markdown stringlengths 0 1.02M | code stringlengths 0 832k | output stringlengths 0 1.02M | license stringlengths 3 36 | path stringlengths 6 265 | repo_name stringlengths 6 127 |
|---|---|---|---|---|---|
stage2を結合 | # 表記揺れの確認
print(np.sort(train['stage'].unique()))
print(np.sort(test['stage'].unique()))
# 「mystery~」はイベント時に解放されるステージ、今回のtrain,testデータには無し
print(np.sort(stage2['key'].unique()))
stage2_r.columns
# 結合のため列名変更
stage2_r = stage2.rename(columns = {'key':'stage', 'area':'stage_size2'})
# 必要カラム
st2_col = ['stage_size2', # ステー... | stage_size2 0
dtype: int64
stage_size2 0
dtype: int64
| MIT | program/lightGBM_base_v0.1.ipynb | tomokoochi/splatoon_competition |
weaponを結合 | # trainのブキ
train_weapon = sorted(list(set(train['A1-weapon'])&set(train['A2-weapon'])&set(train['A3-weapon'])&set(train['A4-weapon'])\
&set(train['B1-weapon'])&set(train['B2-weapon'])&set(train['B3-weapon'])&set(train['B4-weapon'])))
print('{}種類'.format(len(train_weapon)))
print(train_weapon)
# testのブキ
test_weapon = so... | _____no_output_____ | MIT | program/lightGBM_base_v0.1.ipynb | tomokoochi/splatoon_competition |
前処理 | # 欠損値埋める
def fill_all_null(df, num):
for col_name in df.columns[df.isnull().sum()!=0]:
df[col_name] = df[col_name].fillna(num)
# 訓練データ、テストデータの欠損値を-1で補完
fill_all_null(train_input, -1)
fill_all_null(test_input, -1)
# ターゲットエンコーディングの関数定義
## Holdout TSを用いる 変更の余地あり
def change_to_target2(train_df,test_df,input_col... | _____no_output_____ | MIT | program/lightGBM_base_v0.1.ipynb | tomokoochi/splatoon_competition |
データの確認 | # 訓練データとテストデータの列を確認
print(train_input.columns)
print(test_input.columns) | Index(['A1-level', 'A2-level', 'A3-level', 'A4-level', 'B1-level', 'B2-level',
'B3-level', 'B4-level', 'y', 'stage_size1', 'stage_size2', 'enc_period',
'enc_game-ver', 'enc_lobby-mode', 'enc_lobby', 'enc_mode', 'enc_stage',
'enc_A1-weapon', 'enc_A1-rank', 'enc_A2-weapon', 'enc_A2-rank',
'enc... | MIT | program/lightGBM_base_v0.1.ipynb | tomokoochi/splatoon_competition |
学習の準備 | # 訓練データを説明変数と目的変数に分割
target = train_input['y']
train_x = train_input.drop('y',axis=1)
# LGBMのパラメータを設定
params = {
# 二値分類問題
'objective': 'binary',
# 損失関数は二値のlogloss
#'metric': 'auc',
'metric': 'binary_logloss',
# 最大イテレーション回数指定
'num_iterations' : 1000,
# early_stopping 回数指定
'early_stopp... | _____no_output_____ | MIT | program/lightGBM_base_v0.1.ipynb | tomokoochi/splatoon_competition |
学習・予測の実行 | # k-分割交差検証を使って学習＆予測（K=10）
FOLD_NUM = 10
kf = KFold(n_splits=FOLD_NUM,
random_state=42)
#lgbmのラウンド数を定義
num_round = 10000
#検証時のスコアを初期化
scores = []
#テストデータの予測値を初期化
pred_cv = np.zeros(len(test.index))
for i, (tdx, vdx) in enumerate(kf.split(train_x, target)):
print(f'Fold : {i}')
# 訓練用データと検証用データに分割... | _____no_output_____ | MIT | program/lightGBM_base_v0.1.ipynb | tomokoochi/splatoon_competition |
Angle-based Outlier Detector (ABOD) | clf1=ABOD(contamination=outliers_fraction)
clf1.fit(X)
y_pred1=clf1.predict(X)
from sklearn.metrics import confusion_matrix
confusion_matrix(y, y_pred1) | _____no_output_____ | MIT | outlierdetector_lib.ipynb | eaglewarrior/Anamoly-Detection |
Cluster-based Local Outlier Factor (CBLOF) | clf2=CBLOF(contamination=outliers_fraction,check_estimator=False, random_state=random_state)
clf2.fit(X)
y_pred2=clf2.predict(X)
from sklearn.metrics import confusion_matrix
confusion_matrix(y, y_pred2) | _____no_output_____ | MIT | outlierdetector_lib.ipynb | eaglewarrior/Anamoly-Detection |
Feature Bagging | clf3=FeatureBagging(LOF(n_neighbors=35),contamination=outliers_fraction,check_estimator=False,random_state=random_state)
clf3.fit(X)
y_pred3=clf3.predict(X)
from sklearn.metrics import confusion_matrix
confusion_matrix(y, y_pred3) | _____no_output_____ | MIT | outlierdetector_lib.ipynb | eaglewarrior/Anamoly-Detection |
Histogram-base Outlier Detection (HBOS) | clf4=HBOS(alpha=0.1, contamination=0.037, n_bins=10, tol=0.9)
clf4.fit(X)
y_pred4=clf4.predict(X)
from sklearn.metrics import confusion_matrix
confusion_matrix(y, y_pred4) | _____no_output_____ | MIT | outlierdetector_lib.ipynb | eaglewarrior/Anamoly-Detection |
Isolation Forest | clf5=IForest(contamination=outliers_fraction,random_state=random_state)
clf5.fit(X)
y_pred5=clf5.predict(X)
from sklearn.metrics import confusion_matrix
confusion_matrix(y, y_pred5) | _____no_output_____ | MIT | outlierdetector_lib.ipynb | eaglewarrior/Anamoly-Detection |
K Nearest Neighbors (KNN) | clf6=KNN(contamination=outliers_fraction)
clf6.fit(X)
y_pred6=clf6.predict(X)
from sklearn.metrics import confusion_matrix
confusion_matrix(y, y_pred6) | _____no_output_____ | MIT | outlierdetector_lib.ipynb | eaglewarrior/Anamoly-Detection |
Average KNN | clf7=KNN(method='mean',contamination=outliers_fraction)
clf7.fit(X)
y_pred7=clf7.predict(X)
from sklearn.metrics import confusion_matrix
confusion_matrix(y, y_pred7) | _____no_output_____ | MIT | outlierdetector_lib.ipynb | eaglewarrior/Anamoly-Detection |
Exercise 1Add the specified code for each code cell, running the cells _in order_. Create a variable `food` that stores your favorite kind of food. Print or output the variable. | food = "pizza" | _____no_output_____ | MIT | exercise-1/exercise.ipynb | ajm1813/ch4-python-intro |
Create a variable `restaurant` that stores your favorite place to eat that kind of food. | restaurant = "Delfinos pizza" | _____no_output_____ | MIT | exercise-1/exercise.ipynb | ajm1813/ch4-python-intro |
Print the message `"I'm going to RESTAURANT for some FOOD"`, replacing the restaurant and food with your variables. | print ("I'm going to " + restaurant + " for some " + food) | I'm going to Delfinos pizza for some pizza
| MIT | exercise-1/exercise.ipynb | ajm1813/ch4-python-intro |
