qiskit_machine_learning.kernels import QuantumKernel # input classes = ["g", "h"] # image size of 28 x 28 image_size = 28 # Maximum number of sheets to read max_read = 75 num_classes = len(classes) # load the image X = [] Y = [] for index, classlabel in enumerate(classes): images_dir = "./hangul_characters/" + classlabel files = glob.glob(images_dir + "/*.jpg") for i, file in enumerate(files): # Stop if you read more than max_read to make the number of sheets for each class if i >= max_read: break # open the file, read as data, add to X # Repeatedly add labels with the same index to Y image = Image.open(file) image = ImageOps.invert(image) image = image.convert("L") image = image.resize((image_size, image_size)) data = np.asarray(image) X.append(data) Y.append(index) X = np.array(X) Y = np.array(Y) X_train, X_test, y_train, y_test = model_selection.train_test_split(X, Y, test_size=0.2, random_state=42) X_train = X_train.reshape(len(X_train), -1).astype(np.float64) X_test = X_test.reshape(len(X_test), -1).astype(np.float64) fig = plt.figure() LABELS = [0, 1] num_labels = len(LABELS) for i in range(num_labels): ax = fig.add_subplot(1, num_labels, i+1) img = X[Y==LABELS[i]][0].reshape((28,28)) ax.imshow(img, cmap="Greys") # Standardize ss = StandardScaler() X_train = ss.fit_transform(X_train) X_test = ss.transform(X_test) # Reduce dimensions N_DIM = 5 pca = PCA(n_components=N_DIM) X_train = pca.fit_transform(X_train) X_test = pca.transform(X_test) # Normalize mms = MinMaxScaler((-1, 1)) X_train = mms.fit_transform(X_train) X_test = mms.transform(X_test) # 3 features, depth 1, linear entanglement map_zz = ZZFeatureMap(feature_dimension=5, reps=1, entanglement='linear') map_zz.decompose().draw('mpl') zz_kernel = QuantumKernel(feature_map=map_zz, quantum_instance=Aer.get_backend('statevector_simulator')) matrix_train = zz_kernel.evaluate(x_vec=X_train) matrix_val = zz_kernel.evaluate(x_vec=X_test, y_vec=X_train) fig, axs = plt.subplots(1, 2, figsize=(10, 5)) axs[0].imshow(np.asmatrix(matrix_train), interpolation='nearest', origin='upper', cmap='Blues') axs[0].set_title("training kernel matrix") axs[1].imshow(np.asmatrix(matrix_val), interpolation='nearest', origin='upper', cmap='Reds') axs[1].set_title("validation kernel matrix") plt.show() zz_svc = SVC(kernel='precomputed') zz_svc.fit(matrix_train, y_train) zz_score = zz_svc.score(matrix_val, y_test) print(f'Precomputed kernel classification test score: {zz_score}')
https://github.com/jhlee29/quantum-meets-hangul
jhlee29
# from PIL import Image from PIL import Image, ImageOps import os, glob import numpy as np from sklearn import model_selection #cross_validation # from keras.utils import np_utils # General imports import os import gzip import numpy as np import matplotlib.pyplot as plt from pylab import cm import warnings # input classes = ["a", "b"] # image size of 28 x 28 image_size = 28 # image_size = 8 # Maximum number of sheets to read max_read = 100 num_classes = len(classes) # load the image X = [] Y = [] for index, classlabel in enumerate(classes): images_dir = "./hangul_characters/" + classlabel files = glob.glob(images_dir + "/*.jpg") for i, file in enumerate(files): # Stop if you read more than max_read to make the number of sheets for each class if i >= max_read: break # open the file, read as data, add to X # Repeatedly add labels with the same index to Y image = Image.open(file) image = ImageOps.invert(image) image = image.convert("L") image = image.resize((image_size, image_size)) data = np.asarray(image) X.append(data) Y.append(index) X = np.array(X) Y = np.array(Y) X_train, X_test, y_train, y_test = model_selection.train_test_split(X, Y, test_size=0.2, random_state=42) X_train = X_train.reshape(len(X_train), -1).astype(np.float64) X_test = X_test.reshape(len(X_test), -1).astype(np.float64) fig = plt.figure() LABELS = [0, 1] num_labels = len(LABELS) for i in range(num_labels): ax = fig.add_subplot(1, num_labels, i+1) img = X[Y==LABELS[i]][0].reshape((28,28)) ax.imshow(img, cmap="Greys") print(Y) print(X_train[0]) warnings.filterwarnings("ignore") # scikit-learn imports from sklearn import datasets from sklearn.model_selection import train_test_split from sklearn.preprocessing import StandardScaler, MinMaxScaler from sklearn.decomposition import PCA from sklearn.svm import SVC from sklearn.metrics import accuracy_score # Qiskit imports from qiskit import Aer, execute from qiskit.circuit import QuantumCircuit, Parameter, ParameterVector from qiskit.circuit.library import PauliFeatureMap, ZFeatureMap, ZZFeatureMap from qiskit.circuit.library import TwoLocal, NLocal, RealAmplitudes, EfficientSU2 from qiskit.circuit.library import HGate, RXGate, RYGate, RZGate, CXGate, CRXGate, CRZGate from qiskit_machine_learning.kernels import QuantumKernel # Standardize ss = StandardScaler() X_train = ss.fit_transform(X_train) X_test = ss.transform(X_test) #sample_test = ss.transform(sample_test) # Reduce dimensions N_DIM = 5 pca = PCA(n_components=N_DIM) X_train = pca.fit_transform(X_train) X_test = pca.transform(X_test) # Normalize mms = MinMaxScaler((-1, 1)) X_train = mms.fit_transform(X_train) X_test = mms.transform(X_test) #sample_test = mms.transform(sample_test) # 3 features, depth 1, linear entanglement map_zz = ZZFeatureMap(feature_dimension=5, reps=1, entanglement='linear') map_zz.decompose().draw('mpl') zz_kernel = QuantumKernel(feature_map=map_zz, quantum_instance=Aer.get_backend('statevector_simulator')) matrix_train = zz_kernel.evaluate(x_vec=X_train) matrix_val = zz_kernel.evaluate(x_vec=X_test, y_vec=X_train) fig, axs = plt.subplots(1, 2, figsize=(10, 5)) axs[0].imshow(np.asmatrix(matrix_train), interpolation='nearest', origin='upper', cmap='Blues') axs[0].set_title("training kernel matrix") axs[1].imshow(np.asmatrix(matrix_val), interpolation='nearest', origin='upper', cmap='Reds') axs[1].set_title("validation kernel matrix") plt.show() zz_svc = SVC(kernel='precomputed') zz_svc.fit(matrix_train, y_train) zz_score = zz_svc.score(matrix_val, y_test) print(f'Precomputed kernel classification test score: {zz_score}')
https://github.com/jhlee29/quantum-meets-hangul
jhlee29
# from PIL import Image from PIL import Image, ImageOps import os, glob import numpy as np from sklearn import model_selection #cross_validation # General imports import os import gzip import numpy as np import matplotlib.pyplot as plt from pylab import cm import warnings # input classes = ["bul", "pul", "ppul"] # image size of 28 x 28 image_size = 28 # Maximum number of sheets to read max_read = 60 num_classes = len(classes) # load the image X = [] Y = [] T = [] for index, classlabel in enumerate(classes): images_dir = "./triple/" + classlabel files = glob.glob(images_dir + "/*.jpg") for i, file in enumerate(files): # Stop if you read more than max_read to make the number of sheets for each class if i >= max_read: break # open the file, read as data, add to X # Repeatedly add labels with the same index to Y image = Image.open(file) image = ImageOps.invert(image) image = image.convert("L") image = image.resize((image_size, image_size)) data = np.asarray(image) X.append(data) Y.append(index) images_dir = "./triple" files = glob.glob(images_dir + "/*.jpg") for i, file in enumerate(files): # open the file, read as data, add to X # Repeatedly add labels with the same index to Y if i >= 36: break image = Image.open(file) image = ImageOps.invert(image) image = image.convert("L") image = image.resize((image_size, image_size)) data = np.asarray(image) T.append(data) X = np.array(X) Y = np.array(Y) T = np.array(T) print(X.shape,T.shape) X_train, X_test, y_train, y_test = model_selection.train_test_split(X, Y, test_size=0.2, random_state=42) X_train = X_train.reshape(len(X_train), -1).astype(np.float64) X_test = X_test.reshape(len(X_test), -1).astype(np.float64) T = T.reshape(len(T), -1).astype(np.float64) fig = plt.figure() LABELS = [0, 1, 2] num_labels = len(LABELS) for i in range(num_labels): ax = fig.add_subplot(1, num_labels, i+1) img = X[Y==LABELS[i]][0].reshape((28,28)) ax.imshow(img, cmap="Greys") print(Y) print(X_train[0]) warnings.filterwarnings("ignore") # scikit-learn imports from sklearn import datasets from sklearn.model_selection import train_test_split from sklearn.preprocessing import StandardScaler, MinMaxScaler from sklearn.decomposition import PCA from sklearn.svm import SVC from sklearn.metrics import accuracy_score # Qiskit imports from qiskit import Aer, execute from qiskit.circuit import QuantumCircuit, Parameter, ParameterVector from qiskit.circuit.library import PauliFeatureMap, ZFeatureMap, ZZFeatureMap from qiskit.circuit.library import TwoLocal, NLocal, RealAmplitudes, EfficientSU2 from qiskit.circuit.library import HGate, RXGate, RYGate, RZGate, CXGate, CRXGate, CRZGate from qiskit_machine_learning.kernels import QuantumKernel # Standardize ss = StandardScaler() X_train = ss.fit_transform(X_train) X_test = ss.transform(X_test) sample_test = ss.transform(T) print(sample_test) # Reduce dimensions N_DIM = 5 pca = PCA(n_components=N_DIM) X_train = pca.fit_transform(X_train) X_test = pca.transform(X_test) sample_test = pca.transform(sample_test) # Normalize mms = MinMaxScaler((-1, 1)) X_train = mms.fit_transform(X_train) X_test = mms.transform(X_test) sample_test = mms.transform(sample_test) # label 0 labels_train_0 = np.where(y_train==0, 1, 0) labels_val_0 = np.where(y_test==0, 1, 0) print(f'Original validation labels: {y_test}') print(f'Validation labels for 0 vs Rest: {labels_val_0}') pauli_map_0 = PauliFeatureMap(feature_dimension=N_DIM, reps=2, paulis = ['X', 'Y', 'ZZ']) pauli_kernel_0 = QuantumKernel(feature_map=pauli_map_0, quantum_instance=Aer.get_backend('statevector_simulator')) pauli_svc_0 = SVC(kernel='precomputed', probability=True) matrix_train_0 = pauli_kernel_0.evaluate(x_vec=X_train) pauli_svc_0.fit(matrix_train_0, labels_train_0) matrix_val_0 = pauli_kernel_0.evaluate(x_vec=sample_test, y_vec=X_train) pauli_score_0 = pauli_svc_0.score(matrix_val_0, labels_val_0) print(f'Accuracy of discriminating between label 0 and others: {pauli_score_0*100}%') matrix_test_0 = pauli_kernel_0.evaluate(x_vec=sample_test, y_vec=X_train) pred_0 = pauli_svc_0.predict_proba(matrix_test_0)[:, 1] print(f'Probability of label 0: {np.round(pred_0, 2)}') # label 1 labels_train_1 = np.where(y_train==1, 1, 0) labels_val_1 = np.where(y_test==1, 1, 0) print(f'Original validation labels: {y_test}') print(f'Validation labels for 0 vs Rest: {labels_val_1}') zz_map_1 = ZZFeatureMap(feature_dimension=N_DIM, reps=2, entanglement='linear', insert_barriers=True) zz_kernel_1 = QuantumKernel(feature_map=zz_map_1, quantum_instance=Aer.get_backend('statevector_simulator')) zzpc_svc_1 = SVC(kernel='precomputed', probability=True) matrix_train_1 = zz_kernel_1.evaluate(x_vec=X_train) zzpc_svc_1.fit(matrix_train_1, labels_train_1) matrix_val_1 = zz_kernel_1.evaluate(x_vec=X_test, y_vec=X_train) zzpc_score_1 = zzpc_svc_1.score(matrix_val_1, labels_val_1) print(f'Accuracy for validation dataset: {zzpc_score_1}') matrix_test_1 = zz_kernel_1.evaluate(x_vec=sample_test, y_vec=X_train) pred_1 = zzpc_svc_1.predict_proba(matrix_test_1)[:, 1] print(pred_1) # label 2 labels_train_2 = np.where(y_train==2, 1, 0) labels_val_2 = np.where(y_test==2, 1, 0) print(f'Original validation labels: {y_test}') print(f'Validation labels for 0 vs Rest: {labels_val_2}') zz_map_2 = ZZFeatureMap(feature_dimension=N_DIM, reps=2, entanglement='linear', insert_barriers=True) zz_kernel_2 = QuantumKernel(feature_map=zz_map_2, quantum_instance=Aer.get_backend('statevector_simulator')) zzpc_svc_2 = SVC(kernel='precomputed', probability=True) matrix_train_2 = zz_kernel_2.evaluate(x_vec=X_train) zzpc_svc_2.fit(matrix_train_2, labels_train_2) matrix_val_2 = zz_kernel_2.evaluate(x_vec=X_test, y_vec=X_train) zzpc_score_2 = zzpc_svc_2.score(matrix_val_2, labels_val_2) print(f'Accuracy for validation dataset: {zzpc_score_2}') matrix_test_2 = zz_kernel_2.evaluate(x_vec=sample_test, y_vec=X_train) pred_2 = zzpc_svc_2.predict_proba(matrix_test_2)[:, 1] print(pred_2) # prediction pred_test = np.where((pred_0 > pred_1) & (pred_0 > pred_2), 0, np.where(pred_1 > pred_2, 1, 2)) print(f'Prediction: {pred_test}')
https://github.com/andre-juan/qiskit_certificate_sample_test
andre-juan
import numpy as np import pandas as pd import matplotlib.pyplot as plt import seaborn as sns # main classes and functions from qiskit import QuantumRegister, ClassicalRegister, QuantumCircuit, Aer, BasicAer, execute # visualization stuff from qiskit.visualization import (plot_histogram, plot_bloch_multivector, plot_state_city, plot_state_hinton, plot_state_qsphere, plot_state_paulivec, plot_bloch_multivector, plot_bloch_vector, plot_gate_map, plot_error_map) # gates from qiskit.circuit.library import CXGate, MCMT, XGate # quantum info theory stuff from qiskit.quantum_info import Operator, Statevector, state_fidelity, process_fidelity, average_gate_fidelity # other tools from qiskit.tools import job_monitor from qiskit import IBMQ def show_figure(fig): ''' auxiliar function to display plot even if it's not the last command of the cell from: https://github.com/Qiskit/qiskit-terra/issues/1682 ''' new_fig = plt.figure() new_mngr = new_fig.canvas.manager new_mngr.canvas.figure = fig fig.set_canvas(new_mngr.canvas) plt.show(fig) def test_circuit_qasm(qc): ''' auxiliar function, used to run a given quantum circuit with the qasm simulator. measurement is added in this function, so you must pass a circuit with no classical registers and measurements. ''' qc.measure_all() ################################# backend = Aer.get_backend("qasm_simulator") job = execute(qc, backend, shots=1e5, seed_simulator=42) results = job.result() counts = results.get_counts() return plot_histogram(counts, title="Results", figsize=(12, 4)) def test_circuit_sv(qc, print_stuff=True): ''' auxiliar function, used to run a given quantum circuit with the statevector simulator. it optinally prints the state vector components and respective probabilities. ''' backend = Aer.get_backend("statevector_simulator") job = execute(qc, backend, seed_simulator=42) results = job.result() counts = results.get_counts() if print_stuff: sv = results.data(0)['statevector'] probs = sv**2 print(f"Statevector:\t{sv}\n") print(f"Probabilities:\t{probs}") return plot_histogram(counts, title="Results", figsize=(12, 4)) # 1 - A) qc = QuantumCircuit(4, 4) qc.qregs, qc.cregs # 2 - C) qc = QuantumCircuit(1) qc.ry(3*np.pi/4, 0) test_circuit_sv(qc) # 3 - A), B) & D) are possible inp_reg = QuantumRegister(2, name="inp") ancilla = QuantumRegister(1, name="anc") qc = QuantumCircuit(inp_reg, ancilla) qc.h(inp_reg) qc.x(ancilla) qc.barrier() qc.h(inp_reg[0]) qc.h(inp_reg[1]) qc.x(ancilla) qc.barrier() qc.h(inp_reg[0:2]) qc.x(ancilla[0]) qc.draw("mpl") # 4 - A) inp = QuantumRegister(3, name="q") meas = ClassicalRegister(3, name="c") qc = QuantumCircuit(inp, meas) # # this would also be possible, but it's not in options # qc.measure(inp, meas) qc.measure([0, 1, 2], [0, 1, 2]) qc.draw("mpl") # 5 - A) bell = QuantumCircuit(2) # this makes |01> + |10> bell.h(0) bell.x(1) bell.cx(0, 1) show_figure(bell.draw("mpl")) test_circuit_sv(bell) # 6 - A) & C) qc = QuantumCircuit(1, 1) # # both below are equivalent # qc.h(0) qc.ry(np.pi/2, 0) show_figure(qc.draw("mpl")) simulator = Aer.get_backend("statevector_simulator") job = execute(qc, simulator) results = job.result() outputstate = results.get_statevector(qc) print(f"\nState:\t{outputstate}\n") plot_bloch_multivector(outputstate) # 7 - B) # S gate induces a pi/2 phase! (is a P gate with \theta = pi/2, i.e., it's a Rz gate with \theta = pi/2) # tip: S = T^2, hence the "S" stands for "T Squared" (not sure if that's the actual reason for the name, but it's a mnemonic) qc = QuantumCircuit(1) qc.h(0) qc.s(0) show_figure(qc.draw("mpl")) simulator = Aer.get_backend("statevector_simulator") job = execute(qc, simulator) results = job.result() outputstate = results.get_statevector(qc) print(f"\nState:\t{outputstate}\n") plot_bloch_multivector(outputstate) # another way to get the statevector, using the Statevector class qc = QuantumCircuit(1) qc.h(0) qc.s(0) show_figure(qc.draw("mpl")) outputstate = Statevector.from_label("0").evolve(qc) print(f"\nState:\t{outputstate}\n") plot_bloch_multivector(outputstate) # 8 - A), B) qc = QuantumCircuit(2) # # both options below work # v = [1/np.sqrt(2), 0, 0, 1/np.sqrt(2)] # qc.initialize(v, [0, 1]) qc.h(0) qc.cx(0, 1) show_figure(qc.draw("mpl")) simulator = Aer.get_backend("statevector_simulator") result = execute(qc, simulator).result() statevector = result.get_statevector(qc) print(f"\nState:\t{statevector}\n") # 9 - D) # notice how the first tree are equivalent to cnot # the last one is the only multi-qubit gate OTHER THAN a cnot ;) qc = QuantumCircuit(2) ################################ qc.cx(0, 1) qc.barrier() ################################ qc.cnot(0, 1) qc.barrier() ################################ # that's nice! multi-controlled x qc.mct([0], 1) qc.barrier() ################################ qc.cz(0, 1) qc.barrier() ################################ qc.draw("mpl") # 10 - D) # notice how the first tree are equivalent to the toffoli gate # the last one is the only multi-qubit gate OTHER THAN a toffoli ;) qc = QuantumCircuit(3) ################################ qc.ccx(0, 1, 2) qc.barrier() ################################ qc.mct([0, 1], 2) qc.barrier() ################################ # very nice construction! ccx = CXGate().control() qc.append(ccx, [0, 1, 2]) qc.barrier() ################################ qc.cry(0, 1, 2) qc.barrier() ################################ qc.draw("mpl") # another nice way to do toffoli qc = QuantumCircuit(3) mct = MCMT(gate='x', num_ctrl_qubits = 2, num_target_qubits = 1) qc.append(mct, [0,1,2]) qc = qc.decompose() qc.draw("mpl") # 11 - B) & C) qc = QuantumCircuit(3, 3) qc.barrier() qc.barrier([0, 1, 2]) qc.draw("mpl") # 12 - A) qc = QuantumCircuit(1, 1) qc.h(0) qc.t(0) qc.t(0) qc.h(0) qc.measure(0, 0) show_figure(qc.draw("mpl")) print("\nSince S = T^2, the circuit above is equivalent to:") qc = QuantumCircuit(1, 1) qc.h(0) qc.s(0) qc.h(0) qc.measure(0, 0) show_figure(qc.draw("mpl")) # 13 - A) qc = QuantumCircuit(2, 2) qc.h(0) qc.barrier(0) qc.cx(0, 1) qc.barrier([0, 1]) # take a look in the docstring! print(f"\nDepth of the circuit: {qc.depth()}") qc.draw("mpl") # 14 - A) qc = QuantumCircuit(3) qasm_sim = Aer.get_backend("qasm_simulator") # coupling map connecting 3 qubits linearly couple_map = [[0, 1], [1, 2]] job = execute(qc, qasm_sim, shots=1024, coupling_map=couple_map) result = job.result() # 15 - A) backend = BasicAer.get_backend("qasm_simulator") qc = QuantumCircuit(3) execute(qc, backend, shots=1024, coupling_map=[[0, 1], [1, 2]]) # 16 - A), C) & D) # answer here: https://qiskit.org/documentation/apidoc/providers_basicaer.html#simulators qasm = BasicAer.get_backend("qasm_simulator") sv = BasicAer.get_backend("statevector_simulator") unitary = BasicAer.get_backend("unitary_simulator") # 17 - B) backend = BasicAer.get_backend("statevector_simulator") # 18 - C) # info: https://qiskit.org/documentation/stubs/qiskit.quantum_info.Operator.html # # i think the following is possible: # op = Operator([[0, 1], [1, 0]]) qc = QuantumCircuit(1) qc.x(0) op = Operator(qc) op # 19 - C) # info here: https://qiskit.org/documentation/tutorials/circuits_advanced/02_operators_overview.html#Process-Fidelity # namely: # "We may also compare operators using the process_fidelity function from the Quantum Information module. # This is an information theoretic quantity for how close two quantum channels are to each other, # and in the case of unitary operators it does not depend on global phase." op_a = Operator(XGate()) op_b = np.exp(1j * 0.5) * Operator(XGate()) pF = process_fidelity(op_a, op_b) print(f'\nProcess fidelity pF = {pF}') agF = average_gate_fidelity(op_a, op_b) print(f'\nAverage gate fidelity agF = {agF}') # # code below returns an error, because, of course, op_a and op_b are not states, # # but operators (processes, gates) # sF = state_fidelity(op_a, op_b) # print(f'\nState fidelity sF = {sF}') # 20 - B) # BE VERY CAREFUL WITH THE ENDIANNESS!!! qc = QuantumCircuit(2, 2) qc.x(0) # better to visualize order of measurements qc.barrier() qc.measure([0, 1], [0, 1]) show_figure(qc.draw("mpl")) simulator = Aer.get_backend("qasm_simulator") result = execute(qc, simulator, shots=1000).result() counts = result.get_counts(qc) print(counts) # openqasm string qc = QuantumCircuit(2, 2) qc.h(0) qc.cnot(0, 1) qc.barrier() qc.measure([0, 1], [0, 1]) #################################### show_figure(qc.draw("mpl")) #################################### print(qc.qasm()) print("\n") print(qc.qasm(formatted=True)) qasm_str = qc.qasm(filename="qasm_test") qc_test_str = QuantumCircuit.from_qasm_str(qasm_str) qc_test_str.draw("mpl") qc_test_file = QuantumCircuit.from_qasm_file("qasm_test") qc_test_file.draw("mpl") qc = QuantumCircuit(2, 2) qc.h(0) qc.cnot(0, 1) qc.barrier() ########################## backend = Aer.get_backend("unitary_simulator") unitary = execute(qc, backend).result().get_unitary() unitary fig, ax = plt.subplots(1, 2, figsize=(8, 3)) sns.heatmap(np.real(unitary), annot=True, ax=ax[0]) ax[0].set_title("Real part") sns.heatmap(np.imag(unitary), annot=True, ax=ax[1]) ax[1].set_title("Imaginary part") plt.tight_layout() plt.show() qc = QuantumCircuit(3) qc.h(0) qc.cnot(0, 1) qc.cnot(0, 2) show_figure(qc.draw("mpl")) #################################### sv = Statevector.from_label("000") sv = sv.evolve(qc) print(f'Statevector:\n\n{sv.data}') show_figure(plot_state_hinton(sv)) #################################### gate = qc.to_gate() gate.name = "my_gate" #################################### qc2 = QuantumCircuit(3) qc2.append(gate, [0, 1, 2]) print("Gate in the circuit:") show_figure(qc2.draw("mpl")) print("Decomposed gate in the circuit:") show_figure(qc2.decompose().draw("mpl")) # that's more useful when "backend" is a real hardware, of course job = execute(qc, backend) job_monitor(job) bell = QuantumCircuit(2, 2) bell.h(0) bell.cx(0, 1) meas = QuantumCircuit(2, 2) meas.measure([0,1], [0,1]) circ = bell.compose(meas) show_figure(bell.draw("mpl")) show_figure(meas.draw("mpl")) show_figure(circ.draw("mpl")) backend = Aer.get_backend("qasm_simulator") job = execute(circ, backend, shots=1e5, seed_simulator=42) results = job.result() counts = results.get_counts() plot_histogram(counts, title="Results", figsize=(6, 4)) simulator = Aer.get_backend("statevector_simulator") job = execute(bell, simulator) results = job.result() psi = results.get_statevector(bell) print(psi) # "here we see that there is no information about the quantum state in the single qubit space as all vectors are zero." # if it was a separable state, it would be shown! show_figure(plot_bloch_multivector(psi)) show_figure(plot_state_city(psi)) show_figure(plot_state_hinton(psi)) show_figure(plot_state_qsphere(psi)) show_figure(plot_bloch_vector([1,0,0], figsize=(2, 2))) show_figure(plot_bloch_vector([0,1,0], figsize=(2, 2))) show_figure(plot_bloch_vector([0,0,1], figsize=(2, 2))) # already saved my token! # for more info: https://quantum-computing.ibm.com/lab/docs/iql/manage/account/ibmq provider = IBMQ.load_account() accountProvider = IBMQ.get_provider(hub='ibm-q') provider.backends() provider.backends(simulator=False) for bcknd in provider.backends(simulator=False): print(f"\nBackend: {bcknd.name()}") backend = provider.get_backend(bcknd.name()) show_figure(plot_gate_map(backend, figsize=(3, 3))) print(f"Coupling map:\n\n{bcknd._configuration.coupling_map}\n") print("="*50) print("="*50) for bcknd in provider.backends(simulator=False): print(f"\nBackend: {bcknd.name()}") backend = provider.get_backend(bcknd.name()) show_figure(plot_error_map(backend)) print("="*50) print("="*50)
https://github.com/mberna/qce22-qiskit-runtime-tutorial
mberna
# define Hamiltonian from qiskit.opflow import PauliSumOp H = PauliSumOp.from_list([('XYII', 1), ('IYZI', 2), ('IIZX', 3), ('XIII', 4), ('IYII', 5)]) print(H) from qiskit.circuit.library import EfficientSU2 qc = EfficientSU2(num_qubits=H.num_qubits, reps=1) qc_with_meas = qc.measure_all(inplace=False) qc_with_meas.decompose().draw(fold=120) # define a set of (random) parameter values for the ansatz circuit import numpy as np theta = np.random.rand(qc.num_parameters) # use the Sampler to sample from this circuit from qiskit.primitives import Sampler sampler = Sampler([qc_with_meas]) s_result = sampler([qc_with_meas], [theta]) print(s_result) from qiskit.primitives import Estimator estimator = Estimator([qc], [H]) e_result = estimator([qc], [H], [theta]) print(e_result) from qiskit.algorithms import NumPyEigensolver exact_solver = NumPyEigensolver(k=2) # find the first two '2' eigenvalues exact_result = exact_solver.compute_eigenvalues(H) print(exact_result.eigenvalues) # define objective as expectation value of Hamiltonian with circuit ansatz objective = lambda x: estimator([qc], [H], [x]).values[0] # instantiate optimizer from qiskit.algorithms.optimizers import SPSA optimizer = SPSA(maxiter=500) # define initial values for our circuit parameters x0 = np.random.rand(qc.num_parameters) # minimize the objective function result = optimizer.minimize(objective, x0=x0) # store ground state parameters for later ground_state_params = result.x # print the resulting ground state energy print(result.fun) from qiskit.algorithms import MinimumEigensolver, VQEResult class CustomVQE(MinimumEigensolver): def __init__(self, estimator, circuit, optimizer, callback=None): self._estimator = estimator self._circuit = circuit self._optimizer = optimizer self._callback = callback def compute_minimum_eigenvalue(self, operator, aux_operators=None): # define objective def objective(x): e_job = self._estimator([self._circuit], [operator], [x]) value = e_job.values[0] if self._callback: self._callback(value) return value # run optimization x0 = np.random.rand(self._circuit.num_parameters) res = self._optimizer.minimize(objective, x0=x0) # populate results result = VQEResult() result.cost_function_evals = res.nfev result.eigenvalue = res.fun result.optimal_parameters = res.x return result # run the custom VQE function custom_vqe = CustomVQE(estimator, qc, optimizer) result = custom_vqe.compute_minimum_eigenvalue(H) print(result) from qiskit.circuit import ParameterVector class Fidelity(): def __init__(self, sampler): """ Instantiate fidelity estimator. """ #self._sampler = sampler def run(self, circuit_1, circuit_2, param_values_1, param_values_2): """ Estimate fidelity between two states defined by given parameter values.""" # prepare circuit circuit_1_ = circuit_1.remove_final_measurements(inplace=False) circuit_2_ = circuit_2.remove_final_measurements(inplace=False) x = ParameterVector("x", circuit_1_.num_parameters) y = ParameterVector("y", circuit_2_.num_parameters) circuit = circuit_1_.assign_parameters(x).compose(circuit_2_.inverse().assign_parameters(y)) circuit.measure_all() # run circuit and get probabilities param_values = np.append(param_values_1, param_values_2) sampler = Sampler([circuit]) s_job = sampler([circuit], [param_values]) probabilities = s_job.quasi_dists[0] # estimate fidelity: |<0|U_1^dag U_2|0>|^2 clipped to [0, 1] fidelity = np.minimum(1.0, np.maximum(0.0, probabilities.get(0, 0.0))) return fidelity # instantiate fidelity fidelity = Fidelity(sampler) # interpolation between parameters ts = np.linspace(0, 1) fidelities = [fidelity.run(qc, qc, ground_state_params, ground_state_params + t*theta) for t in ts] import matplotlib.pyplot as plt plt.figure(figsize=(8, 4)) plt.plot(ts, fidelities) plt.xlabel('t', fontsize=20) plt.ylabel('fidelity', fontsize=20) plt.show() # set penalty weight for overlap term penalty = 25 # define objective for VQD vqd_objective = lambda x: objective(x) + penalty*fidelity.run(qc, qc, ground_state_params, x) # run optimization to get first excited state result = optimizer.minimize(vqd_objective, ground_state_params) print(result) # determine energy e_job = estimator([qc], [H], [result.x]) print(e_job) fid = fidelity.run(qc, qc, result.x, ground_state_params); print(fid) from qiskit.algorithms import Eigensolver, EigensolverResult class VQD(Eigensolver): def __init__(self, estimator, fidelity, circuit, optimizer, penalty=25, callback=None): # self._estimator = estimator self._fidelity = fidelity self._circuit = circuit self._optimizer = optimizer self._penalty = penalty self._callback = callback def compute_eigenvalues(self, operator, aux_operators=None): # compute ground state estimator = Estimator([self._circuit], [operator]) energy = lambda x: estimator([self._circuit], [operator], [x]).values[0] x0 = np.random.rand(self._circuit.num_parameters) res_1 = self._optimizer.minimize(energy, x0) # compute first excited states overlap = lambda x: self._fidelity.run(self._circuit, self._circuit, res_1.x, x) def objective(x): value = energy(x) + self._penalty * overlap(x) if self._callback: self._callback(value) return value res_2 = self._optimizer.minimize(objective, res_1.x) # populate results result = EigensolverResult() result.eigenvalues = [estimator([self._circuit], [operator], [res_1.x]), estimator([self._circuit], [operator], [res_2.x])] return result vqd = VQD(estimator, fidelity, qc, optimizer=SPSA(maxiter=1000)) result = vqd.compute_eigenvalues(H) print(result) from qiskit_ibm_runtime import (Session, Options, QiskitRuntimeService, Sampler as RuntimeSampler, Estimator as RuntimeEstimator) service = QiskitRuntimeService(channel='ibm_quantum') from qiskit.algorithms.optimizers import COBYLA optimizer = COBYLA(maxiter=400) with Session(service=service, backend='ibmq_qasm_simulator') as session: # prepare primitives rt_estimator = RuntimeEstimator(session=session) rt_sampler = RuntimeSampler(session=session) rt_fidelity = Fidelity(rt_sampler) # set up algorithm rt_vqd = VQD(rt_estimator, rt_fidelity, qc, optimizer) # run algorithm result = rt_vqd.compute_eigenvalues(H) print(result) import qiskit_ibm_runtime qiskit_ibm_runtime.version.get_version_info() from qiskit.tools.jupyter import * %qiskit_version_table %qiskit_copyright
https://github.com/mberna/qce22-qiskit-runtime-tutorial
mberna
# instantiate runtime service from qiskit_ibm_runtime import QiskitRuntimeService QiskitRuntimeService.save_account(channel="ibm_quantum", token="MY_IBM_QUANTUM_TOKEN") service = QiskitRuntimeService(channel="ibm_quantum") # import required libraries and instantiate runtime service from qiskit_ibm_runtime import Session, Estimator, Options from qiskit.circuit.library import RealAmplitudes from qiskit.quantum_info import SparsePauliOp # create circuits and observables psi1 = RealAmplitudes(num_qubits=2, reps=2) psi2 = RealAmplitudes(num_qubits=2, reps=3) H1 = SparsePauliOp.from_list([("II", 1), ("IZ", 2), ("XI", 3)]) H2 = SparsePauliOp.from_list([("IZ", 1)]) H3 = SparsePauliOp.from_list([("ZI", 1), ("ZZ", 1)]) psi1.decompose().draw(output='mpl') psi2.decompose().draw(output='mpl') import numpy as np # generate random theta params np.random.seed(0) theta1 = np.random.rand(6) * np.pi theta2 = np.random.rand(8) * np.pi theta3 = np.random.rand(6) * np.pi print(theta1) print(theta2) print(theta3) with Session(service=service, backend="ibmq_qasm_simulator") as session: estimator = Estimator(session=session) from qiskit_ibm_runtime import Options options = Options(optimization_level=1) options.execution.shots = 1024 with Session(service=service, backend="ibmq_qasm_simulator") as session: estimator = Estimator(session=session, options=options) # calculate [ <psi1(theta1)|H1|psi1(theta1)> ] psi1_H1 = estimator.run(circuits=[psi1], observables=[H1], parameter_values=[theta1]) print(psi1_H1.result()) # You can invoke run() multiple times! # calculate [ <psi1(theta1)|H2|psi1(theta1)>, <psi1(theta1)|H3|psi1(theta1)> ] psi1_H23 = estimator.run(circuits=[psi1, psi1], observables=[H2, H3], parameter_values=[theta1]*2) print(psi1_H23.result()) # Ex. 1: Calculate [ <psi2(theta2)|H2|psi2(theta2)> ] # psi2_H2 = ... # Ex. 2 calculate [ <psi1(theta1)|H1|psi1(theta1)>, # <psi2(theta2)|H2|psi2(theta2)>, # <psi1(theta3)|H3|psi1(theta3)> ] # psi12_H23 = ... # Ex. 3: use optimization_level 3 options = Options() # options... # hint: use autocomplete with Session(service=service, backend="ibmq_qasm_simulator") as session: estimator = Estimator(session=session, options=options) theta4 = [0, 1, 1, 2, 3, 5] # calculate [ <psi1(theta1)|H1|psi1(theta1)> ] psi1_H1 = estimator.run(circuits=[psi1], observables=[H1], parameter_values=[theta4]) print(psi1_H1.result()) # Ex. 4: Apply REM to the following example # Remeber REM is enabled when resilience level is 1 options = Options() # options... # hint: use autocomplete with Session(service=service, backend="ibmq_qasm_simulator",) as session: estimator = Estimator(session=session, options=options) theta4 = [0, 1, 1, 2, 3, 5] psi1_H1 = estimator.run(circuits=[psi1], observables=[H1], parameter_values=[theta4]) print(psi1_H1.result()) from qiskit import QuantumCircuit from qiskit_ibm_runtime import Sampler bell = QuantumCircuit(2) bell.h(0) bell.cx(0, 1) bell.measure_all() with Session(service=service, backend="ibmq_qasm_simulator") as session: sampler = Sampler(session=session, options=options) job = sampler.run(circuits=bell) print(job.result()) # You can invoke run() multiple times. job = sampler.run(circuits=bell) print(job.result()) ## REM sampler from qiskit import QuantumCircuit from qiskit_ibm_runtime import Sampler, Options, Session bell = QuantumCircuit(2) bell.h(0) bell.cx(0, 1) bell.measure_all() options_rem = Options(resilience_level=1) with Session(service=service, backend="ibmq_qasm_simulator") as session: sampler = Sampler(session=session) job = sampler.run(circuits=bell) sampler_rem = Sampler(session=session, options=options_rem) job_rem = sampler_rem.run(circuits=bell) print(f"Quasi_dist without EM {job.result().quasi_dists}") print(f"Quasi_dist with EM {job_rem.result().quasi_dists}")
https://github.com/geduardo/Q-Snake-Qiskitcamp-Europe-2019
geduardo
import numpy as np from qiskit import * def Qand(First_bool,Second_bool): #input two booleans into the function a = int(First_bool) b = int(Second_bool) #quantum circuit that applies a Toffoli gate q = QuantumRegister(3) qc = QuantumCircuit(q) if a is 1: qc.x(0) if b is 1: qc.x(1) qc.ccx(q[0], q[1], q[2]) # Code for giving a True if the 3rd qubit is 1 and False if the 3rd qubit is 0 backend = Aer.get_backend('statevector_simulator') job = execute(qc, backend) result = job.result() outputstate = result.get_statevector(qc, decimals=8) Q_and = bool(int(np.real(outputstate[7]))) return Q_and def Qor(First_bool,Second_bool): #input two booleans into the function a = int(First_bool) b = int(Second_bool) #quantum circuit that applies a Toffoli gate q = QuantumRegister(3) qc = QuantumCircuit(q) if a is 1: qc.x(0) if b is 1: qc.x(1) qc.ccx(q[0], q[1], q[2]) # Code for giving a True at least one of the two control qubits are in state 1 backend = Aer.get_backend('statevector_simulator') job = execute(qc, backend) result = job.result() outputstate = result.get_statevector(qc, decimals=8) if a is 1 and b is 0: Q_or = bool(int(np.real(outputstate[1]))) elif a is 0 and b is 1: Q_or = bool(int(np.real(outputstate[2]))) else: Q_or = bool(int(np.real(outputstate[7]))) return(Q_or)
https://github.com/geduardo/Q-Snake-Qiskitcamp-Europe-2019
geduardo