Create a variable `num_friends` equal to the number of friends you would like to eat with. | num_friends = 5 | _____no_output_____ | MIT | exercise-1/exercise.ipynb | ajm1813/ch4-python-intro |
Print a message `"I'm going with X friends"`, replacing the X with the number of friends. | print ("I'm going with " + str(num_friends) + " friends ") | I'm going with 5 friends
| MIT | exercise-1/exercise.ipynb | ajm1813/ch4-python-intro |
Create a variable `meal_price`, which is how expensive you think one meal at the restaurant would be. This price should be a `float`. | meal_price = 35.90 | _____no_output_____ | MIT | exercise-1/exercise.ipynb | ajm1813/ch4-python-intro |
Update (re-assign) the `meal_price` variable so it includes a 15% tip—that is, so the price is 15% higher. Output the variable. | meal_price = meal_price * 1.15 | _____no_output_____ | MIT | exercise-1/exercise.ipynb | ajm1813/ch4-python-intro |
Create a variable `total_cost` that has the total estimated cost of the bill for you and all of your friends. Output or print the variable | total_cost = meal_price * num_friends | _____no_output_____ | MIT | exercise-1/exercise.ipynb | ajm1813/ch4-python-intro |
Create a variable `budget` representing your spending budget for a night out. | budget = 500 | _____no_output_____ | MIT | exercise-1/exercise.ipynb | ajm1813/ch4-python-intro |
Create a variable `max_friends`, which is the maximum number of friends you can invite, at the estimated meal price, while staying within your budget. Output or print this value.- Be carefully that you only invite whole people! | max_friends = int (budget/meal_price) | _____no_output_____ | MIT | exercise-1/exercise.ipynb | ajm1813/ch4-python-intro |
Bonus: Create a variable `chorus` that is the string `"FOOD time!"` _repeated_ once for each of the friends you are able to bring. _Hint_ use the **`*`** operator. Print out the variable. | print ("food time!\n " * 5) | food time!
food time!
food time!
food time!
food time!
| MIT | exercise-1/exercise.ipynb | ajm1813/ch4-python-intro |
OPTIMIZATION PHASES List of variables Variable Description Comment $B$ Number of full blocks/pages that need the records $\lceil \frac{|T|}{R} \rceil$; $B \ll |T|$ $R$ Number of r... | import math
c_P, c_W, c_V = 10000, 5000, 100000
R_p, R_w, R_v = 12, 10, 20
B_p, B_w, B_v = math.ceil(c_P / R_p), math.ceil(c_W / R_w), math.ceil(c_V / R_v)
print("Cardinality of {}: {}, Records: {}, number of Full Blocks: {}".format('P', c_P, R_p, B_p))
print("Cardinality of {}: {}, Records: {}, number of Full Blocks... | Cardinality of P: 10000, Records: 12, number of Full Blocks: 834
Cardinality of W: 5000, Records: 10, number of Full Blocks: 500
Cardinality of V: 100000, Records: 20, number of Full Blocks: 5000
| MIT | ADSDB/Optimization-Costs.ipynb | MiguelHeCa/miri-notes |
Phase 1. Alternatives generation```sqlSELECT DISTINCT w.strengthFROM wines w, producers p, vintages vWHERE v.wineId = w.wineId AND p.prodId = v.prodId AND p.region = "Priorat" AND v.quantity > 100;```Change selection and join arrangement Phase 2. Intermediate results estimation`... | c = 100
min_v = 10
max_v = 500
SF_v_prime = (max_v - c) / (max_v - min_v)
print("Selectivity factor of V': {} \n".format(SF_v_prime))
C_v_prime = math.floor(SF_v_prime * c_V)
print("Cardinality output of V': {} \n".format(C_v_prime))
R_v_len = 5 + 5
B = 500
R_v_prime = math.floor(B / R_v_len)
print("V' number of reco... | Selectivity factor of V': 0.8163265306122449
Cardinality output of V': 81632
V' number of records per block : 50
Blocks needed for V': 1633
| MIT | ADSDB/Optimization-Costs.ipynb | MiguelHeCa/miri-notes |
**Selection over P: P'**Record length of prodId:$$|R_{P'}| = 5$$Selectivity factor of selection:$$\mathrm{SF}(A = c) = \frac{1}{\text{ndist}(A)}$$Where $c = 'Priorat'$ and the query specifies `p.region = 'Priorat'`, then:$$\text{SF}(\text{region} = \text{Priorat}) = \frac{1}{30} = 0.033333$$Output ... | ndist_region = 30
SF_p_prime = 1 / ndist_region
print("Selectivity factor of P': {} \n".format(SF_p_prime))
C_p_prime = math.floor(SF_p_prime * c_P)
print("Cardinality output of P': {} \n".format(C_p_prime))
R_p_len = 5
B = 500
R_p_prime = math.floor(B / R_p_len)
print("P' number of records per block : {} \n".format(... | Selectivity factor of P': 0.03333333333333333
Cardinality output of P': 333
P' number of records per block : 100
Blocks needed for P': 4
| MIT | ADSDB/Optimization-Costs.ipynb | MiguelHeCa/miri-notes |
**PT1****Join between W and V': WV'**Record length of `strength` and `prodId`:$$|R_{WV'}| = 5 + 5$$Selectivity factor$$\text{SF}_{WV'} = \frac{1}{|W|} = \frac{1}{5000} = 0.0002$$Cardinality ouput of WV'$$|WV'| = SF_{WV'} \cdot |W| \cdot |V'| = \frac{1}{5000} \cdot 5000 \cdot 81632 = 81632$$Number of... | SF_wv_prime = 1 / c_W
print("Selectivity factor of WV': {} \n".format(SF_wv_prime))
C_wv_prime = math.floor(SF_wv_prime * c_W * C_v_prime)
print("Cardinality output of WV': {} \n".format(C_wv_prime))
R_wv_prime_len = 5 + 5
B = 500
R_wv_prime = math.floor(B / R_wv_prime_len)
print("WV' number of records per block : {}... | Selectivity factor of WV': 0.0002
Cardinality output of WV': 81632
WV' number of records per block : 50
Blocks needed for WV': 1633
| MIT | ADSDB/Optimization-Costs.ipynb | MiguelHeCa/miri-notes |
**Join between WV' and P': WV'P'**Record length for `strength`:$$|R_{WV'P'}| = 5$$Selectivity Factor, assuming quantity and region independent$$\text{SF(WV'} \cdot \text{P')} = \frac{1}{|P'|} \cdot \frac{1}{ndist(\text{region})} = \frac{1}{333 \cdot 30} = 10^{-4}$$Cardinality output$$|WV'P'| ... | SF_wvp_prime = (1 / C_p_prime) * (1 / ndist_region)
print("Selectivity factor of WV'P': {} \n".format(SF_wvp_prime))
C_wvp_prime = math.floor(SF_wvp_prime * C_wv_prime * C_p_prime)
print("Cardinality output of WV'P': {} \n".format(C_wvp_prime))
R_wvp_prime_len = 5
B = 500
R_wvp_prime = math.floor(B / R_wvp_prime_len)... | Selectivity factor of WV'P': 0.00010010010010010009
Cardinality output of WV'P': 2721
WV'P' number of records per block : 100
Blocks needed for WV'P': 28
| MIT | ADSDB/Optimization-Costs.ipynb | MiguelHeCa/miri-notes |
**PT2****Join V' and P': V'P'**Assuming independence of variablesRecord length for `wineId`$$|R_{V'P'}| = 5$$Selectivity factor$$\text{SF}_{V'P'} = \frac{1}{ndist(\text{region})} \cdot \frac{1}{|P'|} = \frac{1}{30} \cdot \frac{1}{333} = 10^{-4}$$Output cardinality$$|V'P'| = \text{SF}_{V'P'} \cd... | ndist_region = 30
SF_vp_prime = (1 / ndist_region) * (1 / C_p_prime)
print("Selectivity factor of V'P': {} \n".format(SF_vp_prime))
C_vp_prime = math.floor(SF_vp_prime * C_v_prime * C_p_prime)
print("Cardinality output of V'P': {} \n".format(C_vp_prime))
R_vp_len = 5
B = 500
R_vp_prime = math.floor(B / R_vp_len)
prin... | Selectivity factor of V'P': 0.00010010010010010009
Cardinality output of V'P': 2721
V'P' number of records per block : 100
Blocks needed for V'P': 28
| MIT | ADSDB/Optimization-Costs.ipynb | MiguelHeCa/miri-notes |
**Join between W and V'P': WV'P'**Record length for WV'P'$$|R_{WV'P'}| = 5$$Selectivity Factor for WV'P'$$\text{SF} = \frac{1}{|W|} = \frac{1}{5000} = 0.0002$$Cardinality Output$$|WV'P'| = SF \cdot |W| \cdot |V'P'| = 10^{-4} \cdot 5000 \cdot 2721 = 2721$$Number of records per block$$R_{WV'P'... | SF_wv_pr_p_pr = 1 / c_W
print("Selectivity factor of WV'P': {} \n".format(SF_wv_pr_p_pr))
C_wv_pr_p_pr = math.floor(SF_wv_pr_p_pr * c_W * C_vp_prime)
print("Cardinality output of WV'P': {} \n".format(C_wv_pr_p_pr))
R_wv_pr_p_pr_len = 5
B = 500
R_wv_pr_p_pr = math.floor(B / R_wv_pr_p_pr_len)
print("WV'P' number of rec... | Selectivity factor of WV'P': 0.0002
Cardinality output of WV'P': 2721
WV'P' number of records per block : 100
Blocks needed for WV'P': 28
| MIT | ADSDB/Optimization-Costs.ipynb | MiguelHeCa/miri-notes |
**PT1/PT2****Final result = O**Record length$$|R_O| = 5$$Output cardinality$$|O| = \text{ndist}(\text{strength}) = 100$$Number of records$$R_O = \lfloor \frac{B}{|R_O|} \rfloor = \lfloor \frac{500}{5} \rfloor = 100$$Blocks needed$$B_O = \lceil \frac{|O|}{R_O} \rceil = \lceil \frac{100}{100} \rceil = 1$$ | ndist_strength = 100
C_o = ndist_strength
print("Cardinality output of O: {} \n".format(C_o))
R_o_len = 5
B = 500
R_o = math.floor(B / R_o_len)
print("O number of records per block : {} \n".format(R_o))
B_o = math.ceil(C_o / R_o)
print("Blocks needed for O: {} \n".format(B_o)) | Cardinality output of O: 100
O number of records per block : 100
Blocks needed for O: 1
| MIT | ADSDB/Optimization-Costs.ipynb | MiguelHeCa/miri-notes |
**Map result** Phase 3. Cost estimation for each algorithmRecall:$$u = \frac{2}{3} \cdot 2(75) = 100$$**AP1/AP2****Selection over V: V'**Recall that for Vintages is clustered by wineId and prodIdAvailable access paths: No index$$\text{cost}_{\text{scan}}(V') = \lceil 1.5 B_{V} \rceil \cdot D = \l... | load = 2/3
d = 75
u = load * (2 * d)
h = math.ceil(math.log(c_P, u)) - 1
D = 1
print("load is: {}\nd is: {}\nu is: {}\nh is: {}\nD is: {}\n".format(load, d, u, h, D))
cost_scan_p = math.ceil(1.5 * B_p) * D
cost_bplus_p = (h * D) + ((C_p_prime / u) * D) + (C_p_prime * D)
print("Cost of scan is: {} \nCost of B+ is: {}".f... | load is: 0.6666666666666666
d is: 75
u is: 100.0
h is: 1
D is: 1
Cost of scan is: 1251
Cost of B+ is: 337.33
| MIT | ADSDB/Optimization-Costs.ipynb | MiguelHeCa/miri-notes |
**PT1****Join over W and V': WV'**Available algorithms: Block Nested Loops (BML), Row Nested Loops (RML) and Sort-Match (SM)*Block Nested Loops*Recall:$$M = 4$$$\lceil 1.5 B_{W} \rceil < B_{V'}$ we use the commutative property of joins$$\begin{align}\text{cost}_{\text{BML}}(WV') & = \lceil 1.5 B_{W} \rceil + \lceil \fr... | print("B_p' is {}\nB_wv' is {}".format(B_p_prime, B_wv_prime))
(2 * B_wv_prime * math.ceil(math.log(B_wv_prime, 2))) + (2 * B_p_prime * math.ceil(math.log(B_p_prime, 2))) + B_wv_prime + B_p_prime | _____no_output_____ | MIT | ADSDB/Optimization-Costs.ipynb | MiguelHeCa/miri-notes |
**PT2****Join between V' and P': V'P'Available algorithms: BNL and SM.*Block Nested Loops*$B_{P'} < B_{V'}$ we use the commutative property of joins$$\begin{align}\text{cost}_{\text{BML}}(V'P') & = B_{P'} + \lceil \frac{B_{P'}}{M} \rceil \cdot B_{V'} \\& = 4 + \lceil \frac{4}{4} \rceil \cdot 1633 \\& = 1637\end{align}$... | print("B_p' is {}\nB_v' is {}".format(B_p_prime, B_v_prime)) | B_p' is 4
B_v' is 1633
| MIT | ADSDB/Optimization-Costs.ipynb | MiguelHeCa/miri-notes |
**Join between W and V'P': WV'P'**Available algorithms: Block Nested Loops (BML), Row Nested Loops (RML) and Sort-Match (SM)*Block Nested Loops*$B_{V'P'} < \lceil 1.5 B_{W} \rceil$ we use the commutative property of joins$$\begin{align}\text{cost}_{\text{BML}}(WV'P') & = B_{V'P'} + \lceil \frac{B_{V'P'}}{M} \rceil \cdo... | print("B_v'p' is {}\n1.5*B_w is {}\n|V'P'| is {}".format(B_vp_prime, math.ceil(1.5*B_w), C_vp_prime))
28 + math.ceil(28/4) * 750
28+(2721*(math.ceil(math.log(5000, 100)) - 1 + 1))
Cost_v_prime = 1633 + 7500
Cost_p_prime = 4 + 337
Cost_wv = 1633 + 2383
Cost_vp = 28 + 1637
Cost_wvp_pt1 = 28 + 1637
Cost_wvp_pt2 = 28 + 105... | Total cost of:
PT1: 15408
PT2: 12478
| MIT | ADSDB/Optimization-Costs.ipynb | MiguelHeCa/miri-notes |
Спортивный анализ данных. Платформа Kaggle Урок 1. Введение в спортивный анализ данных, Exploration Data Analysis Домашняя работа к уроку 1 Ссылка на наборы данных: https://drive.google.com/file/d/1j8zuKbI-PW5qKwhybP4S0EtugbPqmeyX/view?usp=sharing Задание 1 Сделать базовый анализ данных: вывести размерность датасет... | # В работе. Как-то все наложилось. Надеюсь на этой неделе все нагнать.