%matplotlib notebook import pew pew.init() screen=pew.Pix() # First we set variables of the game: # The snake will be represented by a list of (x,y) coordinates. The last element of the list (snake[-1]) will # represent the head of the snake. The 0th element of the list will be the last pixel of the tail. # Let's create the list by setting the initial position of the head at (3,3): snake=[(3,3)] # Now we create a variable to store the speed of the movement: game_speed=5 # We also have to define the initial velocity of the snake. Let's set it to move in the x direction by 1 pixel. dx, dy= 1, 0 # First we create and endless loop for the movement of the snake while True: x,y=snake[-1] # We store the current position of the head in two variables x and y. screen.pixel(x, y, 3) # This command turns on the pixel (x,y) with the maximum brightnes (3) to print on # on the screen the current position of the head pew.show(screen) # Lights up the active pixels on the screen. We need to pew.show(screen) each frame of the game. pew.tick(1/game_speed) # Sets the refresh rate. When bigger "game_speed" the lower the refresh rate and the game # runs faster # Now we have to check for user input by looking which keys are pressed. Depending on which keys are pressed the velocity of # of the snake will change. keys=pew.keys() # Return a number telling which keys (or buttons) have been pressed since the # last check. The number can then be filtered with the ``&`` operator and # the ``K_X``, ``K_DOWN``, ``K_LEFT``, ``K_RIGHT``, ``K_UP``, and ``K_O`` # constants to see whether any of the keys was pressed. if keys & pew.K_UP and dy == 0: dx, dy = 0, -1 # When the snake is moving in x and wee press UP the velocity of the snake changes to go up in the screen. # Note that in the pew-pew screen the index of the top row is 0 and the index of the bottom row is 7. # This explains why we have dy=-1 when we press UP. The rest of the keys follow a similar logic. elif keys & pew.K_LEFT and dx == 0: dx, dy = -1, 0 elif keys & pew.K_RIGHT and dx == 0: dx, dy = 1, 0 elif keys & pew.K_DOWN and dy == 0: dx, dy = 0, 1 # Now we calculate the next position of the snake's head: x=(x+dx)%8 y=(y+dy)%8 # We insert %8 to apply mod 8 to the coordinates, so when we pass through one of the walls we restart the counting and # appear in the other side of the screen. # We add to the list representing our snake a the next position of the head. snake.append((x,y)) #We delete from the list the last position of the snake (in this case the head) and turn of that pixel (x,y)=snake.pop(0) screen.pixel(x,y,0) import pew pew.init() screen=pew.Pix() snake=[(3,3)] game_speed=5 dx, dy= 1, 0 while True: x,y=snake[-1] screen.pixel(x, y, 3) pew.show(screen) pew.tick(1/game_speed) keys=pew.keys() if keys & pew.K_UP and dy == 0: dx, dy = 0, -1 elif keys & pew.K_LEFT and dx == 0: dx, dy = -1, 0 elif keys & pew.K_RIGHT and dx == 0: dx, dy = 1, 0 elif keys & pew.K_DOWN and dy == 0: dx, dy = 0, 1 #To implement the crashing mechanism we first have to get rid from the periodic screen, so we remove %8 x=(x+dx) y=(y+dy) #Now add an if condition that breaks the while loop if any of the coordinates for the new head goes out from the grid. if x >7 or y > 7 or x < 0 or y < 0: # We create a loop for turning of all the pixels that forms the screen. for (i,j) in snake: screen.pixel(i,j,0) break snake.append((x,y)) (x,y)=snake.pop(0) screen.pixel(x,y,0) # Now that we are out the loop we show the "Game Over" screen: text = pew.Pix.from_text("Game over!") for dx in range(-8, text.width): screen.blit(text, -dx, 1) pew.show(screen) pew.tick(1 / 12) from aether import QuantumCircuit, simulate def qrand(nbits): # We import the necessary elements form the aether library circ= QuantumCircuit(1,1) # We create a circuit with 1 qubit register and 1 classical bit register circ.h(0) # We apply the Hadamard gate to the qubit (note that by default the qubits are already intializated at 0) circ.measure(0, 0) # We make a measurement in the qubit and we store the result in the classical bit val = simulate(circ, shots=nbits, get='memory') # We simulate three times the circuit and we store the results in a list # We create a string with the binary number to use the function int() to obtain the integer b='' for i in range(nbits): b+= str(val[i]) # Trnsform the string into an int integ=int(b,2) # Return the integer return integ import pew from aether import QuantumCircuit, simulate # We define the function that we just created at the beginning of our code def qrand(nbits): circ= QuantumCircuit(1,1) circ.h(0) circ.measure(0, 0) val = simulate(circ, shots=nbits, get='memory') b='' for i in range(nbits): b+= str(val[i]) integ=int(b,2) return integ pew.init() screen=pew.Pix() snake=[(3,3)] game_speed=5 dx, dy= 1, 0 # We must define the initial position of the apple: apple_x, apple_y = 6, 6 screen.pixel(apple_x, apple_y, 3) while True: x,y=snake[-1] screen.pixel(x, y, 3) pew.show(screen) pew.tick(1/game_speed) keys=pew.keys() if keys & pew.K_UP and dy == 0: dx, dy = 0, -1 elif keys & pew.K_LEFT and dx == 0: dx, dy = -1, 0 elif keys & pew.K_RIGHT and dx == 0: dx, dy = 1, 0 elif keys & pew.K_DOWN and dy == 0: dx, dy = 0, 1 x=(x+dx) y=(y+dy) # We need to add the condition that if the next position of the snake's head is in the snake's body the player dies if (x,y) in snake or x >7 or y > 7 or x < 0 or y < 0: # We create a loop for turning of all the pixels that forms the screen. for (i,j) in snake: screen.pixel(i,j,0) break snake.append((x,y)) # In this loop we implement the generation of the apples each time the next position of the snake's head # lies into the current position of the apple if x == apple_x and y == apple_y: screen.pixel(apple_x, apple_y, 0) # We try the random generation of the apple until the apple appears out of the snake apple_x, apple_y = snake[0] while (apple_x, apple_y) in snake: apple_x=qrand(3) apple_y=qrand(3) screen.pixel(apple_x, apple_y, 2) # We also increase the velocity of the game each time the snake eats an apple gm_sp += 0.2 # Otherwise we remove the tail of the snake else: x, y = snake.pop(0) screen.pixel(x, y, 0) text = pew.Pix.from_text("Game over!") for dx in range(-8, text.width): screen.blit(text, -dx, 1) pew.show(screen) pew.tick(1 / 12) import pew from aether import QuantumCircuit, simulate def qrand(nbits): circ= QuantumCircuit(1,1) circ.h(0) circ.measure(0, 0) val = simulate(circ, shots=nbits, get='memory') b='' for i in range(nbits): b+= str(val[i]) integ=int(b,2) return integ pew.init() screen=pew.Pix() l=True # We initialize a boolean that will help us to reset the initial conditions of the game while True: #Now we put the condition of reseting the initial variables only if l=True if l: game_speed=5 points=0 # We initialize a variable to count the points snake = [(3,3)] dx, dy = 1, 0 apple_x, apple_y = 6, 6 screen.pixel(apple_x, apple_y, 3) l=False while True: x,y=snake[-1] screen.pixel(x, y, 3) pew.show(screen) pew.tick(1/game_speed) keys=pew.keys() if keys & pew.K_UP and dy == 0: dx, dy = 0, -1 elif keys & pew.K_LEFT and dx == 0: dx, dy = -1, 0 elif keys & pew.K_RIGHT and dx == 0: dx, dy = 1, 0 elif keys & pew.K_DOWN and dy == 0: dx, dy = 0, 1 x=(x+dx) y=(y+dy) if (x,y) in snake or x >7 or y > 7 or x < 0 or y < 0: for (i,j) in snake: screen.pixel(i,j,0) break snake.append((x,y)) if x == apple_x and y == apple_y: screen.pixel(apple_x, apple_y, 0) apple_x, apple_y = snake[0] while (apple_x, apple_y) in snake: apple_x=qrand(3) apple_y=qrand(3) screen.pixel(apple_x, apple_y, 2) game_speed += 0.2 # We add one to the points counter each time we eat an apple points=points+1 else: x, y = snake.pop(0) screen.pixel(x, y, 0) l=True # We set l to True to restart to the initial variables in the next iteration #Now let's display the points in the screen: text = pew.Pix.from_text('Points: ' + str(points)) for dx in range(-8, text.width): screen.blit(text, -dx, 1) pew.show(screen) pew.tick(1 / 12) text = pew.Pix.from_text("Game over!") for dx in range(-8, text.width): screen.blit(text, -dx, 1) pew.show(screen) pew.tick(1 / 12)
https://github.com/geduardo/Q-Snake-Qiskitcamp-Europe-2019
geduardo
import pew_tunnel as pew import random import pygame from qiskit import QuantumCircuit, QuantumRegister, ClassicalRegister, execute, Aer import numpy as np ######################################################################### #FUNCTIONS ######################################################################### simulator = Aer.get_backend('statevector_simulator') ################################################################# ################### #IBMQ real backends ################### from qiskit import IBMQ from qiskit.providers.ibmq import least_busy # Load local account information provider = IBMQ.load_account() def ibmq_qrand(nbits, N): # Get the least busy real quantum system backend = least_busy(provider.backends(simulator=False)) print(backend, backend.status().pending_jobs) circ = QuantumCircuit(1, 1) circ.h(0) circ.measure(0, 0) job = execute(circ, backend, memory=True, shots=int(N*nbits)) res = job.result() val = res.get_memory() print(val) integ = np.zeros(N) for k in range(N): #convert val array into bitstring b and then into integer b = '' shift = int(nbits * k) for i in range(nbits): b += str(val[i+shift]) integ[k] = int(b, 2) return integ def get_rand_from_list(counter, list): return list[counter] ############################################################### def qrand(nbits): """generates nbits real random numbers using quantum state measurements in qiskit.""" circ = QuantumCircuit(1, 1) circ.h(0) circ.measure(0, 0) val = np.zeros(nbits) for i in range(nbits): job=execute(circ, simulator) res = job.result() vec = res.get_statevector() val[i] = vec[0] * 1 + vec[1] * 0 #convert val array into bitstring b and then into integer b = '' for i in range(nbits): b += str(int(val[i])) integ = int(b, 2) return integ def Pt(U0, E, L, betac, gamma_sqc): """return tunneling probability for square barrier""" return 1/ (np.cosh(betac * L)**2 + gamma_sqc * np.sinh(betac * L)**2) def beta(U0, E): """supply function for Pt""" return np.sqrt(2* (U0 - E)) def gamma_sq(U0, E): """supply function for Pt""" return 0.25 * ((1 - E/U0)/(E/U0) + (E/U0)/(1-E/U0) - 2) def theta(p_tunnel): """returns rotation angle corresponding to tunneling prob. p_tunnel""" return 2 * np.arcsin(np.sqrt(p_tunnel)) def tunnelres(U0, length_snake, L, betac, gamma_sqc): """returns 0 if tunnel, returns 1 if no tunnel""" P_t = Pt(U0, length_snake, L, betac, gamma_sqc) #get tunneling prob depending on current snake length theta_rot = theta(P_t) #get rot angle #qcirc qr = QuantumRegister(1) cr = ClassicalRegister(1) circ = QuantumCircuit(qr, cr) circ.rx(theta_rot, qr[0]) circ.measure(qr, cr) job = execute(circ, simulator) res = job.result() vec = res.get_statevector() val = vec[0] * 1 + vec[1] * 0 if val == 1: return 0 else: return 1 #r= random.randint(0, 1) #return round(r) ########################################################################## #MAIN ########################################################################## #initialize pew dis = pew.init() screen = pew.Pix() #set size bits = 3 ds= 2**bits #displazsize #set game starting parameters game_speed = 4 snake = [(2, 4)] dx, dy = 1, 0 apple_x, apple_y = 6, 4 screen.pixel(apple_x, apple_y, 2) howmanyapples = 1 #marker for total number of eaten apples, used for scoring #set graphics for probability display pygame.font.init() #gate backkgorund font1 = pygame.font.Font(None, 33) text = font1.render('Probability for tunneling is', True, (255, 0, 0)) dis.blit(text, (20, 330)) font2 = pygame.font.Font(None, 45) text2 = font2.render('100%', True, (255, 0, 0)) dis.blit(text2, (130, 360)) ima = pygame.image.load('pewblack.jpg') #tunneling parameters U0=37 #max snake length = 6x6 = 36 E=1 L=0.05 #optimal barrier size for nice tunneling probabilities num_list = ibmq_qrand(3, 250) #list of random numbers for ibmq function num_list_x = num_list[:125] num_list_y = num_list[125:] #initialize tunneling tracker tunnel=0 #don't see other side as second barrier snakepos=1 #marker of snakepos, 1=head, increase towards tail headtunnel=0 #let the head tunnel again through other even if tail still in process while True: #snake runs #create barrier bar= [] for i in range(ds): screen.pixel(0, i, 1) screen.pixel(ds-1, i, 1) screen.pixel(i, 0, 1) screen.pixel(i, ds-1, 1) bar.append((0, i)) bar.append((ds-1, i)) bar.append((i, 0)) bar.append((i, ds-1)) #find the head if len(snake) > 1: x, y = snake[-2] screen.pixel(x, y, 1) x, y = snake[-1] screen.pixel(x, y, 3) #color the head yellow pew.show(screen) pew.tick(1 / game_speed) #get commands keys = pew.keys() if headtunnel==0: if keys & pew.K_UP and dy == 0: dx, dy = 0, -1 elif keys & pew.K_LEFT and dx == 0: dx, dy = -1, 0 elif keys & pew.K_RIGHT and dx == 0: dx, dy = 1, 0 elif keys & pew.K_DOWN and dy == 0: dx, dy = 0, 1 x = (x + dx) % 8 y = (y + dy) % 8 elif headtunnel==1: #steering not allowed during tunneling of the head (during two rounds) x = (x + dx) % 8 y = (y + dy) % 8 headtunnel=2 elif headtunnel>=2: x = (x + dx) % 8 y = (y + dy) % 8 headtunnel=0 ##TUNNELING PROCESS #snake tail tunnels if tunnel>0 and snakepos<=len(snake): #get segment for tunneling sx, sy = snake[-snakepos] E=len(snake)/2 #divide by two for lower tunnel prob for tail (lower mass->lower energy) tunnels = tunnelres(U0, E, L, beta(U0, E), gamma_sq(U0, E)) if tunnels==1: #tunnels snakepos+=1 else: #does not tunnel del snake[-snakepos] screen.pixel(sx, sy, 0) #reset if last segment tunneled if tunnel>0 and snakepos==(len(snake)+1): tunnel=0 snakepos=1 #snake head tunnels if headtunnel==0 and (x, y) in bar: E=len(snake) tunnel = tunnelres(U0, E, L, beta(U0, E), gamma_sq(U0, E)) if tunnel==0 and len(snake) != 1: #head doesn't tunnel --> game over break else: snakepos+=1 headtunnel+=1 elif headtunnel==1 and (x, y) in bar: headtunnel=0 #display tunneling prob. E = len(snake) if E > 1: prob = Pt(U0, E, L, beta(U0, E), gamma_sq(U0, E)) text3 = font2.render(str(int(round(prob * 100))) + '%', True, (255, 0, 0)) dis.blit(ima, (130, 360)) dis.blit(text3, (130, 360)) else: #if length of snake ==1 (only head), tunneling prob = 100% dis.blit(ima, (130, 360)) #cover the ultimate prob. display dis.blit(text2, (130, 360)) #text2 = '100%' #####TUNNEL END if (x, y) in snake: #exit, game over condition break snake.append((x, y)) #apple generation if x == apple_x and y == apple_y: screen.pixel(apple_x, apple_y, 0) apple_x, apple_y = snake[0] while (apple_x, apple_y) in snake or (apple_x, apple_y) in bar: apple_x = get_rand_from_list(howmanyapples, num_list_x) #random.getrandbits(3) #use this for pseudo random number gen, no qiskit needed apple_y = get_rand_from_list(howmanyapples, num_list_y) #random.getrandbits(3) screen.pixel(apple_x, apple_y, 2) game_speed += 0.2 #increase game speed howmanyapples += 1 #increase number of eaten apples, score +1 else: x, y = snake.pop(0) screen.pixel(x, y, 0) text = pew.Pix.from_text("Game over!") #Game over message and closing for dx in range(-8, text.width): screen.blit(text, -dx, 1) pew.show(screen) pew.tick(1 / 12) text = pew.Pix.from_text("Score:" + str(int(howmanyapples))) #Score message for dx in range(-8, text.width): screen.blit(text, -dx, 1) pew.show(screen) pew.tick(1 / 12)
https://github.com/geduardo/Q-Snake-Qiskitcamp-Europe-2019
geduardo
import pew_tunnel as pew import random import pygame from qiskit import QuantumCircuit, QuantumRegister, ClassicalRegister, execute, Aer import numpy as np ######################################################################### #FUNCTIONS ######################################################################### simulator = Aer.get_backend('statevector_simulator') def qrand(nbits): """generates nbits real random numbers using quantum state measurements in qiskit.""" circ = QuantumCircuit(1, 1) circ.h(0) circ.measure(0, 0) val = np.zeros(nbits) for i in range(nbits): job=execute(circ, simulator) res = job.result() vec = res.get_statevector() val[i] = vec[0] * 1 + vec[1] * 0 #convert val array into bitstring b and then into integer b = '' for i in range(nbits): b += str(int(val[i])) integ= int(b, 2) return integ def Pt(U0, E, L, betac, gamma_sqc): """return tunneling probability for square barrier""" return 1/ (np.cosh(betac * L)**2 + gamma_sqc * np.sinh(betac * L)**2) def beta(U0, E): """supply function for Pt""" return np.sqrt(2* (U0 - E)) def gamma_sq(U0, E): """supply function for Pt""" return 0.25 * ((1 - E/U0)/(E/U0) + (E/U0)/(1-E/U0) - 2) def theta(p_tunnel): """returns rotation angle corresponding to tunneling prob. p_tunnel""" return 2 * np.arcsin(np.sqrt(p_tunnel)) def tunnelres(U0, length_snake, L, betac, gamma_sqc): """returns 0 if tunnel, returns 1 if no tunnel""" P_t = Pt(U0, length_snake, L, betac, gamma_sqc) #get tunneling prob depending on current snake length theta_rot = theta(P_t) #get rot angle #qcirc qr = QuantumRegister(1) cr = ClassicalRegister(1) circ = QuantumCircuit(qr, cr) circ.rx(theta_rot, qr[0]) circ.measure(qr, cr) job = execute(circ, simulator) res = job.result() vec = res.get_statevector() val = vec[0] * 1 + vec[1] * 0 if val == 1: return 0 else: return 1 #r= random.randint(0, 1) #return round(r) ########################################################################## #MAIN ########################################################################## #initialize pew dis = pew.init() screen = pew.Pix() #set size bits = 3 ds= 2**bits #displazsize #set game starting parameters game_speed = 4 snake = [(2, 4)] dx, dy = 1, 0 apple_x, apple_y = 6, 4 screen.pixel(apple_x, apple_y, 2) howmanyapples = 1 #marker for total number of eaten apples, used for scoring #set graphics for probability display pygame.font.init() #gate backkgorund font1 = pygame.font.Font(None, 33) text = font1.render('Probability for tunneling is', True, (255, 0, 0)) dis.blit(text, (20, 330)) font2 = pygame.font.Font(None, 45) text2 = font2.render('100%', True, (255, 0, 0)) dis.blit(text2, (130, 360)) ima = pygame.image.load('pewblack.jpg') #tunneling parameters U0=37 #max snake length = 6x6 = 36 E=1 L=0.05 #optimal barrier size for nice tunneling probabilities #initialize tunneling tracker tunnel=0 #don't see other side as second barrier snakepos=1 #marker of snakepos, 1=head, increase towards tail headtunnel=0 #let the head tunnel again through other even if tail still in process while True: #snake runs #create barrier bar= [] for i in range(ds): screen.pixel(0, i, 1) screen.pixel(ds-1, i, 1) screen.pixel(i, 0, 1) screen.pixel(i, ds-1, 1) bar.append((0, i)) bar.append((ds-1, i)) bar.append((i, 0)) bar.append((i, ds-1)) #find the head if len(snake) > 1: x, y = snake[-2] screen.pixel(x, y, 1) x, y = snake[-1] screen.pixel(x, y, 3) #color the head yellow pew.show(screen) pew.tick(1 / game_speed) #get commands keys = pew.keys() if headtunnel==0: if keys & pew.K_UP and dy == 0: dx, dy = 0, -1 elif keys & pew.K_LEFT and dx == 0: dx, dy = -1, 0 elif keys & pew.K_RIGHT and dx == 0: dx, dy = 1, 0 elif keys & pew.K_DOWN and dy == 0: dx, dy = 0, 1 x = (x + dx) % 8 y = (y + dy) % 8 elif headtunnel==1: #steering not allowed during tunneling of the head (during two rounds) x = (x + dx) % 8 y = (y + dy) % 8 headtunnel=2 elif headtunnel>=2: x = (x + dx) % 8 y = (y + dy) % 8 headtunnel=0 ##TUNNELING PROCESS #snake tail tunnels if tunnel>0 and snakepos<=len(snake): #get segment for tunneling sx, sy = snake[-snakepos] E=len(snake)/2 #divide by two for lower tunnel prob for tail (lower mass->lower energy) tunnels = tunnelres(U0, E, L, beta(U0, E), gamma_sq(U0, E)) if tunnels==1: #tunnels snakepos+=1 else: #does not tunnel del snake[-snakepos] screen.pixel(sx, sy, 0) #reset if last segment tunneled if tunnel>0 and snakepos==(len(snake)+1): tunnel=0 snakepos=1 #snake head tunnels if headtunnel==0 and (x, y) in bar: E=len(snake) tunnel = tunnelres(U0, E, L, beta(U0, E), gamma_sq(U0, E)) if tunnel==0 and len(snake) != 1: #head doesn't tunnel --> game over break else: snakepos+=1 headtunnel+=1 elif headtunnel==1 and (x, y) in bar: headtunnel=0 #display tunneling prob. E = len(snake) if E > 1: prob = Pt(U0, E, L, beta(U0, E), gamma_sq(U0, E)) text3 = font2.render(str(int(round(prob * 100))) + '%', True, (255, 0, 0)) dis.blit(ima, (130, 360)) dis.blit(text3, (130, 360)) else: #if length of snake ==1 (only head), tunneling prob = 100% dis.blit(ima, (130, 360)) #cover the ultimate prob. display dis.blit(text2, (130, 360)) #text2 = '100%' #####TUNNEL END if (x, y) in snake: #exit, game over condition break snake.append((x, y)) #apple generation if x == apple_x and y == apple_y: screen.pixel(apple_x, apple_y, 0) apple_x, apple_y = snake[0] while (apple_x, apple_y) in snake or (apple_x, apple_y) in bar: apple_x = qrand(bits) #random.getrandbits(3) #use this for pseudo random number gen, no qiskit needed apple_y = qrand(bits) #random.getrandbits(3) screen.pixel(apple_x, apple_y, 2) game_speed += 0.2 #increase game speed howmanyapples += 1 #increase number of eaten apples, score +1 else: x, y = snake.pop(0) screen.pixel(x, y, 0) text = pew.Pix.from_text("Game over!") #Game over message and closing for dx in range(-8, text.width): screen.blit(text, -dx, 1) pew.show(screen) pew.tick(1 / 12) text = pew.Pix.from_text("Score:" + str(int(howmanyapples))) #Score message for dx in range(-8, text.width): screen.blit(text, -dx, 1) pew.show(screen) pew.tick(1 / 12)
https://github.com/geduardo/Q-Snake-Qiskitcamp-Europe-2019
geduardo
import pew_tunnel as pew import random import pygame from qiskit import QuantumCircuit, QuantumRegister, ClassicalRegister, execute, Aer import numpy as np ######################################################################### #FUNCTIONS ######################################################################### simulator = Aer.get_backend('statevector_simulator') def qrand(nbits): """generates nbits real random numbers using quantum state measurements in qiskit.""" circ = QuantumCircuit(1, 1) circ.h(0) circ.measure(0, 0) val = np.zeros(nbits) for i in range(nbits): job=execute(circ, simulator) res = job.result() vec = res.get_statevector() val[i] = vec[0] * 1 + vec[1] * 0 #convert val array into bitstring b and then into integer b = '' for i in range(nbits): b += str(int(val[i])) integ= int(b, 2) return integ def Pt(U0, E, L, betac, gamma_sqc): """return tunneling probability for square barrier""" return 1/ (np.cosh(betac * L)**2 + gamma_sqc * np.sinh(betac * L)**2) def beta(U0, E): """supply function for Pt""" return np.sqrt(2* (U0 - E)) def gamma_sq(U0, E): """supply function for Pt""" return 0.25 * ((1 - E/U0)/(E/U0) + (E/U0)/(1-E/U0) - 2) def theta(p_tunnel): """returns rotation angle corresponding to tunneling prob. p_tunnel""" return 2 * np.arcsin(np.sqrt(p_tunnel)) def tunnelres(U0, length_snake, L, betac, gamma_sqc): """returns 0 if tunnel, returns 1 if no tunnel""" P_t = Pt(U0, length_snake, L, betac, gamma_sqc) #get tunneling prob depending on current snake length theta_rot = theta(P_t) #get rot angle #qcirc qr = QuantumRegister(1) cr = ClassicalRegister(1) circ = QuantumCircuit(qr, cr) circ.rx(theta_rot, qr[0]) circ.measure(qr, cr) job = execute(circ, simulator) res = job.result() vec = res.get_statevector() val = vec[0] * 1 + vec[1] * 0 if val == 1: return 0 else: return 1 #r= random.randint(0, 1) #return round(r) def Qand(First_bool,Second_bool): a = int(First_bool) b = int(Second_bool) q = QuantumRegister(3) qc = QuantumCircuit(q) if a is 1: qc.x(0) if b is 1: qc.x(1) qc.ccx(q[0], q[1], q[2]) qc.draw() backend = Aer.get_backend('statevector_simulator') job = execute(qc, backend) result = job.result() outputstate = result.get_statevector(qc, decimals=8) Q_and = bool(int(np.real(outputstate[7]))) return Q_and def Qor(First_bool,Second_bool): a = int(First_bool) b = int(Second_bool) q = QuantumRegister(3) qc = QuantumCircuit(q) if a is 1: qc.x(0) if b is 1: qc.x(1) qc.ccx(q[0], q[1], q[2]) qc.draw() backend = Aer.get_backend('statevector_simulator') job = execute(qc, backend) result = job.result() outputstate = result.get_statevector(qc, decimals=8) if a is 1 and b is 0: Q_or = bool(int(np.real(outputstate[1]))) elif a is 0 and b is 1: Q_or = bool(int(np.real(outputstate[2]))) else: Q_or = bool(int(np.real(outputstate[7]))) return(Q_or) ########################################################################## #MAIN ########################################################################## #initialize pew dis = pew.init() screen = pew.Pix() #set size bits = 3 ds= 2**bits #displazsize #set game starting parameters game_speed = 4 snake = [(2, 4)] dx, dy = 1, 0 apple_x, apple_y = 6, 4 screen.pixel(apple_x, apple_y, 2) howmanyapples = 1 #marker for total number of eaten apples, used for scoring #set graphics for probability display pygame.font.init() #gate backkgorund font1 = pygame.font.Font(None, 33) text = font1.render('Probability for tunneling is', True, (255, 0, 0)) dis.blit(text, (20, 330)) font2 = pygame.font.Font(None, 45) text2 = font2.render('100%', True, (255, 0, 0)) dis.blit(text2, (130, 360)) ima = pygame.image.load('pewblack.jpg') #tunneling parameters U0=37 #max snake length = 6x6 = 36 E=1 L=0.05 #optimal barrier size for nice tunneling probabilities #initialize tunneling tracker tunnel=0 #don't see other side as second barrier snakepos=1 #marker of snakepos, 1=head, increase towards tail headtunnel=0 #let the head tunnel again through other even if tail still in process while True: #snake runs #create barrier bar= [] for i in range(ds): screen.pixel(0, i, 1) screen.pixel(ds-1, i, 1) screen.pixel(i, 0, 1) screen.pixel(i, ds-1, 1) bar.append((0, i)) bar.append((ds-1, i)) bar.append((i, 0)) bar.append((i, ds-1)) #find the head if len(snake) > 1: x, y = snake[-2] screen.pixel(x, y, 1) x, y = snake[-1] screen.pixel(x, y, 3) #color the head yellow pew.show(screen) pew.tick(1 / game_speed) #get commands keys = pew.keys() if headtunnel==0: if Qand(bool(int(keys & pew.K_UP)),dy == 0) is True: dx, dy = 0, -1 #print(bool(int(keys & pew.K_UP))) #print(dy == 0) #print(Qand(bool(int(keys & pew.K_UP)),dy == 0)) elif Qand(bool(int(keys & pew.K_LEFT)),dx == 0) is True: #print(str(keys & pew.K_LEFT) + 'right') dx, dy = -1, 0 elif Qand(bool(int(keys & pew.K_RIGHT)),dx == 0) is True: #print(str(keys & pew.K_RIGHT) + 'right') dx, dy = 1, 0 elif Qand(bool(int(keys & pew.K_DOWN)),dy == 0) is True: #print(str(keys & pew.K_DOWN) + 'down') dx, dy = 0, 1 x = (x + dx) % 8 y = (y + dy) % 8 elif headtunnel==1: #steering not allowed during tunneling of the head (during two rounds) x = (x + dx) % 8 y = (y + dy) % 8 headtunnel=2 elif headtunnel>=2: x = (x + dx) % 8 y = (y + dy) % 8 headtunnel=0 ##TUNNELING PROCESS #snake tail tunnels if Qand(tunnel>0,snakepos<=len(snake)) is True: #get segment for tunneling sx, sy = snake[-snakepos] E=len(snake)/2 #divide by two for lower tunnel prob for tail (lower mass->lower energy) tunnels = tunnelres(U0, E, L, beta(U0, E), gamma_sq(U0, E)) if tunnels==1: #tunnels snakepos+=1 else: #does not tunnel del snake[-snakepos] screen.pixel(sx, sy, 0) #reset if last segment tunneled if Qand(tunnel>0, snakepos==(len(snake)+1)) is True: tunnel=0 snakepos=1 #snake head tunnels if Qand(headtunnel==0, (x, y) in bar) is True: E=len(snake) tunnel = tunnelres(U0, E, L, beta(U0, E), gamma_sq(U0, E)) if Qand(tunnel==0, len(snake) != 1) is True: #head doesn't tunnel --> game over break else: snakepos+=1 headtunnel+=1 elif Qand(headtunnel==1, (x, y) in bar) is True: headtunnel=0 #display tunneling prob. E = len(snake) if E > 1: prob = Pt(U0, E, L, beta(U0, E), gamma_sq(U0, E)) text3 = font2.render(str(int(round(prob * 100))) + '%', True, (255, 0, 0)) dis.blit(ima, (130, 360)) dis.blit(text3, (130, 360)) else: #if length of snake ==1 (only head), tunneling prob = 100% dis.blit(ima, (130, 360)) #cover the ultimate prob. display dis.blit(text2, (130, 360)) #text2 = '100%' #####TUNNEL END if (x, y) in snake: #exit, game over condition break snake.append((x, y)) #apple generation if Qand(x == apple_x, y == apple_y) is True: screen.pixel(apple_x, apple_y, 0) apple_x, apple_y = snake[0] while Qor((apple_x, apple_y) in snake, (apple_x, apple_y) in bar) is True: apple_x = qrand(bits) #random.getrandbits(3) #use this for pseudo random number gen, no qiskit needed apple_y = qrand(bits) #random.getrandbits(3) screen.pixel(apple_x, apple_y, 2) game_speed += 0.2 #increase game speed howmanyapples += 1 #increase number of eaten apples, score +1 else: x, y = snake.pop(0) screen.pixel(x, y, 0) text = pew.Pix.from_text("Game over!") #Game over message and closing for dx in range(-8, text.width): screen.blit(text, -dx, 1) pew.show(screen) pew.tick(1 / 12) text = pew.Pix.from_text("Score:" + str(int(howmanyapples))) #Score message for dx in range(-8, text.width): screen.blit(text, -dx, 1) pew.show(screen) pew.tick(1 / 12) pygame.quit()
https://github.com/geduardo/Q-Snake-Qiskitcamp-Europe-2019
geduardo
import pew_tunnel as pew import pygame from qiskit import QuantumCircuit, QuantumRegister, ClassicalRegister, execute, Aer import numpy as np ######################################################################### #FUNCTIONS ######################################################################### simulator = Aer.get_backend('qasm_simulator') shot=1 def Qand(First_bool,Second_bool): First_bool=bool(First_bool) Second_bool=bool(Second_bool) a = int(First_bool) b = int(Second_bool) qc = QuantumCircuit(3,1) if a == 1: qc.x(0) if b == 1: qc.x(1) qc.ccx(0, 1, 2) #toffoli qc.measure(2,0) job = execute(qc, simulator,shots=shot) counts = job.result().get_counts() state = [key for (key, value) in counts.items() if value ==1] return bool(int(state[0])) def Qnand(First_bool,Second_bool): First_bool=bool(First_bool) Second_bool=bool(Second_bool) a = int(First_bool) b = int(Second_bool) qc = QuantumCircuit(3,1) if a == 1: qc.x(0) if b == 1: qc.x(1) qc.ccx(0, 1, 2) #toffoli qc.x(2) qc.measure(2,0) job = execute(qc, simulator,shots=shot) counts = job.result().get_counts() state = [key for (key, value) in counts.items() if value ==1] return bool(int(state[0])) def Qor(First_bool,Second_bool): return Qnand(Qnand(First_bool,First_bool),Qnand(Second_bool,Second_bool)) def qrand(nbits): """generates nbits real random numbers using quantum state measurements in qiskit.""" circ = QuantumCircuit(1, 1) circ.h(0) circ.measure(0, 0) b=''#string holder for i in range(nbits): job=execute(circ, simulator,shots=shot) counts = job.result().get_counts() state = [key for (key, value) in counts.items() if value == 1] #find the measured state, this is a list b=b+state[0] #state[0] is a string return int(b, 2) def Pt(U0, E, L, betac, gamma_sqc): """return tunneling probability for square barrier""" return 1/ (np.cosh(betac * L)**2 + gamma_sqc * np.sinh(betac * L)**2) def beta(U0, E): """supply function for Pt""" return np.sqrt(2* (U0 - E)) def gamma_sq(U0, E): """supply function for Pt""" return 0.25 * ((1 - E/U0)/(E/U0) + (E/U0)/(1-E/U0) - 2) def theta(p_tunnel): """returns rotation angle corresponding to tunneling prob. p_tunnel""" return 2 * np.arcsin(np.sqrt(p_tunnel)) def tunnelres(U0, length_snake, L, betac, gamma_sqc): """returns 0 if tunnel, returns 1 if no tunnel""" P_t = Pt(U0, length_snake, L, betac, gamma_sqc) #get tunneling prob depending on current snake length theta_rot = theta(P_t) #get rot angle qr = QuantumRegister(1) cr = ClassicalRegister(1) circ = QuantumCircuit(qr, cr) circ.rx(theta_rot, qr[0]) circ.measure(qr, cr) job = execute(circ, simulator,shots=shot) counts = job.result().get_counts() state = [key for (key, value) in counts.items() if value == 1] return int(state[0]) ########################################################################## #MAIN ########################################################################## #initialize pew dis = pew.init() screen = pew.Pix() #set size bits = 3 ds= 2**bits #displazsize #set game starting parameters game_speed = 4 snake = [(2, 4)] dx, dy = 1, 0 apple_x, apple_y = 6, 5 screen.pixel(apple_x, apple_y, 1) howmanyapples = 1 #marker for total number of eaten apples, used for scoring #set graphics for probability display pygame.font.init() #gate backkgorund font1 = pygame.font.Font(None, 33) text = font1.render('Probability for tunneling is', True, (255, 0, 0)) dis.blit(text, (20, 330)) font2 = pygame.font.Font(None, 45) text2 = font2.render('100%', True, (255, 0, 0)) dis.blit(text2, (130, 360)) ima = pygame.image.load('pewblack.jpg') #tunneling parameters U0=37 #max snake length = 6x6 = 36 E=1 L=0.05 #optimal barrier size for nice tunneling probabilities #initialize tunneling tracker tunnel=0 #don't see other side as second barrier snakepos=1 #marker of snakepos, 1=head, increase towards tail headtunnel=0 #let the head tunnel again through other even if tail still in process while True: #snake runs #create barrier bar= [] for i in range(ds): screen.pixel(0, i, 2) screen.pixel(ds-1, i, 2) screen.pixel(i, 0, 2) screen.pixel(i, ds-1, 2) bar.append((0, i)) bar.append((ds-1, i)) bar.append((i, 0)) bar.append((i, ds-1)) #find the head if len(snake) > 1: x, y = snake[-2] screen.pixel(x, y, 1) x, y = snake[-1] screen.pixel(x, y, 3) #color the head yellow pew.show(screen) pew.tick(1 / game_speed) #get commands keys = pew.keys() if headtunnel==0: if keys & pew.K_UP and dy == 0: dx, dy = 0, -1 elif keys & pew.K_LEFT and dx == 0: dx, dy = -1, 0 elif keys & pew.K_RIGHT and dx == 0: dx, dy = 1, 0 elif keys & pew.K_DOWN and dy == 0: dx, dy = 0, 1 elif headtunnel==1: #steering not allowed during tunneling of the head (during two rounds) headtunnel=2 elif headtunnel>=2: headtunnel=0 x = (x + dx) % 8 y = (y + dy) % 8 ##TUNNELING PROCESS #snake tail tunnels if Qand(tunnel>0 ,snakepos<=len(snake)): #get segment for tunneling sx, sy = snake[-snakepos] E=len(snake)/2 #divide by two for lower tunnel prob for tail (lower mass->lower energy) tunnels = tunnelres(U0, E, L, beta(U0, E), gamma_sq(U0, E)) if tunnels==1: #tunnels snakepos+=1 else: #does not tunnel del snake[-snakepos] screen.pixel(sx, sy, 0) #reset if last segment tunneled if Qand(tunnel>0 ,snakepos==(len(snake)+1)): tunnel=0 snakepos=1 #snake head tunnels if Qand(headtunnel==0, (x, y) in bar): E=len(snake) tunnel = tunnelres(U0, E, L, beta(U0, E), gamma_sq(U0, E)) if Qand(tunnel==0, len(snake) != 1): #head doesn't tunnel --> game over break else: snakepos+=1 headtunnel+=1 elif headtunnel==1 and (x, y) in bar: headtunnel=0 #display tunneling prob. E = len(snake) if E > 1: prob = Pt(U0, E, L, beta(U0, E), gamma_sq(U0, E)) text3 = font2.render(str(int(round(prob * 100))) + '%', True, (255, 0, 0)) dis.blit(ima, (130, 360)) dis.blit(text3, (130, 360)) else: #if length of snake ==1 (only head), tunneling prob = 100% dis.blit(ima, (130, 360)) #cover the ultimate prob. display dis.blit(text2, (130, 360)) #text2 = '100%' #####TUNNEL END if (x, y) in snake: #exit, game over condition break snake.append((x, y)) #apple generation if Qand(x == apple_x, y == apple_y): screen.pixel(apple_x, apple_y, 0) apple_x, apple_y = snake[0] while Qor((apple_x, apple_y) in snake , (apple_x, apple_y) in bar): apple_x = qrand(bits) #random.getrandbits(3) #use this for pseudo random number gen, no qiskit needed apple_y = qrand(bits) #random.getrandbits(3) screen.pixel(apple_x, apple_y, 1) game_speed += 0.2 #increase game speed howmanyapples += 1 #increase number of eaten apples, score +1 else: x, y = snake.pop(0) screen.pixel(x, y, 0) text = pew.Pix.from_text("Game over!") #Game over message and closing for dx in range(-8, text.width): screen.blit(text, -dx, 1) pew.show(screen) pew.tick(1 / 12) text = pew.Pix.from_text("Score:" + str(int(howmanyapples))) #Score message for dx in range(-8, text.width): screen.blit(text, -dx, 1) pew.show(screen) pew.tick(1 / 12) pygame.quit()
https://github.com/geduardo/Q-Snake-Qiskitcamp-Europe-2019
geduardo
import pew_circuit as pew from qiskit import QuantumCircuit, execute, Aer #from aether import QuantumCircuit, simulate import pygame import numpy as np simulator = Aer.get_backend('qasm_simulator') shot=1 def Qand(First_bool,Second_bool): First_bool=bool(First_bool) Second_bool=bool(Second_bool) a = int(First_bool) b = int(Second_bool) qc = QuantumCircuit(3,1) if a == 1: qc.x(0) if b == 1: qc.x(1) qc.ccx(0, 1, 2) #toffoli qc.measure(2,0) job = execute(qc, simulator,shots=shot) counts = job.result().get_counts() state = [key for (key, value) in counts.items() if value ==1] return bool(int(state[0])) def Qnand(First_bool,Second_bool): First_bool=bool(First_bool) Second_bool=bool(Second_bool) a = int(First_bool) b = int(Second_bool) qc = QuantumCircuit(3,1) if a == 1: qc.x(0) if b == 1: qc.x(1) qc.ccx(0, 1, 2) #toffoli qc.x(2) qc.measure(2,0) job = execute(qc, simulator,shots=shot) counts = job.result().get_counts() state = [key for (key, value) in counts.items() if value ==1] return bool(int(state[0])) def Qor(First_bool,Second_bool): return Qnand(Qnand(First_bool,First_bool),Qnand(Second_bool,Second_bool)) def qrand(nbits): """generates nbits real random numbers using quantum state measurements in qiskit.""" circ = QuantumCircuit(1, 1) circ.h(0) circ.measure(0, 0) b=''#string holder for i in range(nbits): job=execute(circ, simulator,shots=shot) counts = job.result().get_counts() state = [key for (key, value) in counts.items() if value == 1] #find the measured state, this is a list b=b+state[0] #state[0] is a string return int(b, 2) dis = pew.init() screen = pew.Pix() #Selecting game speed init_speed =3 game_speed = init_speed #circuit graphics pygame.font.init() #add backgorund pygame.draw.rect(dis,(255,255,255),(0,320,320,100)) #gate backkgorund #add gate, change from H to Z to Meas by key stroke font1 = pygame.font.Font(None, 70) text = font1.render('H', True, (255, 0, 0)) dis.blit(text, (10, 330)) g=0 #add gate, change from H to Z to Meas by key stroke gates = np.zeros(3, dtype=str) gates[0] = 'H' gates[1] = 'Z' gates[2] = 'M' corr = [0, 1, 0, 2] #correct sequence of gates images = {} for i in range(4): imag_ = pygame.image.load(str(i+1) + 'c.png') x_size = int((i+1)*65) images[i] = imag_#pygame.transform.scale(imag_,(x_size,85)) g=0 #tracks selected gate currg=0 #tracks current position in circuit success=False #tracks whether or not circuit is successfully constructed ima3 = pygame.image.load('1c.png') #ima3 = pygame.transform.scale(ima3,(65,50)) dis.blit(ima3, (45, 325)) #Selecting initial position of the snake snake = [(3,3)] #Selecting initial velocity of the snake dx, dy = 1, 0 #Selecting initial position of the apple apple_x, apple_y = 6, 4 screen.pixel(apple_x, apple_y, 1) #Selecting the initial position of the first noise nx=3 ny=4 noise=[(nx, ny)] while True: screen.pixel(noise[0][0], noise[0][1],2) #Here we print the head of the snake if len(snake) > 1: x, y = snake[-2] screen.pixel(x, y, 1) x, y = snake[-1] screen.pixel(x, y, 3) pew.show(screen) pew.tick(1 / game_speed) #Here we change the velocity of the snake depending on the key input keys = pew.keys() if keys & pew.K_UP and dy == 0: dx, dy = 0, -1 elif keys & pew.K_LEFT and dx == 0: dx, dy = -1, 0 elif keys & pew.K_RIGHT and dx == 0: dx, dy = 1, 0 elif keys & pew.K_DOWN and dy == 0: dx, dy = 0, 1 #Gate switch if keys & pew.K_O: pygame.draw.rect(dis,(255,255,255),(0,320,50,50)) g = (g+1) % 3 #cycle between gates font1 = pygame.font.Font(None, 70) text = font1.render(gates[g], True, (255, 0, 0)) dis.blit(text, (10, 330)) #Now are going to update the position of the head of the snake #We define the next position of the head depending on the velocity x = (x + dx) %8 y = (y + dy) %8 #Now we define a loop to end the loop (and the game) if the next position # of the head is in the snake or it goes out of the grid if Qor((x, y) in snake, (x,y) in noise): #Here we turn of all the pixles from the snake and the apple for (i,j) in snake: screen.pixel(i,j,0) screen.pixel(apple_x,apple_y,0) for (i,j) in noise: screen.pixel(i,j,0) break #If none of those thing happens the head of the snake gets updated to the new position snake.append((x, y)) #Now we create a conditional loop for changing the size and spawning new apples depending # on if the snake eats the apple or not if Qand(x == apple_x, y == apple_y): if g != corr[currg]: break else: if currg == len(corr)-1: #if the required sequence is accomplished success=True break dis.blit(images[currg+1], (45, 325)) currg+=1 #If the snake eats the apple we turn of the pixel of the apple screen.pixel(apple_x, apple_y, 0) #Now we define the coordinates of the apple to lie inside the snake for the loop apple_x, apple_y = snake[0] #We create a loop to generate an apple ouside the snake while (apple_x, apple_y) in snake or (apple_x, apple_y) in noise: apple_x=qrand(3) apple_y=qrand(3) #We light the pixels of the new apple screen.pixel(apple_x, apple_y, 1) # k=k+1 nx=snake[0][0] ny=snake[0][1] while Qor(Qor(Qor((nx,ny) in snake, (Qand(nx==apple_x, ny==apple_y))) ,(nx,ny) in noise) , (nx,ny)==(x,y)): nx=qrand(3) #as opposed to random.getrandbits(3) ny=qrand(3) noise.append((nx,ny)) screen.pixel(nx, ny,2) #We increase the speed of the game game_speed += 0.2 #If the snake eats the apple we don't delete the last pixel of the snake. Otherwise we remove # the last pixel to not increase the size of the snake else: x, y = snake.pop(0) screen.pixel(x, y, 0) def blit_screen(screen,text): for dx in range(-8, text.width): screen.blit(text, -dx, 1) pew.show(screen) pew.tick(1 / 12) #When the loop is finished we print the game over screen #turn off all pixels if success == True: text = pew.Pix.from_text("You win!") for dx in range(-8, text.width): screen.blit(text, -dx, 1) pew.show(screen) pew.tick(1 / 12) else: text = pew.Pix.from_text("Game over!") blit_screen(screen,text) for dx in range(-8, text.width): screen.blit(text, -dx, 1) pew.show(screen) pew.tick(1 / 12) pygame.quit()
https://github.com/geduardo/Q-Snake-Qiskitcamp-Europe-2019
geduardo
import numpy as np from qiskit import * q = QuantumRegister(1) # Create a Quantum Circuit acting on a quantum register of n qubits qc = QuantumCircuit(q) qc.draw(filename="0.png") qc.h(0) qc.draw() qc.z(0) qc.draw() qc.h(0) qc.draw() qc.measure(0,0) qc.draw() from qiskit.visualization import plot_bloch_vector plot_bloch_vector([0,0,1], title='Intial state') plot_bloch_vector([1,0,0], title='H gate') plot_bloch_vector([-1,0,0], title='Z gate') plot_bloch_vector([0,0,-1], title='H gate')
https://github.com/geduardo/Q-Snake-Qiskitcamp-Europe-2019
geduardo
import numpy as np from qiskit import * q = QuantumRegister(1) # Create a Quantum Circuit acting on a quantum register of n qubits qc = QuantumCircuit(q) diag = qc.draw(output="mpl") diag.savefig("1.png", format="png") qc.h(0) diag = qc.draw(output="mpl") diag.savefig("2.png", format="png") qc.z(0) diag = qc.draw(output="mpl") diag.savefig("3.png", format="png") qc.h(0) diag = qc.draw(output="mpl") diag.savefig("4.png", format="png") #Cropping images !pip3 install Pillow crp = [0,7,17,25] from PIL import Image for i in range(1,5): image = Image.open(str(i)+".png") w,h = image.size image.crop((20+crp[i-1],20,w,h-20)).save(str(i)+"c.png") qc.measure(0,0) qc.draw() from qiskit.visualization import plot_bloch_vector plot_bloch_vector([0,0,1], title='Intial state') plot_bloch_vector([1,0,0], title='H gate') plot_bloch_vector([-1,0,0], title='Z gate') plot_bloch_vector([0,0,-1], title='H gate')
https://github.com/GlazeDonuts/QSVM
GlazeDonuts
import matplotlib.pyplot as plt import numpy as np from qiskit import Aer from qiskit.ml.datasets import ad_hoc_data, sample_ad_hoc_data, breast_cancer from qiskit.circuit.library import ZZFeatureMap from qiskit.aqua.utils import split_dataset_to_data_and_labels, map_label_to_class_name from qiskit.aqua import QuantumInstance from qiskit.aqua.algorithms import QSVM feature_dim = 2 # Number of input features sample_total, train_data, test_data, labels = ad_hoc_data(training_size=20, test_size=10, n=feature_dim, gap=0.3, plot_data=True) sample_test_data = sample_ad_hoc_data(sample_total, 10, n=feature_dim) data_pts, class2label = split_dataset_to_data_and_labels(sample_test_data) print("Classes and corresponding labels are:") for c in class2label: print(f"Class: {c}, Label: {class2label[c]}") seed = 10598 # Setting seed to ensure reproducable results feature_map = ZZFeatureMap(feature_dimension=feature_dim, reps=2, entanglement='linear') qsvm = QSVM(feature_map, train_data, test_data, data_pts[0]) backend = Aer.get_backend('qasm_simulator') quantum_instance = QuantumInstance(backend, shots=1024, seed_simulator=seed, seed_transpiler=seed) result = qsvm.run(quantum_instance) print(f"Testing Accuracy: {result['testing_accuracy'] * 100}%") print("Prediction on Datapoints:") print(f"Ground Truth: {map_label_to_class_name(data_pts[1], qsvm.label_to_class)}") print(f"Predictions: {result['predicted_classes']}") print("Trained Kernel Matrix:") kernel_matrix = result['kernel_matrix_training'] img = plt.imshow(np.asmatrix(kernel_matrix),interpolation='nearest',origin='upper',cmap='bone_r') plt.show() feature_dim = 2 # Number of input features sample_total_bc, train_data_bc, test_data_bc, labels_bc = breast_cancer(training_size=20, test_size=10, n=feature_dim, plot_data=True) seed = 10598 # Setting seed to ensure reproducable results feature_map_bc = ZZFeatureMap(feature_dimension=feature_dim, reps=2, entanglement='linear') qsvm_bc = QSVM(feature_map_bc, train_data_bc, test_data_bc) backend = Aer.get_backend('qasm_simulator') quantum_instance_bc = QuantumInstance(backend, shots=1024, seed_simulator=seed, seed_transpiler=seed) result_bc = qsvm_bc.run(quantum_instance_bc) print(f"Testing Accuracy: {result_bc['testing_accuracy'] * 100}%") print("Trained Kernel Matrix:") kernel_matrix_bc = result_bc['kernel_matrix_training'] img = plt.imshow(np.asmatrix(kernel_matrix_bc),interpolation='nearest',origin='upper',cmap='bone_r') plt.show()
https://github.com/petr-ivashkov/qiskit-cert-workbook
petr-ivashkov
#initialization import matplotlib.pyplot as plt import numpy as np #ignore deprecation warnings because they are annoying (not recommened generally) from warnings import simplefilter simplefilter(action='ignore', category=DeprecationWarning) # Importing standard Qiskit libraries from qiskit import * from qiskit.tools.jupyter import * from qiskit.visualization import * from qiskit.quantum_info import * from qiskit.circuit.library import * from ibm_quantum_widgets import * NUM_QUBITS = 2 # let's construct a multi-qubit quantum register qr = QuantumRegister(NUM_QUBITS, 'q') qc = QuantumCircuit(qr, name='my-circuit') # let's create a Bell's state qc.h(0) qc.cx(0,1) qc.measure_all() qc.draw() # let's construct a multi-bit classical register cr = ClassicalRegister(NUM_QUBITS, 'c') qc = QuantumCircuit(qr, cr, name='my-circuit') # explicitly measure qubits [0,1] into classical bits [0,1] qc.measure(qr, cr) # alternatively: qc.measure([0,1], [0,1]) qc.draw() # Bell state 0 bell_0 = QuantumCircuit(2) bell_0.h(0) bell_0.cx(0,1) sv = Statevector.from_label('00') #evolve the initial state through the circuit sv_bell1 = sv.evolve(bell_0) sv_bell1.draw('latex') # Bell state 1 bell_1 = QuantumCircuit(2) bell_1.x(0) bell_1.h(0) bell_1.cx(0,1) sv = Statevector.from_label('00') #evolve the initial state through the circuit sv_bell2 = sv.evolve(bell_1) sv_bell2.draw('latex') # Bell state 2 bell_2 = QuantumCircuit(2) bell_2.x(0) bell_2.h(1) bell_2.cx(1,0) sv = Statevector.from_label('00') #evolve the initial state through the circuit sv_bell3 = sv.evolve(bell_2) sv_bell3.draw('latex') # create the last remaining Bell state on your own bell_3 = QuantumCircuit(2) ### ADD CODE BELOW ### DON'T ADD CODE BELOW sv = Statevector.from_label('00') #evolve the initial state through the circuit sv_bell3 = sv.evolve(bell_3) sv_bell3.draw('latex') # state of equal superposition of basis states max_sp = QuantumCircuit(2) max_sp.h(0) max_sp.h(1) sv = Statevector.from_label('00') # evolve the initial state through the circuit ### ADD CODE BELOW ### DON'T ADD CODE BELOW qc = QuantumCircuit(1) qc.h(0) qc.h(0) qc.draw('mpl') backend = BasicAer.get_backend('qasm_simulator') qc = QuantumCircuit(1) qc.h(0) qc.h(0) transpile(qc, backend=backend).draw('mpl') backend = BasicAer.get_backend('qasm_simulator') qc = QuantumCircuit(1) qc.h(0) qc.barrier() qc.h(0) transpile(qc, backend=backend).draw('mpl') # try it out yourself with different gates: instead of Hadamard, try Pauili gates my_qc = QuantumCircuit(1) ### ADD CODE BELOW ### DON'T ADD CODE BELOW transpile(my_qc, backend=backend).draw('mpl') # did you get what you expected? qc = QuantumCircuit(1) qc.x(0) qc.depth() qc = QuantumCircuit(2) qc.x([0,1]) qc.x(1) qc.depth() # draw the circuit yourself to see why the depth has increased now ### ADD CODE BELOW ### DON'T ADD CODE BELOW qc = QuantumCircuit(2) qc.x(0) qc.barrier(0) qc.h(0) qc.cx(0,1) qc.depth() # draw the circuit yourself to see why the depth has increased now # hint: the longest path is not always the lenth of the longest sequence along one channel qc.draw('mpl') # the second qubit only has a CX gate applied to it, but has to "wait" for the first qubit to provide the control # also, barier doesn't count # in the following we will use the implementation of Grover's algorithm in Qiskit from qiskit.algorithms import Grover, AmplificationProblem # let's construct an oracle with a single solution sv = Statevector.from_label('010') problem = AmplificationProblem(oracle=sv) grover = Grover(iterations=2).construct_circuit(problem=problem) # the original circuit contains unitaries # transpile the circuit in terms of basis gates to see what it is "made of" t_grover = transpile(grover, basis_gates=['cx', 'u3']) t_grover.draw('mpl', fold=-1) # draw the original cicuit yourself grover.draw('mpl', fold=-1) # lets's try to evolve the initial state through the circuit like we have done for Bell states sv = Statevector.from_label('000') sv_ev = sv.evolve(grover) sv_ev.draw('latex') # it is obvious from statevector evolutions, that the state 010 # has the best chances to be measured, as desired # now let's obtain the same result from the QASM simulation # for that we need to measure the qubits in the end of our circuit grover.measure_all() backend= BasicAer.get_backend('qasm_simulator') job = execute(grover, backend, shots=1024) result = job.result() plot_histogram(result.get_counts()) # finally let's see how our circuit performs on actual hardware # you will need your IBM Quantum account IBMQ.load_account() provider = IBMQ.get_provider(hub='ibm-q') # sometimes you might need to wait when running on real devices # let's monitor the status of our job from qiskit.tools import job_monitor # to call a real backend let's pick the least busy one, which satisfies out requirments from qiskit.providers.ibmq import least_busy suitable_devices = provider.backends(filters = lambda b: b.configuration().n_qubits >= 3 and not b.configuration().simulator and b.status().operational==True) backend = least_busy(backends=suitable_devices) print("We are runnig our circuit on ", backend) job = execute(grover, backend=backend) job_monitor(job) # we see that results from a real device also match our expectations # the algorithm successfully finds the marked string result = job.result() counts = result.get_counts() # plot the histogramm yourself ### ADD CODE BELOW ### DON'T ADD CODE BELOW # let's again use our transpiled Grover circuit to demonstarte the idea # Notice: QASM doesn't look "nice". This is a low-level language which can be understood by quantum hardware. # If you want to take a look at the QASM code make sure to specify 'formatted=True' which makes the code readable by a human being. t_grover.qasm(formatted=True) # store the circuit to a file in the follwing way qasm_code = t_grover.qasm(filename='grover.qasm') # to get the circuit back from QASM file execute the following my_qc = QuantumCircuit.from_qasm_file('grover.qasm') my_qc.draw('mpl', fold=-1) # let's create an arbitrary Operator from a matrix A = np.array([[0,1,0], [1,0,0], [0,1,1]]) op = Operator(A) # Operator contains information about its matrix representation and input and output dimensions op # create an Operator from a circuit yourself qc = QuantumCircuit(2) qc.h([0,1]) qc.cx(0,1) op = Operator(qc) op A = Operator(Pauli(label='X')) B = Operator(Pauli(label='Z')) A.expand(B) # B x A A = Operator(Pauli(label='X')) B = Operator(Pauli(label='Z')) A.tensor(B) # result A x B III = Operator(np.eye(2 ** 3)) XX = Operator(Pauli(label='XX')) XIX = III.compose(XX, qargs=[0,2]) XIX # the resulting operator is X x I x X # compose an XIXZ Operator from IIII and XX and Z Operators ### ADD CODE BELOW ### DON'T ADD CODE BELOW # with Qiskit you can find the state fidelity between two quantum states # let's investigate Bell states state_fidelity(sv_bell3, sv_bell1) basis_state_00 = Statevector.from_label('00') state_fidelity(basis_state_00, sv_bell1) basis_state_00 = Statevector.from_label('00') state_fidelity(basis_state_00, sv_bell3) # let's find process fidelity of X and H Operators op_1 = Operator([[0,1],[0,1]]) op_2 = Operator([[1,1],[1,-1]])/np.sqrt(2) F = process_fidelity(op_1, op_2) F # what about fidelity of X Operator with itself multiplied with a complex phase op_1 = Operator([[0,1],[0,1]]) op_2 = op_1*np.exp(0.5j) # calculate the process fidelity yourself ### ADD CODE BELOW ### DON'T ADD CODE BELOW # as you expected, the fidelity is independent of global phase # encountered a compatibility error? Check which version of Qiskit you are using import qiskit qiskit.__qiskit_version__ # display the entire version table %qiskit_version_table # Let's have an overview information on all the IBMQ backends that are available %qiskit_backend_overview # now let's pick one device and investigate it from qiskit.visualization import plot_gate_map backend = provider.get_backend('ibmq_bogota') plot_gate_map(backend, plot_directed=True) plot_error_map(backend) # let's plot the layout of our Bell state circuit transpiled for our target backend transpiled_bell = transpile(bell_0, backend=backend) plot_circuit_layout(transpiled_bell, backend=backend) # display the layout of transpiled Grover's circuit on our target backend transpiled_grover = transpile(grover, backend=backend) plot_circuit_layout(transpiled_grover, backend=backend) # what if we want to specify a custom mapping of our circuit qubits to device qubits? # let's create an initial layout initial_layout = [2,3,4] # virtual to physical: virtual qubits are ordered transpiled_qc = transpile(grover, backend=backend, initial_layout=initial_layout) plot_circuit_layout(transpiled_qc, backend=backend) # for example, let's draw our Bell's state circuit in text mode bell_0.draw('text') # qubits in reversed order bell_0.draw('mpl', reverse_bits=True) # More customization style = {'linecolor': 'red', 'backgroundcolor': 'grey'} bell_0.draw('mpl',scale = 1.2, style=style, fold=-1) plot_bloch_vector([0,1,0]) plot_state_city(bell_0) # to understand why the 'city' looks like this consider the density matrix of this state plot_state_hinton(bell_0) plot_state_qsphere(bell_0) plot_state_paulivec(bell_0) plot_bloch_multivector(bell_0) # here, the bloch spheres represent maximally entangled qubits as zero-length vectors. Hence, no arrows. # a more informative example psi = Statevector.from_label('+1') plot_bloch_multivector(psi) # the first qubit is in state |1> # the second qubit is in the superposition of |0> and |1> rho = DensityMatrix.from_instruction(bell_0) rho.draw('latex', prefix='\\rho_{Bell_0} = ') plot_state_city(rho.data, title='Density Matrix') # prepare a density matrix for the state of the equal superposition of 2 basis states my_qc = QuantumCircuit(2) my_qc.h([0,1]) my_rho = DensityMatrix.from_instruction(my_qc) my_rho.draw('latex', prefix='\\rho_{equal} = ') # prepare a density matrix for a mixed one-qubit state in an equal mixture of |0> and |1> ### ADD CODE BELOW ### DON'T ADD CODE BELOW qr = QuantumRegister(3, name='q') qc = QuantumCircuit(qr) qc.ccx(control_qubit1=qr[1], control_qubit2=qr[2], target_qubit=qr[0]) qc.draw('mpl') # Toffoli gate has the following matrix representation for the above configuration matrix = Operator(qc).data array_to_latex(matrix, prefix="CCX = ") qr = QuantumRegister(2, name='q') qc = QuantumCircuit(qr) qc.swap(qubit1=qr[0], qubit2=[1]) qc.draw('mpl') # SWAP gate has the following matrix representation for the above configuration matrix = Operator(qc).data # display the matrix in latex array_to_latex(matrix, prefix="SWAP = ")
https://github.com/petr-ivashkov/qiskit-cert-workbook
petr-ivashkov
#initialization import matplotlib.pyplot as plt import numpy as np #ignore deprecation warnings because they are annoying (not recommened generally) from warnings import simplefilter simplefilter(action='ignore', category=DeprecationWarning) # Importing standard Qiskit libraries from qiskit import * from qiskit.tools.jupyter import * from qiskit.visualization import * from qiskit.quantum_info import * from qiskit.circuit.library import * from ibm_quantum_widgets import * NUM_QUBITS = 2 # let's construct a multi-qubit quantum register qr = QuantumRegister(NUM_QUBITS, 'q') qc = QuantumCircuit(qr, name='my-circuit') # let's create a Bell's state qc.h(0) qc.cx(0,1) qc.measure_all() qc.draw() # let's construct a multi-bit classical register cr = ClassicalRegister(NUM_QUBITS, 'c') qc = QuantumCircuit(qr, cr, name='my-circuit') # explicitly measure qubits [0,1] into classical bits [0,1] qc.measure(qr, cr) # alternatively: qc.measure([0,1], [0,1]) qc.draw() # Bell state 0 bell_0 = QuantumCircuit(2) bell_0.h(0) bell_0.cx(0,1) sv = Statevector.from_label('00') #evolve the initial state through the circuit sv_bell1 = sv.evolve(bell_0) sv_bell1.draw('latex') # Bell state 1 bell_1 = QuantumCircuit(2) bell_1.x(0) bell_1.h(0) bell_1.cx(0,1) sv = Statevector.from_label('00') #evolve the initial state through the circuit sv_bell2 = sv.evolve(bell_1) sv_bell2.draw('latex') # Bell state 2 bell_2 = QuantumCircuit(2) bell_2.x(0) bell_2.h(1) bell_2.cx(1,0) sv = Statevector.from_label('00') #evolve the initial state through the circuit sv_bell3 = sv.evolve(bell_2) sv_bell3.draw('latex') # create the last remaining Bell state on your own bell_3 = QuantumCircuit(2) ### ADD CODE BELOW bell_3.x(0) bell_3.x(1) bell_3.h(1) bell_3.cx(1,0) ### DON'T ADD CODE BELOW sv = Statevector.from_label('00') #evolve the initial state through the circuit sv_bell3 = sv.evolve(bell_3) sv_bell3.draw('latex') # state of equal superposition of basis states max_sp = QuantumCircuit(2) max_sp.h(0) max_sp.h(1) sv = Statevector.from_label('00') # evolve the initial state through the circuit ### ADD CODE BELOW sv_ev = sv.evolve(max_sp) sv_ev.draw('latex') ### DON'T ADD CODE BELOW qc = QuantumCircuit(1) qc.h(0) qc.h(0) qc.draw('mpl') backend = BasicAer.get_backend('qasm_simulator') qc = QuantumCircuit(1) qc.h(0) qc.h(0) transpile(qc, backend=backend).draw('mpl') backend = BasicAer.get_backend('qasm_simulator') qc = QuantumCircuit(1) qc.h(0) qc.barrier() qc.h(0) transpile(qc, backend=backend).draw('mpl') # try it out yourself with different gates: instead of Hadamard, try Pauili gates my_qc = QuantumCircuit(1) ### ADD CODE BELOW my_qc.x(0) my_qc.barrier() my_qc.x(0) ### DON'T ADD CODE BELOW transpile(my_qc, backend=backend).draw('mpl') # did you get what you expected? qc = QuantumCircuit(1) qc.x(0) qc.depth() qc = QuantumCircuit(2) qc.x([0,1]) qc.x(1) qc.depth() # draw the circuit yourself to see why the depth has increased now ### ADD CODE BELOW qc.draw() ### DON'T ADD CODE BELOW qc = QuantumCircuit(2) qc.x(0) qc.barrier(0) qc.h(0) qc.cx(0,1) qc.depth() # draw the circuit yourself to see why the depth has increased now # hint: the longest path is not always the lenth of the longest sequence along one channel qc.draw('mpl') # the second qubit only has a CX gate applied to it, but has to "wait" for the first qubit to provide the control # also, barier doesn't count # in the following we will use the implementation of Grover's algorithm in Qiskit from qiskit.algorithms import Grover, AmplificationProblem # let's construct an oracle with a single solution sv = Statevector.from_label('010') problem = AmplificationProblem(oracle=sv) grover = Grover(iterations=2).construct_circuit(problem=problem) # the original circuit contains unitaries # transpile the circuit in terms of basis gates to see what it is "made of" t_grover = transpile(grover, basis_gates=['cx', 'u3']) t_grover.draw('mpl', fold=-1) # draw the original cicuit yourself grover.draw('mpl', fold=-1) # lets's try to evolve the initial state through the circuit like we have done for Bell states sv = Statevector.from_label('000') sv_ev = sv.evolve(grover) sv_ev.draw('latex') # it is obvious from statevector evolutions, that the state 010 # has the best chances to be measured, as desired # now let's obtain the same result from the QASM simulation # for that we need to measure the qubits in the end of our circuit grover.measure_all() backend= BasicAer.get_backend('qasm_simulator') job = execute(grover, backend, shots=1024) result = job.result() plot_histogram(result.get_counts()) # finally let's see how our circuit performs on actual hardware # you will need your IBM Quantum account IBMQ.load_account() provider = IBMQ.get_provider(hub='ibm-q') # sometimes you might need to wait when running on real devices # let's monitor the status of our job from qiskit.tools import job_monitor # to call a real backend let's pick the least busy one, which satisfies out requirments from qiskit.providers.ibmq import least_busy suitable_devices = provider.backends(filters = lambda b: b.configuration().n_qubits >= 3 and not b.configuration().simulator and b.status().operational==True) backend = least_busy(backends=suitable_devices) print("We are runnig our circuit on ", backend) #job = execute(grover, backend=backend) #job_monitor(job) # we see that results from a real device also match our expectations # the algorithm successfully finds the marked string result = job.result() counts = result.get_counts() ### ADD CODE BELOW # plot the histogramm yourself plot_histogram(counts) ### DON'T ADD CODE BELOW # let's again use our transpiled Grover circuit to demonstarte the idea # Notice: QASM doesn't look "nice". This is a low-level language which can be understood by quantum hardware. # If you want to take a look at the QASM code make sure to specify 'formatted=True' which makes the code readable by a human being. t_grover.qasm(formatted=True) # store the circuit to a file in the follwing way qasm_code = t_grover.qasm(filename='grover.qasm') # to get the circuit back from QASM file execute the following my_qc = QuantumCircuit.from_qasm_file('grover.qasm') my_qc.draw('mpl', fold=-1) # let's create an arbitrary Operator from a matrix A = np.array([[0,1,0], [1,0,0], [0,1,1]]) op = Operator(A) # Operator contains information about its matrix representation and input and output dimensions op # create an Operator from a circuit yourself qc = QuantumCircuit(2) qc.h([0,1]) qc.cx(0,1) op = Operator(qc) op A = Operator(Pauli(label='X')) B = Operator(Pauli(label='Z')) A.expand(B) # B x A A = Operator(Pauli(label='X')) B = Operator(Pauli(label='Z')) A.tensor(B) # result A x B III = Operator(np.eye(2 ** 3)) XX = Operator(Pauli(label='XX')) XIX = III.compose(XX, qargs=[0,2]) XIX # the resulting operator is X x I x X # compose an XIXZ Operator from IIII and XX and Z Operators ### ADD CODE BELOW IIII = Operator(np.eye(2 ** 4)) XX = Operator(Pauli(label='XX')) Z = Operator(Pauli('Z')) XIXI = IIII.compose(XX, qargs=[1,3]) XIXZ = XIXI.compose(Z, qargs=[0]) ### DON'T ADD CODE BELOW print(XIXZ) # with Qiskit you can find the state fidelity between two quantum states # let's investigate Bell states state_fidelity(sv_bell3, sv_bell1) basis_state_00 = Statevector.from_label('00') state_fidelity(basis_state_00, sv_bell1) basis_state_00 = Statevector.from_label('00') state_fidelity(basis_state_00, sv_bell3) # let's find process fidelity of X and H Operators op_1 = Operator([[0,1],[0,1]]) op_2 = Operator([[1,1],[1,-1]])/np.sqrt(2) F = process_fidelity(op_1, op_2) F # what about fidelity of X Operator with itself multiplied with a complex phase op_1 = Operator([[0,1],[0,1]]) op_2 = op_1*np.exp(0.5j) # calculate the process fidelity yourself ### ADD CODE BELOW F = process_fidelity(op_1, op_2) F ### DON'T ADD CODE BELOW # as you expected, the fidelity is independent of global phase # encountered a compatibility error? Check which version of Qiskit you are using import qiskit qiskit.__qiskit_version__ # display the entire version table %qiskit_version_table # Let's have an overview information on all the IBMQ backends that are available %qiskit_backend_overview # now let's pick one device and investigate it from qiskit.visualization import plot_gate_map backend = provider.get_backend('ibmq_bogota') plot_gate_map(backend, plot_directed=True) plot_error_map(backend) # let's plot the layout of our Bell state circuit transpiled for our target backend transpiled_bell = transpile(bell_0, backend=backend) plot_circuit_layout(transpiled_bell, backend=backend) # display the layout of transpiled Grover's circuit on our target backend transpiled_grover = transpile(grover, backend=backend) plot_circuit_layout(transpiled_grover, backend=backend) # what if we want to specify a custom mapping of our circuit qubits to device qubits? # let's create an initial layout initial_layout = [2,3,4] # virtual to physical: virtual qubits are ordered transpiled_qc = transpile(grover, backend=backend, initial_layout=initial_layout) plot_circuit_layout(transpiled_qc, backend=backend) # for example, let's draw our Bell's state circuit in text mode bell_0.draw('text') # qubits in reversed order bell_0.draw('mpl', reverse_bits=True) # More customization style = {'linecolor': 'red', 'backgroundcolor': 'grey'} bell_0.draw('mpl',scale = 1.2, style=style, fold=-1) plot_bloch_vector([0,1,0]) plot_state_city(bell_0) # to understand why the 'city' looks like this consider the density matrix of this state plot_state_hinton(bell_0) plot_state_qsphere(bell_0) plot_state_paulivec(bell_0) plot_bloch_multivector(bell_0) # here, the bloch spheres represent maximally entangled qubits as zero-length vectors. Hence, no arrows. # a more informative example psi = Statevector.from_label('+1') plot_bloch_multivector(psi) # the first qubit is in state |1> # the second qubit is in the superposition of |0> and |1> rho = DensityMatrix.from_instruction(bell_0) rho.draw('latex', prefix='\\rho_{Bell_0} = ') plot_state_city(rho.data, title='Density Matrix') # prepare a density matrix for the state of the equal superposition of 2 basis states my_qc = QuantumCircuit(2) my_qc.h([0,1]) my_rho = DensityMatrix.from_instruction(my_qc) my_rho.draw('latex', prefix='\\rho_{equal} = ') # prepare a density matrix for a mixed one-qubit state in an equal mixture of |0> and |1> ### ADD CODE BELOW my_rho_M = np.array([[0.5, 0], [0, 0.5]]) my_rho = DensityMatrix(my_rho_M) my_rho.draw('latex', prefix='\\rho_{mixture} = ') ### DON'T ADD CODE BELOW qr = QuantumRegister(3, name='q') qc = QuantumCircuit(qr) qc.ccx(control_qubit1=qr[1], control_qubit2=qr[2], target_qubit=qr[0]) qc.draw('mpl') # Toffoli gate has the following matrix representation for the above configuration matrix = Operator(qc).data array_to_latex(matrix, prefix="CCX = ") qr = QuantumRegister(2, name='q') qc = QuantumCircuit(qr) qc.swap(qubit1=qr[0], qubit2=[1]) qc.draw('mpl') # SWAP gate has the following matrix representation for the above configuration matrix = Operator(qc).data # display the matrix in latex array_to_latex(matrix, prefix="SWAP = ")
https://github.com/qclib/qclib
qclib
# This code is part of Qiskit. # # (C) Copyright IBM 2023 # # This code is licensed under the Apache License, Version 2.0. You may # obtain a copy of this license in the LICENSE.txt file in the root directory # of this source tree or at http://www.apache.org/licenses/LICENSE-2.0. # # Any modifications or derivative works of this code must retain this # copyright notice, and modified files need to carry a notice indicating # that they have been altered from the originals. # pylint: disable=missing-docstring,invalid-name,no-member # pylint: disable=attribute-defined-outside-init # pylint: disable=unused-argument from qiskit import QuantumRegister, QuantumCircuit from qiskit.compiler import transpile from qiskit.quantum_info.random import random_unitary class IsometryTranspileBench: params = ([0, 1, 2, 3], [3, 4, 5, 6]) param_names = ["number of input qubits", "number of output qubits"] def setup(self, m, n): q = QuantumRegister(n) qc = QuantumCircuit(q) if not hasattr(qc, "iso"): raise NotImplementedError iso = random_unitary(2**n, seed=0).data[:, 0 : 2**m] if len(iso.shape) == 1: iso = iso.reshape((len(iso), 1)) qc.iso(iso, q[:m], q[m:]) self.circuit = qc def track_cnot_counts_after_mapping_to_ibmq_16_melbourne(self, *unused): coupling = [ [1, 0], [1, 2], [2, 3], [4, 3], [4, 10], [5, 4], [5, 6], [5, 9], [6, 8], [7, 8], [9, 8], [9, 10], [11, 3], [11, 10], [11, 12], [12, 2], [13, 1], [13, 12], ] circuit = transpile( self.circuit, basis_gates=["u1", "u3", "u2", "cx"], coupling_map=coupling, seed_transpiler=0, ) counts = circuit.count_ops() cnot_count = counts.get("cx", 0) return cnot_count
https://github.com/qclib/qclib
qclib
# This code is part of Qiskit. # # (C) Copyright IBM 2017, 2019. # # This code is licensed under the Apache License, Version 2.0. You may # obtain a copy of this license in the LICENSE.txt file in the root directory # of this source tree or at http://www.apache.org/licenses/LICENSE-2.0. # # Any modifications or derivative works of this code must retain this # copyright notice, and modified files need to carry a notice indicating # that they have been altered from the originals. """ Arbitrary unitary circuit instruction. """ import numpy from qiskit.circuit import Gate, ControlledGate from qiskit.circuit import QuantumCircuit from qiskit.circuit import QuantumRegister, Qubit from qiskit.circuit.exceptions import CircuitError from qiskit.circuit._utils import _compute_control_matrix from qiskit.circuit.library.standard_gates import UGate from qiskit.extensions.quantum_initializer import isometry from qiskit.quantum_info.operators.predicates import matrix_equal from qiskit.quantum_info.operators.predicates import is_unitary_matrix from qiskit.quantum_info.synthesis.one_qubit_decompose import OneQubitEulerDecomposer from qiskit.quantum_info.synthesis.two_qubit_decompose import two_qubit_cnot_decompose from qiskit.extensions.exceptions import ExtensionError _DECOMPOSER1Q = OneQubitEulerDecomposer("U") class UnitaryGate(Gate): """Class quantum gates specified by a unitary matrix. Example: We can create a unitary gate from a unitary matrix then add it to a quantum circuit. The matrix can also be directly applied to the quantum circuit, see :meth:`.QuantumCircuit.unitary`. .. code-block:: python from qiskit import QuantumCircuit from qiskit.extensions import UnitaryGate matrix = [[0, 0, 0, 1], [0, 0, 1, 0], [1, 0, 0, 0], [0, 1, 0, 0]] gate = UnitaryGate(matrix) circuit = QuantumCircuit(2) circuit.append(gate, [0, 1]) """ def __init__(self, data, label=None): """Create a gate from a numeric unitary matrix. Args: data (matrix or Operator): unitary operator. label (str): unitary name for backend [Default: None]. Raises: ExtensionError: if input data is not an N-qubit unitary operator. """ if hasattr(data, "to_matrix"): # If input is Gate subclass or some other class object that has # a to_matrix method this will call that method. data = data.to_matrix() elif hasattr(data, "to_operator"): # If input is a BaseOperator subclass this attempts to convert # the object to an Operator so that we can extract the underlying # numpy matrix from `Operator.data`. data = data.to_operator().data # Convert to numpy array in case not already an array data = numpy.array(data, dtype=complex) # Check input is unitary if not is_unitary_matrix(data): raise ExtensionError("Input matrix is not unitary.") # Check input is N-qubit matrix input_dim, output_dim = data.shape num_qubits = int(numpy.log2(input_dim)) if input_dim != output_dim or 2**num_qubits != input_dim: raise ExtensionError("Input matrix is not an N-qubit operator.") # Store instruction params super().__init__("unitary", num_qubits, [data], label=label) def __eq__(self, other): if not isinstance(other, UnitaryGate): return False if self.label != other.label: return False # Should we match unitaries as equal if they are equal # up to global phase? return matrix_equal(self.params[0], other.params[0], ignore_phase=True) def __array__(self, dtype=None): """Return matrix for the unitary.""" # pylint: disable=unused-argument return self.params[0] def inverse(self): """Return the adjoint of the unitary.""" return self.adjoint() def conjugate(self): """Return the conjugate of the unitary.""" return UnitaryGate(numpy.conj(self.to_matrix())) def adjoint(self): """Return the adjoint of the unitary.""" return self.transpose().conjugate() def transpose(self): """Return the transpose of the unitary.""" return UnitaryGate(numpy.transpose(self.to_matrix())) def _define(self): """Calculate a subcircuit that implements this unitary.""" if self.num_qubits == 1: q = QuantumRegister(1, "q") qc = QuantumCircuit(q, name=self.name) theta, phi, lam, global_phase = _DECOMPOSER1Q.angles_and_phase(self.to_matrix()) qc._append(UGate(theta, phi, lam), [q[0]], []) qc.global_phase = global_phase self.definition = qc elif self.num_qubits == 2: self.definition = two_qubit_cnot_decompose(self.to_matrix()) else: from qiskit.quantum_info.synthesis.qsd import ( # pylint: disable=cyclic-import qs_decomposition, ) self.definition = qs_decomposition(self.to_matrix()) def control(self, num_ctrl_qubits=1, label=None, ctrl_state=None): """Return controlled version of gate Args: num_ctrl_qubits (int): number of controls to add to gate (default=1) label (str): optional gate label ctrl_state (int or str or None): The control state in decimal or as a bit string (e.g. '1011'). If None, use 2**num_ctrl_qubits-1. Returns: UnitaryGate: controlled version of gate. Raises: QiskitError: Invalid ctrl_state. ExtensionError: Non-unitary controlled unitary. """ mat = self.to_matrix() cmat = _compute_control_matrix(mat, num_ctrl_qubits, ctrl_state=None) iso = isometry.Isometry(cmat, 0, 0) return ControlledGate( "c-unitary", num_qubits=self.num_qubits + num_ctrl_qubits, params=[mat], label=label, num_ctrl_qubits=num_ctrl_qubits, definition=iso.definition, ctrl_state=ctrl_state, base_gate=self.copy(), ) def _qasm2_decomposition(self): """Return an unparameterized version of ourselves, so the OQ2 exporter doesn't choke on the non-standard things in our `params` field.""" out = self.definition.to_gate() out.name = self.name return out def validate_parameter(self, parameter): """Unitary gate parameter has to be an ndarray.""" if isinstance(parameter, numpy.ndarray): return parameter else: raise CircuitError(f"invalid param type {type(parameter)} in gate {self.name}") def unitary(self, obj, qubits, label=None): """Apply unitary gate specified by ``obj`` to ``qubits``. Args: obj (matrix or Operator): unitary operator. qubits (Union[int, Tuple[int]]): The circuit qubits to apply the transformation to. label (str): unitary name for backend [Default: None]. Returns: QuantumCircuit: The quantum circuit. Raises: ExtensionError: if input data is not an N-qubit unitary operator. Example: Apply a gate specified by a unitary matrix to a quantum circuit .. code-block:: python from qiskit import QuantumCircuit matrix = [[0, 0, 0, 1], [0, 0, 1, 0], [1, 0, 0, 0], [0, 1, 0, 0]] circuit = QuantumCircuit(2) circuit.unitary(matrix, [0, 1]) """ gate = UnitaryGate(obj, label=label) if isinstance(qubits, QuantumRegister): qubits = qubits[:] # for single qubit unitary gate, allow an 'int' or a 'list of ints' as qubits. if gate.num_qubits == 1: if isinstance(qubits, (int, Qubit)) or len(qubits) > 1: qubits = [qubits] return self.append(gate, qubits, []) QuantumCircuit.unitary = unitary
https://github.com/qclib/qclib
qclib
import json import logging import numpy as np import warnings from functools import wraps from typing import Any, Callable, Optional, Tuple, Union from qiskit import IBMQ, QuantumCircuit, assemble from qiskit.circuit import Barrier, Gate, Instruction, Measure from qiskit.circuit.library import UGate, U3Gate, CXGate from qiskit.providers.ibmq import AccountProvider, IBMQProviderError from qiskit.providers.ibmq.job import IBMQJob def get_provider() -> AccountProvider: with warnings.catch_warnings(): warnings.simplefilter('ignore') ibmq_logger = logging.getLogger('qiskit.providers.ibmq') current_level = ibmq_logger.level ibmq_logger.setLevel(logging.ERROR) # get provider try: provider = IBMQ.get_provider() except IBMQProviderError: provider = IBMQ.load_account() ibmq_logger.setLevel(current_level) return provider def get_job(job_id: str) -> Optional[IBMQJob]: try: job = get_provider().backends.retrieve_job(job_id) return job except Exception: pass return None def circuit_to_json(qc: QuantumCircuit) -> str: class _QobjEncoder(json.encoder.JSONEncoder): def default(self, obj: Any) -> Any: if isinstance(obj, np.ndarray): return obj.tolist() if isinstance(obj, complex): return (obj.real, obj.imag) return json.JSONEncoder.default(self, obj) return json.dumps(circuit_to_dict(qc), cls=_QobjEncoder) def circuit_to_dict(qc: QuantumCircuit) -> dict: qobj = assemble(qc) return qobj.to_dict() def get_job_urls(job: Union[str, IBMQJob]) -> Tuple[bool, Optional[str], Optional[str]]: try: job_id = job.job_id() if isinstance(job, IBMQJob) else job download_url = get_provider()._api_client.account_api.job(job_id).download_url()['url'] result_url = get_provider()._api_client.account_api.job(job_id).result_url()['url'] return download_url, result_url except Exception: return None, None def cached(key_function: Callable) -> Callable: def _decorator(f: Any) -> Callable: f.__cache = {} @wraps(f) def _decorated(*args: Any, **kwargs: Any) -> int: key = key_function(*args, **kwargs) if key not in f.__cache: f.__cache[key] = f(*args, **kwargs) return f.__cache[key] return _decorated return _decorator def gate_key(gate: Gate) -> Tuple[str, int]: return gate.name, gate.num_qubits @cached(gate_key) def gate_cost(gate: Gate) -> int: if isinstance(gate, (UGate, U3Gate)): return 1 elif isinstance(gate, CXGate): return 10 elif isinstance(gate, (Measure, Barrier)): return 0 return sum(map(gate_cost, (g for g, _, _ in gate.definition.data))) def compute_cost(circuit: Union[Instruction, QuantumCircuit]) -> int: print('Computing cost...') circuit_data = None if isinstance(circuit, QuantumCircuit): circuit_data = circuit.data elif isinstance(circuit, Instruction): circuit_data = circuit.definition.data else: raise Exception(f'Unable to obtain circuit data from {type(circuit)}') return sum(map(gate_cost, (g for g, _, _ in circuit_data))) def uses_multiqubit_gate(circuit: QuantumCircuit) -> bool: circuit_data = None if isinstance(circuit, QuantumCircuit): circuit_data = circuit.data elif isinstance(circuit, Instruction) and circuit.definition is not None: circuit_data = circuit.definition.data else: raise Exception(f'Unable to obtain circuit data from {type(circuit)}') for g, _, _ in circuit_data: if isinstance(g, (Barrier, Measure)): continue elif isinstance(g, Gate): if g.num_qubits > 1: return True elif isinstance(g, (QuantumCircuit, Instruction)) and uses_multiqubit_gate(g): return True return False