# Посмотрел. Очень серьезный курс, темы сложные. Зря его вынесли во вне четверти. | _____no_output_____ | MIT | webinar_1/Lesson 1.ipynb | superbe/KagglePlatform |
Задание 2 Сделать базовый анализ целевой переменной, сделать выводы; | # В работе | _____no_output_____ | MIT | webinar_1/Lesson 1.ipynb | superbe/KagglePlatform |
Задание 3 Построить распределение признаков в зависимости от значения целевой переменной и распределение признаков для обучающей и тестовой выборки (если машина не позволяет построить распределение для всех признаков, то выполнить задание для признаков var_0, var_1, var_2, var_5, var_9, var_10, var_13, var_20, var_26,... | # В работе | _____no_output_____ | MIT | webinar_1/Lesson 1.ipynb | superbe/KagglePlatform |
Задание 4 Построить распределение основных статистики признаков (среднее, стандартное отклонение) в разрезе целевой переменной и распределение основных статистик обучающей и тестовой выборки, сделать выводы; | # В работе | _____no_output_____ | MIT | webinar_1/Lesson 1.ipynb | superbe/KagglePlatform |
Задание 5 Построить распределение коэффициентов корреляции между признаками. Есть ли зависимость между признаками (будем считать, что связь между признаками отсутствует, если коэффициент корреляции < 0.2)? | # В работе | _____no_output_____ | MIT | webinar_1/Lesson 1.ipynb | superbe/KagglePlatform |
Задание 6 Выявить 10 признаков, которые обладают наибольшей нелинейной связью с целевой переменной. | # В работе | _____no_output_____ | MIT | webinar_1/Lesson 1.ipynb | superbe/KagglePlatform |
Задание 7 Провести анализ идентичности распределения признаков на обучающей и тестовой выборках, сделать выводы. | # В работе | _____no_output_____ | MIT | webinar_1/Lesson 1.ipynb | superbe/KagglePlatform |
Euler Problem 206=================Find the unique positive integer whose square has the form 1_2_3_4_5_6_7_8_9_0,where each "&95;" is a single digit. | from itertools import product
for a, b, c, d in product(range(10), repeat=4):
N = 10203040596979899
N += a*10**15 + b*10**13 + c*10**11 + d*10**9
sqrtN = int(N**0.5)
s = str(sqrtN**2)
if s[0:17:2] == '123456789':
print(sqrtN * 10)
break | 1389019170
| MIT | Euler 206 - Concealed square.ipynb | Radcliffe/project-euler |
BLU15 - Model CSI Intro:It often happens that your data distribution changes with time. More than that, sometimes you don't know how a model was trained and what was the original training data. In this learning unit we're going to try to identify whether an existing model meets our expectations and redeploy it. Pro... | import joblib
import pandas as pd
import json
import joblib
import pickle
from sklearn.metrics import precision_score, recall_score, roc_auc_score
from sklearn.metrics import confusion_matrix
import requests
import matplotlib.pyplot as plt
import matplotlib.image as mpimg
from sklearn.metrics import precision_recall_cu... | _____no_output_____ | MIT | S06 - DS in the Real World/BLU15 - Model CSI/BLU15 - Learning Unit - Model CSI.ipynb | LDSSA/batch4-students |
Let's start from sending all those requests and comparing the model prediction results with the target values. The model is already prepared to convert our observations to the format its expecting, the only thing we need to change is making department and intervention location names lowercase, and we're good to extract... | # lowercaes departments and location names
df['Department Name'] = df['Department Name'].apply(lambda x: str(x).lower())
df['InterventionLocationName'] = df['InterventionLocationName'].apply(lambda x: str(x).lower())
url = "http://127.0.0.1:5000/predict"
headers = {'Content-Type': 'application/json'}
def send_request(i... | _____no_output_____ | MIT | S06 - DS in the Real World/BLU15 - Model CSI/BLU15 - Learning Unit - Model CSI.ipynb | LDSSA/batch4-students |
**NOTE:** We could also load the model and make predictions locally (without using the api), but:1. I wanted to show you how you might send requests in a similar situation2. If you have a running API and some model file, you always need to understand how the API works (if it makes any kind of data preprocessing), which... | confusion_matrix(df['ContrabandIndicator'], df['prediction']) | _____no_output_____ | MIT | S06 - DS in the Real World/BLU15 - Model CSI/BLU15 - Learning Unit - Model CSI.ipynb | LDSSA/batch4-students |
If you're not familiar with confusion matrixes, **here is an explanation of the values:** These values don't seem to be good. Let's once again take a look on the client's requirements and see if we still meet them: > A minimum 50% success rate for searches (when a car is searched, it should be at least 50% likely that... | def verify_success_rate_above(y_true, y_pred, min_success_rate=0.5):
"""
Verifies the success rate on a test set is above a provided minimum
"""
precision = precision_score(y_true, y_pred, pos_label=True)
is_satisfied = (precision >= min_success_rate)
return is_satisfied, precisi... | _____no_output_____ | MIT | S06 - DS in the Real World/BLU15 - Model CSI/BLU15 - Learning Unit - Model CSI.ipynb | LDSSA/batch4-students |
 > The largest possible amount of contraband found, given the constraints above.As the client says, their model recall was 0.893. And what now? | def verify_amount_found(y_true, y_pred):
"""
Verifies the amout of contraband found in the test dataset - a.k.a the recall in our test set
"""
recall = recall_score(y_true, y_pred, pos_label=True)
return recall
verify_amount_found(df['ContrabandIndicator'], df['prediction']) | _____no_output_____ | MIT | S06 - DS in the Real World/BLU15 - Model CSI/BLU15 - Learning Unit - Model CSI.ipynb | LDSSA/batch4-students |
Okay, relax, it happens. Let's start from checking different thresholds. Maybe the selected threshold was to specific and doesn't work anymore. What about 0.25? | threshold = 0.25
df['prediction'] = [1 if p >= threshold else 0 for p in df['proba']]
verify_success_rate_above(df['ContrabandIndicator'], df['prediction'], 0.5)