https://github.com/qclib/qclib
qclib
# Copyright 2021 qclib project. # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # http://www.apache.org/licenses/LICENSE-2.0 # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """ Linear-depth Multicontrolled Special Unitary """ from typing import Union, List import numpy as np from qiskit import QuantumCircuit, QuantumRegister from qiskit.circuit.library import RZGate, RYGate from qiskit.circuit.library import UnitaryGate from qiskit.synthesis import OneQubitEulerDecomposer from qiskit.circuit import Gate, Qubit from qclib.gates.mcx import LinearMcx, McxVchainDirty from qclib.gates.util import check_su2, apply_ctrl_state, isclose # pylint: disable=protected-access class Ldmcsu(Gate): """ Linear depth Multi-Controlled Gate for Special Unitary ------------------------------------------------ Multicontrolled gate decomposition with linear cost. `unitary` must be a SU(2) matrix. """ def __init__(self, unitary, num_controls, ctrl_state: str = None): check_su2(unitary) self.unitary = unitary self.controls = QuantumRegister(num_controls) self.target = QuantumRegister(1) self.num_controls = num_controls + 1 self.ctrl_state = ctrl_state super().__init__("ldmcsu", self.num_controls, [], "ldmcsu") def _define(self): self.definition = QuantumCircuit(self.controls, self.target) is_main_diag_real = isclose(self.unitary[0, 0].imag, 0.0) and isclose( self.unitary[1, 1].imag, 0.0 ) is_secondary_diag_real = isclose(self.unitary[0, 1].imag, 0.0) and isclose( self.unitary[1, 0].imag, 0.0 ) if not is_main_diag_real and not is_secondary_diag_real: # U = V D V^-1, where the entries of the diagonal D are the eigenvalues # `eig_vals` of U and the column vectors of V are the eigenvectors # `eig_vecs` of U. These columns are orthonormal and the main diagonal # of V is real-valued. eig_vals, eig_vecs = np.linalg.eig(self.unitary) x_vecs, z_vecs = self._get_x_z(eig_vecs) self.half_linear_depth_mcv( x_vecs, z_vecs, self.controls, self.target, self.ctrl_state, inverse=True, ) self.linear_depth_mcv( np.diag(eig_vals), self.controls, self.target, self.ctrl_state, general_su2_optimization=True, ) self.half_linear_depth_mcv( x_vecs, z_vecs, self.controls, self.target, self.ctrl_state ) else: if not is_secondary_diag_real: self.definition.h(self.target) self.linear_depth_mcv(self.unitary, self.controls, self.target, self.ctrl_state) if not is_secondary_diag_real: self.definition.h(self.target) @staticmethod def _get_x_z(su2): is_secondary_diag_real = isclose(su2[0, 1].imag, 0.0) and isclose( su2[1, 0].imag, 0.0 ) if is_secondary_diag_real: x_value = su2[0, 1] z_value = su2[1, 1] else: x_value = -su2[0, 1].real z_value = su2[1, 1] - su2[0, 1].imag * 1.0j return x_value, z_value @staticmethod def _compute_gate_a(x_value, z_value): if x_value == 0: alpha = (z_value + 0j) ** (1 / 4) beta = 0.0 else: alpha_r = np.sqrt((np.sqrt((z_value.real + 1.0) / 2.0) + 1.0) / 2.0) alpha_i = z_value.imag / ( 2.0 * np.sqrt((z_value.real + 1.0) * (np.sqrt((z_value.real + 1.0) / 2.0) + 1.0)) ) alpha = alpha_r + 1.0j * alpha_i beta = x_value / ( 2.0 * np.sqrt((z_value.real + 1.0) * (np.sqrt((z_value.real + 1.0) / 2.0) + 1.0)) ) s_op = np.array([[alpha, -np.conj(beta)], [beta, np.conj(alpha)]]) return s_op def linear_depth_mcv( self, su2_unitary, controls: Union[QuantumRegister, List[Qubit]], target: Qubit, ctrl_state: str = None, general_su2_optimization=False, ): """ Theorem 1 - https://arxiv.org/pdf/2302.06377.pdf """ # S gate definition x_value, z_value = self._get_x_z(su2_unitary) op_a = Ldmcsu._compute_gate_a(x_value, z_value) gate_a = UnitaryGate(op_a) num_ctrl = len(controls) k_1 = int(np.ceil(num_ctrl / 2.0)) k_2 = int(np.floor(num_ctrl / 2.0)) ctrl_state_k_1 = None ctrl_state_k_2 = None if ctrl_state is not None: ctrl_state_k_1 = ctrl_state[::-1][:k_1][::-1] ctrl_state_k_2 = ctrl_state[::-1][k_1:][::-1] if not general_su2_optimization: mcx_1 = McxVchainDirty(k_1, ctrl_state=ctrl_state_k_1).definition self.definition.append( mcx_1, controls[:k_1] + controls[k_1 : 2 * k_1 - 2] + [target] ) self.definition.append(gate_a, [target]) mcx_2 = McxVchainDirty( k_2, ctrl_state=ctrl_state_k_2, action_only=general_su2_optimization ).definition self.definition.append( mcx_2.inverse(), controls[k_1:] + controls[k_1 - k_2 + 2 : k_1] + [target] ) self.definition.append(gate_a.inverse(), [target]) mcx_3 = McxVchainDirty(k_1, ctrl_state=ctrl_state_k_1).definition self.definition.append( mcx_3, controls[:k_1] + controls[k_1 : 2 * k_1 - 2] + [target] ) self.definition.append(gate_a, [target]) mcx_4 = McxVchainDirty(k_2, ctrl_state=ctrl_state_k_2).definition self.definition.append( mcx_4, controls[k_1:] + controls[k_1 - k_2 + 2 : k_1] + [target] ) self.definition.append(gate_a.inverse(), [target]) def half_linear_depth_mcv( self, x_value, z_value, controls: Union[QuantumRegister, List[Qubit]], target: Qubit, ctrl_state: str = None, inverse: bool = False, ): """ Theorem 4 - https://arxiv.org/pdf/2302.06377.pdf """ alpha_r = np.sqrt((z_value.real + 1.0) / 2.0) alpha_i = z_value.imag / np.sqrt(2 * (z_value.real + 1.0)) alpha = alpha_r + 1.0j * alpha_i beta = x_value / np.sqrt(2 * (z_value.real + 1.0)) s_op = np.array([[alpha, -np.conj(beta)], [beta, np.conj(alpha)]]) # S gate definition s_gate = UnitaryGate(s_op) # Hadamard equivalent definition h_gate = UnitaryGate(np.array([[-1, 1], [1, 1]]) * 1 / np.sqrt(2)) num_ctrl = len(controls) k_1 = int(np.ceil(num_ctrl / 2.0)) k_2 = int(np.floor(num_ctrl / 2.0)) ctrl_state_k_1 = None ctrl_state_k_2 = None if ctrl_state is not None: ctrl_state_k_1 = ctrl_state[::-1][:k_1][::-1] ctrl_state_k_2 = ctrl_state[::-1][k_1:][::-1] if inverse: self.definition.h(target) self.definition.append(s_gate, [target]) mcx_2 = McxVchainDirty( k_2, ctrl_state=ctrl_state_k_2, action_only=True ).definition self.definition.append( mcx_2, controls[k_1:] + controls[k_1 - k_2 + 2 : k_1] + [target] ) self.definition.append(s_gate.inverse(), [target]) self.definition.append(h_gate, [target]) else: mcx_1 = McxVchainDirty(k_1, ctrl_state=ctrl_state_k_1).definition self.definition.append( mcx_1, controls[:k_1] + controls[k_1 : 2 * k_1 - 2] + [target] ) self.definition.append(h_gate, [target]) self.definition.append(s_gate, [target]) mcx_2 = McxVchainDirty(k_2, ctrl_state=ctrl_state_k_2).definition self.definition.append( mcx_2, controls[k_1:] + controls[k_1 - k_2 + 2 : k_1] + [target] ) self.definition.append(s_gate.inverse(), [target]) self.definition.h(target) @staticmethod def ldmcsu( circuit, unitary, controls: Union[QuantumRegister, List[Qubit]], target: Qubit, ctrl_state: str = None, ): """ Apply multi-controlled SU(2) https://arxiv.org/abs/2302.06377 """ circuit.append( Ldmcsu(unitary, len(controls), ctrl_state=ctrl_state), [*controls, target] ) class LdMcSpecialUnitary(Gate): """ Linear-depth Multicontrolled Special Unitary -------------------------------------------- Implements the gate decompostion of any gate in SU(2) with linear depth (Ld) presented in Lemma 7.9 in Barenco et al., 1995 (arXiv:quant-ph/9503016) with optimizations from Theorem 5 of Iten et al., 2016 (arXiv:1501.06911) """ def __init__(self, unitary, num_controls, ctrl_state=None): if not check_su2(unitary): raise ValueError("Operator must be in SU(2)") self.unitary = np.array(unitary, dtype=complex) if num_controls > 0: self.control_qubits = QuantumRegister(num_controls) else: self.control_qubits = [] self.target_qubit = QuantumRegister(1) self.num_qubits = num_controls + 1 self.ctrl_state = ctrl_state if self.ctrl_state is None: self.ctrl_state = "1" * num_controls super().__init__("ldmc_su2", self.num_qubits, [], "LdMcSu2") @staticmethod def get_abc_operators(beta, gamma, delta): """ Creates A,B and C matrices such that ABC = I """ # A a_rz = RZGate(beta).to_matrix() a_ry = RYGate(gamma / 2).to_matrix() a_matrix = a_rz.dot(a_ry) # B b_ry = RYGate(-gamma / 2).to_matrix() b_rz = RZGate(-(delta + beta) / 2).to_matrix() b_matrix = b_ry.dot(b_rz) # C c_matrix = RZGate((delta - beta) / 2).to_matrix() a_gate = UnitaryGate(a_matrix, label="A") b_gate = UnitaryGate(b_matrix, label="B") c_gate = UnitaryGate(c_matrix, label="C") return a_gate, b_gate, c_gate def _define(self): self.definition = QuantumCircuit(self.control_qubits, self.target_qubit) if len(self.control_qubits) > 0: self._apply_ctrl_state() theta, phi, lamb, _ = OneQubitEulerDecomposer._params_zyz(self.unitary) a_gate, b_gate, c_gate = LdMcSpecialUnitary.get_abc_operators( phi, theta, lamb ) self._apply_abc(a_gate, b_gate, c_gate) self._apply_ctrl_state() else: self.unitary(self.unitary, self.target_qubit) def _apply_abc(self, a_gate: UnitaryGate, b_gate: UnitaryGate, c_gate: UnitaryGate): """ Applies ABC matrices to the quantum circuit according to theorem 5 of Iten et al. 2016 (arXiv:1501.06911). Parameters ---------- a_gate, b_gate and c_gate expceted to be special unitary gates """ if len(self.control_qubits) < 3: self.definition.append(c_gate, [self.target_qubit]) self.definition.mcx(self.control_qubits, self.target_qubit) self.definition.append(b_gate, [self.target_qubit]) self.definition.mcx(self.control_qubits, self.target_qubit) self.definition.append(a_gate, [self.target_qubit]) else: ancilla = self.control_qubits[-1] action_only = True if len(self.control_qubits) < 6: action_only = False # decompose A, B and C to use their optimized controlled versions theta_a, phi_a, lam_a, _ = OneQubitEulerDecomposer._params_zyz( a_gate.to_matrix() ) theta_b, phi_b, lam_b, _ = OneQubitEulerDecomposer._params_zyz( b_gate.to_matrix() ) theta_c, phi_c, lam_c, _ = OneQubitEulerDecomposer._params_zyz( c_gate.to_matrix() ) a_a, b_a, c_a = LdMcSpecialUnitary.get_abc_operators(phi_a, theta_a, lam_a) a_b, b_b, c_b = LdMcSpecialUnitary.get_abc_operators(phi_b, theta_b, lam_b) a_c, b_c, c_c = LdMcSpecialUnitary.get_abc_operators(phi_c, theta_c, lam_c) # definition of left mcx, which will also be inverted as the right mcx mcx_gate = LinearMcx( len(self.control_qubits[:-1]), action_only=action_only ).definition # decomposed controlled C self.definition.unitary(c_c, self.target_qubit) self.definition.cx(ancilla, self.target_qubit) self.definition.unitary(b_c, self.target_qubit) self.definition.cx(ancilla, self.target_qubit) self.definition.unitary(a_c, self.target_qubit) self.definition.append( mcx_gate, self.control_qubits[:-1] + [self.target_qubit] + [ancilla] ) # decomposed controlled B self.definition.unitary(c_b, self.target_qubit) self.definition.cx(ancilla, self.target_qubit) self.definition.unitary(b_b, self.target_qubit) self.definition.cx(ancilla, self.target_qubit) self.definition.unitary(a_b, self.target_qubit) self.definition.append( mcx_gate.inverse(), self.control_qubits[:-1] + [self.target_qubit] + [ancilla], ) # decomposed A self.definition.unitary(c_a, self.target_qubit) self.definition.cx(ancilla, self.target_qubit) self.definition.unitary(b_a, self.target_qubit) self.definition.cx(ancilla, self.target_qubit) self.definition.unitary(a_a, self.target_qubit) @staticmethod def ldmcsu(circuit, unitary, controls, target, ctrl_state=None): """ Linear-depth Multicontrolled Special Unitary -------------------------------------------- Implements the gate decompostion of any gate in SU(2) with linear depth (Ld) presented in Lemma 7.9 in Barenco et al., 1995 (arXiv:quant-ph/9503016) with optimizations from Theorem 5 of Iten et al., 2016 (arXiv:1501.06911) """ circuit.append( LdMcSpecialUnitary(unitary, len(controls), ctrl_state), [*controls, target] ) LdMcSpecialUnitary._apply_ctrl_state = apply_ctrl_state
https://github.com/qclib/qclib
qclib
# Copyright 2021 qclib project. # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # http://www.apache.org/licenses/LICENSE-2.0 # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """ n-qubit controlled gate """ from collections import namedtuple import numpy as np import qiskit from qiskit.circuit import Gate from qiskit import QuantumCircuit, QuantumRegister from .util import check_u2, apply_ctrl_state # pylint: disable=maybe-no-member # pylint: disable=protected-access class Ldmcu(Gate): """ Linear Depth Multi-Controlled Unitary ----------------------------------------- Implements gate decomposition of a multi-controlled operator in U(2) according to https://arxiv.org/abs/2203.11882 https://journals.aps.org/pra/abstract/10.1103/PhysRevA.106.042602. """ def __init__(self, unitary, num_controls, ctrl_state: str = None): check_u2(unitary) self.unitary = unitary if num_controls > 0: self.control_qubits = QuantumRegister(num_controls) else: self.control_qubits = [] self.target_qubit = QuantumRegister(1) self.num_qubits = num_controls + 1 self.ctrl_state = ctrl_state super().__init__("Ldmcu", self.num_qubits, [], "Ldmcu") def _define(self): if len(self.control_qubits) > 0: self.definition = QuantumCircuit(self.control_qubits, self.target_qubit) self._apply_ctrl_state() qubits_indexes = list(range(self.num_qubits)) gate_circuit = qiskit.QuantumCircuit(self.num_qubits, name="T" + str(0)) self._c1c2(self.unitary, self.num_qubits, gate_circuit) self._c1c2(self.unitary, self.num_qubits, gate_circuit, step=-1) self._c1c2(self.unitary, self.num_qubits - 1, gate_circuit, False) self._c1c2(self.unitary, self.num_qubits - 1, gate_circuit, False, step=-1) self.definition.append(gate_circuit, [*self.control_qubits, self.target_qubit]) self._apply_ctrl_state() else: self.definition = QuantumCircuit(self.target_qubit) self.definition.unitary(self.unitary, 0) def _c1c2(self, unitary, n_qubits, gate_circ, first=True, step=1): pairs = namedtuple("pairs", ["control", "target"]) if step == 1: start = 0 reverse = True else: start = 1 reverse = False qubit_pairs = [ pairs(control, target) for target in range(n_qubits) for control in range(start, target) ] qubit_pairs.sort(key=lambda e: e.control + e.target, reverse=reverse) for pair in qubit_pairs: exponent = pair.target - pair.control if pair.control == 0: exponent = exponent - 1 param = 2 ** exponent signal = -1 if (pair.control == 0 and not first) else 1 signal = step * signal if pair.target == n_qubits - 1 and first: csqgate = Ldmcu._gate_u(unitary, param, signal) gate_circ.compose(csqgate, qubits=[pair.control, pair.target], inplace=True) else: gate_circ.crx(signal * np.pi / param, pair.control, pair.target) @staticmethod def _gate_u(agate, coef, signal): param = 1 / np.abs(coef) values, vectors = np.linalg.eig(agate) gate = np.power(values[0] + 0j, param) * vectors[:, [0]] @ vectors[:, [0]].conj().T gate = ( gate + np.power(values[1] + 0j, param) * vectors[:, [1]] @ vectors[:, [1]].conj().T ) if signal < 0: gate = np.linalg.inv(gate) sqgate = QuantumCircuit(1, name="U^1/" + str(coef)) sqgate.unitary(gate, 0) # pylint: disable=maybe-no-member csqgate = sqgate.control(1) return csqgate @staticmethod def ldmcu(circuit, unitary, controls, target, ctrl_state=None): circuit.append( Ldmcu(unitary, len(controls), ctrl_state=ctrl_state), [*controls, target] ) Ldmcu._apply_ctrl_state = apply_ctrl_state
https://github.com/qclib/qclib
qclib
# Copyright 2021 qclib project. # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # http://www.apache.org/licenses/LICENSE-2.0 # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """Multicontrolled gate decompositions for unitaries in U(2) and SU(2)""" from qiskit import QuantumCircuit, QuantumRegister from qiskit.circuit import Gate from .ldmcu import Ldmcu from .ldmcsu import Ldmcsu from .util import check_u2, check_su2, u2_to_su2 # pylint: disable=maybe-no-member # pylint: disable=protected-access class Mcg(Gate): """ Selects the most cost-effective multicontrolled gate decomposition. """ def __init__( self, unitary, num_controls, ctrl_state: str=None, up_to_diagonal: bool=False ): check_u2(unitary) self.unitary = unitary self.controls = QuantumRegister(num_controls) self.target = QuantumRegister(1) self.num_qubits = num_controls + 1 self.ctrl_state = ctrl_state self.up_to_diagonal = up_to_diagonal super().__init__("mcg", self.num_qubits, [], "mcg") def _define(self): self.definition = QuantumCircuit(self.controls, self.target) num_ctrl = len(self.controls) if num_ctrl == 0: self.definition.unitary(self.unitary, [self.target]) elif num_ctrl == 1: u_gate = QuantumCircuit(1) u_gate.unitary(self.unitary, 0) self.definition.append( u_gate.control(num_ctrl, ctrl_state=self.ctrl_state), [*self.controls, self.target] ) else: if check_su2(self.unitary): Ldmcsu.ldmcsu( self.definition, self.unitary, self.controls, self.target, ctrl_state=self.ctrl_state ) else: if self.up_to_diagonal: su_2, _ = u2_to_su2(self.unitary) self.mcg(su_2, self.controls, self.target, self.ctrl_state) else: Ldmcu.ldmcu(self.definition, self.unitary, self.controls[:], self.target[0], self.ctrl_state) @staticmethod def mcg( circuit, unitary, controls, target, ctrl_state: str = None ): circuit.append( Mcg(unitary, len(controls), ctrl_state=ctrl_state), [*controls, target] )
https://github.com/qclib/qclib
qclib
# Copyright 2021 qclib project. # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # http://www.apache.org/licenses/LICENSE-2.0 # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """ n-qubit controlled gate """ from collections import namedtuple import numpy as np import qiskit from qiskit.circuit import Gate from qiskit import QuantumCircuit, QuantumRegister from qclib.gates.util import check_u2, apply_ctrl_state from qclib.gates.multitargetmcsu2 import MultiTargetMCSU2 from qclib.gates.ldmcu import Ldmcu # pylint: disable=protected-access class MCU(Gate): """ Approximated Multi-Controlled Unitary Gate https://arxiv.org/abs/2310.14974 ----------------------------------------- Implements gate decomposition of a multi-controlled operator in U(2) """ def __init__(self, unitary, num_controls, error=0, ctrl_state: str = None): """ Parameters ---------- unitary: U(2) gate num_controls (int): Number of controls error (float): max_error ctrl_state (str or int): Control state in decimal or as a bitstring """ check_u2(unitary) self.unitary = unitary self.error = error if num_controls > 0: self.control_qubits = QuantumRegister(num_controls) else: self.control_qubits = [] self.target_qubit = QuantumRegister(1) self.n_ctrl_base = self._get_num_base_ctrl_qubits(self.unitary, self.error) if self.n_ctrl_base == 0: raise ValueError("The number of base qubits is 0") if self.n_ctrl_base > num_controls: raise ValueError("The number of control qubits is too low") self.ctrl_state = ctrl_state super().__init__("McuApprox", num_controls + 1, [], "McuApprox") def _define(self): if len(self.control_qubits) > 0: self.definition = QuantumCircuit(self.control_qubits, self.target_qubit) self._apply_ctrl_state() gate_circuit = qiskit.QuantumCircuit(self.num_qubits, name="T" + str(0)) self._c1c2(self.num_qubits, gate_circuit) self._c1c2(self.num_qubits, gate_circuit, step=-1) self._c1c2(self.num_qubits - 1, gate_circuit, False) self._c1c2(self.num_qubits - 1, gate_circuit, False, -1) self.definition.append( gate_circuit, [*self.control_qubits, self.target_qubit] ) self._apply_ctrl_state() else: self.definition = QuantumCircuit(self.target_qubit) self.definition.unitary(self.unitary, 0) @staticmethod def _get_num_base_ctrl_qubits(unitary, error): """ Get the baseline number of control qubits for the approximation given an error args: unitary: Unitary to be approximated error: Error of the approximation """ eig_vals, _ = np.linalg.eig(unitary) angles = np.angle(eig_vals) if (1 - np.cos(angles[0])) >= (1 - np.cos(angles[1])): angle = angles[0] else: angle = angles[1] quotient = angle / np.arccos(1 - error**2 / 2) return int(np.ceil(np.log2(quotient))) + 1 def get_n_base(self, unitary, error): return self._get_num_base_ctrl_qubits(unitary, error) def _c1c2(self, n_qubits, gate_circ, first=True, step=1): extra_q, n_qubits_base = self._calc_extra_qubits(first, n_qubits) qubit_pairs = self._compute_qubit_pairs(n_qubits_base, step) unitary_list = [] targets = [] for pair in qubit_pairs: param = self._compute_param(pair) signal = -1 if (pair.control == 0 and not first) else 1 signal = step * signal # Perform a translation of the qubits by extra_q qubits in all cases # When target == last qubit and first==true apply the U gates, except # when control==0, in which case we don't do anything if pair.target == n_qubits_base - 1 and first: if pair.control != 0: csqgate = MCU._gate_u(self.unitary, param, signal) gate_circ.compose( csqgate, qubits=[pair.control + extra_q, pair.target + extra_q], inplace=True, ) # For the controlled rotations, when control==0, apply a multi-controlled # rotation with the extra control qubits else: if pair.control == 0 and extra_q >= 1: # Apply a multi-controlled Rx gate with the additional control qubits control_list = np.array(range(0, extra_q + 1)) unitary_list.append(self._compute_rx_matrix(param, signal)) targets.append(pair.target + extra_q) if pair.target == 1: MultiTargetMCSU2.multi_target_mcsu2( gate_circ, unitary_list, control_list, targets ) else: gate_circ.crx( signal * np.pi / param, pair.control + extra_q, pair.target + extra_q, ) def _calc_extra_qubits(self, first, n_qubits): if first: n_qubits_base = self.n_ctrl_base + 1 else: n_qubits_base = self.n_ctrl_base extra_q = n_qubits - n_qubits_base return extra_q, n_qubits_base @staticmethod def _compute_param(pair): exponent = pair.target - pair.control if pair.control == 0: exponent = exponent - 1 param = 2**exponent return param @staticmethod def _compute_rx_matrix(param, signal): theta = signal * np.pi / param rx_matrix = np.array( [ [np.cos(theta / 2), (-1j) * np.sin(theta / 2)], [(-1j) * np.sin(theta / 2), np.cos(theta / 2)], ] ) return rx_matrix @staticmethod def _compute_qubit_pairs(n_qubits_base, step): pairs = namedtuple("pairs", ["control", "target"]) if step == 1: start = 0 reverse = True else: start = 1 reverse = False qubit_pairs = [ pairs(control, target) for target in range(n_qubits_base) for control in range(start, target) ] qubit_pairs.sort(key=lambda e: e.control + e.target, reverse=reverse) return qubit_pairs @staticmethod def _gate_u(a_gate, coefficient, signal): param = 1 / np.abs(coefficient) values, vectors = np.linalg.eig(a_gate) gate = ( np.power(values[0] + 0j, param) * vectors[:, [0]] @ vectors[:, [0]].conj().T ) gate = ( gate + np.power(values[1] + 0j, param) * vectors[:, [1]] @ vectors[:, [1]].conj().T ) if signal < 0: gate = np.linalg.inv(gate) s_q_gate = QuantumCircuit(1, name="U^1/" + str(coefficient)) s_q_gate.unitary(gate, 0) # pylint: disable=maybe-no-member c_s_q_gate = s_q_gate.control(1) return c_s_q_gate @staticmethod def mcu(circuit, unitary, controls, target, error, ctrl_state=None): """ Approximated Multi-Controlled Unitary Gate https://arxiv.org/abs/2310.14974 """ if error == 0: circuit.append( Ldmcu(unitary, len(controls), ctrl_state=ctrl_state), [*controls, target] ) else: circuit.append( MCU(unitary, len(controls), error, ctrl_state=ctrl_state), [*controls, target], ) MCU._apply_ctrl_state = apply_ctrl_state
https://github.com/qclib/qclib
qclib
# Copyright 2021 qclib project. # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # http://www.apache.org/licenses/LICENSE-2.0 # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """ linear-depth n-qubit controlled X with ancilla """ import numpy as np from qiskit import QuantumRegister, QuantumCircuit from qiskit.circuit.library import C3XGate, C4XGate from qiskit.circuit import Gate from qclib.gates.toffoli import Toffoli from qclib.gates.util import apply_ctrl_state # pylint: disable=protected-access class McxVchainDirty(Gate): """ Implementation based on lemma 8 of Iten et al. (2016) arXiv:1501.06911. Decomposition of a multicontrolled X gate with at least k <= ceil(n/2) ancilae for n as the total number of qubits in the system. It also includes optimizations using approximated Toffoli gates up to a diagonal. """ def __init__( self, num_controls: int, num_target_qubit=1, ctrl_state=None, relative_phase=False, action_only=False, ): """ Parameters ---------- num_target_qubit num_controls ctrl_state relative_phase action_only """ self.control_qubits = QuantumRegister(num_controls) self.target_qubits = QuantumRegister(num_target_qubit) self.ctrl_state = ctrl_state self.relative_phase = relative_phase self.action_only = action_only num_ancilla = 0 self.ancilla_qubits = [] if num_controls - 2 > 0: num_ancilla = num_controls - 2 self.ancilla_qubits = QuantumRegister(num_ancilla) super().__init__( "mcx_vc_dirty", num_controls + num_ancilla + 1, [], "mcx_vc_dirty" ) @staticmethod def toffoli_multi_target(num_targets, side=None): """ " """ size = 2 + num_targets circuit = QuantumCircuit(size) if side == "l": for i in range(num_targets - 1): circuit.name = 'toff_left' circuit.cx(size - i - 2, size - i - 1) circuit.ccx(0, 1, 2) return circuit elif side == "r": circuit.name = 'toff_right' circuit.ccx(0, 1, 2) for i in range(num_targets - 1): circuit.cx(i + 2, i + 3) return circuit elif side is None: circuit.name = 'toff_l_and_r' for i in range(num_targets - 1): circuit.cx(size - i - 2, size - i - 1) circuit.ccx(0, 1, 2) for i in range(num_targets - 1): circuit.cx(i + 2, i + 3) return circuit def _define(self): self.definition = QuantumCircuit( self.control_qubits, self.ancilla_qubits, self.target_qubits ) num_ctrl = len(self.control_qubits) num_target = len(self.target_qubits) num_ancilla = num_ctrl - 2 targets_aux = self.target_qubits[0:1] + self.ancilla_qubits[:num_ancilla][::-1] self._apply_ctrl_state() if num_ctrl == 2: self.definition.append( self.toffoli_multi_target(len(self.target_qubits)), [*self.control_qubits, *self.target_qubits], ) elif num_ctrl == 1: for k, _ in enumerate(self.target_qubits): self.definition.mcx( control_qubits=self.control_qubits, target_qubit=self.target_qubits[k], mode="noancilla", ) elif not self.relative_phase and num_ctrl == 3 and num_target < 2: for k, _ in enumerate(self.target_qubits): self.definition.append( C3XGate(), [*self.control_qubits[:], self.target_qubits[k]], [] ) else: side = 'l' for j in range(2): self._action_circuit(j, num_ancilla, num_ctrl, targets_aux, side) side = 'r' for i, _ in enumerate(self.ancilla_qubits[1:]): # reset part controls = [self.control_qubits[2 + i], self.ancilla_qubits[i]] self.definition.append( Toffoli(cancel="left"), [*controls, self.ancilla_qubits[i + 1]] ) if self.action_only: control_1 = self.control_qubits[-1] control_2 = self.ancilla_qubits[-1] targets = self.target_qubits num_targets = len(targets) self.definition.append( self.toffoli_multi_target(num_targets, side), [control_1, control_2, *targets], ) break self._apply_ctrl_state() def _action_circuit(self, j, num_ancilla, num_ctrl, targets_aux, side): for i, _ in enumerate(self.control_qubits): # action part if i < num_ctrl - 2: if ( targets_aux[i] not in self.target_qubits or self.relative_phase ): # gate cancelling controls = [ self.control_qubits[num_ctrl - i - 1], self.ancilla_qubits[num_ancilla - i - 1], ] # cancel rightmost gates of action part # with leftmost gates of reset part if ( self.relative_phase and targets_aux[i] in self.target_qubits and j == 1 ): self.definition.append( Toffoli(cancel="left"), [*controls, targets_aux[i]] ) else: self.definition.append( Toffoli(cancel="right"), [*controls, targets_aux[i]] ) else: control_1 = self.control_qubits[num_ctrl - i - 1] control_2 = self.ancilla_qubits[num_ancilla - i - 1] targets = self.target_qubits num_targets = len(targets) self.definition.append( self.toffoli_multi_target(num_targets, side), [control_1, control_2, *targets], ) else: controls = [ self.control_qubits[num_ctrl - i - 2], self.control_qubits[num_ctrl - i - 1], ] self.definition.append(Toffoli(), [*controls, targets_aux[i]]) break @staticmethod def mcx_vchain_dirty( circuit, controls=None, target=None, ctrl_state=None, relative_phase=False, action_only=False, ): """ Implementation based on lemma 8 of Iten et al. (2016) arXiv:1501.06911. Decomposition of a multicontrolled X gate with at least k <= ceil(n/2) ancilae for n as the total number of qubits in the system. It also includes optimizations using approximated Toffoli gates up to a diagonal. """ circuit.append( McxVchainDirty(len(controls), ctrl_state, relative_phase, action_only), [*controls, target], ) McxVchainDirty._apply_ctrl_state = apply_ctrl_state class LinearMcx(Gate): """ Implementation based on lemma 9 of Iten et al. (2016) arXiv:1501.06911. Decomposition of a multicontrolled X gate with a dirty ancilla by splitting it into two sequences of two alternating multicontrolled X gates on k1 = ceil((n+1)/2) and k2 = floor((n+1)/2) qubits. For n the total number of qubits in the system. Where it also reuses some optimizations available """ def __init__(self, num_controls, ctrl_state=None, action_only=False): self.action_only = action_only self.ctrl_state = ctrl_state num_qubits = num_controls + 2 self.control_qubits = list(range(num_qubits - 2)) self.target_qubit = (num_qubits - 2,) self.ancilla_qubit = num_qubits - 1 super().__init__("linear_mcx", num_controls + 2, [], "mcx") def _define(self): self.definition = QuantumCircuit(self.num_qubits) self._apply_ctrl_state() if self.num_qubits < 5: self.definition.mcx( control_qubits=self.control_qubits, target_qubit=self.target_qubit, mode="noancilla", ) elif self.num_qubits == 5: self.definition.append( C3XGate(), [*self.control_qubits[:], self.target_qubit], [] ) elif self.num_qubits == 6: self.definition.append( C4XGate(), [*self.control_qubits[:], self.target_qubit], [] ) elif self.num_qubits == 7: self.definition.append( C3XGate(), [*self.control_qubits[:3], self.ancilla_qubit], [] ) self.definition.append( C3XGate(), [*self.control_qubits[3:], self.ancilla_qubit, self.target_qubit], [], ) self.definition.append( C3XGate(), [*self.control_qubits[:3], self.ancilla_qubit], [] ) self.definition.append( C3XGate(), [*self.control_qubits[3:], self.ancilla_qubit, self.target_qubit], [], ) else: # split controls to halve the number of qubits used for each mcx num_ctrl = len(self.control_qubits) k_2 = int(np.ceil(self.num_qubits / 2.0)) k_1 = num_ctrl - k_2 + 1 first_gate = McxVchainDirty(k_1, relative_phase=True).definition second_gate = McxVchainDirty(k_2).definition self.definition.append( first_gate, self.control_qubits[:k_1] + self.control_qubits[k_1 : k_1 + k_1 - 2] + [self.ancilla_qubit], ) self.definition.append( second_gate, [*self.control_qubits[k_1:], self.ancilla_qubit] + self.control_qubits[k_1 - k_2 + 2 : k_1] + [self.target_qubit], ) self.definition.append( first_gate, self.control_qubits[:k_1] + self.control_qubits[k_1 : k_1 + k_1 - 2] + [self.ancilla_qubit], ) last_gate = McxVchainDirty(k_2, action_only=self.action_only).definition self.definition.append( last_gate, [*self.control_qubits[k_1:], self.ancilla_qubit] + self.control_qubits[k_1 - k_2 + 2 : k_1] + [self.target_qubit], ) self._apply_ctrl_state() @staticmethod def mcx(circuit, controls=None, target=None, ctrl_state=None, action_only=False): """ Implementation based on lemma 9 of Iten et al. (2016) arXiv:1501.06911. Decomposition of a multicontrolled X gate with a dirty ancilla by splitting it into two sequences of two alternating multicontrolled X gates on k1 = ceil((n+1)/2) and k2 = floor((n+1)/2) qubits. For n the total number of qubits in the system. Where it also reuses some optimizations available """ circuit.append( LinearMcx(len(controls), ctrl_state, action_only), [*controls, target] ) LinearMcx._apply_ctrl_state = apply_ctrl_state
https://github.com/qclib/qclib
qclib
# Copyright 2021 qclib project. # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # http://www.apache.org/licenses/LICENSE-2.0 # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """ Linear-depth Multicontrolled Special Unitary """ from typing import Union, List import numpy as np from qiskit import QuantumCircuit, QuantumRegister from qiskit.circuit.library import UnitaryGate from qiskit.circuit import Gate, Qubit from qclib.gates.ldmcsu import Ldmcsu from qclib.gates.mcx import McxVchainDirty from qclib.gates.util import check_su2, isclose # pylint: disable=protected-access class MultiTargetMCSU2(Gate): """ Multi-target Multi-Controlled Gate for Special Unitary ------------------------------------------------ Linear decomposition of approximate multi-controlled single qubit gates https://arxiv.org/abs/2310.14974 Lemma 7 """ def __init__(self, unitaries, num_controls, num_target=1, ctrl_state: str = None): """ Parameters ---------- unitaries (list): List of SU(2) matrices num_controls (int): Number of control gastes num_target (int): Number of target gates ctrl_state (str): Control state in decimal or as a bitstring """ if isinstance(unitaries, list): for unitary in unitaries: check_su2(unitary) else: check_su2(unitaries) self.unitaries = unitaries self.controls = QuantumRegister(num_controls) self.target = QuantumRegister(num_target) self.num_controls = num_controls + 1 self.ctrl_state = ctrl_state super().__init__("ldmcsu", self.num_controls, [], "ldmcsu") def _define(self): if isinstance(self.unitaries, list): self.definition = QuantumCircuit(self.controls, self.target) is_main_diags_real = [] is_secondary_diags_real = [] for unitary in self.unitaries: is_main_diags_real.append( isclose(unitary[0, 0].imag, 0.0) and isclose(unitary[1, 1].imag, 0.0) ) is_secondary_diags_real.append( isclose(unitary[0, 1].imag, 0.0) and isclose(unitary[1, 0].imag, 0.0) ) for idx, unitary in enumerate(self.unitaries): if not is_secondary_diags_real[idx] and is_main_diags_real[idx]: self.definition.h(self.target[idx]) self.clinear_depth_mcv() for idx, unitary in enumerate(self.unitaries): if not is_secondary_diags_real[idx] and is_main_diags_real[idx]: self.definition.h(self.target[idx]) else: self.definition = Ldmcsu(self.unitaries, self.num_controls) def clinear_depth_mcv(self, general_su2_optimization=False): """ Multi-target version of Theorem 1 - https://arxiv.org/pdf/2302.06377.pdf """ # S gate definition gates_a = self._s_gate_definition(self.unitaries) num_ctrl = len(self.controls) target_size = len(self.target) k_1 = int(np.ceil(num_ctrl / 2.0)) k_2 = int(np.floor(num_ctrl / 2.0)) ctrl_state_k_1 = None ctrl_state_k_2 = None if self.ctrl_state is not None: ctrl_state_k_1 = self.ctrl_state[::-1][:k_1][::-1] ctrl_state_k_2 = self.ctrl_state[::-1][k_1:][::-1] if not general_su2_optimization: mcx_1 = McxVchainDirty( k_1, num_target_qubit=target_size, ctrl_state=ctrl_state_k_1 ).definition self.definition.append( mcx_1, self.controls[:k_1] + self.controls[k_1: 2 * k_1 - 2] + [*self.target], ) for idx, gate_a in enumerate(gates_a): self.definition.append(gate_a, [num_ctrl + idx]) mcx_2 = McxVchainDirty( k_2, num_target_qubit=target_size, ctrl_state=ctrl_state_k_2, action_only=general_su2_optimization, ).definition self.definition.append( mcx_2.inverse(), self.controls[k_1:] + self.controls[k_1 - k_2 + 2: k_1] + [*self.target], ) for idx, gate_a in enumerate(gates_a): self.definition.append(gate_a.inverse(), [num_ctrl + idx]) mcx_3 = McxVchainDirty( k_1, num_target_qubit=target_size, ctrl_state=ctrl_state_k_1 ).definition self.definition.append( mcx_3, self.controls[:k_1] + self.controls[k_1: 2 * k_1 - 2] + [*self.target], ) for idx, gate_a in enumerate(gates_a): self.definition.append(gate_a, [num_ctrl + idx]) mcx_4 = McxVchainDirty( k_2, num_target_qubit=target_size, ctrl_state=ctrl_state_k_2 ).definition self.definition.append( mcx_4, self.controls[k_1:] + self.controls[k_1 - k_2 + 2: k_1] + [*self.target], ) for idx, gate_a in enumerate(gates_a): self.definition.append(gate_a.inverse(), [num_ctrl + idx]) def _s_gate_definition(self, su2_unitaries): gates_a = [] for su2_unitary in su2_unitaries: x_value, z_value = Ldmcsu._get_x_z(su2_unitary) op_a = Ldmcsu._compute_gate_a(x_value, z_value) gates_a.append(UnitaryGate(op_a)) return gates_a @staticmethod def multi_target_mcsu2( circuit, unitary, controls: Union[QuantumRegister, List[Qubit]], target: Qubit, ctrl_state: str = None, ): """ Multi-target version of multi-controlled SU(2) https://arxiv.org/abs/2302.06377 """ if isinstance(unitary, list): num_target = len(unitary) circuit.append( MultiTargetMCSU2(unitary, len(controls), num_target=num_target).definition, [*controls, *target], ) else: circuit.append( Ldmcsu(unitary, len(controls), ctrl_state=ctrl_state), [*controls, target], )
https://github.com/qclib/qclib
qclib
# Copyright 2021 qclib project. # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # http://www.apache.org/licenses/LICENSE-2.0 # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """ Quadratic-depth Multicontrolled Special Unitary """ from typing import Union, List from qiskit import QuantumCircuit, QuantumRegister from qiskit.circuit import Qubit from qiskit.circuit import Gate from numpy import sqrt, outer from numpy.linalg import eig from .mcx import LinearMcx from .util import check_u2 # pylint: disable=protected-access class Qdmcu(Gate): """ Quandratic Depth Multi-Controlled Unitary ----------------------------------------- Implements gate decomposition of a munticontrolled operator in U(2) according to Theorem 4 of Iten et al. (2016) arXiv:1501.06911. """ def __init__( self, unitary, num_controls, ctrl_state: str=None ): """ Parameters ---------- unitary : numpy.ndarray 2 x 2 unitary matrix controls : Either qiskit.QuantumRegister or list of qiskit.Qubit containing the qubits to be used as control gates. target : qiskit.Qubit on wich the unitary operation is to be applied ctrl_state : String of binary digits describing the basis state used as control """ check_u2(unitary) self.unitary = unitary self.controls = QuantumRegister(num_controls) self.target = QuantumRegister(1) self.num_qubits = num_controls + 1 self.ctrl_state = ctrl_state super().__init__("qdmcu", self.num_qubits, [], "qdmcu") def _define(self): self.definition = QuantumCircuit(self.controls, self.target) num_ctrl = len(self.controls) if num_ctrl == 1: u_gate = QuantumCircuit(1) u_gate.unitary(self.unitary, 0) self.definition.append( u_gate.control( num_ctrl_qubits=num_ctrl, ctrl_state=self.ctrl_state ), [*self.controls, self.target] ) else: if self.ctrl_state is None: self.ctrl_state = '1' * num_ctrl # Notice that `ctrl_state`` is reversed with respect to `controls``. v_op = Qdmcu.custom_sqrtm(self.unitary) v_gate = QuantumCircuit(1, name="V") v_gate.unitary(v_op, 0) v_gate_dag = QuantumCircuit(1, name="V^dag") v_gate_dag.unitary(v_op.T.conj(), 0) linear_mcx_gate = LinearMcx( num_controls=num_ctrl-1, ctrl_state=self.ctrl_state[1:], action_only=True ).definition self.definition.append( v_gate.control(1, ctrl_state=self.ctrl_state[:1]), [self.controls[-1], self.target] ) self.definition.append( linear_mcx_gate, [*self.controls[:-1], self.controls[-1], self.target] ) self.definition.append( v_gate_dag.control(1, ctrl_state=self.ctrl_state[:1]), [self.controls[-1], self.target] ) self.definition.append( linear_mcx_gate.inverse(), [*self.controls[:-1], self.controls[-1], self.target] ) self.qdmcu( self.definition, v_op, self.controls[:-1], self.target, self.ctrl_state[1:] ) @staticmethod def custom_sqrtm(unitary): eig_vals, eig_vecs = eig(unitary) first_eig = sqrt(eig_vals[0]) * outer(eig_vecs[:, 0], eig_vecs[:, 0].conj()) second_eig = sqrt(eig_vals[1]) * outer(eig_vecs[:, 1], eig_vecs[:, 1].conj()) return first_eig + second_eig @staticmethod def qdmcu( circuit, unitary, controls: Union[QuantumRegister, List[Qubit]], target: Qubit, ctrl_state: str=None ): circuit.append( Qdmcu(unitary, len(controls), ctrl_state=ctrl_state), [*controls, target] )
https://github.com/qclib/qclib
qclib
# Copyright 2021 qclib project. # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # http://www.apache.org/licenses/LICENSE-2.0 # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. ''' Toffoli decomposition explained in Lemma 8 from Quantum Circuits for Isometries. https://arxiv.org/abs/1501.06911 ''' from numpy import pi from qiskit import QuantumCircuit from qiskit.circuit import Gate class Toffoli(Gate): def __init__(self, cancel=None): self.cancel = cancel super().__init__('toffoli', 3, [], "Toffoli") def _define(self): self.definition = QuantumCircuit(3) theta = pi / 4. control_qubits = self.definition.qubits[:2] target_qubit = self.definition.qubits[-1] if self.cancel != 'left': self.definition.u(theta=-theta, phi=0., lam=0., qubit=target_qubit) self.definition.cx(control_qubits[0], target_qubit) self.definition.u(theta=-theta, phi=0., lam=0., qubit=target_qubit) self.definition.cx(control_qubits[1], target_qubit) if self.cancel != 'right': self.definition.u(theta=theta, phi=0., lam=0., qubit=target_qubit) self.definition.cx(control_qubits[0], target_qubit) self.definition.u(theta=theta, phi=0., lam=0., qubit=target_qubit) @staticmethod def ccx(circuit, controls=None, target=None, cancel=None): if controls is None or target is None: circuit.append(Toffoli(cancel), circuit.qubits[:3]) else: circuit.append(Toffoli(cancel), [*controls, target])
https://github.com/qclib/qclib
qclib
# Copyright 2021 qclib project. # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # http://www.apache.org/licenses/LICENSE-2.0 # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """ Constructs a multiplexor gate. """ from math import log2 from typing import List, Union, Type import numpy as np from qiskit import QuantumCircuit, QuantumRegister from qiskit.circuit.library import RZGate, RYGate, CXGate, CZGate def ucr( r_gate: Union[Type[RZGate], Type[RYGate]], angles: List[float], c_gate: Union[Type[CXGate], Type[CZGate]] = CXGate, last_control=True, ) -> QuantumCircuit: """ Constructs a multiplexor rotation gate. Synthesis of Quantum Logic Circuits https://arxiv.org/abs/quant-ph/0406176 """ size = len(angles) n_qubits = int(log2(size)) + 1 reg = QuantumRegister(n_qubits) circuit = QuantumCircuit(reg) target = reg[0] control = reg[n_qubits - 1] if n_qubits == 1: if abs(angles[0]) > 10**-8: circuit.append(r_gate(angles[0]), [target]) return circuit angle_multiplexor = np.kron( [[0.5, 0.5], [0.5, -0.5]], np.identity(2 ** (n_qubits - 2)) ) multiplexed_angles = angle_multiplexor.dot(angles) # Figure 2 from Synthesis of Quantum Logic Circuits: # The recursive decomposition of a multiplexed Rz gate. # The boxed CNOT gates may be canceled. # This is why "last_cnot=False" in both calls of "rotation_multiplexor()" and # also why the multiplexer in the second "circuit.append()" is reversed. mult = ucr(r_gate, multiplexed_angles[: size // 2], c_gate, False) circuit.append(mult.to_instruction(), reg[0:-1]) circuit.append(c_gate(), [control, target]) mult = ucr(r_gate, multiplexed_angles[size // 2 :], c_gate, False) circuit.append(mult.reverse_ops().to_instruction(), reg[0:-1]) # The following condition allows saving CNOTs when two multiplexors are used # in sequence. Any multiplexor can have its operation reversed. Therefore, if # the second multiplexor is reverted, its last CNOT will be cancelled by the # last CNOT of the first multiplexer. In this condition, both last CNOTs are # unnecessary. if last_control: circuit.append(c_gate(), [control, target]) return circuit
https://github.com/qclib/qclib
qclib
# Copyright 2021 qclib project. # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # http://www.apache.org/licenses/LICENSE-2.0 # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """ Bounded Approximation version of Plesch's algorithm. https://arxiv.org/abs/2111.03132 """ from dataclasses import dataclass from qiskit import QuantumCircuit from qclib.gates.initialize import Initialize from qclib.state_preparation.util.baa import adaptive_approximation from .lowrank import LowRankInitialize @dataclass class _OptParams: def __init__(self, opt_params): if opt_params is None: self.max_fidelity_loss = 0.0 self.isometry_scheme = "ccd" self.unitary_scheme = "qsd" self.strategy = "greedy" self.max_combination_size = 0 self.use_low_rank = False else: self.max_fidelity_loss = 0.0 if opt_params.get("max_fidelity_loss") is None \ else opt_params.get("max_fidelity_loss") self.isometry_scheme = "ccd" if opt_params.get("iso_scheme") is None else \ opt_params.get("iso_scheme") self.unitary_scheme = "qsd" if opt_params.get("unitary_scheme") is None else \ opt_params.get("unitary_scheme") self.strategy = "greedy" if opt_params.get("strategy") is None else \ opt_params.get("strategy") self.max_combination_size = 0 if opt_params.get("max_combination_size") is None else \ opt_params.get("max_combination_size") self.use_low_rank = False if opt_params.get("use_low_rank") is None else \ opt_params.get("use_low_rank") if self.max_fidelity_loss < 0 or self.max_fidelity_loss > 1: self.max_fidelity_loss = 0.0 class BaaLowRankInitialize(Initialize): """ State preparation using the bounded approximation algorithm via Schmidt decomposition arXiv:1003.5760 https://arxiv.org/abs/2111.03132 This class implements a state preparation gate. """ def __init__(self, params, label=None, opt_params=None): """ Parameters ---------- params: list of complex A unit vector representing a quantum state. Values are amplitudes. opt_params: Dictionary max_fidelity_loss: float ``state`` allowed (fidelity) error for approximation (0<=``max_fidelity_loss``<=1). If ``max_fidelity_loss`` is not in the valid range, it will be ignored. isometry_scheme: string Scheme used to decompose isometries. Possible values are ``'knill'`` and ``'ccd'`` (column-by-column decomposition). Default is ``isometry_scheme='ccd'``. unitary_scheme: string Scheme used to decompose unitaries. Possible values are ``'csd'`` (cosine-sine decomposition) and ``'qsd'`` (quantum Shannon decomposition). Default is ``unitary_scheme='qsd'``. strategy: string Method to search for the best approximation (``'brute_force'`` or ``'greedy'``). For states larger than 2**8, the greedy strategy should preferably be used. Default is ``strategy='greedy'``. max_combination_size: int Maximum size of the combination ``C(n_qubits, max_combination_size)`` between the qubits of an entangled subsystem of length ``n_qubits`` to produce the possible bipartitions (1 <= ``max_combination_size`` <= ``n_qubits``//2). For example, if ``max_combination_size``==1, there will be ``n_qubits`` bipartitions between 1 and ``n_qubits``-1 qubits. The default value is 0 (the size will be maximum for each level). use_low_rank: bool If set to True, ``rank``>1 approximations are also considered. This is fine tuning for high-entanglement states and is slower. The default value is False. """ self._name = "baa-lrsp" self._get_num_qubits(params) self.node = None self.opt_params = _OptParams(opt_params) if label is None: self._label = "BAASP" super().__init__(self._name, self.num_qubits, params, label=label) def _define(self): self.definition = self._define_initialize() def _define_initialize(self): self.node = adaptive_approximation( self.params, self.opt_params.max_fidelity_loss, self.opt_params.strategy, self.opt_params.max_combination_size, self.opt_params.use_low_rank, ) circuit = QuantumCircuit(self.num_qubits) for vector, qubits, rank, partition in zip( self.node.vectors, self.node.qubits, self.node.ranks, self.node.partitions ): opt_params = { "iso_scheme": self.opt_params.isometry_scheme, "unitary_scheme": self.opt_params.unitary_scheme, "partition": partition, "lr": rank, } gate = LowRankInitialize(vector, opt_params=opt_params) circuit.compose(gate, qubits[::-1], inplace=True) # qiskit little-endian. return circuit.reverse_bits() @staticmethod def initialize(q_circuit, state, qubits=None, opt_params=None): """ Appends a BaaLowRankInitialize gate into the q_circuit """ if qubits is None: q_circuit.append( BaaLowRankInitialize(state, opt_params=opt_params), q_circuit.qubits ) else: q_circuit.append(BaaLowRankInitialize(state, opt_params=opt_params), qubits)
https://github.com/qclib/qclib
qclib
# Copyright 2021 qclib project. # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # http://www.apache.org/licenses/LICENSE-2.0 # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """ Bidirectional state preparation https://arxiv.org/abs/2108.10182 """ from math import ceil, log2 import numpy as np from qiskit import QuantumCircuit from qclib.gates.initialize import Initialize from qclib.state_preparation.util.state_tree_preparation import ( Amplitude, state_decomposition, ) from qclib.state_preparation.util.angle_tree_preparation import create_angles_tree from qclib.state_preparation.util.tree_register import add_register from qclib.state_preparation.util.tree_walk import top_down, bottom_up class BdspInitialize(Initialize): """ Configurable sublinear circuits for quantum state preparation https://arxiv.org/abs/2108.10182 This class implements a state preparation gate. """ def __init__(self, params, label=None, opt_params=None): """ Parameters ---------- params: list of complex A unit vector representing a quantum state. Values are amplitudes. opt_params: {'split': split} split: int Level (enumerated from bottom to top, where 1 ≤ s ≤ n) at which the angle tree is split. Default value is ``ceil(n_qubits/2)`` (sublinear). """ if opt_params is None: self.split = int(ceil(log2(len(params)) / 2)) # sublinear else: if opt_params.get("split") is None: self.split = int(ceil(log2(len(params)) / 2)) # sublinear else: self.split = opt_params.get("split") self._name = "bdsp" self._get_num_qubits(params) if label is None: label = "BDSP" super().__init__(self._name, self.num_qubits, params, label=label) def _define(self): self.definition = self._define_initialize() def _define_initialize(self): n_qubits = int(np.log2(len(self.params))) data = [Amplitude(i, a) for i, a in enumerate(self.params)] state_tree = state_decomposition(n_qubits, data) angle_tree = create_angles_tree(state_tree) circuit = QuantumCircuit() add_register(circuit, angle_tree, n_qubits - self.split) top_down(angle_tree, circuit, n_qubits - self.split) bottom_up(angle_tree, circuit, n_qubits - self.split) return circuit def _get_num_qubits(self, params): n_qubits = log2(len(params)) if not n_qubits.is_integer(): Exception("The number of amplitudes is not a power of 2") n_qubits = int(n_qubits) self.num_qubits = (self.split + 1) * 2 ** (n_qubits - self.split) - 1 @staticmethod def initialize(q_circuit, state, qubits=None, opt_params=None): """ Appends a BdspInitialize gate into the q_circuit """ if qubits is None: q_circuit.append( BdspInitialize(state, opt_params=opt_params), q_circuit.qubits ) else: q_circuit.append(BdspInitialize(state, opt_params=opt_params), qubits)
https://github.com/qclib/qclib
qclib
# Copyright 2021 qclib project. # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # http://www.apache.org/licenses/LICENSE-2.0 # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """ Black-box state preparation Grover, Lov K. "Synthesis of quantum superpositions by quantum computation." Physical review letters 85.6 (2000): 1334. Gate U2 in PRL 85.6 (2000) is implemented with uniformly controlled rotations """ import numpy as np from qiskit import QuantumCircuit from qiskit.circuit.library import UCRYGate, UCRZGate from qiskit.circuit.library import UnitaryGate from qclib.gates.initialize import Initialize class BlackBoxInitialize(Initialize): """ Black-box state preparation Grover, Lov K. "Synthesis of quantum superpositions by quantum computation." Physical review letters 85.6 (2000): 1334. Gate U2 in PRL 85.6 (2000) is implemented with uniformly controlled rotations """ def __init__(self, params, label=None): self._name = "blackbox" self._get_num_qubits(params) self.num_qubits += 1 if label is None: label = "BBSP" super().__init__(self._name, self.num_qubits, params, label=label) def _define(self): self.definition = self._define_initialize() def _define_initialize(self): n_amplitudes = len(self.params) theta = 2 * np.arccos(np.abs(self.params)) phi = -2 * np.angle(self.params) ury_gate = UCRYGate(list(theta)) urz_gate = UCRZGate(list(phi)) gate_u = QuantumCircuit(self.num_qubits, name="U") gate_u.h(gate_u.qubits[1:]) gate_u.append(ury_gate, gate_u.qubits) gate_u.append(urz_gate, gate_u.qubits) gate_u = gate_u.to_instruction() it_matrix = [[-1, 0], [0, 1]] gate_it = UnitaryGate(it_matrix) gate_it.name = "I_t" gate_is = gate_it.control(self.num_qubits - 1, ctrl_state=0) gate_is.name = "I_s" repetitions = (np.pi / 4) * ( np.sqrt(n_amplitudes) / np.linalg.norm(self.params) ) repetitions = int(repetitions) q_circuit = QuantumCircuit(self.num_qubits) for _ in range(repetitions): q_circuit.append(gate_u, q_circuit.qubits) q_circuit.append(gate_it, q_circuit.qubits[0:1]) q_circuit.append(gate_u.inverse(), q_circuit.qubits) q_circuit.append(gate_is, q_circuit.qubits) q_circuit.append(gate_u, q_circuit.qubits) if repetitions % 2 == 1: q_circuit.global_phase = np.pi return q_circuit @staticmethod def initialize(q_circuit, state, qubits=None): if qubits is None: q_circuit.append(BlackBoxInitialize(state), q_circuit.qubits) else: q_circuit.append(BlackBoxInitialize(state), qubits)
https://github.com/qclib/qclib
qclib
# Copyright 2021 qclib project. # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # http://www.apache.org/licenses/LICENSE-2.0 # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """ cvo-qram """ import numpy as np from qiskit import QuantumCircuit, QuantumRegister from qiskit.circuit.library.standard_gates import UGate from qiskit.quantum_info import Operator from qclib.util import _compute_matrix_angles from qclib.gates.initialize_sparse import InitializeSparse from qclib.gates.mcg import Mcg from qclib.gates.ldmcsu import LdMcSpecialUnitary # pylint: disable=maybe-no-member class CvoqramInitialize(InitializeSparse): """ Initializing the Amplitude Distribution of a Quantum State This class implements a sparse state preparation gate. """ def __init__(self, params, label=None, opt_params=None): self._get_num_qubits(params) self.norm = 1 self.anc = None self.aux = None self.memory = None self.control = None default_with_aux = True if opt_params is None: self.with_aux = default_with_aux self.mcg_method = 'linear' else: if opt_params.get("with_aux") is None: self.with_aux = default_with_aux else: self.with_aux = opt_params.get("with_aux") if opt_params.get("mcg_method") is None: self.mcg_method = 'linear' else: self.mcg_method = opt_params.get("mcg_method") if label is None: label = "CVOSP" super().__init__("cvoqram", self.num_qubits, params.items(), label=label) def _define(self): self.definition = self._define_initialize() def _define_initialize(self): """Initialize quantum registers""" self.anc = QuantumRegister(self.num_qubits - 1, name="anc") self.aux = QuantumRegister(1, name="u") self.memory = QuantumRegister(self.num_qubits, name="m") if self.with_aux: circuit = QuantumCircuit(self.aux, self.anc, self.memory) else: self.anc = None circuit = QuantumCircuit(self.aux, self.memory) self.norm = 1 circuit.x(self.aux[0]) for k, (binary_string, feature) in enumerate(self.params): self.control = self._select_controls(binary_string) self._flip_flop(circuit) self._load_superposition(circuit, feature, self.control, self.memory) if k < len(self.params) - 1: self._flip_flop(circuit) else: break return circuit def _flip_flop(self, circuit): for k in self.control: circuit.cx(self.aux[0], self.memory[k]) @staticmethod def _select_controls(binary_string): control = [] for k, bit in enumerate(binary_string[::-1]): if bit == "1": control.append(k) return control def _mcuvchain(self, circuit, alpha, beta, phi): """ N-qubit controlled-unitary gate """ lst_ctrl = self.control lst_ctrl_reversed = list(reversed(lst_ctrl)) circuit.rccx( self.memory[lst_ctrl_reversed[0]], self.memory[lst_ctrl_reversed[1]], self.anc[self.num_qubits - 2], ) tof = {} i = self.num_qubits - 1 for ctrl in lst_ctrl_reversed[2:]: circuit.rccx(self.anc[i - 1], self.memory[ctrl], self.anc[i - 2]) tof[ctrl] = [i - 1, i - 2] i -= 1 circuit.cu(alpha, beta, phi, 0, self.anc[i - 1], self.aux[0]) for ctrl in lst_ctrl[:-2]: circuit.rccx(self.anc[tof[ctrl][0]], self.memory[ctrl], self.anc[tof[ctrl][1]]) circuit.rccx( self.memory[lst_ctrl[-1]], self.memory[lst_ctrl[-2]], self.anc[self.num_qubits - 2] ) def _load_superposition(self, circuit, feature, control, memory): """ Load pattern in superposition """ theta, phi, lam = _compute_matrix_angles(feature, self.norm) if len(control) == 0: circuit.u(theta, phi, lam, self.aux[0]) elif len(control) == 1: circuit.cu(theta, phi, lam, 0, memory[control[0]], self.aux[0]) else: if self.with_aux: self._mcuvchain(circuit, theta, phi, lam) else: gate = UGate(theta, phi, lam) if self.mcg_method == 'qiskit': circuit.append(gate.control(len(control)), memory[control] + [self.aux[0]]) elif self.mcg_method == 'barenco': gate_op = Operator(gate).data LdMcSpecialUnitary.ldmcsu(circuit, gate_op, memory[control], [self.aux[0]]) else: gate_op = Operator(gate).data Mcg.mcg(circuit, gate_op, memory[control], [self.aux[0]]) self.norm = self.norm - np.absolute(np.power(feature, 2)) @staticmethod def initialize(q_circuit, state, qubits=None, opt_params=None): if qubits is None: q_circuit.append( CvoqramInitialize(state, opt_params=opt_params), q_circuit.qubits ) else: q_circuit.append(CvoqramInitialize(state, opt_params=opt_params), qubits)
https://github.com/qclib/qclib
qclib
# Copyright 2021 qclib project. # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # http://www.apache.org/licenses/LICENSE-2.0 # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """ https://arxiv.org/abs/2011.07977 """ import numpy as np from qiskit import QuantumCircuit, QuantumRegister from qclib.util import _compute_matrix_angles from qclib.gates.initialize_sparse import InitializeSparse # pylint: disable=maybe-no-member class CvqramInitialize(InitializeSparse): """ Initializing the Amplitude Distribution of a Quantum State https://arxiv.org/abs/2011.07977 This class implements a sparse state preparation gate. """ def __init__(self, params, label=None, opt_params=None): self._name = "cvqram" self._get_num_qubits(params) self.norm = 1 default_mode = "v-chain" if opt_params is None: self.mode = default_mode else: if opt_params.get("mode") is None: self.mode = default_mode else: self.mode = opt_params.get("mode") if label is None: label = "CVSP" super().__init__(self._name, self.num_qubits, params.items(), label=label) def _define(self): self.definition = self._define_initialize() def _define_initialize(self): memory = QuantumRegister(self.num_qubits, name="m") if self.mode == "mct": aux = None qr_u1 = QuantumRegister(2, name="u1") circuit = QuantumCircuit(qr_u1, memory) circuit.x(qr_u1[1]) elif self.mode == "v-chain": aux = QuantumRegister(self.num_qubits - 1, name="anc") qr_u1 = QuantumRegister(1, name="u1") qr_u2 = QuantumRegister(1, name="u2") circuit = QuantumCircuit( qr_u1, qr_u2, memory, aux, ) circuit.x(qr_u2[0]) self.norm = 1 control = range(self.num_qubits) for binary_string, amplitude in self.params: self._load_binary(circuit, binary_string, self.mode, memory, qr_u1) self._load_superposition( circuit, amplitude, self.mode, control, memory, qr_u1, qr_u2, aux ) self._load_binary(circuit, binary_string, self.mode, memory, qr_u1) return circuit @staticmethod def _load_binary(circuit, binary_string, mode, memory, qr_u1): for bit_index, bit in enumerate(binary_string): if bit == "1": if mode == "v-chain": circuit.cx(qr_u1[0], memory[bit_index]) elif mode == "mct": circuit.cx(qr_u1[1], memory[bit_index]) elif bit == "0": circuit.x(memory[bit_index]) def _load_superposition( self, circuit, feature, mode, control, memory, qr_u1, qr_u2, aux ): """ Load pattern in superposition """ alpha, beta, phi = _compute_matrix_angles(feature, self.norm) if mode == "mct": circuit.mct(memory, qr_u1[0]) circuit.cu3(alpha, beta, phi, qr_u1[0], qr_u1[1]) circuit.mct(memory, qr_u1[0]) elif mode == "v-chain": circuit.rccx(memory[control[0]], memory[control[1]], aux[0]) for j in range(2, len(control)): circuit.rccx(memory[control[j]], aux[j - 2], aux[j - 1]) circuit.cx(aux[len(control) - 2], qr_u1[0]) circuit.cu3(alpha, beta, phi, qr_u1[0], qr_u2[0]) circuit.cx(aux[len(control) - 2], qr_u1[0]) for j in reversed(range(2, len(control))): circuit.rccx(memory[control[j]], aux[j - 2], aux[j - 1]) circuit.rccx(memory[control[0]], memory[control[1]], aux[0]) self.norm = self.norm - np.absolute(np.power(feature, 2)) @staticmethod def initialize(q_circuit, state, qubits=None, opt_params=None): if qubits is None: q_circuit.append( CvqramInitialize(state, opt_params=opt_params), q_circuit.qubits ) else: q_circuit.append(CvqramInitialize(state, opt_params=opt_params), qubits)
https://github.com/qclib/qclib
qclib
# Copyright 2021 qclib project. # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # http://www.apache.org/licenses/LICENSE-2.0 # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """ Divide-and-conquer state preparation https://doi.org/10.1038/s41598-021-85474-1 https://arxiv.org/abs/2108.10182 """ from math import log2 from qiskit import QuantumCircuit from qclib.gates.initialize import Initialize from qclib.state_preparation.util.state_tree_preparation import ( Amplitude, state_decomposition, ) from qclib.state_preparation.util.angle_tree_preparation import create_angles_tree from qclib.state_preparation.util.tree_register import add_register from qclib.state_preparation.util.tree_walk import bottom_up class DcspInitialize(Initialize): """ A divide-and-conquer algorithm for quantum state preparation https://doi.org/10.1038/s41598-021-85474-1 This class implements a state preparation gate. """ def __init__(self, params, label=None): """ Parameters ---------- params: list of complex A unit vector representing a quantum state. Values are amplitudes. """ self._name = "dcsp" self._get_num_qubits(params) if label is None: self._label = "DCSP" super().__init__(self._name, self.num_qubits, params, label=label) def _define(self): self.definition = self._define_initialize() def _define_initialize(self): n_qubits = int(log2(len(self.params))) data = [Amplitude(i, a) for i, a in enumerate(self.params)] state_tree = state_decomposition(n_qubits, data) angle_tree = create_angles_tree(state_tree) circuit = QuantumCircuit() add_register(circuit, angle_tree, n_qubits - 1) bottom_up(angle_tree, circuit, n_qubits) return circuit def _get_num_qubits(self, params): if not log2(len(params)).is_integer(): Exception("The number of amplitudes is not a power of 2") self.num_qubits = len(params) - 1 @staticmethod def initialize(q_circuit, state, qubits=None): """ Appends a DcspInitialize gate into the q_circuit """ if qubits is None: q_circuit.append(DcspInitialize(state), q_circuit.qubits) else: q_circuit.append(DcspInitialize(state), qubits)
https://github.com/qclib/qclib
qclib
# Copyright 2021 qclib project. # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # http://www.apache.org/licenses/LICENSE-2.0 # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """ https://arxiv.org/abs/quant-ph/9807054 """ import numpy as np from qiskit import QuantumCircuit, QuantumRegister from qclib.gates.initialize_sparse import InitializeSparse # pylint: disable=maybe-no-member class FnPointsInitialize(InitializeSparse): """ Initializing the Amplitude Distribution of a Quantum State https://arxiv.org/abs/quant-ph/9807054 This class implements a state preparation gate. """ def __init__(self, params, label=None, opt_params=None): """State preparation using Ventura and Martinez algorithm quant-ph/9807054 Algorithm that requires a polynomial number of elementary operations for initializing a quantum system to represent only the m known points of a function f (m = len(state)). The result is a quantum superposition with m nonzero coefficients -- the creation of which is a nontrivial task compared to creating a superposition of all basis states. The amplitudes of modulus "1/sqrt(m)" will be "2 pi / N" radians apart from each other on the complex plane. Binary output function case: f:z->s with z \\in {0,1}^n and s \\in {0, 1} General case: f:z->s with z \\in {0,1}^n and s \\in {0, 1, ..., N-1} For instance, to initialize the state 1/sqrt(3)|01> + 1/sqrt(3)*e^(1*i*2pi/N)|10> + 1/sqrt(3)*e^(2*i*2pi/N)c|11> $ state = {1: 0, 2: 1, 3: 2} $ circuit = initialize(state, n=2, N=3) Parameters ---------- params: dict of {int:float} A unit vector representing a quantum state. Keys are function binary input values and values are function output values. opt_params: Dictionary n_output_values: int Number of possible output values N (Ex.: n_output_values=2 for a binary function). Default value is the max value in ``params`` minus 1. """ self._name = "fn-points" self._get_num_qubits(params) default_n_output_values = max(params.values()) - 1 if opt_params is None: self.n_output_values = default_n_output_values else: if opt_params.get("n_output_values") is None: self.n_output_values = default_n_output_values else: self.n_output_values = opt_params.get("n_output_values") self.n_output_values = max(self.n_output_values, default_n_output_values) if label is None: self._label = "FNSP" super().__init__(self._name, self.num_qubits, params.items(), label) def _define(self): self.definition = self._define_initialize() def _define_initialize(self): reg_x = QuantumRegister(self.num_qubits, "x") reg_g = QuantumRegister(self.num_qubits - 1, "g") reg_c = QuantumRegister(2, "c") circuit = QuantumCircuit(reg_x, reg_g, reg_c) reg_x = reg_x[::-1] # qiskit reverse (qiskit little-endian) bits_z0 = [int(k) for k in f"{0:0{self.num_qubits}b}"] for idx_p, (input_z, output_s) in list(enumerate(self.params))[::-1]: bits_z = [int(k) for k in input_z] circuit.x(reg_c[1]) for j, k in enumerate(bits_z): if bits_z0[j] != k: circuit.cx(reg_c[1], reg_x[j]) bits_z0 = bits_z circuit.cx(reg_c[1], reg_c[0]) circuit.x(reg_c[1]) self._apply_smatrix( circuit, idx_p, output_s.real, reg_c ) self._flipflop01(bits_z, circuit, reg_x) circuit.ccx(reg_x[0], reg_x[1], reg_g[0]) self._flipflop01(bits_z, circuit, reg_x) for k in range(2, self.num_qubits): if bits_z[k] == 0: circuit.x(reg_x[k]) circuit.ccx(reg_x[k], reg_g[k - 2], reg_g[k - 1]) if bits_z[k] == 0: circuit.x(reg_x[k]) circuit.cx(reg_g[self.num_qubits - 2], reg_c[0]) for k in range(self.num_qubits - 1, 1, -1): if bits_z[k] == 0: circuit.x(reg_x[k]) circuit.ccx(reg_x[k], reg_g[k - 2], reg_g[k - 1]) if bits_z[k] == 0: circuit.x(reg_x[k]) self._flipflop01(bits_z, circuit, reg_x) circuit.ccx(reg_x[0], reg_x[1], reg_g[0]) self._flipflop01(bits_z, circuit, reg_x) circuit.x(reg_c[1]) return circuit def _apply_smatrix(self, circuit, idx_p, output_s, reg_c): theta = -2 * np.arccos(np.sqrt(idx_p / (idx_p + 1))) # This sign is here for the smaller values of "s" to be represented by # negative amplitudes and the larger ones by positive amplitudes. # In the paper this negative sign is missing. Without it the matrix S # is not unitary. lamb = -output_s * 2 * np.pi / self.n_output_values phi = -lamb circuit.cu(theta, phi, lamb, 0, reg_c[0], reg_c[1]) @staticmethod def _flipflop01(bits_z, circuit, reg_x): if bits_z[0] == 0: circuit.x(reg_x[0]) if bits_z[1] == 0: circuit.x(reg_x[1]) @staticmethod def initialize(q_circuit, state, qubits=None, opt_params=None): if qubits is None: q_circuit.append( FnPointsInitialize(state, opt_params=opt_params), q_circuit.qubits ) else: q_circuit.append(FnPointsInitialize(state, opt_params=opt_params), qubits)
https://github.com/qclib/qclib
qclib
# This code is part of Qiskit. # # (C) Copyright IBM 2023 # # This code is licensed under the Apache License, Version 2.0. You may # obtain a copy of this license in the LICENSE.txt file in the root directory # of this source tree or at http://www.apache.org/licenses/LICENSE-2.0. # # Any modifications or derivative works of this code must retain this # copyright notice, and modified files need to carry a notice indicating # that they have been altered from the originals. # pylint: disable=missing-docstring,invalid-name,no-member # pylint: disable=attribute-defined-outside-init # pylint: disable=unused-argument from qiskit import QuantumRegister, QuantumCircuit from qiskit.compiler import transpile from qiskit.quantum_info.random import random_unitary class IsometryTranspileBench: params = ([0, 1, 2, 3], [3, 4, 5, 6]) param_names = ["number of input qubits", "number of output qubits"] def setup(self, m, n): q = QuantumRegister(n) qc = QuantumCircuit(q) if not hasattr(qc, "iso"): raise NotImplementedError iso = random_unitary(2**n, seed=0).data[:, 0 : 2**m] if len(iso.shape) == 1: iso = iso.reshape((len(iso), 1)) qc.iso(iso, q[:m], q[m:]) self.circuit = qc def track_cnot_counts_after_mapping_to_ibmq_16_melbourne(self, *unused): coupling = [ [1, 0], [1, 2], [2, 3], [4, 3], [4, 10], [5, 4], [5, 6], [5, 9], [6, 8], [7, 8], [9, 8], [9, 10], [11, 3], [11, 10], [11, 12], [12, 2], [13, 1], [13, 12], ] circuit = transpile( self.circuit, basis_gates=["u1", "u3", "u2", "cx"], coupling_map=coupling, seed_transpiler=0, ) counts = circuit.count_ops() cnot_count = counts.get("cx", 0) return cnot_count
https://github.com/qclib/qclib
qclib
# Copyright 2021 qclib project. # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # http://www.apache.org/licenses/LICENSE-2.0 # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """ Implements the state preparation defined at https://arxiv.org/abs/1003.5760. """ import numpy as np from qiskit import QuantumCircuit from qclib.unitary import unitary as decompose_unitary, cnot_count as cnots_unitary from qclib.isometry import decompose as decompose_isometry, cnot_count as cnots_isometry from qclib.gates.initialize import Initialize from qclib.entanglement import schmidt_decomposition, _to_qubits from .topdown import TopDownInitialize # pylint: disable=maybe-no-member class LowRankInitialize(Initialize): """ Approximated quantum-state preparation with entanglement dependent complexity https://arxiv.org/abs/2111.03132 This class implements a state preparation gate. """ def __init__(self, params, label=None, opt_params=None): """ Parameters ---------- params: list of complex A unit vector representing a quantum state. Values are amplitudes. opt_params: {'lr': low_rank, 'iso_scheme': isometry_scheme, 'unitary_scheme': unitary_scheme, 'partition': partition} low_rank: int ``state`` low-rank approximation (1 <= ``low_rank`` < 2**(n_qubits//2)). If ``low_rank`` is not in the valid range, it will be ignored. This parameter limits the rank of the Schmidt decomposition. If the Schmidt rank of the state decomposition is greater than ``low_rank``, a low-rank approximation is applied. iso_scheme: string Scheme used to decompose isometries. Possible values are ``'knill'`` and ``'ccd'`` (column-by-column decomposition). Default is ``isometry_scheme='ccd'``. unitary_scheme: string Scheme used to decompose unitaries. Possible values are ``'csd'`` (cosine-sine decomposition) and ``'qsd'`` (quantum Shannon decomposition). Default is ``unitary_scheme='qsd'``. partition: list of int Set of qubit indices that represent a part of the bipartition. The other partition will be the relative complement of the full set of qubits with respect to the set ``partition``. The valid range for indexes is ``0 <= index < n_qubits``. The number of indexes in the partition must be greater than or equal to ``1`` and less than or equal to ``n_qubits//2`` (``n_qubits//2+1`` if ``n_qubits`` is odd). Default is ``partition=list(range(n_qubits//2 + odd))``. svd: string Function to compute the SVD, acceptable values are 'auto' (default), 'regular', and 'randomized'. 'auto' sets `svd='randomized'` for `n_qubits>=14 and rank==1`. """ self._name = "low_rank" self._get_num_qubits(params) if opt_params is None: self.isometry_scheme = "ccd" self.unitary_scheme = "qsd" self.low_rank = 0 self.partition = None self.svd = "auto" else: self.low_rank = 0 if opt_params.get("lr") is None else opt_params.get("lr") self.partition = opt_params.get("partition") if opt_params.get("iso_scheme") is None: self.isometry_scheme = "ccd" else: self.isometry_scheme = opt_params.get("iso_scheme") if opt_params.get("unitary_scheme") is None: self.unitary_scheme = "qsd" else: self.unitary_scheme = opt_params.get("unitary_scheme") if opt_params.get("svd") is None: self.svd = "auto" else: self.svd = opt_params.get("svd") if label is None: label = "LRSP" super().__init__(self._name, self.num_qubits, params, label=label) def _define(self): self.definition = self._define_initialize() def _define_initialize(self): if self.num_qubits < 2: return TopDownInitialize(self.params).definition circuit, reg_a, reg_b = self._create_quantum_circuit() # Schmidt decomposition rank, svd_u, singular_values, svd_v = schmidt_decomposition( self.params, reg_a, rank=self.low_rank, svd=self.svd ) # Schmidt measure of entanglement e_bits = _to_qubits(rank) # Phase 1. Encodes the singular values. if e_bits > 0: reg_sv = reg_b[:e_bits] singular_values = singular_values / np.linalg.norm(singular_values) self._encode(singular_values.reshape(rank, 1), circuit, reg_sv) # Phase 2. Entangles only the necessary qubits, according to rank. for j in range(e_bits): circuit.cx(reg_b[j], reg_a[j]) # Phase 3 and 4 encode gates U and V.T self._encode(svd_u, circuit, reg_b) self._encode(svd_v.T, circuit, reg_a) return circuit.reverse_bits() @staticmethod def initialize(q_circuit, state, qubits=None, opt_params=None): """ Appends a LowRankInitialize gate into the q_circuit """ if qubits is None: q_circuit.append( LowRankInitialize(state, opt_params=opt_params), q_circuit.qubits ) else: q_circuit.append(LowRankInitialize(state, opt_params=opt_params), qubits) def _encode(self, data, circuit, reg): """ Encodes data using the most appropriate method. """ if data.shape[1] == 1: # state preparation gate_u = LowRankInitialize(data[:, 0], opt_params={ "iso_scheme": self.isometry_scheme, "unitary_scheme": self.unitary_scheme, "svd": self.svd }) elif data.shape[0] // 2 == data.shape[1]: # isometry 2^(n-1) to 2^n. gate_u = decompose_isometry(data, scheme="csd") elif data.shape[0] > data.shape[1]: gate_u = decompose_isometry(data, scheme=self.isometry_scheme) else: gate_u = decompose_unitary(data, decomposition=self.unitary_scheme) # Apply gate U to the register reg circuit.compose(gate_u, reg, inplace=True) def _create_quantum_circuit(self): if self.partition is None: self.partition = _default_partition(self.num_qubits) complement = sorted(set(range(self.num_qubits)).difference(set(self.partition))) circuit = QuantumCircuit(self.num_qubits) return circuit, self.partition[::-1], complement[::-1] def _default_partition(n_qubits): odd = n_qubits % 2 return list(range(n_qubits // 2 + odd)) def cnot_count( state_vector, low_rank=0, isometry_scheme="ccd", unitary_scheme="qsd", partition=None, method = "estimate", svd="auto" ): """ Estimate the number of CNOTs to build the state preparation circuit. """ n_qubits = _to_qubits(len(state_vector)) if n_qubits < 2: return 0 if partition is None: partition = _default_partition(n_qubits) cnots = 0 rank, svd_u, singular_values, svd_v = schmidt_decomposition( state_vector, partition, rank=low_rank, svd=svd ) # Schmidt measure of entanglement ebits = _to_qubits(rank) # Phase 1. if ebits > 0: singular_values = singular_values / np.linalg.norm(singular_values) cnots += _cnots( singular_values.reshape(rank, 1), isometry_scheme, unitary_scheme, method, svd ) # Phase 2. cnots += ebits # Phases 3 and 4. cnots += _cnots(svd_u, isometry_scheme, unitary_scheme, method, svd) cnots += _cnots(svd_v.T, isometry_scheme, unitary_scheme, method, svd) return cnots def _cnots(data, iso_scheme="ccd", uni_scheme="qsd", method="estimate", svd="auto"): if data.shape[1] == 1: return cnot_count( data[:, 0], isometry_scheme=iso_scheme, unitary_scheme=uni_scheme, method=method, svd=svd ) if data.shape[0] // 2 == data.shape[1]: return cnots_isometry(data, scheme="csd", method=method) if data.shape[0] > data.shape[1]: return cnots_isometry(data, scheme=iso_scheme, method=method) return cnots_unitary(data, decomposition=uni_scheme, method=method)
https://github.com/qclib/qclib