verify_amount_found(df['ContrabandIndicator'], df['prediction']) | _____no_output_____ | MIT | S06 - DS in the Real World/BLU15 - Model CSI/BLU15 - Learning Unit - Model CSI.ipynb | LDSSA/batch4-students |
Okay, let's try the same technique to identify the best threshold as they originally did. Maybe we find something good enough. It's not a good idea to verify such things on the test data, but we're going to use it just to confirm the model's performance, not to select the threshold. | precision, recall, thresholds = precision_recall_curve(df['ContrabandIndicator'], df['proba'])
precision = precision[:-1]
recall = recall[:-1]
fig=plt.figure()
ax1 = plt.subplot(211)
ax2 = plt.subplot(212)
ax1.hlines(y=0.5,xmin=0, xmax=1, colors='red')
ax1.plot(thresholds,precision)
ax2.plot(thresholds,recall)
ax1.get_... | _____no_output_____ | MIT | S06 - DS in the Real World/BLU15 - Model CSI/BLU15 - Learning Unit - Model CSI.ipynb | LDSSA/batch4-students |
So what do we see? There is some threshold value (around 0.6) that gives us precision >= 0.5. But the threshold is so big, that the recall at this point is really-really low. Let's calculate the exact values: | min_index = [i for i, prec in enumerate(precision) if prec >= 0.5][0]
print(min_index)
thresholds[min_index]
precision[min_index]
recall[min_index] | _____no_output_____ | MIT | S06 - DS in the Real World/BLU15 - Model CSI/BLU15 - Learning Unit - Model CSI.ipynb | LDSSA/batch4-students |
Before we move on, we need to understand why this happens, so that we can decide what kind of action to perform. Let's try to analyze the changes in data and discuss different things we might want to do. | old_df = pd.read_csv('./data/train_searched.csv')
old_df.head() | _____no_output_____ | MIT | S06 - DS in the Real World/BLU15 - Model CSI/BLU15 - Learning Unit - Model CSI.ipynb | LDSSA/batch4-students |
We're going to apply the same changes to the dataset as in the original model notebook unit to understand what was the original data like and how the current dataset differs. | old_df = old_df[(old_df['VehicleSearchedIndicator']==True)]
# lowercaes departments and location names
old_df['Department Name'] = old_df['Department Name'].apply(lambda x: str(x).lower())
old_df['InterventionLocationName'] = old_df['InterventionLocationName'].apply(lambda x: str(x).lower())
train_features = old_df.col... | _____no_output_____ | MIT | S06 - DS in the Real World/BLU15 - Model CSI/BLU15 - Learning Unit - Model CSI.ipynb | LDSSA/batch4-students |
Looks like we got a bit more contraband now, and it's already a good sign:if the training data had a different target feature distribution than the test set, the model's predictions might have a different distribution as well. It's a good practice to have the same target feature distribution both in training and test s... | new_department_names = df['Department Name'].unique()
old_department_names = old_df['Department Name'].unique()
unknown_departments = [department for department in new_department_names if department not in old_department_names]
len(unknown_departments)
df[df['Department Name'].isin(unknown_departments)].shape | _____no_output_____ | MIT | S06 - DS in the Real World/BLU15 - Model CSI/BLU15 - Learning Unit - Model CSI.ipynb | LDSSA/batch4-students |
So we have 10 departments that the original model was not trained on, but they are only 23 rows from the test set. Let's repeat the same thing for the Intervention Location names | new_location_names = df['InterventionLocationName'].unique()
old_location_names = old_df['InterventionLocationName'].unique()
unknown_locations = [location for location in new_location_names if location not in old_location_names]
len(unknown_locations)
df[df['InterventionLocationName'].isin(unknown_locations)].shape[0... | unknown locations: 5.3 %
| MIT | S06 - DS in the Real World/BLU15 - Model CSI/BLU15 - Learning Unit - Model CSI.ipynb | LDSSA/batch4-students |
Alright, a bit more of unknown locations. We don't know if the feature was important for the model, so these 5.3% of unknown locations might be important or not.But it's worth keeping it in mind. **Here are a few ideas of what we could try to do:**1. Reanalyze the filtered locations, e.g. filter more rare ones.2. Creat... | common_departments = df['Department Name'].value_counts().head(20).index
departments_new = df[df['Department Name'].isin(common_departments)]
departments_old = old_df[old_df['Department Name'].isin(common_departments)]
pd.crosstab(departments_new['ContrabandIndicator'], departments_new['Department Name'], normalize="co... | _____no_output_____ | MIT | S06 - DS in the Real World/BLU15 - Model CSI/BLU15 - Learning Unit - Model CSI.ipynb | LDSSA/batch4-students |
We can clearly see that some departments got a huge difference in the contraband indicator.E.g. Bridgeport used to have 93% of False contrabands, and now has only 62%.Similar situation with Danbury and New Haven. Why? Hard to say. There are really a lot of variables here. Maybe the departments got instructed on how to ... | common_location = df['InterventionLocationName'].value_counts().head(20).index
locations_new = df[df['InterventionLocationName'].isin(common_location)]
locations_old = old_df[old_df['InterventionLocationName'].isin(common_location)]
pd.crosstab(locations_new['ContrabandIndicator'], locations_new['InterventionLocationNa... | _____no_output_____ | MIT | S06 - DS in the Real World/BLU15 - Model CSI/BLU15 - Learning Unit - Model CSI.ipynb | LDSSA/batch4-students |
What do we see? First of all, the InterventionLocationName and the Department Name are often same.It sounds pretty logic, as probably policeman's usually work in the area of their department. But we could try to create a feature saying whether InterventionLocationName is equal to the Department Name.Or maybe we could j... | pd.crosstab(df['ContrabandIndicator'], df['InterventionReasonCode'], normalize="columns")
pd.crosstab(old_df['ContrabandIndicator'], old_df['InterventionReasonCode'], normalize="columns") | _____no_output_____ | MIT | S06 - DS in the Real World/BLU15 - Model CSI/BLU15 - Learning Unit - Model CSI.ipynb | LDSSA/batch4-students |
There are some small changes, but they don't seem to be significant. Especially that all the 3 values have around 33% of Contraband.Time for officers: | df['ReportingOfficerIdentificationID'].value_counts()
filter_values(df, 'ReportingOfficerIdentificationID', 2)['ReportingOfficerIdentificationID'].nunique() | _____no_output_____ | MIT | S06 - DS in the Real World/BLU15 - Model CSI/BLU15 - Learning Unit - Model CSI.ipynb | LDSSA/batch4-students |
Well, looks like there are a lot of unique values for the officer id (1166 for 2000 records), and there are not so many common ones (only 206 officers have more than 2 rows in the dataset) so it doesn't make much sense to analyze it. Let's quickly go throw the rest of the columns: | df.columns
rest = ['ResidentIndicator', 'SearchAuthorizationCode',
'StatuteReason', 'SubjectEthnicityCode',
'SubjectRaceCode', 'SubjectSexCode','TownResidentIndicator']
for col in rest:
display(pd.crosstab(df['ContrabandIndicator'], df[col], normalize="columns"))
display(pd.crosstab(old_df['Con... | _____no_output_____ | MIT | S06 - DS in the Real World/BLU15 - Model CSI/BLU15 - Learning Unit - Model CSI.ipynb | LDSSA/batch4-students |
Reviewing Automated Machine Learning ExplanationsAs machine learning becomes more and more and more prevelant, the predictions made by models have greater influence over many aspects of our society. For example, machine learning models are an increasingly significant factor in how banks decide to grant loans or doctor... | import azureml.core
from azureml.core import Workspace
# Load the workspace from the saved config file
ws = Workspace.from_config()
print('Ready to use Azure ML {} to work with {}'.format(azureml.core.VERSION, ws.name)) | _____no_output_____ | MIT | 09A - Reviewing Automated Machine Learning Explanations.ipynb | LucaSavio/DP100 |
Run an Automated Machine Learning ExperimentTo reduce time in this lab, you'll run an automated machine learning experiment with only three iterations.Note that the **model_explainability** configuration option is set to **True**. | import pandas as pd
from azureml.train.automl import AutoMLConfig
from azureml.core.experiment import Experiment
from azureml.widgets import RunDetails
from azureml.core.compute import ComputeTarget, AmlCompute
from azureml.core.compute_target import ComputeTargetException
from azureml.core import Dataset
cluster_name... | _____no_output_____ | MIT | 09A - Reviewing Automated Machine Learning Explanations.ipynb | LucaSavio/DP100 |
View Feature ImportanceWhen the experiment has completed in the widget above, click the run that produced the best result to see its details. Then scroll to the bottom of the visualizations to see the relative feature importance.You can also view feature importance for the best model produced by the experiment by usin... | from azureml.contrib.interpret.explanation.explanation_client import ExplanationClient
from azureml.core.run import Run
# Wait for the best model explanation run to complete
model_explainability_run_id = automl_run.get_properties().get('ModelExplainRunId')
print(model_explainability_run_id)
if model_explainability_run... | _____no_output_____ | MIT | 09A - Reviewing Automated Machine Learning Explanations.ipynb | LucaSavio/DP100 |
Classifying Fashion-MNISTNow it's your turn to build and train a neural network. You'll be using the [Fashion-MNIST dataset](https://github.com/zalandoresearch/fashion-mnist), a drop-in replacement for the MNIST dataset. MNIST is actually quite trivial with neural networks where you can easily achieve better than 97% ... | import torch
from torchvision import datasets, transforms
import helper
# Define a transform to normalize the data
transform = transforms.Compose([transforms.ToTensor(),
transforms.Normalize((0.5,), (0.5,))])
# Download and load the training data
trainset = datasets.FashionMNIST('~/.pyt... | _____no_output_____ | MIT | intro-to-pytorch/Part 4 - Fashion-MNIST (Solution).ipynb | sizigia/deep-learning-v2-pytorch |
Here we can see one of the images. | image, label = next(iter(trainloader))
helper.imshow(image[0,:]); | _____no_output_____ | MIT | intro-to-pytorch/Part 4 - Fashion-MNIST (Solution).ipynb | sizigia/deep-learning-v2-pytorch |
Building the networkHere you should define your network. As with MNIST, each image is 28x28 which is a total of 784 pixels, and there are 10 classes. You should include at least one hidden layer. We suggest you use ReLU activations for the layers and to return the logits or log-softmax from the forward pass. It's up t... | from torch import nn, optim
import torch.nn.functional as F
# TODO: Define your network architecture here
class Classifier(nn.Module):
def __init__(self):
super().__init__()
self.fc1 = nn.Linear(784, 256)
self.fc2 = nn.Linear(256, 128)
self.fc3 = nn.Linear(128, 64)
self.fc4 =... | _____no_output_____ | MIT | intro-to-pytorch/Part 4 - Fashion-MNIST (Solution).ipynb | sizigia/deep-learning-v2-pytorch |
Train the networkNow you should create your network and train it. First you'll want to define [the criterion](http://pytorch.org/docs/master/nn.htmlloss-functions) (something like `nn.CrossEntropyLoss` or `nn.NLLLoss`) and [the optimizer](http://pytorch.org/docs/master/optim.html) (typically `optim.SGD` or `optim.Adam... | # TODO: Create the network, define the criterion and optimizer
model = Classifier()
criterion = nn.NLLLoss()
optimizer = optim.Adam(model.parameters(), lr=0.003)
# TODO: Train the network here
epochs = 5
for e in range(epochs):
running_loss = 0
for images, labels in trainloader:
log_ps = model(images)
... | _____no_output_____ | MIT | intro-to-pytorch/Part 4 - Fashion-MNIST (Solution).ipynb | sizigia/deep-learning-v2-pytorch |