qclib
# Copyright 2021 qclib project. # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # http://www.apache.org/licenses/LICENSE-2.0 # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """ An Efficient Algorithm for Sparse Quantum State Preparation https://ieeexplore.ieee.org/document/9586240 """ import numpy as np from qiskit import QuantumCircuit, QuantumRegister from qiskit.circuit.library import UGate from qclib.gates.ldmcu import Ldmcu from qclib.gates.initialize_sparse import InitializeSparse class MergeInitialize(InitializeSparse): """ An Efficient Algorithm for Sparse Quantum State Preparation https://ieeexplore.ieee.org/document/9586240 """ def __init__(self, params, label=None): """ Classical algorithm that creates a quantum circuit C that loads a sparse quantum state, applying a sequence of operations maping the desired state |sigma> to |0>. And then inverting C to obtain the mapping of |0> to the desired state |sigma>. Args: params: A dictionary with the non-zero amplitudes corresponding to each state in format { '000': <value>, ... , '111': <value> } Returns: Creates a quantum gate that maps |0> to the desired state |params> """ self._name = "merge" if label is None: label = "MERGESP" # Parameters need to be validated first by superclass self._get_num_qubits(params) super().__init__(self._name, self.num_qubits, params.items(), label=label) def _define(self): self.definition = self._define_initialize() def _define_initialize(self): state_dict = dict(self.params) b_strings = list(state_dict.keys()) n_qubits = len(b_strings[0]) quantum_register = QuantumRegister(n_qubits) quantum_circuit = QuantumCircuit(quantum_register) while len(b_strings) > 1: bitstr1, bitstr2, dif, dif_qubits = self._select_strings(state_dict) bitstr1, bitstr2, state_dict, quantum_circuit = self._preprocess_states( bitstr1, bitstr2, dif, dif_qubits, state_dict, quantum_circuit ) state_dict, quantum_circuit = self._merge( state_dict, quantum_circuit, bitstr1, bitstr2, dif_qubits, dif ) b_strings = list(state_dict.keys()) b_string = b_strings.pop() for (bit_idx, bit) in enumerate(b_string): if bit == "1": quantum_circuit.x(bit_idx) return quantum_circuit.reverse_ops() @staticmethod def initialize(q_circuit, state, qubits=None): if qubits is None: q_circuit.append(MergeInitialize(state), q_circuit.qubits) else: q_circuit.append(MergeInitialize(state), qubits) @staticmethod def _maximizing_difference_bit_search(b_strings, dif_qubits): """ Splits the set of bit strings into two (t_0 and t_1), by setting t_0 as the set of bit_strings with 0 in the bit_index position, and t_1 as the set of bit_strings with 1 in the bit_index position. Searching for the bit_index not in dif_qubits that maximizes the difference between the size of the nonempty t_0 and t_1. Args: b_string: A list of bit strings eg.: ['000', '011', ...,'101'] dif_qubits: A list of previous qubits found to maximize the difference Returns: bit_index: The qubit index that maximizes abs(len(t_0)-len(t_1)) t_0: List of binary strings with 0 on the bit_index qubit t_1: List of binary strings with 1 on the bit_index qubit """ t_0 = [] t_1 = [] bit_index = 0 set_difference = -1 bit_search_space = list(set(range(len(b_strings[0]))) - set(dif_qubits)) for bit in bit_search_space: temp_t0 = [x for x in b_strings if x[bit] == "0"] temp_t1 = [x for x in b_strings if x[bit] == "1"] if temp_t0 and temp_t1: temp_difference = np.abs(len(temp_t0) - len(temp_t1)) if temp_difference > set_difference: t_0 = temp_t0 t_1 = temp_t1 bit_index = bit set_difference = temp_difference return bit_index, t_0, t_1 @staticmethod def _build_bit_string_set(b_strings, dif_qubits, dif_values): """ Creates a new set of bit strings from b_strings, where the bits in the indexes in dif_qubits match the values in dif_values. Args: b_strings: list of bit strings eg.: ['000', '011', ...,'101'] dif_qubits: list of integers with the bit indexes dif_values: list of integers values containing the values each bit with index in dif_qubits shoud have Returns: A new list of bit_strings, with matching values in dif_values on indexes dif_qubits """ bit_string_set = [] for b_string in b_strings: if [b_string[i] for i in dif_qubits] == dif_values: bit_string_set.append(b_string) return bit_string_set def _bit_string_search(self, b_strings, dif_qubits, dif_values): """ Searches for the bit strings with unique qubit values in `dif_values` on indexes `dif_qubits`. Args: b_strings: List of binary strings where the search is to be performed e.g.: ['000', '010', '101', '111'] dif_qubits: List of indices on a binary string of size N e.g.: [1, 3, 5] dif_values: List of values each qubit must have on indexes stored in dif_qubits [0, 1, 1] Returns: b_strings: One size list with the string found, to have values dif_values on indexes dif_qubits dif_qubits: Updated list with new indexes dif_values: Updated list with new values """ temp_strings = b_strings while len(temp_strings) > 1: bit, t_0, t_1 = self._maximizing_difference_bit_search( temp_strings, dif_qubits ) dif_qubits.append(bit) if len(t_0) < len(t_1): dif_values.append("0") temp_strings = t_0 else: dif_values.append("1") temp_strings = t_1 return temp_strings, dif_qubits, dif_values def _select_strings(self, state_dict): """ Searches for the states described by the bit strings bitstr1 and bitstr2 to be merged Args: state_dict: A dictionary with the non-zero amplitudes associated to their corresponding binary strings as keys e.g.: {'001': <value>, '101': <value>} Returns: bitstr1: First binary string bitstr2: Second binary string dif_qubit: Qubit index to be used as target for the merging operation dif_qubits: List of qubit indexes where bitstr1 and bitstr2 must be equal, because the correspondig qubits of those indexes are to be used as control for the merging operation """ # Initialization dif_qubits = [] dif_values = [] b_strings1 = b_strings2 = list(state_dict.keys()) # Searching for bitstr1 (b_strings1, dif_qubits, dif_values) = self._bit_string_search( b_strings1, dif_qubits, dif_values ) dif_qubit = dif_qubits.pop() dif_values.pop() bitstr1 = b_strings1[0] # Searching for bitstr2 b_strings2.remove(bitstr1) b_strings1 = self._build_bit_string_set(b_strings2, dif_qubits, dif_values) (b_strings1, dif_qubits, dif_values) = self._bit_string_search( b_strings1, dif_qubits, dif_values ) bitstr2 = b_strings1[0] return bitstr1, bitstr2, dif_qubit, dif_qubits @staticmethod def _apply_operation_to_bit_string(b_string, operation, qubit_indexes): """ Applies changes on binary strings according to the operation Args: b_string: Binary string '00110' operation: Operation to be applied to the string qubit_indexes: Indexes of the qubits on the binary strings where the operations are to be applied Returns: Updated binary string """ assert operation in ["x", "cx"] if operation == "x": compute = _compute_op_x else: compute = _compute_op_cx return compute(b_string, qubit_indexes) @staticmethod def _update_state_dict_according_to_operation( state_dict, operation, qubit_indexes, merge_strings=None ): """ Updates the keys of the state_dict according to the operation being applied to the circuit Args: state_dict: A dictionary with the non-zero amplitudes associated to their corresponding binary strings as keys e.g.: {'001': <value>, '101': <value>} operation: Operation to be applied to the states, it must be ['x', 'cx', 'merge'] qubit_indexes: Indexes of the qubits on the binary strings where the operations are to be applied merge_strings: Binary strings associated ot the states on the quantum processor to be merge e.g.:['01001', '10110'] Returns: A state_dict with the updated states """ assert operation in ["x", "cx", "merge"] state_list = list(state_dict.items()) new_state_dict = {} if operation == "merge": assert merge_strings is not None # Computes the norm of bitstr1 and bitstr2 new_state_dict = state_dict.copy() norm = np.linalg.norm( [new_state_dict[merge_strings[0]], new_state_dict[merge_strings[1]]] ) new_state_dict.pop(merge_strings[1], None) new_state_dict[merge_strings[0]] = norm else: for (bit_string, value) in state_list: temp_bstring = MergeInitialize._apply_operation_to_bit_string( bit_string, operation, qubit_indexes ) new_state_dict[temp_bstring] = value return new_state_dict @staticmethod def _equalize_bit_string_states( bitstr1, bitstr2, dif, state_dict, quantum_circuit ): """ Applies operations to the states represented by bit strings bitstr1 and bitstr2 equalizing them at every qubit except the one in the dif index. And alters the bit strings and state_dict accordingly. Args: bitstr1: First bit string bitstr2: Second bit string dif: index where both bitstr1 and bitstr2 must be different state_dict: A dictionary with the non-zero amplitudes associated to their corresponding binary strings as keys e.g.: {'001': <value>, '101': <value>} quantum_circuit: Qiskit's quantum circuit's object with the gates applied to the circuit Returns: Updated bitstr1, bitstr2, state_dict and quantum_circuit """ b_index_list = list(range(len(bitstr1))) b_index_list.remove(dif) for b_index in b_index_list: if bitstr1[b_index] != bitstr2[b_index]: quantum_circuit.cx(dif, b_index) bitstr1 = MergeInitialize._apply_operation_to_bit_string( bitstr1, "cx", [dif, b_index] ) bitstr2 = MergeInitialize._apply_operation_to_bit_string( bitstr2, "cx", [dif, b_index] ) state_dict = MergeInitialize._update_state_dict_according_to_operation( state_dict, "cx", [dif, b_index] ) return bitstr1, bitstr2, state_dict, quantum_circuit @staticmethod def _apply_not_gates_to_qubit_index_list( bitstr1, bitstr2, dif_qubits, state_dict, quantum_circuit ): """ Applies quantum not gate at the qubit at a given index, where the state represented by the bit string bitstr2 is different than '1' at index in diff_qubits. Args: bitstr1: First bit string bitstr2: Second bit string dif_qubits: indexes where both bitstr1 and bitstr2 are equal state_dict: A dictionary with the non-zero amplitudes associated to their corresponding binary strings as keys e.g.: {'001': <value>, '101': <value>} quantum_circuit: Qiskit's quantum circuit's object with the gates applied to the circuit Returns: Updated bitstr1, bitstr2, state_dict and quantum_circuit """ for b_index in dif_qubits: if bitstr2[b_index] != "1": quantum_circuit.x(b_index) bitstr1 = MergeInitialize._apply_operation_to_bit_string(bitstr1, "x", b_index) bitstr2 = MergeInitialize._apply_operation_to_bit_string(bitstr2, "x", b_index) state_dict = MergeInitialize._update_state_dict_according_to_operation( state_dict, "x", b_index ) return bitstr1, bitstr2, state_dict, quantum_circuit def _preprocess_states( self, bitstr1, bitstr2, dif, dif_qubits, state_dict, quantum_circuit ): """ Apply the operations on the basis states to prepare for merging bitstr1 and bitstr2. Args: state_dict: A dictionary with the non-zero amplitudes associated to their corresponding binary strings as keys e.g.: {'001': <value>, '101': <value>} bitstr1: First binary string to be merged bitstr2: Second binary string to be merged dif_qubits: List of qubit indexes on the binary strings dif: Target qubit index where the merge operation is to be applied quantum_circuit: Qiskit's QuantumCircuit object where the operations are to be called Returns: state_dict: Updated state dict bitstr1: First updated binary string to be merged bitstr2: Second updated binary string to be merged quantum_circuit: Qiskit's quantum circuit's object with the gates applied to the circuit """ if bitstr1[dif] != "1": quantum_circuit.x(dif) bitstr1 = MergeInitialize._apply_operation_to_bit_string(bitstr1, "x", dif) bitstr2 = MergeInitialize._apply_operation_to_bit_string(bitstr2, "x", dif) state_dict = self._update_state_dict_according_to_operation( state_dict, "x", dif ) ( bitstr1, bitstr2, state_dict, quantum_circuit, ) = self._equalize_bit_string_states( bitstr1, bitstr2, dif, state_dict, quantum_circuit ) ( bitstr1, bitstr2, state_dict, quantum_circuit, ) = self._apply_not_gates_to_qubit_index_list( bitstr1, bitstr2, dif_qubits, state_dict, quantum_circuit ) return bitstr1, bitstr2, state_dict, quantum_circuit @staticmethod def _compute_angles(amplitude_1, amplitude_2): """ Computes the angles for the adjoint of the merge matrix M that is going to map the dif qubit to zero e.g.: M(a|0> + b|1>) -> |1> Args: amplitude_1: A complex/real value, associated with the string with 1 on the dif qubit amplitude_2: A complex/real value, associated with the string with 0 on the dif qubit Returns: The angles theta, lambda and phi for the U operator """ norm = np.linalg.norm([amplitude_1, amplitude_2]) phi = 0 lamb = 0 # there is no minus on the theta because the intetion is to compute the inverse if isinstance(amplitude_1, complex) or isinstance(amplitude_2, complex): amplitude_1 = ( complex(amplitude_1) if not isinstance(amplitude_1, complex) else amplitude_1 ) amplitude_2 = ( complex(amplitude_2) if not isinstance(amplitude_2, complex) else amplitude_2 ) theta = -2 * np.arcsin(np.abs(amplitude_2 / norm)) lamb = np.log(amplitude_2 / norm).imag phi = np.log(amplitude_1 / norm).imag - lamb else: theta = -2 * np.arcsin(amplitude_2 / norm) return theta, phi, lamb def _merge(self, state_dict, quantum_circuit, bitstr1, bitstr2, dif_qubits, dif): theta, phi, lamb = self._compute_angles( state_dict[bitstr1], state_dict[bitstr2] ) # Applying merge operation merge_gate = UGate(theta, phi, lamb, label="U") if not dif_qubits: quantum_circuit.append(merge_gate, dif_qubits + [dif], []) else: gate_definition = UGate(theta, phi, lamb, label="U").to_matrix() Ldmcu.ldmcu(quantum_circuit, gate_definition, dif_qubits, dif) state_dict = self._update_state_dict_according_to_operation( state_dict, "merge", None, merge_strings=[bitstr1, bitstr2] ) return state_dict, quantum_circuit def _compute_op_cx(xlist, idx): return ( f'{xlist[: idx[1]]}{(not int(xlist[idx[1]])) * 1}{xlist[idx[1] + 1:]}' if xlist[idx[0]] == "1" else xlist ) def _compute_op_x(xlist, idx): return ( xlist[:idx] + "1" + xlist[idx + 1:] if xlist[idx] == "0" else xlist[0:idx] + "0" + xlist[idx + 1:] )
https://github.com/qclib/qclib
qclib
# Copyright 2021 qclib project. # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # http://www.apache.org/licenses/LICENSE-2.0 # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """ Initializes a mixed quantum state. """ from math import log2, ceil, isclose import numpy as np from qiskit import QuantumCircuit, QuantumRegister from qclib.gates.initialize import Initialize from qclib.gates.initialize_mixed import InitializeMixed from qclib.gates.initialize_sparse import InitializeSparse from qclib.state_preparation import LowRankInitialize # pylint: disable=maybe-no-member class MixedInitialize(InitializeMixed): """ This class implements a mixed state preparation gate. """ def __init__( self, params, initializer=LowRankInitialize, opt_params=None, probabilities=None, label=None, reset=True, classical = True ): """ Parameters ---------- params: list of list of complex A list of unit vectors, each representing a quantum state. Values are amplitudes. initializer: Initialize or InitializeSparse Type of the class that will be applied to prepare pure states. Default is ``LowRankInitialize``. opt_params: dictionary Optional parameters of the class of type ``initializer``. reset: bool Indicates whether the auxiliary qubits should be reset or not. Default is ``True``. classical: bool Indicates whether the purification is done classically or in-circuit (quantically). Default is ``True``. """ if ( not issubclass(initializer, Initialize) and not issubclass(initializer, InitializeSparse) ): raise TypeError("The value of initializer should be Initialize or InitializeSparse.") if probabilities is None: probabilities = [1/len(params)] * len(params) elif any(i < 0.0 for i in probabilities): raise ValueError("All probabilities must greater than or equal to 0.") elif any(i > 1.0 for i in probabilities): raise ValueError("All probabilities must less than or equal to 1.") elif not isclose(sum(probabilities), 1.0): raise ValueError("The sum of the probabilities must be 1.0.") self._name = "mixed" self._get_num_qubits(params) self._initializer = initializer self._reset = reset self._opt_params = opt_params self._probabilities = probabilities self._classical = classical self._num_ctrl_qubits = int(ceil(log2(len(params)))) self._num_data_qubits = initializer(params[0]).num_qubits self._list_params = params if label is None: label = "Mixed" super().__init__(self._name, self.num_qubits, np.array(params).reshape(-1), label=label) def _define(self): self.definition = self._define_initialize() def _define_initialize(self): purified_circuit = QuantumCircuit(self.num_qubits) if self._classical: # Calculates the pure state classically. pure_state = np.zeros(2**(self._num_qubits), dtype=complex) for index, (state_vector, prob) in enumerate( zip(self._list_params, self._probabilities) ): basis = np.zeros(2**self._num_ctrl_qubits) basis[index] = 1 pure_state += np.kron(np.sqrt(prob) * state_vector, basis) purified_circuit = self._initializer( pure_state, opt_params=self._opt_params ).definition else: # Calculates the pure state quantically. aux_state = np.concatenate(( np.sqrt(self._probabilities), [0] * (2**(self._num_ctrl_qubits) - len(self._probabilities)) )) sub_circuit = self._initializer( aux_state, opt_params=self._opt_params ).definition sub_circuit.name = 'aux. space' purified_circuit.append(sub_circuit, range(self._num_ctrl_qubits)) for index, state in enumerate(self._list_params): sub_circuit = self._initializer( state, opt_params=self._opt_params ).definition sub_circuit.name = f'state {index}' sub_circuit = sub_circuit.control( num_ctrl_qubits=self._num_ctrl_qubits, ctrl_state = f"{index:0{self._num_ctrl_qubits}b}" ) purified_circuit.compose(sub_circuit, purified_circuit.qubits, inplace=True) purified_circuit.name = 'purified state' circuit = QuantumCircuit() circuit.add_register(QuantumRegister(self._num_ctrl_qubits, 'aux')) circuit.add_register(QuantumRegister(self._num_data_qubits, 'rho')) circuit.append(purified_circuit, circuit.qubits) if self._reset: circuit.reset(range(self._num_ctrl_qubits)) return circuit @staticmethod def initialize(q_circuit, ensemble, qubits=None, opt_params=None, probabilities=None): """ Appends a MixedInitialize gate into the q_circuit """ if qubits is None: q_circuit.append( MixedInitialize( ensemble, opt_params=opt_params, probabilities=probabilities ) , q_circuit.qubits ) else: q_circuit.append( MixedInitialize( ensemble, opt_params=opt_params, probabilities=probabilities ) , qubits )
https://github.com/qclib/qclib
qclib
# Copyright 2021 qclib project. # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # http://www.apache.org/licenses/LICENSE-2.0 # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """ sparse state preparation """ import numpy as np import qiskit from qiskit import QuantumCircuit from qclib.gates.initialize_sparse import InitializeSparse from .lowrank import LowRankInitialize # pylint: disable=maybe-no-member class PivotInitialize(InitializeSparse): """Pivot State Preparation arXiv:2006.00016""" def __init__(self, params, label=None, opt_params=None): self._get_num_qubits(params) default_aux = False if opt_params is None: self.aux = default_aux else: if opt_params.get("aux") is None: self.aux = default_aux else: self.aux = opt_params.get("aux") self.non_zero = len(params) self.register = qiskit.QuantumRegister(self.num_qubits, name="q") self.index_differ = None self.ctrl_state = None if label is None: self.label = "PivotSP" super().__init__("PivotInitialize", self.num_qubits, params.items(), label=label) def _define(self): self.definition = self._define_initialize() def _define_initialize(self): target_size = np.log2(self.non_zero) target_size = np.ceil(target_size) target_size = int(target_size) if self.aux: anc, n_anci, pivot_circuit = self._circuit_with_ancilla(target_size) else: pivot_circuit = qiskit.QuantumCircuit(self.register) next_state = self.params.copy() index_nonzero = self._get_index_nz(self.num_qubits - target_size, next_state) while index_nonzero is not None: index_zero = self._get_index_zero(self.non_zero, next_state) circ, next_state = self._pivoting( index_nonzero, target_size, index_zero, next_state ) pivot_circuit.compose(circ, pivot_circuit.qubits, inplace=True) index_nonzero = self._get_index_nz( self.num_qubits - target_size, next_state ) dense_state = np.zeros(2 ** target_size, dtype=complex) for key, value in next_state: dense_state[int(key, 2)] = value if self.non_zero <= 2: initialize_circ = qiskit.QuantumCircuit(1) LowRankInitialize.initialize(initialize_circ, dense_state) else: initialize_circ = QuantumCircuit(target_size) LowRankInitialize.initialize(initialize_circ, dense_state) if self.aux: circuit = qiskit.QuantumCircuit(anc, self.register) nun_aux = n_anci - 1 circuit.compose( initialize_circ, circuit.qubits[nun_aux : nun_aux + target_size], inplace=True, ) circuit.barrier() circuit.compose(pivot_circuit.reverse_bits().reverse_ops(), inplace=True) else: circuit = qiskit.QuantumCircuit(self.num_qubits) circuit.compose(initialize_circ, circuit.qubits[:target_size], inplace=True) circuit.compose( pivot_circuit.reverse_bits().reverse_ops(), circuit.qubits, inplace=True ) return circuit def _circuit_with_ancilla(self, target_size): remain = list(range(self.num_qubits - target_size, self.num_qubits)) n_anci = len(remain) anc = qiskit.QuantumRegister(n_anci - 1, name="anc") pivot_circuit = qiskit.QuantumCircuit(anc, self.register) return anc, n_anci, pivot_circuit def _next_state( self, remain, target_cx, index_zero, next_state ): tab = {"0": "1", "1": "0"} new_state = {} for index, amp in next_state: if index[self.index_differ] == self.ctrl_state: n_index = "" for k, value in enumerate(index): if k in target_cx: n_index = n_index + tab[value] else: n_index = n_index + value else: n_index = index if n_index[remain[0] :] == index_zero[remain[0] :]: n_index = ( n_index[:self.index_differ] + tab[index[self.index_differ]] + n_index[self.index_differ + 1 :] ) new_state[n_index] = amp return new_state.items() def _pivoting(self, index_nonzero, target_size, index_zero, next_state): """pivot amplitudes of index_nonzero and self.index_zero""" target = list(range(self.num_qubits - target_size)) remain = list(range(self.num_qubits - target_size, self.num_qubits)) memory = qiskit.QuantumRegister(self.num_qubits) anc, circuit = self._initialize_circuit(memory, remain) self.index_differ = 0 for k in target: if index_nonzero[k] != index_zero[k]: self.index_differ = k break target_cx = [] self.ctrl_state = index_nonzero[self.index_differ] for k in target: if self.index_differ != k and index_nonzero[k] != index_zero[k]: circuit.cx(self.index_differ, k, ctrl_state=self.ctrl_state) target_cx.append(k) for k in remain: if index_nonzero[k] != index_zero[k]: circuit.cx(self.index_differ, k, ctrl_state=self.ctrl_state) target_cx.append(k) for k in remain: if index_zero[k] == "0": circuit.x(k) if self.aux: # apply mcx using mode v-chain self._mcxvchain(circuit, memory, anc, remain, self.index_differ) else: circuit.mcx(remain, self.index_differ) for k in remain: if index_zero[k] == "0": circuit.x(k) next_state = self._next_state( remain, target_cx, index_zero, next_state ) return circuit, next_state def _initialize_circuit(self, memory, remain): if self.aux: n_anci = len(remain) anc = qiskit.QuantumRegister(n_anci - 1, name="anc") circuit = qiskit.QuantumCircuit(memory, anc) else: circuit = qiskit.QuantumCircuit(memory) anc = None return anc, circuit def _get_index_zero(self, non_zero, state): index_zero = None for k in range(2**non_zero): index = f"{k:0{self.num_qubits}b}" not_exists = sum(1 for v in state if v[0] == index) == 0 if not_exists: index_zero = index break return index_zero @staticmethod def _get_index_nz(target_size, next_state): index_nonzero = None for index, _ in next_state: if index[:target_size] != target_size * "0": index_nonzero = index break return index_nonzero @staticmethod def _mcxvchain(circuit, memory, anc, lst_ctrl, tgt): """multi-controlled x gate with working qubits""" circuit.rccx(memory[lst_ctrl[0]], memory[lst_ctrl[1]], anc[0]) for j in range(2, len(lst_ctrl)): circuit.rccx(memory[lst_ctrl[j]], anc[j - 2], anc[j - 1]) circuit.cx(anc[len(lst_ctrl) - 2], tgt) for j in reversed(range(2, len(lst_ctrl))): circuit.rccx(memory[lst_ctrl[j]], anc[j - 2], anc[j - 1]) circuit.rccx(memory[lst_ctrl[0]], memory[lst_ctrl[1]], anc[0]) @staticmethod def initialize(q_circuit, state, qubits=None, opt_params=None): if qubits is None: q_circuit.append( PivotInitialize(state, opt_params=opt_params), q_circuit.qubits ) else: q_circuit.append(PivotInitialize(state, opt_params=opt_params), qubits)
https://github.com/qclib/qclib
qclib
# Copyright 2021 qclib project. # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # http://www.apache.org/licenses/LICENSE-2.0 # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """ Implements the state preparation defined at https://arxiv.org/abs/1003.5760. """ import numpy as np from qiskit import QuantumRegister, QuantumCircuit from qclib.unitary import unitary from qclib.gates.initialize import Initialize from .topdown import TopDownInitialize class SVDInitialize(Initialize): """ https://arxiv.org/abs/1003.5760 This class implements a state preparation gate. """ def __init__(self, params, label=None): """ Parameters ---------- params: list of complex A unit vector representing a quantum state. Values are amplitudes. """ self._name = "svd" self._get_num_qubits(params) if label is None: label = "SVDSP" super().__init__(self._name, self.num_qubits, params, label=label) def _define(self): self.definition = self._define_initialize() def _define_initialize(self): """ State preparation using Schmidt decomposition arXiv:1003.5760 """ state = np.copy(self.params) n_qubits = self.num_qubits odd = n_qubits % 2 state.shape = (int(2 ** (n_qubits // 2)), int(2 ** (n_qubits // 2 + odd))) matrix_u, matrix_d, matrix_v = np.linalg.svd(state) matrix_d = matrix_d / np.linalg.norm(matrix_d) reg_a = QuantumRegister(n_qubits // 2 + odd) reg_b = QuantumRegister(n_qubits // 2) circuit = QuantumCircuit(reg_a, reg_b) if len(matrix_d) > 2: gate = SVDInitialize(matrix_d) circuit.append(gate, reg_b) else: gate = TopDownInitialize(matrix_d) circuit.append(gate, reg_b) for k in range(int(n_qubits // 2)): circuit.cx(reg_b[k], reg_a[k]) # apply gate U to the first register gate_u = unitary(matrix_u) circuit.append(gate_u, reg_b) # apply gate V to the second register gate_v = unitary(matrix_v.T) circuit.append(gate_v, reg_a) return circuit @staticmethod def initialize(q_circuit, state, qubits=None): """ Appends a SVDInitialize gate into the q_circuit """ if qubits is None: q_circuit.append(SVDInitialize(state), q_circuit.qubits) else: q_circuit.append(SVDInitialize(state), qubits)
https://github.com/qclib/qclib
qclib
# Copyright 2021 qclib project. # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # http://www.apache.org/licenses/LICENSE-2.0 # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """ https://arxiv.org/abs/quant-ph/0407010 https://arxiv.org/abs/2108.10182 """ import numpy as np from qiskit import QuantumCircuit, QuantumRegister from qclib.gates.initialize import Initialize from qclib.state_preparation.util.state_tree_preparation import ( Amplitude, state_decomposition, ) from qclib.state_preparation.util.angle_tree_preparation import create_angles_tree from qclib.state_preparation.util.tree_register import add_register from qclib.state_preparation.util.tree_walk import top_down class TopDownInitialize(Initialize): """ Top-down state preparation https://arxiv.org/abs/quant-ph/0407010 https://arxiv.org/abs/2108.10182 This class implements a state preparation gate. """ def __init__(self, params, label=None, opt_params=None): """ Parameters ---------- params: list of complex A unit vector representing a quantum state. Values are amplitudes. opt_params: {'global_phase': global_phase, 'lib': lib} global_phase: bool If ``True``, corrects the global phase. Default value is ``True``. lib: str Library to be used. Default value is ``'qclib'``. """ self._name = "top-down" self._get_num_qubits(params) if opt_params is None: self.global_phase = True self.lib = "qclib" else: if opt_params.get("global_phase") is None: self.global_phase = True else: self.global_phase = opt_params.get("global_phase") if opt_params.get("lib") is None: self.lib = "qclib" else: self.lib = opt_params.get("lib") if label is None: label = "TDSP" super().__init__(self._name, self.num_qubits, params, label=label) def _define(self): self.definition = self._define_initialize() def _define_initialize(self): if self.lib == "qiskit": reg = QuantumRegister(self.num_qubits) circuit = QuantumCircuit(reg) # pylint: disable=maybe-no-member circuit.initialize(self.params) return circuit data = [Amplitude(i, a) for i, a in enumerate(self.params)] state_tree = state_decomposition(self.num_qubits, data) angle_tree = create_angles_tree(state_tree) circuit = QuantumCircuit() add_register(circuit, angle_tree, 0) top_down(angle_tree, circuit, 0) if self.global_phase: circuit.global_phase += sum(np.angle(self.params)) / len(self.params) return circuit @staticmethod def initialize(q_circuit, state, qubits=None, opt_params=None): """ Appends a TopDownInitialize gate into the q_circuit """ if qubits is None: q_circuit.append( TopDownInitialize(state, opt_params=opt_params), q_circuit.qubits ) else: q_circuit.append(TopDownInitialize(state, opt_params=opt_params), qubits)
https://github.com/qclib/qclib
qclib
# Copyright 2021 qclib project. # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # http://www.apache.org/licenses/LICENSE-2.0 # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """ This module implements the state preparation proposed by Bergholm et al (2005) available in: https://journals.aps.org/pra/abstract/10.1103/PhysRevA.71.052330 """ import numpy as np import numpy.linalg as la from qiskit import QuantumCircuit, QuantumRegister from qiskit.circuit.library import UCGate from qclib.gates.initialize import Initialize class UCGInitialize(Initialize): """ Quantum circuits with uniformly controlled one-qubit gates https://doi.org/10.48550/arXiv.quant-ph/0410066 """ def __init__(self, params, label=None, opt_params=None): self._name = "ucg_initialize" self._get_num_qubits(params) self.register = QuantumRegister(self.num_qubits) self.circuit = QuantumCircuit(self.register) self.target_state = 0 if opt_params is None else opt_params.get("target_state") self.str_target = bin(self.target_state)[2:].zfill(self.num_qubits)[::-1] self.preserve = False if opt_params is None else opt_params.get("preserve_previous") if label is None: label = "ucg_initialize" super().__init__(self._name, self.num_qubits, params, label=label) def _define(self): self.definition = self._define_initialize() def _define_initialize(self): children = self.params parent = self._update_parent(children) tree_level = self.num_qubits r_gate = self.target_state // 2 while tree_level > 0: bit_target, ucg = self._disentangle_qubit(children, parent, r_gate, tree_level) children = self._apply_diagonal(bit_target, parent, ucg) parent = self._update_parent(children) # prepare next iteration r_gate = r_gate // 2 tree_level -= 1 return self.circuit.inverse() def _disentangle_qubit(self, children: 'list[float]', parent: 'list[float]', r_gate: int, tree_level: int): """ Apply UCGate to disentangle qubit target""" bit_target = self.str_target[self.num_qubits - tree_level] mult, mult_controls, target = self._define_mult(children, parent, tree_level) if self.preserve: self._preserve_previous(mult, mult_controls, r_gate, target) ucg = self._apply_ucg(mult, mult_controls, target) return bit_target, ucg def _define_mult(self, children: 'list[float]', parent: 'list[float]', tree_level: int): current_level_mux = self._build_multiplexor(parent, children, self.str_target) mult_controls, target = self._get_ctrl_targ(tree_level) return current_level_mux, mult_controls, target def _apply_ucg(self, current_level_mux: 'list[np.ndarray]', mult_controls: 'list[int]', target: int): """ Creates and applies multiplexer """ ucg = UCGate(current_level_mux, up_to_diagonal=True) if len(current_level_mux) != 1: self.circuit.append(ucg, [target] + mult_controls) else: self.circuit.unitary(current_level_mux[0], target) # pylint: disable=maybe-no-member return ucg def _preserve_previous(self, mux: 'list[np.ndarray]', mult_controls: 'list[int]', r_gate: int, target: int): """ Remove one gate from mux and apply separately to avoid changing previous base vectors """ out_gate = mux[r_gate] qc_gate = QuantumCircuit(1) qc_gate.unitary(out_gate, 0) # pylint: disable=maybe-no-member mux[r_gate] = np.eye(2) out_gate_ctrl = list(range(0, target)) + list(range(target + 1, self.num_qubits)) ctrl_state = self.str_target[0:target][::-1] if len(ctrl_state) < self.num_qubits - 1: ctrl_state = bin(r_gate)[2:].zfill(len(mult_controls)) + ctrl_state gate = qc_gate.control(self.num_qubits - 1, ctrl_state=ctrl_state) self.circuit.compose(gate, out_gate_ctrl + [target], inplace=True) @staticmethod def _update_parent(children): size = len(children) // 2 parent = [la.norm([children[2 * k], children[2 * k + 1]]) for k in range(size)] return parent @staticmethod def _apply_diagonal(bit_target: str, parent: 'list[float]', ucg: UCGate): children = parent if bit_target == '1': diagonal = np.conj(ucg._get_diagonal())[1::2] # pylint: disable=protected-access else: diagonal = np.conj(ucg._get_diagonal())[::2] # pylint: disable=protected-access children = children * diagonal return children def _get_ctrl_targ(self, tree_level: int): controls = list(range(self.num_qubits - tree_level + 1, self.num_qubits)) target = self.num_qubits - tree_level return controls, target def _build_multiplexor(self, parent_amplitudes: 'list[float]', children_amplitudes: 'list[float]', str_target: str): """ Infers the unitary to be used in the uniformily controlled multiplexor defined by Bergholm et al (2005). Args: parent_amplitudes: list of amplitudes children_amplitudes: children of the parent amplitudes Returns: list of 2-by-2 numpy arrays with the desired unitaries to be used in the multiplexor """ tree_lvl = int(np.log2(len(children_amplitudes))) bit_target = str_target[self.num_qubits - tree_lvl] gates = [] len_pnodes = len(parent_amplitudes) len_snodes = len(children_amplitudes) for parent_idx, sibling_idx in zip(range(len_pnodes), range(0, len_snodes, 2)): if parent_amplitudes[parent_idx] != 0: amp_ket0 = (children_amplitudes[sibling_idx] / parent_amplitudes[parent_idx]) amp_ket1 = (children_amplitudes[sibling_idx + 1] / parent_amplitudes[parent_idx]) if amp_ket0 != 0: gates += [self._get_branch_operator(amp_ket0, amp_ket1, bit_target)] else: gates += [self._get_diagonal_operator(amp_ket1, bit_target)] else: gates += [np.eye(2)] return gates @staticmethod def _get_branch_operator(amplitude_ket0, amplitude_ket1, target='0'): """ Returns the matrix operator that is going to split the qubit in to two components of a superposition Args: amplitude_ket0: Complex amplitude of the |0> component of the superpoisition amplitude_ket1: Complex amplitude of the |1> component of the superpoisition Returns: A 2x2 numpy array defining the desired operator inferred from the amplitude argument """ if target == '0': operator = np.array([[amplitude_ket0, -np.conj(amplitude_ket1)], [amplitude_ket1, np.conj(amplitude_ket0)]]) else: operator = np.array([[-np.conj(amplitude_ket1), amplitude_ket0], [np.conj(amplitude_ket0), amplitude_ket1]]) return np.conj(operator).T @staticmethod def initialize(q_circuit, state, qubits=None, opt_params=None): gate = UCGInitialize(state, opt_params=opt_params) if qubits is None: q_circuit.append(gate.definition, q_circuit.qubits) else: q_circuit.append(gate.definition, qubits) @staticmethod def _get_diagonal_operator(amplitude_ket1, target): if target == '0': operator = np.array([[0, 1], [amplitude_ket1, 0]]) else: operator = np.array([[1, 0], [0, amplitude_ket1]]) return np.conj(operator).T
https://github.com/qclib/qclib
qclib
# Copyright 2021 qclib project. # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # http://www.apache.org/licenses/LICENSE-2.0 # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """ todo """ import numpy as np import qiskit from qiskit.circuit.library import UCGate from qiskit.quantum_info import Operator from qclib.state_preparation.ucg import UCGInitialize def _repetition_verify(base, d, mux, mux_cpy): i = 0 next_base = base + d while i < d: if not np.allclose(mux[base], mux[next_base]): return False mux_cpy[next_base] = None base, next_base, i = base + 1, next_base + 1, i + 1 return True def _repetition_search(mux, n, mux_cpy): dont_carry = [] for i in range(1, len(mux) // 2 + 1): d = i entanglement = False if np.log2(d).is_integer() and np.allclose(mux[i], mux[0]): mux_org = mux_cpy[:] repetitions = len(mux) // (2 * d) base = 0 while repetitions: repetitions -= 1 valid = _repetition_verify(base, d, mux, mux_cpy) base += 2 * d if not valid: mux_cpy[:] = mux_org break if repetitions == 0: entanglement = True if entanglement: dont_carry.append(n + int(np.log2(d)) + 1) return dont_carry class UCGEInitialize(UCGInitialize): """ todo """ def __init__(self, params, label=None, opt_params=None): super().__init__(params, label=label, opt_params=opt_params) def _define_initialize(self): children = self.params parent = self._update_parent(children) tree_level = self.num_qubits r_gate = self.target_state // 2 while tree_level > 0: bit_target, ucg = self._disentangle_qubit( children, parent, r_gate, tree_level ) children = self._apply_diagonal(bit_target, parent, ucg) parent = self._update_parent(children) # prepare next iteration r_gate = r_gate // 2 tree_level -= 1 return self.circuit.inverse() def _disentangle_qubit( self, children: "list[float]", parent: "list[float]", r_gate: int, tree_level: int, ): """Apply UCGate to disentangle qubit target""" bit_target = self.str_target[self.num_qubits - tree_level] old_mult, old_controls, target = self._define_mult(children, parent, tree_level) nc, mult = self._simplify(old_mult, tree_level) mult_controls = [x for x in old_controls if x not in nc] if self.preserve: self._preserve_previous(mult, mult_controls, r_gate, target) ucg = self._apply_ucg(mult, mult_controls, target) ucg.dont_carry = nc ucg.controls = mult_controls return bit_target, ucg def _simplify(self, mux, level): mux_cpy = mux.copy() dont_carry = [] if len(mux) > 1: n = self.num_qubits - level dont_carry = _repetition_search(mux, n, mux_cpy) new_mux = [matrix for matrix in mux_cpy if matrix is not None] return dont_carry, new_mux def _apply_diagonal( self, bit_target: str, parent: "list[float]", ucg: UCGate ): children = parent if bit_target == "1": diagonal = np.conj(ucg._get_diagonal())[ 1::2 ] # pylint: disable=protected-access else: diagonal = np.conj(ucg._get_diagonal())[ ::2 ] # pylint: disable=protected-access if ucg.dont_carry: ucg.controls.reverse() size_required = len(ucg.dont_carry) + len(ucg.controls) ctrl_qc = [self.num_qubits - 1 - x for x in ucg.controls] unitary_diagonal = np.diag(diagonal) qc = qiskit.QuantumCircuit(size_required) qc.unitary(unitary_diagonal, ctrl_qc) matrix = Operator(qc).to_matrix() diagonal = np.diag(matrix) children = children * diagonal return children @staticmethod def initialize(q_circuit, state, qubits=None, opt_params=None): gate = UCGEInitialize(state, opt_params=opt_params) if qubits is None: q_circuit.append(gate.definition, q_circuit.qubits) else: q_circuit.append(gate.definition, qubits)
https://github.com/qclib/qclib
qclib
# Copyright 2021 qclib project. # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # http://www.apache.org/licenses/LICENSE-2.0 # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """ https://arxiv.org/abs/2108.10182 """ import qiskit from qclib.state_preparation.util.tree_utils import children def output(angle_tree, output_qubits): """Define output qubits""" if angle_tree: output_qubits.insert(0, angle_tree.qubit) # qiskit little-endian if angle_tree.left: output(angle_tree.left, output_qubits) else: output(angle_tree.right, output_qubits) def _add_register(angle_tree, qubits, start_level): if angle_tree: angle_tree.qubit = qubits.pop(0) if angle_tree.level < start_level: _add_register(angle_tree.left, qubits, start_level) _add_register(angle_tree.right, qubits, start_level) else: if angle_tree.left: _add_register(angle_tree.left, qubits, start_level) else: _add_register(angle_tree.right, qubits, start_level) def add_register(circuit, angle_tree, start_level): """ Organize qubit registers, grouping by "output" and "ancilla" types. """ level = 0 level_nodes = [] nodes = [angle_tree] while len(nodes) > 0: # count nodes per level level_nodes.append(len(nodes)) nodes = children(nodes) level += 1 noutput = level # one output qubits per level nqubits = sum(level_nodes[:start_level]) # bottom-up qubits # top-down qubits: (number of sub-states) * (number of qubits per sub-state) nqubits += level_nodes[start_level] * (noutput - start_level) nancilla = nqubits - noutput output_register = qiskit.QuantumRegister(noutput, name="output") circuit.add_register(output_register) qubits = [*output_register[::-1]] if nancilla > 0: ancilla_register = qiskit.QuantumRegister(nancilla, name="ancilla") circuit.add_register(ancilla_register) qubits.extend([*ancilla_register[::-1]]) _add_register(angle_tree, qubits, start_level)
https://github.com/qclib/qclib
qclib
# Copyright 2021 qclib project. # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # http://www.apache.org/licenses/LICENSE-2.0 # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """ Tests for the baa.py module. """ from test.test_baa_lowrank import TestBaaLowRank import datetime import os from unittest import TestCase import numpy as np import qiskit from qiskit import QiskitError from qiskit.circuit.random import random_circuit from qclib.entanglement import geometric_entanglement, \ meyer_wallach_entanglement from qclib.state_preparation import BaaLowRankInitialize from qclib.state_preparation.lowrank import cnot_count as schmidt_cnots from qclib.state_preparation.util.baa import adaptive_approximation from qclib.util import get_state # pylint: disable=missing-function-docstring # pylint: disable=missing-class-docstring use_parallel = os.getenv('QLIB_TEST_PARALLEL', 'False') == 'True' def get_vector(e_lower: float, e_upper: float, num_qubits: int, start_depth_multiplier=1, measure='meyer_wallach'): entanglement = -1.0 if isinstance(start_depth_multiplier, int): multiplier = start_depth_multiplier elif isinstance(start_depth_multiplier, (tuple, list, np.ndarray)): assert len(start_depth_multiplier) > 1 multiplier = np.random.randint(start_depth_multiplier[0], start_depth_multiplier[1]) else: raise ValueError("start_depth_multiplier must be either an int or an array-like of int.") iteration = 0 entanglements = [] vector = np.ndarray(shape=(0,)) while e_lower > entanglement or entanglement > e_upper: np.random.seed() q_circuit: qiskit.QuantumCircuit = random_circuit(num_qubits, multiplier * num_qubits) vector = get_state(q_circuit) if measure == 'geometric': entanglement = geometric_entanglement(vector) elif measure == 'meyer_wallach': entanglement = meyer_wallach_entanglement(vector) else: raise ValueError(f'Entanglement Measure {measure} unknown.') iteration += 1 if iteration > 100: multiplier += 1 iteration = 0 if not use_parallel: print( f'{multiplier} ({np.min(entanglements):.4f}-{np.max(entanglements):.4f})', end='\n', flush=True ) entanglements = [] else: entanglements.append(entanglement) if not use_parallel: print('.', end='', flush=True) if not use_parallel: print( f'Final {multiplier} ({np.min(entanglements):.4f}-{np.max(entanglements):.4f})', end='\n', flush=True ) return vector, entanglement, multiplier * num_qubits def initialize_loss(fidelity_loss, state_vector=None, n_qubits=5, strategy='brute_force', use_low_rank=False): if state_vector is None: state_vector = np.random.rand(2**n_qubits) + np.random.rand(2**n_qubits) * 1j state_vector = state_vector / np.linalg.norm(state_vector) opt_params = {'max_fidelity_loss': fidelity_loss, 'strategy': strategy, 'use_low_rank': use_low_rank} circuit = BaaLowRankInitialize(state_vector, opt_params=opt_params).definition state = get_state(circuit) fidelity = TestBaaLowRank.fidelity(state_vector, state) node = adaptive_approximation(state_vector, max_fidelity_loss=fidelity_loss, strategy=strategy, use_low_rank=use_low_rank) try: basis_circuit = qiskit.transpile( circuit, basis_gates=['rx', 'ry', 'rz', 'cx'], optimization_level=0 ) cnots = len([d[0] for d in basis_circuit.data if d[0].name == 'cx']) depth = basis_circuit.depth() except QiskitError as ex: print(ex) return -1, -1, -1, node return cnots, depth, round(1 - fidelity, 4), node def execute_experiment(exp_idx, num_qubits, entanglement_bounds, max_fidelity_losses, return_state=False, start_depth_multiplier=1): print(f"Starting {exp_idx, num_qubits, entanglement_bounds, max_fidelity_losses}") # State Generation state_vector, entanglement, depth = get_vector( *entanglement_bounds, num_qubits, start_depth_multiplier=start_depth_multiplier, measure='geometric' ) mw_entanglement = meyer_wallach_entanglement(state_vector) geo_entanglement = geometric_entanglement(state_vector) cnots = schmidt_cnots(state_vector) if not use_parallel: print( f"Found state for entanglement bounds {entanglement} in {entanglement_bounds}. " f"State preparation needs {cnots}." ) # Benchmark against real Algorithm real_cnots_benchmark, real_depth_benchmark, real_fidelity_loss_benchmark, _ = initialize_loss( state_vector=state_vector, fidelity_loss=0.0, use_low_rank=False ) data_result = [] for max_fidelity_loss in max_fidelity_losses: for use_low_rank in [False, True]: for strategy in ['brute_force', 'greedy']: if not use_parallel: print( f"[{max_fidelity_loss}] {strategy.upper()} " f"{'With' if use_low_rank else 'No'} Low Rank Processing....", end='' ) start_time = datetime.datetime.now() real_cnots, real_depth, real_fidelity_loss, node = initialize_loss( state_vector=state_vector, fidelity_loss=max_fidelity_loss, use_low_rank=use_low_rank, strategy=strategy ) end_time = datetime.datetime.now() duration = (end_time - start_time).total_seconds() vector_data = list( zip(node.ranks, [list(v.shape) for v in node.vectors]) ) data = [ exp_idx, use_low_rank, strategy, num_qubits, depth, cnots, geo_entanglement, mw_entanglement, max_fidelity_loss, node.total_saved_cnots, node.total_fidelity_loss, vector_data, real_cnots, real_cnots_benchmark, real_depth, real_depth_benchmark, real_fidelity_loss, real_fidelity_loss_benchmark, duration ] if not use_parallel: print(f"in {duration} secs") data_result.append(data) # Experiment transcription data_result = np.asarray(data_result, dtype=object) print(f"Done {exp_idx, num_qubits, entanglement_bounds, max_fidelity_losses}") if return_state: return data_result, state_vector return data_result class TestBaa(TestCase): def test_node_state_vector(self): """ The method Node.state_vector() is an important function for analytics, but it was wrong. The ordering of the qubits must be taken into account. One example, visible to the human eye, was a probability distribution, the log-normal. For a max-fidelity-loss of 0.09, the ordering is computed with qubits [(1,), (3,), (0, 2, 4, 5, 6)] and partitions [None, None, (1, 4)]. The resulting state has an actual fidelity loss of about 0.0859261062108. The fix is to take the ordering into account. Here we test that for this example, the fidelity is actually correct. """ from qclib.state_preparation.util import baa # geometric measure = 0.11600417225836746 state = [0.07790067, 0.12411293, 0.10890448, 0.09848761, 0.05027826, 0.05027438, 0.03067742, 0.11846638, 0.09868947, 0.10711022, 0.01826258, 0.12535963, 0.11613662, 0.05865527, 0.05427737, 0.05451262, 0.07021045, 0.09220849, 0.08365776, 0.06869252, 0.09956702, 0.04754114, 0.0688004, 0.07704547, 0.08596225, 0.11279128, 0.05687908, 0.09127936, 0.09797266, 0.02743385, 0.09921588, 0.05256358, 0.03246542, 0.1239935, 0.12508287, 0.11444701, 0.07025331, 0.03978115, 0.10529168, 0.08444882, 0.04446722, 0.08957199, 0.02360472, 0.12138094, 0.06475262, 0.10360775, 0.07106703, 0.09179565, 0.09411756, 0.05472767, 0.12533861, 0.1120676, 0.12337869, 0.12040975, 0.0984252, 0.12221594, 0.03786564, 0.05635093, 0.02707025, 0.07260295, 0.07935725, 0.06630651, 0.11587787, 0.07602843, 0.06746749, 0.09377139, 0.04778426, 0.11400727, 0.03475503, 0.12645201, 0.11185863, 0.05674245, 0.00945899, 0.11494597, 0.10701826, 0.10868207, 0.11178804, 0.03463689, 0.07621068, 0.04332871, 0.11825607, 0.10049395, 0.07322158, 0.03209064, 0.0709839, 0.07258655, 0.10872672, 0.10163697, 0.11989633, 0.08747055, 0.04401971, 0.10750071, 0.11102557, 0.09536318, 0.11176607, 0.08944697, 0.09203053, 0.0832302, 0.02029422, 0.04181051, 0.02256621, 0.10154549, 0.07136789, 0.0907753, 0.12126382, 0.06355451, 0.08154299, 0.110643, 0.06088612, 0.03531675, 0.06851802, 0.05110969, 0.12273343, 0.11442741, 0.10130534, 0.11882721, 0.11411204, 0.05498105, 0.12025703, 0.09348119, 0.11437924, 0.12049476, 0.07178074, 0.04222706, 0.06077118, 0.08318802, 0.11512578, 0.1180934] node = baa.adaptive_approximation(state, use_low_rank=True, max_fidelity_loss=0.09, strategy='brute_force') fidelity = np.vdot(state, node.state_vector())**2 self.assertAlmostEqual(node.total_fidelity_loss, 1 - fidelity, places=4)
https://github.com/qclib/qclib
qclib
# Copyright 2021 qclib project. # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # http://www.apache.org/licenses/LICENSE-2.0 # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """ Tests for the baa_lowrank.py module. """ from unittest import TestCase import numpy as np from qiskit import ClassicalRegister, transpile from qiskit_aer import AerSimulator from qclib.util import get_state from qclib.state_preparation import BaaLowRankInitialize # pylint: disable=missing-function-docstring # pylint: disable=missing-class-docstring class TestBaaLowRank(TestCase): @staticmethod def fidelity(state1, state2): bra = np.conj(state1) ket = state2 return np.power(np.abs(bra.dot(ket)), 2) @staticmethod def get_counts(circuit): n_qubits = circuit.num_qubits classical_reg = ClassicalRegister(n_qubits) circuit.add_register(classical_reg) circuit.measure(list(range(n_qubits)), classical_reg) backend = AerSimulator() counts = backend.run( transpile(circuit, backend), shots=8192 ).result().get_counts() counts_with_zeros = {} for i in range(2**n_qubits): pattern = f'{i:0{n_qubits}b}' if pattern in counts: counts_with_zeros[pattern] = counts[pattern] else: counts_with_zeros[pattern] = 0.0 sum_values = sum(counts.values()) return [ value/sum_values for (key, value) in counts_with_zeros.items() ] def _test_initialize_loss(self, fidelity_loss, state_vector=None, n_qubits=5, strategy='brute_force', use_low_rank=False): if state_vector is None: state_vector = np.random.rand(2**n_qubits) + np.random.rand(2**n_qubits) * 1j state_vector = state_vector / np.linalg.norm(state_vector) opt_params = {'max_fidelity_loss':fidelity_loss, 'strategy' : strategy, 'use_low_rank':use_low_rank} circuit = BaaLowRankInitialize(state_vector, opt_params=opt_params).definition state = get_state(circuit) fidelity = TestBaaLowRank.fidelity(state_vector, state) self.assertTrue(round(fidelity,2)>=round(1-fidelity_loss,2)*0.99) def test_initialize_loss_brute_force(self): for loss in range(5, 15): self._test_initialize_loss(loss/100, n_qubits=5, strategy='brute_force') def test_initialize_loss_brute_force_low_rank(self): for loss in range(5, 15): self._test_initialize_loss(loss/100, n_qubits=5, strategy='brute_force', use_low_rank=True) def test_initialize_loss_greedy(self): for loss in range(5, 15): self._test_initialize_loss(loss/100, n_qubits=5, strategy='greedy') def test_initialize_loss_greedy_low_rank(self): for loss in range(5, 15): self._test_initialize_loss(loss/100, n_qubits=5, strategy='greedy', use_low_rank=True) def test_initialize_loss_fixed_n3(self): state_vector = [-0.33*1j,0,-0.44-0.44*1j,0.24+0.23*1j,0,0,0,0.62-0.01*1j] state_vector = state_vector/np.linalg.norm(state_vector) for loss in [0.1, 0.28, 0.9]: self._test_initialize_loss(loss, state_vector=state_vector, strategy='brute_force') self._test_initialize_loss(loss, state_vector=state_vector, strategy='greedy') def test_initialize_loss_fixed_n4(self): state_vector = [0.04214906+0.25870366j, 0.18263984+0.05596082j, 0.17202687+0.1843925j , 0.24972444+0.04666321j, 0.03311006+0.28233458j, 0.26680588+0.22211721j, 0.07205056+0.04556719j, 0.27982261+0.01626855j, 0.22908475+0.25461504j, 0.14290823+0.2425394j , 0.14213592+0.08282699j, 0.0068727 +0.03378424j, 0.2016483 +0.298073j , 0.07520782+0.0639856j , 0.01026576+0.07669651j, 0.31755857+0.09279232j] state_vector = state_vector/np.linalg.norm(state_vector) for loss in [0.1, 0.15, 0.18, 0.2]: self._test_initialize_loss(loss, state_vector=state_vector, strategy='brute_force') self._test_initialize_loss(loss, state_vector=state_vector, strategy='greedy') def test_initialize_loss_fixed_n5(self): state_vector = [0.17777766+0.10171662j, 0.19896424+0.10670792j, 0.07982054+0.19653055j, 0.18155708+0.05746777j, 0.04259147+0.17093567j, 0.21551328+0.08246133j, 0.09549255+0.1117806j , 0.20562749+0.12218064j, 0.16191832+0.01653411j, 0.12255337+0.14109365j, 0.20090638+0.11119666j, 0.19851901+0.04543331j, 0.06842539+0.16671467j, 0.03209685+0.16839388j, 0.01707365+0.20060943j, 0.03853768+0.08183117j, 0.00073591+0.10084589j, 0.09524694+0.18785593j, 0.06005853+0.06977443j, 0.01553849+0.05363906j, 0.10294799+0.12558734j, 0.20142903+0.06801796j, 0.05282011+0.20879126j, 0.11257846+0.20746226j, 0.17737416+0.03461382j, 0.01689154+0.06600272j, 0.06428148+0.06199636j, 0.1163249 +0.160533j , 0.14177201+0.10456823j, 0.03156739+0.04567818j, 0.02078566+0.02023752j, 0.18967059+0.03469463j] state_vector = state_vector/np.linalg.norm(state_vector) for loss in [0.1, 0.12, 0.14, 0.16, 0.2]: self._test_initialize_loss(loss, state_vector=state_vector, strategy='brute_force') self._test_initialize_loss(loss, state_vector=state_vector, strategy='greedy') def test_initialize_no_loss(self): state_vector = np.random.rand(32) + np.random.rand(32) * 1j state_vector = state_vector / np.linalg.norm(state_vector) circuit = BaaLowRankInitialize(state_vector).definition state = get_state(circuit) self.assertTrue(np.allclose(state_vector, state)) def test_initialize_ame(self): """ Test initialization of a absolutely maximally entangled state""" state_vector = [1, 1, 1, 1,1,-1,-1, 1, 1,-1,-1, 1, 1, 1,1,1, 1, 1,-1,-1,1,-1, 1,-1,-1, 1,-1, 1,-1,-1,1,1] state_vector = state_vector / np.linalg.norm(state_vector) circuit = BaaLowRankInitialize(state_vector).definition state = get_state(circuit) self.assertTrue(np.allclose(state_vector, state)) def test_measurement_no_loss(self): state_vector = np.random.rand(32) + np.random.rand(32) * 1j state_vector = state_vector / np.linalg.norm(state_vector) circuit = BaaLowRankInitialize(state_vector).definition state = TestBaaLowRank.get_counts(circuit) self.assertTrue(np.allclose( np.power(np.abs(state_vector),2), state, rtol=1e-01, atol=0.005)) def test_compare_strategies(self): fidelities1 = [] fidelities2 = [] for n_qubits in range(3,7): state_vector = np.random.rand(2**n_qubits) + np.random.rand(2**n_qubits) * 1j state_vector = state_vector / np.linalg.norm(state_vector) for loss in range(10, 20): opt_params = {'max_fidelity_loss' : loss/100, 'strategy' : 'brute_force'} circuit = BaaLowRankInitialize(state_vector, opt_params=opt_params).definition state = get_state(circuit) fidelity1 = TestBaaLowRank.fidelity(state_vector, state) opt_params = {'max_fidelity_loss' : loss/100, 'strategy' : 'greedy'} circuit = BaaLowRankInitialize(state_vector, opt_params=opt_params).definition state = get_state(circuit) fidelity2 = TestBaaLowRank.fidelity(state_vector, state) fidelities1.append(fidelity1) fidelities2.append(fidelity2) self.assertTrue(np.allclose(fidelities1, fidelities2, rtol=0.14, atol=0.0)) def test_large_state(self): # Builds a separable state. n_qubits = 16 state_vector1 = np.random.rand(2**(n_qubits//2)) + np.random.rand(2**(n_qubits//2)) * 1j state_vector1 = state_vector1 / np.linalg.norm(state_vector1) state_vector2 = np.random.rand(2**(n_qubits//2)) + np.random.rand(2**(n_qubits//2)) * 1j state_vector2 = state_vector2 / np.linalg.norm(state_vector2) state_vector = np.kron(state_vector1, state_vector2) opt_params = {'max_fidelity_loss' : 0.1, 'strategy' : 'brute_force'} circuit = BaaLowRankInitialize(state_vector, opt_params=opt_params).definition state = get_state(circuit) self.assertTrue(np.allclose(state_vector, state))
https://github.com/qclib/qclib
qclib
# Copyright 2021 qclib project. # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # http://www.apache.org/licenses/LICENSE-2.0 # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """ Test bidirectional state preparation """ from unittest import TestCase import numpy as np from qiskit import ClassicalRegister from qiskit_aer import AerSimulator from qclib.state_preparation import BdspInitialize from qclib.util import measurement backend = AerSimulator() SHOTS = 8192 class TestBdsp(TestCase): """ Testing bdsp """ @staticmethod def bdsp_experiment(state, split=None): """ Run bdsp experiment """ opt_params = {'split': split} circuit = BdspInitialize(state, opt_params=opt_params).definition n_qubits = int(np.log2(len(state))) classical_register = ClassicalRegister(n_qubits) circuit.add_register(classical_register) return measurement(circuit, n_qubits, classical_register, backend, SHOTS) def test_bottom_up(self): """ Testing bdsp """ vector = np.random.rand(16) + np.random.rand(16) * 1j vector = vector / np.linalg.norm(vector) state = TestBdsp.bdsp_experiment(vector, 1) self.assertTrue(np.allclose(np.power(np.abs(vector), 2), state, rtol=1e-01, atol=0.005)) def test_top_down(self): """ Testing bdsp """ vector = np.random.rand(16) + np.random.rand(16) * 1j vector = vector / np.linalg.norm(vector) state = TestBdsp.bdsp_experiment(vector, int(np.log2(len(vector)))) self.assertTrue(np.allclose(np.power(np.abs(vector), 2), state, rtol=1e-01, atol=0.005)) def test_sublinear(self): """ Testing bdsp """ vector = np.random.rand(16) + np.random.rand(16) * 1j vector = vector / np.linalg.norm(vector) state = TestBdsp.bdsp_experiment(vector) self.assertTrue(np.allclose(np.power(np.abs(vector), 2), state, rtol=1e-01, atol=0.005))
https://github.com/qclib/qclib
qclib
# Copyright 2021 qclib project. # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # http://www.apache.org/licenses/LICENSE-2.0 # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """ Test blackbox state preparation """ from unittest import TestCase import numpy as np from qiskit import QuantumCircuit from qclib.state_preparation.blackbox import BlackBoxInitialize from qclib.util import get_state class TestBlackbox(TestCase): """ Test blackbox state preparation """ def test_blackbox(self): """ Run blackbox state preparation """ initialize = BlackBoxInitialize.initialize state = np.random.rand(16) - 0.5 + (np.random.rand(16) - 0.5) * 1j state = state / np.linalg.norm(state) q_circuit = QuantumCircuit(5) initialize(q_circuit, state.tolist()) out = get_state(q_circuit) out = out.reshape((len(out)//2, 2)) out = out[:, 0] self.assertTrue(np.allclose(state, out, atol=0.02))
https://github.com/qclib/qclib
qclib
# Copyright 2021 qclib project. # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # http://www.apache.org/licenses/LICENSE-2.0 # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """ Test divide-and-conquer state preparation """ from unittest import TestCase import numpy as np from qiskit import ClassicalRegister from qiskit_aer import AerSimulator from qclib.state_preparation import DcspInitialize from qclib.util import measurement backend = AerSimulator() SHOTS = 8192 class TestInitialize(TestCase): """ Testing divide-and-conquer state preparation """ @staticmethod def dcsp_experiment(state): """ Run divide-and-conquer state preparation """ circuit = DcspInitialize(state).definition n_qubits = int(np.log2(len(state))) classical_register = ClassicalRegister(n_qubits) circuit.add_register(classical_register) return measurement(circuit, n_qubits, classical_register, backend, SHOTS) def test_dcsp(self): """ Testing divide-and-conquer state preparation """ vector = np.random.rand(16) + np.random.rand(16) * 1j vector = vector / np.linalg.norm(vector) state = TestInitialize.dcsp_experiment(vector) self.assertTrue(np.allclose(np.power(np.abs(vector), 2), state, rtol=1e-01, atol=0.005))
https://github.com/qclib/qclib
qclib
# This code is part of Qiskit. # # (C) Copyright IBM 2019. # # This code is licensed under the Apache License, Version 2.0. You may # obtain a copy of this license in the LICENSE.txt file in the root directory # of this source tree or at http://www.apache.org/licenses/LICENSE-2.0. # # Any modifications or derivative works of this code must retain this # copyright notice, and modified files need to carry a notice indicating # that they have been altered from the originals. """Isometry tests.""" import unittest import numpy as np from ddt import ddt, data from qiskit.quantum_info.random import random_unitary from qiskit import BasicAer from qiskit import QuantumCircuit from qiskit import QuantumRegister from qiskit import execute from qiskit.test import QiskitTestCase from qiskit.compiler import transpile from qiskit.quantum_info import Operator from qiskit.extensions.quantum_initializer.isometry import Isometry @ddt class TestIsometry(QiskitTestCase): """Qiskit isometry tests.""" @data( np.eye(2, 2), random_unitary(2, seed=868540).data, np.eye(4, 4), random_unitary(4, seed=16785).data[:, 0], np.eye(4, 4)[:, 0:2], random_unitary(4, seed=660477).data, np.eye(4, 4)[:, np.random.RandomState(seed=719010).permutation(4)][:, 0:2], np.eye(8, 8)[:, np.random.RandomState(seed=544326).permutation(8)], random_unitary(8, seed=247924).data[:, 0:4], random_unitary(8, seed=765720).data, random_unitary(16, seed=278663).data, random_unitary(16, seed=406498).data[:, 0:8], ) def test_isometry(self, iso): """Tests for the decomposition of isometries from m to n qubits""" if len(iso.shape) == 1: iso = iso.reshape((len(iso), 1)) num_q_output = int(np.log2(iso.shape[0])) num_q_input = int(np.log2(iso.shape[1])) q = QuantumRegister(num_q_output) qc = QuantumCircuit(q) qc.isometry(iso, q[:num_q_input], q[num_q_input:]) # Verify the circuit can be decomposed self.assertIsInstance(qc.decompose(), QuantumCircuit) # Decompose the gate qc = transpile(qc, basis_gates=["u1", "u3", "u2", "cx", "id"]) # Simulate the decomposed gate simulator = BasicAer.get_backend("unitary_simulator") result = execute(qc, simulator).result() unitary = result.get_unitary(qc) iso_from_circuit = unitary[::, 0 : 2**num_q_input] iso_desired = iso self.assertTrue(np.allclose(iso_from_circuit, iso_desired)) @data( np.eye(2, 2), random_unitary(2, seed=99506).data, np.eye(4, 4), random_unitary(4, seed=673459).data[:, 0], np.eye(4, 4)[:, 0:2], random_unitary(4, seed=124090).data, np.eye(4, 4)[:, np.random.RandomState(seed=889848).permutation(4)][:, 0:2], np.eye(8, 8)[:, np.random.RandomState(seed=94795).permutation(8)], random_unitary(8, seed=986292).data[:, 0:4], random_unitary(8, seed=632121).data, random_unitary(16, seed=623107).data, random_unitary(16, seed=889326).data[:, 0:8], ) def test_isometry_tolerance(self, iso): """Tests for the decomposition of isometries from m to n qubits with a custom tolerance""" if len(iso.shape) == 1: iso = iso.reshape((len(iso), 1)) num_q_output = int(np.log2(iso.shape[0])) num_q_input = int(np.log2(iso.shape[1])) q = QuantumRegister(num_q_output) qc = QuantumCircuit(q) # Compute isometry with custom tolerance qc.isometry(iso, q[:num_q_input], q[num_q_input:], epsilon=1e-3) # Verify the circuit can be decomposed self.assertIsInstance(qc.decompose(), QuantumCircuit) # Decompose the gate qc = transpile(qc, basis_gates=["u1", "u3", "u2", "cx", "id"]) # Simulate the decomposed gate simulator = BasicAer.get_backend("unitary_simulator") result = execute(qc, simulator).result() unitary = result.get_unitary(qc) iso_from_circuit = unitary[::, 0 : 2**num_q_input] self.assertTrue(np.allclose(iso_from_circuit, iso)) @data( np.eye(2, 2), random_unitary(2, seed=272225).data, np.eye(4, 4), random_unitary(4, seed=592640).data[:, 0], np.eye(4, 4)[:, 0:2], random_unitary(4, seed=714210).data, np.eye(4, 4)[:, np.random.RandomState(seed=719934).permutation(4)][:, 0:2], np.eye(8, 8)[:, np.random.RandomState(seed=284469).permutation(8)], random_unitary(8, seed=656745).data[:, 0:4], random_unitary(8, seed=583813).data, random_unitary(16, seed=101363).data, random_unitary(16, seed=583429).data[:, 0:8], ) def test_isometry_inverse(self, iso): """Tests for the inverse of isometries from m to n qubits""" if len(iso.shape) == 1: iso = iso.reshape((len(iso), 1)) num_q_output = int(np.log2(iso.shape[0])) q = QuantumRegister(num_q_output) qc = QuantumCircuit(q) qc.append(Isometry(iso, 0, 0), q) qc.append(Isometry(iso, 0, 0).inverse(), q) result = Operator(qc) np.testing.assert_array_almost_equal(result.data, np.identity(result.dim[0])) if __name__ == "__main__": unittest.main()
https://github.com/qclib/qclib
qclib
# Copyright 2021 qclib project. # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # http://www.apache.org/licenses/LICENSE-2.0 # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """ Tests for the lowrank.py module. """ from unittest import TestCase from itertools import combinations import numpy as np from qiskit import QuantumCircuit, ClassicalRegister, transpile from qiskit_aer import AerSimulator from qclib.state_preparation import LowRankInitialize from qclib.state_preparation.lowrank import cnot_count from qclib.util import get_state # pylint: disable=missing-function-docstring # pylint: disable=missing-class-docstring class TestLowRank(TestCase): @staticmethod def mae(state, ideal): """ Mean Absolute Error """ return np.sum(np.abs(state-ideal))/len(ideal) @staticmethod def get_counts(circuit): n_qubits = circuit.num_qubits classical_reg = ClassicalRegister(n_qubits) circuit.add_register(classical_reg) circuit.measure(list(range(n_qubits)), classical_reg) backend = AerSimulator() counts = backend.run( transpile(circuit, backend), shots=8192 ).result().get_counts() counts_with_zeros = {} for i in range(2**n_qubits): pattern = f'{i:0{n_qubits}b}' if pattern in counts: counts_with_zeros[pattern] = counts[pattern] else: counts_with_zeros[pattern] = 0.0 sum_values = sum(counts.values()) return [value/sum_values for (key, value) in counts_with_zeros.items()] def _test_initialize_mae(self, rank=0, max_mae=10**-15): n_qubits = 5 state_vector = np.random.rand(2**n_qubits) + np.random.rand(2**n_qubits) * 1j state_vector = state_vector / np.linalg.norm(state_vector) qubits = list(range(n_qubits)) partitions = combinations(qubits, (n_qubits+1)//2) for partition in partitions: circuit = QuantumCircuit(n_qubits) opt_params = {'lr': rank, 'partition': partition} LowRankInitialize.initialize(circuit, state_vector, opt_params=opt_params) state = get_state(circuit) self.assertTrue(TestLowRank.mae(state, state_vector) < max_mae) def _test_initialize_full_rank(self, n_qubits): state_vector = np.random.rand(2**n_qubits) + np.random.rand(2**n_qubits) * 1j state_vector = state_vector / np.linalg.norm(state_vector) qubits = list(range(n_qubits)) partitions = combinations(qubits, (n_qubits+1)//2) for partition in partitions: circuit = QuantumCircuit(n_qubits) opt_params = {'partition': partition} LowRankInitialize.initialize(circuit, state_vector, opt_params=opt_params) state = get_state(circuit) self.assertTrue(np.allclose(state_vector, state)) def test_initialize_full_rank_6(self): self._test_initialize_full_rank(6) def test_initialize_rank_5(self): state_vector = np.random.rand(32) + np.random.rand(32) * 1j state_vector = state_vector / np.linalg.norm(state_vector) circuit = QuantumCircuit(5) LowRankInitialize.initialize(circuit, state_vector, opt_params={'lr': 5}) state = get_state(circuit) self.assertTrue(np.allclose(state_vector, state)) def test_initialize_real(self): n_qubits = 5 state_vector = np.random.rand(2**n_qubits) state_vector = state_vector / np.linalg.norm(state_vector) qubits = list(range(n_qubits)) partitions = combinations(qubits, (n_qubits+1)//2) for partition in partitions: circuit = QuantumCircuit(n_qubits) opt_params = {'partition': partition} LowRankInitialize.initialize(circuit, state_vector, opt_params=opt_params) state = get_state(circuit) self.assertTrue(np.allclose(state_vector, state)) def test_initialize_rank_4(self): self._test_initialize_mae(4, max_mae=10**-10) def test_initialize_rank_3(self): self._test_initialize_mae(3, max_mae=0.04) def test_initialize_rank_2(self): self._test_initialize_mae(2, max_mae=0.06) def test_initialize_rank_1(self): self._test_initialize_mae(1, max_mae=0.09) def test_cnot_count(self): n_qubits = 7 state_vector = np.random.rand(2**n_qubits) + np.random.rand(2**n_qubits) * 1j state_vector = state_vector / np.linalg.norm(state_vector) circuit = QuantumCircuit(n_qubits) LowRankInitialize.initialize(circuit, state_vector) transpiled_circuit = transpile(circuit, basis_gates=['u', 'cx'], optimization_level=0) n_cx = transpiled_circuit.count_ops()['cx'] self.assertTrue(cnot_count(state_vector) == n_cx) def test_cnot_count_rank_1(self): # Builds a rank 1 state. state_vector = [1] for _ in range(5): vec = np.random.rand(2) + np.random.rand(2) * 1j vec = vec / np.linalg.norm(vec) state_vector = np.kron(state_vector, vec) circuit = QuantumCircuit(5) LowRankInitialize.initialize(circuit, state_vector) transpiled_circuit = transpile(circuit, basis_gates=['u', 'cx'], optimization_level=0) self.assertTrue('cx' not in transpiled_circuit.count_ops()) def test_inverse(self): n_qubits = 2 state_vector = np.random.rand(2**n_qubits) + np.random.rand(2**n_qubits) * 1j state_vector = state_vector / np.linalg.norm(state_vector) gate = LowRankInitialize(state_vector) circuit = QuantumCircuit(n_qubits) circuit.append(gate, list(range(n_qubits))) circuit.append(gate.inverse(), list(range(n_qubits))) state = get_state(circuit) zero_state = [1]+[0]*(2**n_qubits-1) self.assertTrue(np.allclose(zero_state, state)) def test_large_state(self): # Builds a separable state. n_qubits = 16 state_vector1 = np.random.rand(2**(n_qubits//2)) + np.random.rand(2**(n_qubits//2)) * 1j state_vector1 = state_vector1 / np.linalg.norm(state_vector1) state_vector2 = np.random.rand(2**(n_qubits//2)) + np.random.rand(2**(n_qubits//2)) * 1j state_vector2 = state_vector2 / np.linalg.norm(state_vector2) state_vector = np.kron(state_vector1, state_vector2) circuit = QuantumCircuit(n_qubits) LowRankInitialize.initialize(circuit, state_vector, opt_params={'lr': 1, 'svd':'auto'}) state = get_state(circuit) self.assertTrue(np.allclose(state_vector, state))
https://github.com/qclib/qclib
qclib
# Copyright 2021 qclib project. # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # http://www.apache.org/licenses/LICENSE-2.0 # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """ Tests for the merge.py module. """ import unittest import numpy as np from qiskit import QuantumCircuit from qclib.state_preparation import MergeInitialize from qclib.util import get_state, build_state_dict from qiskit import transpile from scipy.sparse import random # pylint: disable=missing-function-docstring # pylint: disable=missing-class-docstring class TestMergeInitialize(unittest.TestCase): """ Teste MergeInitialize Gate""" def test_two_states_uniform(self): state_vector = 1 / np.sqrt(2) * np.array([1, 0, 0, 0, 0, 1, 0, 0]) state_dict = build_state_dict(state_vector) initialize = MergeInitialize.initialize circ = QuantumCircuit(3) initialize(circ, state_dict) state = get_state(circ) self.assertTrue(np.allclose(state_vector, state)) def test_three_states_superposition(self): state_vector = 1 / np.sqrt(168) * np.array([0, 2, 0, 0, 8, 0, 0, 10]) state_dict = build_state_dict(state_vector) initialize = MergeInitialize.initialize circ = QuantumCircuit(3) initialize(circ, state_dict) state = get_state(circ) self.assertTrue(np.allclose(state_vector, state)) def test_three_states_uniform_superposition(self): state_vector = 1 / np.sqrt(3) * np.array([0, 1, 0, 0, 1, 0, 0, 1]) state_dict = build_state_dict(state_vector) initialize = MergeInitialize.initialize circ = QuantumCircuit(3) initialize(circ, state_dict) state = get_state(circ) self.assertTrue(np.allclose(state_vector, state)) def test_three_states_superposition_with_complex_features(self): state_vector = np.array([0, complex(np.sqrt(0.1), np.sqrt(0.1)), 0, 0, complex(np.sqrt(0.1), np.sqrt(0.2)), 0, 0, np.sqrt(0.5)]) state_dict = build_state_dict(state_vector) initialize = MergeInitialize.initialize circ = QuantumCircuit(3) initialize(circ, state_dict) state = get_state(circ) self.assertTrue(np.allclose(state_vector, state)) def test_raises_error_input_not_dict(self): initialize = MergeInitialize.initialize circ = QuantumCircuit() with self.assertRaises(Exception): initialize(circ, []) def test_8qb_sparse(self): state_dict = { '01100000': 0.11496980229422502, '10010000': 0.2012068017595738, '11110000': 0.2552406117427385, '11001000': 0.24483730989689545, '00011000': 0.08064396530053637, '01111000': 0.06609205232425505, '10000100': 0.2567251902135311, '00100010': 0.279179786457501, '11110010': 0.14972323818181424, '00000110': 0.054570286103576615, '10101110': 0.1953959409811345, '00011110': 0.2476316976925628, '00111110': 0.2460713287965397, '00010001': 0.2880964575493704, '00011001': 0.11697558413298771, '11100101': 0.15657582325155645, '00101101': 0.05059343291713247, '10011101': 0.21260965910383026, '11100011': 0.16144719592639006, '01110011': 0.24224885089395568, '10011011': 0.07542653172823867, '01111011': 0.2840232568261471, '00100111': 0.2719803407586484, '01100111': 0.14940066988379283, '11010111': 0.2025293455684502, '01001111': 0.06045929196877916 } initialize = MergeInitialize.initialize qc = QuantumCircuit(8) initialize(qc, state_dict) t_circuit = transpile(qc, basis_gates=['cx', 'u']) def test_several_qubit_sizes(self): for n_qubits in range(4, 12): state_vector = random(1, 2 ** n_qubits, density=0.1, random_state=42).A.reshape(-1) state_vector = state_vector / np.linalg.norm(state_vector) state_dict = build_state_dict(state_vector) initialize = MergeInitialize.initialize qc = QuantumCircuit(n_qubits) initialize(qc, state_dict) state = get_state(qc) self.assertTrue(np.allclose(state_vector, state))
https://github.com/qclib/qclib