Einstein Tensor calculations using Symbolic module | import numpy as np
import pytest
import sympy
from sympy import cos, simplify, sin, sinh, tensorcontraction
from einsteinpy.symbolic import EinsteinTensor, MetricTensor, RicciScalar
sympy.init_printing() | _____no_output_____ | Apache-2.0 | EinsteinPy/Einstein Tensor symbolic calculation.ipynb | IsaacW4/Advanced-GR |
Defining the Anti-de Sitter spacetime Metric | syms = sympy.symbols("t chi theta phi")
t, ch, th, ph = syms
m = sympy.diag(-1, cos(t) ** 2, cos(t) ** 2 * sinh(ch) ** 2, cos(t) ** 2 * sinh(ch) ** 2 * sin(th) ** 2).tolist()
metric = MetricTensor(m, syms) | _____no_output_____ | Apache-2.0 | EinsteinPy/Einstein Tensor symbolic calculation.ipynb | IsaacW4/Advanced-GR |
Calculating the Einstein Tensor (with both indices covariant) | einst = EinsteinTensor.from_metric(metric)
einst.tensor() | _____no_output_____ | Apache-2.0 | EinsteinPy/Einstein Tensor symbolic calculation.ipynb | IsaacW4/Advanced-GR |
Ex1 - Filtering and Sorting Data This time we are going to pull data directly from the internet.Special thanks to: https://github.com/justmarkham for sharing the dataset and materials. Step 1. Import the necessary libraries | import pandas as pd | _____no_output_____ | BSD-3-Clause | 02_Filtering_&_Sorting/Chipotle/Exercises_with_solutions.ipynb | duongv/pandas_exercises |
Step 2. Import the dataset from this [address](https://raw.githubusercontent.com/justmarkham/DAT8/master/data/chipotle.tsv). Step 3. Assign it to a variable called chipo. | url = 'https://raw.githubusercontent.com/justmarkham/DAT8/master/data/chipotle.tsv'
chipo = pd.read_csv(url, sep = '\t') | _____no_output_____ | BSD-3-Clause | 02_Filtering_&_Sorting/Chipotle/Exercises_with_solutions.ipynb | duongv/pandas_exercises |
Step 4. How many products cost more than $10.00? | # clean the item_price column and transform it in a float
prices = [float(value[1 : -1]) for value in chipo.item_price]
# reassign the column with the cleaned prices
chipo.item_price = prices
# delete the duplicates in item_name and quantity
chipo_filtered = chipo.drop_duplicates(['item_name','quantity'])
# select o... | _____no_output_____ | BSD-3-Clause | 02_Filtering_&_Sorting/Chipotle/Exercises_with_solutions.ipynb | duongv/pandas_exercises |
Step 5. What is the price of each item? print a data frame with only two columns item_name and item_price | # delete the duplicates in item_name and quantity
# chipo_filtered = chipo.drop_duplicates(['item_name','quantity'])
chipo[(chipo['item_name'] == 'Chicken Bowl') & (chipo['quantity'] == 1)]
# select only the products with quantity equals to 1
# chipo_one_prod = chipo_filtered[chipo_filtered.quantity == 1]
# select on... | _____no_output_____ | BSD-3-Clause | 02_Filtering_&_Sorting/Chipotle/Exercises_with_solutions.ipynb | duongv/pandas_exercises |
Step 6. Sort by the name of the item | chipo.item_name.sort_values()
# OR
chipo.sort_values(by = "item_name") | _____no_output_____ | BSD-3-Clause | 02_Filtering_&_Sorting/Chipotle/Exercises_with_solutions.ipynb | duongv/pandas_exercises |
Step 7. What was the quantity of the most expensive item ordered? | chipo.sort_values(by = "item_price", ascending = False).head(1) | _____no_output_____ | BSD-3-Clause | 02_Filtering_&_Sorting/Chipotle/Exercises_with_solutions.ipynb | duongv/pandas_exercises |
Step 8. How many times were a Veggie Salad Bowl ordered? | chipo_salad = chipo[chipo.item_name == "Veggie Salad Bowl"]
len(chipo_salad) | _____no_output_____ | BSD-3-Clause | 02_Filtering_&_Sorting/Chipotle/Exercises_with_solutions.ipynb | duongv/pandas_exercises |
Step 9. How many times people orderd more than one Canned Soda? | chipo_drink_steak_bowl = chipo[(chipo.item_name == "Canned Soda") & (chipo.quantity > 1)]
len(chipo_drink_steak_bowl) | _____no_output_____ | BSD-3-Clause | 02_Filtering_&_Sorting/Chipotle/Exercises_with_solutions.ipynb | duongv/pandas_exercises |
window.dataLayer = window.dataLayer || []; function gtag(){dataLayer.push(arguments);} gtag('js', new Date()); gtag('config', 'UA-59152712-8'); Start-to-Finish Example: Unit Testing `GiRaFFE_NRPy`: $A_k$ to $B^i$ Author: Patrick Nelson This module Validates the A-to-B routine for `GiRaFFE`.**Notebook Status:** Val... | import shutil, os, sys # Standard Python modules for multiplatform OS-level functions
# First, we'll add the parent directory to the list of directories Python will check for modules.
nrpy_dir_path = os.path.join("..")
if nrpy_dir_path not in sys.path:
sys.path.append(nrpy_dir_path)
from outputC import *... | _____no_output_____ | BSD-2-Clause | in_progress/Tutorial-Start_to_Finish_UnitTest-GiRaFFE_NRPy-A2B.ipynb | Steve-Hawk/nrpytutorial |
Step 1.a: Polynomial vector potential \[Back to [top](toc)\]$$\label{polynomial}$$We will start with the simplest case - testing the second-order solver. In second-order finite-differencing, we use a three-point stencil that can exactly differentiate polynomials up to quadratic. So, we will use cubic functions three v... | if not Use_Gaussian_Data:
is_gaussian = par.Cparameters("int",thismodule,"is_gaussian",0)
par.set_parval_from_str("reference_metric::CoordSystem","Cartesian")
rfm.reference_metric()
x = rfm.xxCart[0]
y = rfm.xxCart[1]
z = rfm.xxCart[2]
AD[0] = a*x**3 + b*y**3 + c*z**3 + d*y**2 + e*z**2 + f... | _____no_output_____ | BSD-2-Clause | in_progress/Tutorial-Start_to_Finish_UnitTest-GiRaFFE_NRPy-A2B.ipynb | Steve-Hawk/nrpytutorial |
Step 1.b: Gaussian vector potential \[Back to [top](toc)\]$$\label{gaussian}$$Alternatively, we might want to use different functions for the vector potential. Here, we'll give some 3D Gaussians:\begin{align}A_x &= a e^{-((x-b)^2+(y-c)^2+(z-d)^2)} \\A_y &= f e^{-((x-g)^2+(y-h)^2+(z-l)^2)} \\A_z &= m e^{-((x-n)^2+(y-o)... | if Use_Gaussian_Data:
is_gaussian = par.Cparameters("int",thismodule,"is_gaussian",1)
par.set_parval_from_str("reference_metric::CoordSystem","Cartesian")
rfm.reference_metric()
x = rfm.xxCart[0]