qclib
# Copyright 2021 qclib project. # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # http://www.apache.org/licenses/LICENSE-2.0 # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """ Test Probabilistic Quantum Memory""" from unittest import TestCase import numpy as np from qiskit import ClassicalRegister, QuantumRegister, QuantumCircuit from qclib.memory.pqm import initialize as pqm from qclib.state_preparation import LowRankInitialize from qclib.util import get_counts class TestPQM(TestCase): """ Testing qclib.memory.pqm""" @staticmethod def _run_pqm(is_classical_pattern): """ run PQM with classical or quantum input""" # pqm memory data data = [0, 1, 0, 0] data = data / np.linalg.norm(data) # initialize quantum registers memory = QuantumRegister(2, 'm') aux = QuantumRegister(1, 'c') output = ClassicalRegister(1) circ = QuantumCircuit(memory, aux, output) # initialize data init_gate = LowRankInitialize(data) circ.append(init_gate, memory) # initialize input pattern bin_input = [1, 0] if is_classical_pattern: # run pqm recovery algorithm pqm(circ, bin_input, memory, aux, is_classical_pattern=True) else: q_bin_input = QuantumRegister(2) circ.add_register(q_bin_input) # Pattern basis encoding for k, bit in enumerate(bin_input): if bit == 1: circ.x(q_bin_input[k]) # run pqm recovery algorithm pqm(circ, q_bin_input, memory, aux, is_classical_pattern=False) # measure output and verify results circ.measure(aux, output) # pylint: disable=maybe-no-member counts = get_counts(circ) return counts def test_classical_input(self): """ Testing PQM with classical input """ # classical input pattern counts = TestPQM._run_pqm(True) self.assertTrue(counts['0'] / 1024 == 1) def test_quantum_input(self): """ Testing PQM with quantum input """ # quantum input pattern counts = TestPQM._run_pqm(False) self.assertTrue(counts['0'] / 1024 == 1)
https://github.com/qclib/qclib
qclib
# Copyright 2021 qclib project. # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # http://www.apache.org/licenses/LICENSE-2.0 # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """ Tests for the svd.py module. """ from unittest import TestCase import numpy as np from qiskit import QuantumCircuit from qclib.state_preparation.svd import SVDInitialize from qclib.util import get_state # pylint: disable=missing-function-docstring # pylint: disable=missing-class-docstring class TestSVDInitialize(TestCase): def _test_initialize(self, n_qubits): state_vector = np.random.rand(2**n_qubits) + np.random.rand(2**n_qubits) * 1j state_vector = state_vector / np.linalg.norm(state_vector) circuit = QuantumCircuit(n_qubits) SVDInitialize.initialize(circuit, state_vector) state = get_state(circuit) self.assertTrue(np.allclose(state_vector, state)) def test_initialize_6(self): self._test_initialize(6) def test_initialize_5(self): self._test_initialize(5)
https://github.com/qclib/qclib
qclib
# Copyright 2021 qclib project. # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # http://www.apache.org/licenses/LICENSE-2.0 # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """ Tests for the topdown.py module. """ from unittest import TestCase import numpy as np from qiskit import ClassicalRegister, transpile from qiskit_aer import AerSimulator from qclib.state_preparation import TopDownInitialize from qclib.util import get_state, measurement # pylint: disable=missing-function-docstring # pylint: disable=missing-class-docstring backend = AerSimulator() SHOTS = 8192 class TestTopDown(TestCase): @staticmethod def measurement(circuit, n_qubits, classical_reg): circuit.measure(list(range(n_qubits)), classical_reg) job = backend.run( transpile(circuit, backend), shots=SHOTS, optimization_level=3 ) counts = job.result().get_counts(circuit) sum_values = sum(counts.values()) counts2 = {} for i in range(2**n_qubits): pattern = f'{i:0{n_qubits}b}' if pattern in counts: counts2[pattern] = counts[pattern] else: counts2[pattern] = 0.0 return [ value/sum_values for (key, value) in counts2.items() ] @staticmethod def topdown_experiment(state): circuit = TopDownInitialize(state).definition n_qubits = int(np.log2(len(state))) classical_reg = ClassicalRegister(n_qubits) circuit.add_register(classical_reg) return measurement(circuit, n_qubits, classical_reg, backend, SHOTS) def test_topdown_state_real(self): state_vector = np.random.rand(32) state_vector = state_vector / np.linalg.norm(state_vector) circuit = TopDownInitialize(state_vector).definition state = get_state(circuit) self.assertTrue(np.allclose(state_vector, state)) def test_topdown_state_complex(self): state_vector = np.random.rand(32) + np.random.rand(32) * 1j state_vector = state_vector / np.linalg.norm(state_vector) circuit = TopDownInitialize(state_vector).definition state = get_state(circuit) self.assertTrue(np.allclose(state_vector, state)) def test_topdown_measure(self): state_vector = np.random.rand(32) + np.random.rand(32) * 1j state_vector = state_vector / np.linalg.norm(state_vector) state = TestTopDown.topdown_experiment(state_vector) self.assertTrue(np.allclose(np.power(np.abs(state_vector), 2), state, rtol=1e-01, atol=0.005)) def test_topdown_fixed_state(self): state_vector = [ 0, np.sqrt(2 / 8) * np.exp(-1.0j * np.pi / 7), np.sqrt(3 / 8) * np.exp(-1.0j * np.pi / 3), 0, 0, 0, np.sqrt(3 / 8), 0 ] circuit = TopDownInitialize(state_vector).definition state = get_state(circuit) self.assertTrue(np.allclose(state_vector, state))
https://github.com/qclib/qclib
qclib
# Copyright 2021 qclib project. # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # http://www.apache.org/licenses/LICENSE-2.0 # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """ Tests for the bergholm state preparation """ from unittest import TestCase import random import numpy as np from qiskit import QuantumCircuit, transpile from qiskit.quantum_info import Statevector from qclib.state_preparation import UCGInitialize from qclib.state_preparation import UCGEInitialize class TestUCGEInitialize(TestCase): """Test UCGEInitialize""" @staticmethod def _logical_swap(num_qubits, input_vector, new_order): qubit_shape = [2] * num_qubits reshaped_state = input_vector.reshape(qubit_shape) swapped_vector = np.moveaxis(reshaped_state, new_order, range(len(new_order))).reshape(-1,) return swapped_vector def _test_compare_ucg_bipartition(self, num_qubits, input_vector1, input_vector2): qubit_order = list(range(num_qubits)) random.shuffle(qubit_order) input_vector = np.kron(input_vector1, input_vector2) input_vector = self._logical_swap(num_qubits, input_vector, qubit_order) circuit1 = QuantumCircuit(num_qubits) UCGInitialize.initialize(circuit1, input_vector) state1 = Statevector(circuit1) circuit2 = QuantumCircuit(num_qubits) UCGEInitialize.initialize(circuit2, input_vector) state2 = Statevector(circuit2) circuit1_tranpiled = transpile(circuit1, basis_gates=['u', 'cx']) circuit2_transpiled = transpile(circuit2, basis_gates=['u', 'cx']) self.assertTrue(np.allclose(state1, state2)) self.assertTrue(circuit1_tranpiled.depth() >= circuit2_transpiled.depth()) def test_compare_ucg_bipartition_real(self): num_qubits = 8 input_vector1 = np.random.rand(2 ** (num_qubits//2)) input_vector1 = input_vector1 / np.linalg.norm(input_vector1) input_vector2 = np.random.rand(2 ** (num_qubits//2)) input_vector2 = input_vector2 / np.linalg.norm(input_vector2) self._test_compare_ucg_bipartition(num_qubits, input_vector1, input_vector2) def test_compare_ucg_bipartition_complex(self): num_qubits = 8 real_part = np.random.rand(2 ** (num_qubits//2)) imag_part = np.random.rand(2 ** (num_qubits//2)) input_vector1 = real_part + 1j * imag_part input_vector1 = input_vector1 / np.linalg.norm(input_vector1) real_part = np.random.rand(2 ** (num_qubits//2)) imag_part = np.random.rand(2 ** (num_qubits//2)) input_vector2 = real_part + 1j * imag_part input_vector2 = input_vector2 / np.linalg.norm(input_vector2) self._test_compare_ucg_bipartition(num_qubits, input_vector1, input_vector2)
https://github.com/qclib/qclib
qclib
# Copyright 2021 qclib project. # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # http://www.apache.org/licenses/LICENSE-2.0 # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """ Tests for the bergholm state preparation """ from unittest import TestCase import numpy as np from qiskit import QuantumCircuit from qclib.state_preparation import UCGInitialize from qclib.util import get_state class TestUCGInitialize(TestCase): """Test UCGInitialize""" def _test_ucg(self, n_qubits): state = np.random.rand(2 ** n_qubits) + np.random.rand(2 ** n_qubits) * 1j state = state / np.linalg.norm(state) for target_state in range(2 ** n_qubits): gate = UCGInitialize(state.tolist(), opt_params={ "target_state": target_state } ).definition circuit = QuantumCircuit(n_qubits) for j, bit in enumerate(f'{target_state:0{n_qubits}b}'[::-1]): if bit == '1': circuit.x(j) circuit.append(gate, circuit.qubits) output_state = get_state(circuit) self.assertTrue(np.allclose(state, output_state)) def _test_ucg_preserve(self, n_qubits): state = np.random.rand(2 ** n_qubits) + np.random.rand(2 ** n_qubits) * 1j for target_state in range(1, 2 ** n_qubits): state[target_state - 1] = 0 state = state / np.linalg.norm(state) gate = UCGInitialize(state.tolist(), opt_params={ "target_state": target_state, "preserve_previous": True } ).definition circuit = QuantumCircuit(n_qubits) for j, bit in enumerate(f'{target_state:0{n_qubits}b}'[::-1]): if bit == '1': circuit.x(j) circuit.append(gate, circuit.qubits) output_state = get_state(circuit) self.assertTrue(np.allclose(output_state, state)) def test_ucg(self): """Test UCGInitialize""" for n_qubits in range(3, 5): self._test_ucg(n_qubits) def test_ucg_preserve(self): """Test UCGInitialize with `preserve_previous`""" for n_qubits in range(3, 5): self._test_ucg_preserve(n_qubits) def test_real(self): """Test UCGInitialize with four qubits and index 10""" n_qubits = 4 state = np.random.rand(2 ** n_qubits) state[0] = 0 state[1] = 0 state = state / np.linalg.norm(state) initialize = UCGInitialize.initialize circuit = QuantumCircuit(n_qubits) circuit.x(1) circuit.x(3) initialize(circuit, state.tolist(), opt_params={"target_state": 10}) output_state = get_state(circuit) self.assertTrue(np.allclose(output_state, state))
https://github.com/qclib/qclib
qclib
# This code is part of Qiskit. # # (C) Copyright IBM 2017, 2019. # # This code is licensed under the Apache License, Version 2.0. You may # obtain a copy of this license in the LICENSE.txt file in the root directory # of this source tree or at http://www.apache.org/licenses/LICENSE-2.0. # # Any modifications or derivative works of this code must retain this # copyright notice, and modified files need to carry a notice indicating # that they have been altered from the originals. """UnitaryGate tests""" import json import numpy from numpy.testing import assert_allclose import qiskit from qiskit.extensions.unitary import UnitaryGate from qiskit.test import QiskitTestCase from qiskit import QuantumRegister, ClassicalRegister, QuantumCircuit from qiskit.transpiler import PassManager from qiskit.converters import circuit_to_dag, dag_to_circuit from qiskit.quantum_info.random import random_unitary from qiskit.quantum_info.operators import Operator from qiskit.transpiler.passes import CXCancellation class TestUnitaryGate(QiskitTestCase): """Tests for the Unitary class.""" def test_set_matrix(self): """Test instantiation""" try: UnitaryGate([[0, 1], [1, 0]]) # pylint: disable=broad-except except Exception as err: self.fail(f"unexpected exception in init of Unitary: {err}") def test_set_matrix_raises(self): """test non-unitary""" try: UnitaryGate([[1, 1], [1, 0]]) # pylint: disable=broad-except except Exception: pass else: self.fail("setting Unitary with non-unitary did not raise") def test_set_init_with_unitary(self): """test instantiation of new unitary with another one (copy)""" uni1 = UnitaryGate([[0, 1], [1, 0]]) uni2 = UnitaryGate(uni1) self.assertEqual(uni1, uni2) self.assertFalse(uni1 is uni2) def test_conjugate(self): """test conjugate""" ymat = numpy.array([[0, -1j], [1j, 0]]) uni = UnitaryGate([[0, 1j], [-1j, 0]]) self.assertTrue(numpy.array_equal(uni.conjugate().to_matrix(), ymat)) def test_adjoint(self): """test adjoint operation""" uni = UnitaryGate([[0, 1j], [-1j, 0]]) self.assertTrue(numpy.array_equal(uni.adjoint().to_matrix(), uni.to_matrix())) class TestUnitaryCircuit(QiskitTestCase): """Matrix gate circuit tests.""" def test_1q_unitary(self): """test 1 qubit unitary matrix""" qr = QuantumRegister(1) cr = ClassicalRegister(1) qc = QuantumCircuit(qr, cr) matrix = numpy.array([[1, 0], [0, 1]]) qc.x(qr[0]) qc.append(UnitaryGate(matrix), [qr[0]]) # test of qasm output self.log.info(qc.qasm()) # test of text drawer self.log.info(qc) dag = circuit_to_dag(qc) dag_nodes = dag.named_nodes("unitary") self.assertTrue(len(dag_nodes) == 1) dnode = dag_nodes[0] self.assertIsInstance(dnode.op, UnitaryGate) self.assertEqual(dnode.qargs, (qr[0],)) assert_allclose(dnode.op.to_matrix(), matrix) def test_2q_unitary(self): """test 2 qubit unitary matrix""" qr = QuantumRegister(2) cr = ClassicalRegister(2) qc = QuantumCircuit(qr, cr) sigmax = numpy.array([[0, 1], [1, 0]]) sigmay = numpy.array([[0, -1j], [1j, 0]]) matrix = numpy.kron(sigmax, sigmay) qc.x(qr[0]) uni2q = UnitaryGate(matrix) qc.append(uni2q, [qr[0], qr[1]]) passman = PassManager() passman.append(CXCancellation()) qc2 = passman.run(qc) # test of qasm output self.log.info(qc2.qasm()) # test of text drawer self.log.info(qc2) dag = circuit_to_dag(qc) nodes = dag.two_qubit_ops() self.assertEqual(len(nodes), 1) dnode = nodes[0] self.assertIsInstance(dnode.op, UnitaryGate) self.assertEqual(dnode.qargs, (qr[0], qr[1])) assert_allclose(dnode.op.to_matrix(), matrix) qc3 = dag_to_circuit(dag) self.assertEqual(qc2, qc3) def test_3q_unitary(self): """test 3 qubit unitary matrix on non-consecutive bits""" qr = QuantumRegister(4) qc = QuantumCircuit(qr) sigmax = numpy.array([[0, 1], [1, 0]]) sigmay = numpy.array([[0, -1j], [1j, 0]]) matrix = numpy.kron(sigmay, numpy.kron(sigmax, sigmay)) qc.x(qr[0]) uni3q = UnitaryGate(matrix) qc.append(uni3q, [qr[0], qr[1], qr[3]]) qc.cx(qr[3], qr[2]) # test of text drawer self.log.info(qc) dag = circuit_to_dag(qc) nodes = dag.multi_qubit_ops() self.assertEqual(len(nodes), 1) dnode = nodes[0] self.assertIsInstance(dnode.op, UnitaryGate) self.assertEqual(dnode.qargs, (qr[0], qr[1], qr[3])) assert_allclose(dnode.op.to_matrix(), matrix) def test_1q_unitary_int_qargs(self): """test single qubit unitary matrix with 'int' and 'list of ints' qubits argument""" sigmax = numpy.array([[0, 1], [1, 0]]) sigmaz = numpy.array([[1, 0], [0, -1]]) # new syntax qr = QuantumRegister(2) qc = QuantumCircuit(qr) qc.unitary(sigmax, 0) qc.unitary(sigmax, qr[1]) qc.unitary(sigmaz, [0, 1]) # expected circuit qc_target = QuantumCircuit(qr) qc_target.append(UnitaryGate(sigmax), [0]) qc_target.append(UnitaryGate(sigmax), [qr[1]]) qc_target.append(UnitaryGate(sigmaz), [[0, 1]]) self.assertEqual(qc, qc_target) def test_qobj_with_unitary_matrix(self): """test qobj output with unitary matrix""" qr = QuantumRegister(4) qc = QuantumCircuit(qr) sigmax = numpy.array([[0, 1], [1, 0]]) sigmay = numpy.array([[0, -1j], [1j, 0]]) matrix = numpy.kron(sigmay, numpy.kron(sigmax, sigmay)) qc.rx(numpy.pi / 4, qr[0]) uni = UnitaryGate(matrix) qc.append(uni, [qr[0], qr[1], qr[3]]) qc.cx(qr[3], qr[2]) qobj = qiskit.compiler.assemble(qc) instr = qobj.experiments[0].instructions[1] self.assertEqual(instr.name, "unitary") assert_allclose(numpy.array(instr.params[0]).astype(numpy.complex64), matrix) # check conversion to dict qobj_dict = qobj.to_dict() class NumpyEncoder(json.JSONEncoder): """Class for encoding json str with complex and numpy arrays.""" def default(self, obj): if isinstance(obj, numpy.ndarray): return obj.tolist() if isinstance(obj, complex): return (obj.real, obj.imag) return json.JSONEncoder.default(self, obj) # check json serialization self.assertTrue(isinstance(json.dumps(qobj_dict, cls=NumpyEncoder), str)) def test_labeled_unitary(self): """test qobj output with unitary matrix""" qr = QuantumRegister(4) qc = QuantumCircuit(qr) sigmax = numpy.array([[0, 1], [1, 0]]) sigmay = numpy.array([[0, -1j], [1j, 0]]) matrix = numpy.kron(sigmax, sigmay) uni = UnitaryGate(matrix, label="xy") qc.append(uni, [qr[0], qr[1]]) qobj = qiskit.compiler.assemble(qc) instr = qobj.experiments[0].instructions[0] self.assertEqual(instr.name, "unitary") self.assertEqual(instr.label, "xy") def test_qasm_unitary_only_one_def(self): """test that a custom unitary can be converted to qasm and the definition is only written once""" qr = QuantumRegister(2, "q0") cr = ClassicalRegister(1, "c0") qc = QuantumCircuit(qr, cr) matrix = numpy.array([[1, 0], [0, 1]]) unitary_gate = UnitaryGate(matrix) qc.x(qr[0]) qc.append(unitary_gate, [qr[0]]) qc.append(unitary_gate, [qr[1]]) expected_qasm = ( "OPENQASM 2.0;\n" 'include "qelib1.inc";\n' "gate unitary q0 { u(0,0,0) q0; }\n" "qreg q0[2];\ncreg c0[1];\n" "x q0[0];\n" "unitary q0[0];\n" "unitary q0[1];\n" ) self.assertEqual(expected_qasm, qc.qasm()) def test_qasm_unitary_twice(self): """test that a custom unitary can be converted to qasm and that if the qasm is called twice it is the same every time""" qr = QuantumRegister(2, "q0") cr = ClassicalRegister(1, "c0") qc = QuantumCircuit(qr, cr) matrix = numpy.array([[1, 0], [0, 1]]) unitary_gate = UnitaryGate(matrix) qc.x(qr[0]) qc.append(unitary_gate, [qr[0]]) qc.append(unitary_gate, [qr[1]]) expected_qasm = ( "OPENQASM 2.0;\n" 'include "qelib1.inc";\n' "gate unitary q0 { u(0,0,0) q0; }\n" "qreg q0[2];\ncreg c0[1];\n" "x q0[0];\n" "unitary q0[0];\n" "unitary q0[1];\n" ) self.assertEqual(expected_qasm, qc.qasm()) self.assertEqual(expected_qasm, qc.qasm()) def test_qasm_2q_unitary(self): """test that a 2 qubit custom unitary can be converted to qasm""" qr = QuantumRegister(2, "q0") cr = ClassicalRegister(1, "c0") qc = QuantumCircuit(qr, cr) matrix = numpy.asarray([[0, 0, 0, 1], [0, 0, 1, 0], [0, 1, 0, 0], [1, 0, 0, 0]]) unitary_gate = UnitaryGate(matrix) qc.x(qr[0]) qc.append(unitary_gate, [qr[0], qr[1]]) qc.append(unitary_gate, [qr[1], qr[0]]) expected_qasm = ( "OPENQASM 2.0;\n" 'include "qelib1.inc";\n' "gate unitary q0,q1 { u(pi,-pi/2,pi/2) q0; u(pi,pi/2,-pi/2) q1; }\n" "qreg q0[2];\n" "creg c0[1];\n" "x q0[0];\n" "unitary q0[0],q0[1];\n" "unitary q0[1],q0[0];\n" ) self.assertEqual(expected_qasm, qc.qasm()) def test_qasm_unitary_noop(self): """Test that an identity unitary can be converted to OpenQASM 2""" qc = QuantumCircuit(QuantumRegister(3, "q0")) qc.unitary(numpy.eye(8), qc.qubits) expected_qasm = ( "OPENQASM 2.0;\n" 'include "qelib1.inc";\n' "gate unitary q0,q1,q2 { }\n" "qreg q0[3];\n" "unitary q0[0],q0[1],q0[2];\n" ) self.assertEqual(expected_qasm, qc.qasm()) def test_unitary_decomposition(self): """Test decomposition for unitary gates over 2 qubits.""" qc = QuantumCircuit(3) qc.unitary(random_unitary(8, seed=42), [0, 1, 2]) self.assertTrue(Operator(qc).equiv(Operator(qc.decompose()))) def test_unitary_decomposition_via_definition(self): """Test decomposition for 1Q unitary via definition.""" mat = numpy.array([[0, 1], [1, 0]]) self.assertTrue(numpy.allclose(Operator(UnitaryGate(mat).definition).data, mat)) def test_unitary_decomposition_via_definition_2q(self): """Test decomposition for 2Q unitary via definition.""" mat = numpy.array([[0, 0, 1, 0], [0, 0, 0, -1], [1, 0, 0, 0], [0, -1, 0, 0]]) self.assertTrue(numpy.allclose(Operator(UnitaryGate(mat).definition).data, mat)) def test_unitary_control(self): """Test parameters of controlled - unitary.""" mat = numpy.array([[0, 1], [1, 0]]) gate = UnitaryGate(mat).control() self.assertTrue(numpy.allclose(gate.params, mat)) self.assertTrue(numpy.allclose(gate.base_gate.params, mat))
https://github.com/qclib/qclib
qclib
# Copyright 2021 qclib project. # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # http://www.apache.org/licenses/LICENSE-2.0 # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """ Test qclib.gate.mcg.LdMcSpecialUnitary """ from unittest import TestCase import numpy as np from qiskit import QuantumRegister, QuantumCircuit, transpile from qiskit.circuit.library import RXGate from qiskit.quantum_info import Operator from qclib.gates.multitargetmcsu2 import MultiTargetMCSU2 class TestCMcSpecialUnitary(TestCase): """ Test cases for the decomposition of Multicontrolled Special Unitary with linear depth https://arxiv.org/pdf/2302.06377.pdf """ def _build_qiskit_circ( self, num_controls ): """"default mode: target = controls""" controls_list = list(range(num_controls)) target = num_controls qiskit_circ = QuantumCircuit(num_controls + 3) qiskit_circ.mcrx(0.7, controls_list, target) qiskit_circ.mcrx(0.13, controls_list, target+1) qiskit_circ.mcrx(0.5, controls_list, target+2) return qiskit_circ def _build_ldmcsu_circ(self, unitary_list, num_controls): """" default mode: target = controls """ controls_list = list(range(num_controls)) target = num_controls ldmcsu_circ = QuantumCircuit(num_controls + 3) MultiTargetMCSU2.multi_target_mcsu2(ldmcsu_circ, unitary_list[0], controls_list, target) MultiTargetMCSU2.multi_target_mcsu2(ldmcsu_circ, unitary_list[1], controls_list, target + 1) MultiTargetMCSU2.multi_target_mcsu2(ldmcsu_circ, unitary_list[2], controls_list, target + 2) return ldmcsu_circ def _build_cldmcsu_circ(self, unitary_list, num_controls): """" default mode: target = controls """ controls = QuantumRegister(num_controls) target = QuantumRegister(len(unitary_list)) cldmcsu_circ = QuantumCircuit(controls, target) MultiTargetMCSU2.multi_target_mcsu2(cldmcsu_circ, unitary_list, controls, target) return cldmcsu_circ def test_clcmcsu_3targets(self): """ Test for comparison of a cascade uf 3 multi-controlled SU(2) using qiskit and cldmcsu implementations. """ num_controls = 7 unitary_list = [RXGate(0.7).to_matrix(), RXGate(0.13).to_matrix(), RXGate(0.5).to_matrix()] qiskit_circ = self._build_qiskit_circ(num_controls) cldmcsu_circ = self._build_cldmcsu_circ(unitary_list, num_controls) qiskitt = transpile(qiskit_circ, basis_gates=['u', 'cx'], optimization_level=3) cldmcsut = transpile(cldmcsu_circ, basis_gates=['u', 'cx'], optimization_level=3) qiskit_ops = qiskitt.count_ops() qclib_ops = cldmcsut.count_ops() self.assertTrue(qiskit_ops['cx'] > qclib_ops['cx']) cldmcsu_op = Operator(cldmcsu_circ).data qiskit_op = Operator(qiskit_circ).data self.assertTrue(np.allclose(cldmcsu_op, qiskit_op))
https://github.com/qclib/qclib
qclib
# Copyright 2021 qclib project. # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # http://www.apache.org/licenses/LICENSE-2.0 # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """ Test qclib.gate.mcg.LdMcSpecialUnitary """ from unittest import TestCase import numpy as np from qiskit import QuantumCircuit from qiskit.circuit.library import RXGate from qiskit.quantum_info import Operator from qiskit.circuit.library import UnitaryGate from scipy.stats import unitary_group from qclib.gates.ldmcsu import LdMcSpecialUnitary, Ldmcsu from qclib.gates.util import u2_to_su2 from qclib.util import get_cnot_count NUM_CTRL = 6 def _generate_su_2(): """ Returns random SU(2) matrix """ u_2 = unitary_group.rvs(2) su_2, _ = u2_to_su2(u_2) return su_2 class TestLcMcSpecialUnitary(TestCase): """ Test cases for the decomposition of Multicontrolled Special Unitary with linear depth by Barenco et al. """ def _build_qiskit_circuit(self, su2, num_controls, ctrl_state=None): su2_gate = UnitaryGate(su2) controls_list = list(range(num_controls)) target = num_controls qiskit_circ = QuantumCircuit(num_controls + 1) qiskit_circ.append(su2_gate.control(num_controls, ctrl_state=ctrl_state), [*controls_list, target]) return qiskit_circ def _compute_bound(self, num_qubits): if num_qubits % 2 == 0: return 28 * num_qubits - 88 return 28 * num_qubits - 92 def test_lcmcsu_op_for_trivial_control_state(self): """ Test LdMcSpecialUnitary open controls """ su2 = _generate_su_2() num_controls = NUM_CTRL ldmcsu_circ = LdMcSpecialUnitary(su2, num_controls).definition qiskit_circ = self._build_qiskit_circuit(su2, num_controls) ldmcsu_op = Operator(ldmcsu_circ).data qiskit_op = Operator(qiskit_circ).data self.assertTrue(np.allclose(ldmcsu_op, qiskit_op)) def test_lcmcsu_op_for_all_zero_control_states(self): """ Test LdMcSpecialUnitary with open controls """ su_2 = _generate_su_2() num_controls = NUM_CTRL ctrl_state = '0' * num_controls ldmcsu_circ = LdMcSpecialUnitary(su_2, num_controls, ctrl_state=ctrl_state).definition qiskit_circ = self._build_qiskit_circuit(su_2, num_controls, ctrl_state=ctrl_state) ldmcsu_op = Operator(ldmcsu_circ).data qiskit_op = Operator(qiskit_circ).data self.assertTrue(np.allclose(ldmcsu_op, qiskit_op)) def test_lcmcsu_cnot_count(self): """ Test LdMcSpecialUnitary cx count """ su_2 = _generate_su_2() for num_controls in range(8, 10): ldmcsu_circ = LdMcSpecialUnitary(su_2, num_controls).definition ldmcsu_count = get_cnot_count(ldmcsu_circ) self.assertLessEqual(ldmcsu_count, self._compute_bound(num_controls+1)) class TestMcSpecialUnitary(TestCase): """ Test cases for the decomposition of Multicontrolled Special Unitary with linear depth https://arxiv.org/pdf/2302.06377.pdf """ def _build_qiskit_circuit(self, su2, num_controls, ctrl_state=None): su2_gate = UnitaryGate(su2) controls_list = list(range(num_controls)) target = num_controls qiskit_circ = QuantumCircuit(num_controls + 1) qiskit_circ.append(su2_gate.control(num_controls, ctrl_state=ctrl_state), [*controls_list, target]) return qiskit_circ def _compute_bound(self, num_qubits): if num_qubits % 2 == 0: return 28 * num_qubits - 88 return 28 * num_qubits - 92 def test_lcmcsu_op_for_trivial_control_state(self): """ Test LdMcSpecialUnitary open controls """ su2 = _generate_su_2() num_controls = NUM_CTRL ldmcsu_circ = Ldmcsu(su2, num_controls).definition qiskit_circ = self._build_qiskit_circuit(su2, num_controls) ldmcsu_op = Operator(ldmcsu_circ).data qiskit_op = Operator(qiskit_circ).data self.assertTrue(np.allclose(ldmcsu_op, qiskit_op)) def test_lcmcsu_op_for_all_zero_control_states(self): """ Test LdMcSpecialUnitary with open controls """ su_2 = _generate_su_2() num_controls = NUM_CTRL ctrl_state = '0' * num_controls ldmcsu_circ = Ldmcsu(su_2, num_controls, ctrl_state=ctrl_state).definition qiskit_circ = self._build_qiskit_circuit(su_2, num_controls, ctrl_state=ctrl_state) ldmcsu_op = Operator(ldmcsu_circ).data qiskit_op = Operator(qiskit_circ).data self.assertTrue(np.allclose(ldmcsu_op, qiskit_op)) def test_lcmcsu_cnot_count(self): """ Test LdMcSpecialUnitary cx count """ su_2 = _generate_su_2() for num_controls in range(8, 10): ldmcsu_circ = Ldmcsu(su_2, num_controls).definition ldmcsu_count = get_cnot_count(ldmcsu_circ) self.assertTrue(ldmcsu_count <= 20 * (num_controls + 1)-38) def test_lcmcsu_cnot_count_real_diagonal(self): """ Test LdMcSpecialUnitary cx count """ su_2 = RXGate(0.3).to_matrix() for num_controls in range(8, 10): ldmcsu_circ = Ldmcsu(su_2, num_controls).definition ldmcsu_count = get_cnot_count(ldmcsu_circ) self.assertTrue(ldmcsu_count <= 16 * (num_controls + 1)-40) def test_lcmcsu_op_for_exception_unitary(self): """ Test Ldmcsu with Z gate """ su2 = np.array([[-1., 0.], [0., -1.]]) num_controls = 6 ldmcsu_circ = Ldmcsu(su2, num_controls).definition qiskit_circ = self._build_qiskit_circuit(su2, num_controls) ldmcsu_op = Operator(ldmcsu_circ).data qiskit_op = Operator(qiskit_circ).data self.assertTrue(np.allclose(ldmcsu_op, qiskit_op)) def test_lcmcsu_op_for_exception_unitary_2(self): """ Test Ldmcsu diagonal gate """ su2 = np.array([[np.e**(-1j*0.3), 0], [0, np.e**(1j*0.3)]]) num_controls = 6 ldmcsu_circ = Ldmcsu(su2, num_controls).definition qiskit_circ = self._build_qiskit_circuit(su2, num_controls) ldmcsu_op = Operator(ldmcsu_circ).data qiskit_op = Operator(qiskit_circ).data self.assertTrue(np.allclose(ldmcsu_op, qiskit_op))
https://github.com/qclib/qclib
qclib
# Copyright 2021 qclib project. # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # http://www.apache.org/licenses/LICENSE-2.0 # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """ Test qclib.gate.mc_gate """ from unittest import TestCase import numpy as np from scipy.stats import unitary_group import qiskit from qiskit import QuantumRegister, QuantumCircuit from qiskit.circuit.library import MCXGate import qclib.util from qclib.gates.ldmcu import Ldmcu class TestLinearToffoli(TestCase): """ Testing qclib.gate.toffoli """ def test_controlled_gate(self): """ Testing multi controlled gate """ gate_u = unitary_group.rvs(2) controls = QuantumRegister(4) target = QuantumRegister(1) circuit = qiskit.QuantumCircuit(controls, target) circuit.x(0) circuit.x(1) circuit.x(2) circuit.x(3) Ldmcu.ldmcu(circuit, gate_u, controls, target) state = qclib.util.get_state(circuit) self.assertTrue(np.isclose(state[15], gate_u[0, 0])) self.assertTrue(np.isclose(state[31], gate_u[1, 0])) controls2 = QuantumRegister(4) target2 = QuantumRegister(1) circuit2 = QuantumCircuit(controls2, target2) circuit2.x(0) circuit2.x(1) circuit2.x(2) circuit2.x(3) circuit2.x(4) Ldmcu.ldmcu(circuit2, gate_u, controls2, target2) state = qclib.util.get_state(circuit2) self.assertTrue(np.isclose(state[15], gate_u[0, 1])) self.assertTrue(np.isclose(state[31], gate_u[1, 1])) def test_linear_toffoli3(self): """ Testing Toffoli control 111""" gate_x = np.array([[0, 1], [1, 0]]) controls = QuantumRegister(3) target = QuantumRegister(1) circuit = QuantumCircuit(controls, target) circuit.x(1) circuit.x(2) circuit.x(3) circuit.x(0) Ldmcu.ldmcu(circuit, gate_x, controls, target) state = qclib.util.get_state(circuit) exp_state = np.zeros(16, dtype=complex) exp_state[7] = 1 self.assertTrue(np.allclose(state, exp_state)) def test_linear_toffoli2(self): """ Testing Toffoli control 110""" gate_x = np.array([[0, 1], [1, 0]]) controls2 = QuantumRegister(3) target2 = QuantumRegister(1) circuit2 = QuantumCircuit(controls2, target2) circuit2 = qiskit.QuantumCircuit(4) circuit2.x(2) circuit2.x(3) circuit2.x(0) state1 = qclib.util.get_state(circuit2) controls1 = QuantumRegister(3) target1 = QuantumRegister(1) circuit1 = qiskit.QuantumCircuit(controls1, target1) Ldmcu.ldmcu(circuit1, gate_x, controls1, target1, ctrl_state='110') circuit2.compose(circuit1, circuit2.qubits, inplace=True) state2 = qclib.util.get_state(circuit2) self.assertTrue(np.allclose(state1, state2)) def test_linear_toffoli1(self): """ Testing Toffoli control 100""" gate_x = np.array([[0, 1], [1, 0]]) circuit2 = qiskit.QuantumCircuit(4) circuit2.x(0) state1 = qclib.util.get_state(circuit2) controls = QuantumRegister(3) target = QuantumRegister(1) circuit = qiskit.QuantumCircuit(controls, target) Ldmcu.ldmcu(circuit, gate_x, controls, target, ctrl_state='100') circuit2.compose(circuit, circuit2.qubits, inplace=True) state2 = qclib.util.get_state(circuit2) self.assertTrue(np.allclose(state1, state2)) def test_linear_toffoli0(self): """ Testing Toffoli control 000""" gate_x = np.array([[0, 1], [1, 0]]) controls = QuantumRegister(3) target = QuantumRegister(1) mcgate_circuit = qiskit.QuantumCircuit(controls, target) Ldmcu.ldmcu(mcgate_circuit, gate_x, controls, target, ctrl_state="000") controls_2 = QuantumRegister(3) target_2 = QuantumRegister(1) qiskit_circuit = QuantumCircuit(controls_2, target_2) qiskit_circuit.append(MCXGate(len(controls_2), ctrl_state='000'), [*controls_2, target_2]) state_qiskit = qclib.util.get_state(qiskit_circuit) state_mcgate = qclib.util.get_state(mcgate_circuit) self.assertTrue(np.allclose(state_qiskit, state_mcgate)) def test_mct_toffoli(self): """ compare qiskit.mct and toffoli depth with 7 qubits """ gate_x = np.array([[0, 1], [1, 0]]) qcirc1 = qiskit.QuantumCircuit(6) qcirc1.mcx([0, 1, 2, 3, 4], 5) t_qcirc1 = qiskit.transpile(qcirc1, basis_gates=['u', 'cx']) controls = QuantumRegister(5) target = QuantumRegister(1) qcirc2 = qiskit.QuantumCircuit(controls, target) Ldmcu.ldmcu(qcirc2, gate_x, controls, target) t_qcirc2 = qiskit.transpile(qcirc2, basis_gates=['u', 'cx']) self.assertTrue(t_qcirc2.depth() < t_qcirc1.depth())
https://github.com/qclib/qclib
qclib
# Copyright 2021 qclib project. # Licensed under the Apache License, Version 2.0 (the "License"); # you may not use this file except in compliance with the License. # You may obtain a copy of the License at # http://www.apache.org/licenses/LICENSE-2.0 # Unless required by applicable law or agreed to in writing, software # distributed under the License is distributed on an "AS IS" BASIS, # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. # See the License for the specific language governing permissions and # limitations under the License. """ Test qclib.gate.mc_gate """ from unittest import TestCase import numpy as np import qiskit from qiskit import QuantumRegister, QuantumCircuit from qiskit.quantum_info import Operator import qclib.util from qclib.gates.ldmcu import Ldmcu from qclib.gates.mcu import MCU class TestLinearU2(TestCase): def _get_result_unitary(self, unitary, n_ctrl_base): exponent = n_ctrl_base - 1 param = (2**exponent - 1) / 2**exponent values, vectors = np.linalg.eig(unitary) gate = ( np.power(values[0] + 0j, param) * vectors[:, [0]] @ vectors[:, [0]].conj().T ) gate = ( gate + np.power(values[1] + 0j, param) * vectors[:, [1]] @ vectors[:, [1]].conj().T ) return gate def test_get_num_qubits(self): error = 0.1 u1 = np.array([[0, 1], [1, 0]]) ldmcu_approx = MCU(u1, num_controls=10, error=error) self.assertTrue(ldmcu_approx._get_num_base_ctrl_qubits(u1, error), 4) def test_mcz(self): error = 0.1 u = np.array([[1, 0], [0, -1]]) ldmcu_approx_test = MCU(u, num_controls=100, error=error) base_ctrl_qubits = ldmcu_approx_test._get_num_base_ctrl_qubits(u, error) controls = QuantumRegister(base_ctrl_qubits) target = QuantumRegister(1) circuit = qiskit.QuantumCircuit(controls, target) N = len(controls) for i in range(N): circuit.x(i) MCU.mcu(circuit, u, controls, target, error) state = qclib.util.get_state(circuit) res_u = self._get_result_unitary(u, base_ctrl_qubits) self.assertTrue(np.isclose(state[2**N - 1], res_u[0, 0])) self.assertTrue(np.isclose(state[2 ** (N + 1) - 1], res_u[1, 0])) controls = QuantumRegister(base_ctrl_qubits + 2) target = QuantumRegister(1) circuit = qiskit.QuantumCircuit(controls, target) N = len(controls) for i in range(N): circuit.x(i) MCU.mcu(circuit, u, controls, target, error) state = qclib.util.get_state(circuit) res_u = self._get_result_unitary(u, base_ctrl_qubits) self.assertTrue(np.isclose(state[2**N - 1], res_u[0, 0])) self.assertTrue(np.isclose(state[2 ** (N + 1) - 1], res_u[1, 0])) def test_mcx(self): error = 0.1 u = np.array([[0, 1], [1, 0]]) ldmcu_approx_test = MCU(u, num_controls=100, error=error) base_ctrl_qubits = ldmcu_approx_test._get_num_base_ctrl_qubits(u, error) controls = QuantumRegister(base_ctrl_qubits) target = QuantumRegister(1) circuit = qiskit.QuantumCircuit(controls, target) N = len(controls) for i in range(N): circuit.x(i) MCU.mcu(circuit, u, controls, target, error) state = qclib.util.get_state(circuit) res_u = self._get_result_unitary(u, base_ctrl_qubits) self.assertTrue(np.isclose(state[2**N - 1], res_u[0, 0])) self.assertTrue(np.isclose(state[2 ** (N + 1) - 1], res_u[1, 0])) controls = QuantumRegister(base_ctrl_qubits + 2) target = QuantumRegister(1) circuit = qiskit.QuantumCircuit(controls, target) N = len(controls) for i in range(N): circuit.x(i) MCU.mcu(circuit, u, controls, target, error) state = qclib.util.get_state(circuit) res_u = self._get_result_unitary(u, base_ctrl_qubits) self.assertTrue(np.isclose(state[2**N - 1], res_u[0, 0])) self.assertTrue(np.isclose(state[2 ** (N + 1) - 1], res_u[1, 0])) def test_mcz_cnot_count(self): error = 0.01 u = np.array([[1, 0], [0, -1]]) ldmcu_approx_test = MCU(u, num_controls=100, error=error) base_ctrl_qubits = ldmcu_approx_test._get_num_base_ctrl_qubits(u, error) for n_controls in range(base_ctrl_qubits + 60, base_ctrl_qubits + 62): controls = QuantumRegister(n_controls) target = QuantumRegister(1) circuit_approx = qiskit.QuantumCircuit(controls, target) MCU.mcu(circuit_approx, u, controls, target, error) cnot_approx = qclib.util.get_cnot_count(circ=circuit_approx) controls = QuantumRegister(n_controls) target = QuantumRegister(1) circuit_og = qiskit.QuantumCircuit(controls, target) Ldmcu.ldmcu(circuit_og, u, controls, target) cnot_og = qclib.util.get_cnot_count(circ=circuit_og) self.assertLessEqual(cnot_approx, cnot_og) def test_mcx_cnot_count(self): error = 0.01 u = np.array([[0, 1], [1, 0]]) ldmcu_approx_test = MCU(u, num_controls=100, error=error) base_ctrl_qubits = ldmcu_approx_test._get_num_base_ctrl_qubits(u, error) for n_controls in range(base_ctrl_qubits + 60, base_ctrl_qubits + 62): controls = QuantumRegister(n_controls) target = QuantumRegister(1) circuit_approx = qiskit.QuantumCircuit(controls, target) MCU.mcu(circuit_approx, u, controls, target, error) cnot_approx = qclib.util.get_cnot_count(circ=circuit_approx) controls = QuantumRegister(n_controls) target = QuantumRegister(1) circuit_og = qiskit.QuantumCircuit(controls, target) Ldmcu.ldmcu(circuit_og, u, controls, target) cnot_og = qclib.util.get_cnot_count(circ=circuit_og) self.assertLessEqual(cnot_approx, cnot_og) def test_to_compare_ldmcu_and_ldmcu_approx(self): unitary = np.array([[0, 1], [1, 0]]) error = 1e-2 ldmcu_approx_test = MCU(unitary, num_controls=100, error=error) base_ctrl_qubits = ldmcu_approx_test._get_num_base_ctrl_qubits(unitary, error) controls = QuantumRegister(base_ctrl_qubits) target = QuantumRegister(1) ldmcu_circ = QuantumCircuit(controls, target) Ldmcu.ldmcu(ldmcu_circ, unitary, controls, target) ldmcu_approx_circ = QuantumCircuit(controls, target) MCU.mcu(ldmcu_approx_circ, unitary, controls, target, error) ldmcu_op = Operator(ldmcu_circ).data ldmcu_approx_op = Operator(ldmcu_approx_circ).data # absolute(a - b) <= (atol + r_tol * absolute(b) self.assertTrue(np.allclose(ldmcu_op, ldmcu_approx_op, atol=error))