y = rfm.xxCart[1]
z = rfm.xxCart[2]
AD[0] = a * sp.exp(-((x-b)**2 + (y-c)**2 + (z-d)**2))
... | _____no_output_____ | BSD-2-Clause | in_progress/Tutorial-Start_to_Finish_UnitTest-GiRaFFE_NRPy-A2B.ipynb | Steve-Hawk/nrpytutorial |
Step 1.c: The magnetic field $B^i$ \[Back to [top](toc)\]$$\label{magnetic}$$Next, we'll let NRPy+ compute derivatives analytically according to $$B^i = \frac{[ijk]}{\sqrt{\gamma}} \partial_j A_k.$$ Then we can carry out two separate tests to verify the numerical derivatives. First, we will verify that when we let the... | par.set_parval_from_str("reference_metric::CoordSystem","Cartesian")
rfm.reference_metric()
x = rfm.xxCart[0]
y = rfm.xxCart[1]
z = rfm.xxCart[2]
gammaDD[0][0] = a*x**3 + b*y**3 + c*z**3 + d*y**2 + e*z**2 + sp.sympify(1)
gammaDD[1][1] = g*x**3 + h*y**3 + l*z**3 + m*x**2 + n*z**2 + sp.sympify(1)
gammaDD[2][2] = p*x**3 ... | Output C function calculate_metric_gfs() to file Validation/calculate_metric_gfs.h
| BSD-2-Clause | in_progress/Tutorial-Start_to_Finish_UnitTest-GiRaFFE_NRPy-A2B.ipynb | Steve-Hawk/nrpytutorial |
We also should write a function that will use the analytic formulae for $B^i$. | B_analyticU_to_print = [\
lhrh(lhs=gri.gfaccess("out_gfs","B_analyticU0"),rhs=B_analyticU[0]),\
lhrh(lhs=gri.gfaccess("out_gfs","B_analyticU1"),rhs=B_analyticU[1]),\
lhrh(lhs=gri.gfaccess("out_gfs","B_analyticU2"),rhs=B_analyticU[2]),\
... | Output C function calculate_exact_BU() to file Validation/calculate_exact_BU.h
| BSD-2-Clause | in_progress/Tutorial-Start_to_Finish_UnitTest-GiRaFFE_NRPy-A2B.ipynb | Steve-Hawk/nrpytutorial |
Step 1.d: The vector potential $A_k$ \[Back to [top](toc)\]$$\label{vector_potential}$$We'll now write a function to set the vector potential $A_k$. This simply uses NRPy+ to generate most of the code from the expressions we wrote at the beginning. Then, we'll need to call the function from the module `GiRaFFE_NRPy_A2... | AD_to_print = [\
lhrh(lhs=gri.gfaccess("out_gfs","AD0"),rhs=AD[0]),\
lhrh(lhs=gri.gfaccess("out_gfs","AD1"),rhs=AD[1]),\
lhrh(lhs=gri.gfaccess("out_gfs","AD2"),rhs=AD[2]),\
]
desc = "Calculate the vector potential"
name = "calculate_AD"
outCfunction(
outfi... | Output C function calculate_AD() to file Validation/calculate_AD.h
| BSD-2-Clause | in_progress/Tutorial-Start_to_Finish_UnitTest-GiRaFFE_NRPy-A2B.ipynb | Steve-Hawk/nrpytutorial |
Step 1.e: Set free parameters in the code \[Back to [top](toc)\]$$\label{free_parameters}$$We also need to create the files that interact with NRPy's C parameter interface. | # Step 3.d.i: Generate declare_Cparameters_struct.h, set_Cparameters_default.h, and set_Cparameters[-SIMD].h
# par.generate_Cparameters_Ccodes(os.path.join(out_dir))
# Step 3.d.ii: Set free_parameters.h
with open(os.path.join(out_dir,"free_parameters.h"),"w") as file:
file.write("""
// Override parameter defaults ... | _____no_output_____ | BSD-2-Clause | in_progress/Tutorial-Start_to_Finish_UnitTest-GiRaFFE_NRPy-A2B.ipynb | Steve-Hawk/nrpytutorial |
Step 2: `A2B_unit_test.c`: The Main C Code \[Back to [top](toc)\]$$\label{mainc}$$Now that we have our vector potential and analytic magnetic field to compare against, we will start writing our unit test. We'll also import common C functionality, define `REAL`, the number of ghost zones, and the faces, and set the sta... | %%writefile $out_dir/A2B_unit_test.c
// These are common packages that we are likely to need.
#include "stdio.h"
#include "stdlib.h"
#include "math.h"
#include "string.h" // Needed for strncmp, etc.
#include "stdint.h" // Needed for Windows GCC 6.x compatibility
#include <time.h> // Needed to set a random seed.
#def... | Overwriting Validation//A2B_unit_test.c
| BSD-2-Clause | in_progress/Tutorial-Start_to_Finish_UnitTest-GiRaFFE_NRPy-A2B.ipynb | Steve-Hawk/nrpytutorial |
We'll now define the gridfunction names. | %%writefile -a $out_dir/A2B_unit_test.c
// Let's also #define the NRPy+ gridfunctions
#define AD0GF 0
#define AD1GF 1
#define AD2GF 2
#define NUM_EVOL_GFS 3
#define GAMMADD00GF 0
#define GAMMADD01GF 1
#define GAMMADD02GF 2
#define GAMMADD11GF 3
#define GAMMADD12GF 4
#define GAMMADD22GF 5
#define B_ANALYTICU0GF 6
#defi... | Appending to Validation//A2B_unit_test.c
| BSD-2-Clause | in_progress/Tutorial-Start_to_Finish_UnitTest-GiRaFFE_NRPy-A2B.ipynb | Steve-Hawk/nrpytutorial |
Now, we'll handle the different A2B codes. There are several things to do here. First, we'll add `include`s to the C code so that we have access to the functions we want to test, as generated above. We will choose to do this in the subfolder `A2B` relative to this tutorial. | %%writefile -a $out_dir/A2B_unit_test.c
#include "A2B/driver_AtoB.h" // This file contains both functions we need.
#include "calculate_exact_BU.h"
#include "calculate_AD.h"
#include "calculate_metric_gfs.h"
| Appending to Validation//A2B_unit_test.c
| BSD-2-Clause | in_progress/Tutorial-Start_to_Finish_UnitTest-GiRaFFE_NRPy-A2B.ipynb | Steve-Hawk/nrpytutorial |
Now, we'll write the main method. First, we'll set up the grid. In this test, we cannot use only one point. As we are testing a three-point stencil, we can get away with a minimal $3 \times 3 \times 3$ grid. Then, we'll write the A fields. After that, we'll calculate the magnetic field two ways. | %%writefile -a $out_dir/A2B_unit_test.c
int main(int argc, const char *argv[]) {
paramstruct params;
#include "set_Cparameters_default.h"
// Let the last argument be the test we're doing. 1 = coarser grid, 0 = finer grid.
int do_quadratic_test = atoi(argv[4]);
// Step 0c: Set free parameters, over... | Appending to Validation//A2B_unit_test.c
| BSD-2-Clause | in_progress/Tutorial-Start_to_Finish_UnitTest-GiRaFFE_NRPy-A2B.ipynb | Steve-Hawk/nrpytutorial |
Step 2.a: Compile and run the code \[Back to [top](toc)\]$$\label{compile_run}$$Now that we have our file, we can compile it and run the executable. | import time
print("Now compiling, should take ~2 seconds...\n")
start = time.time()
cmd.C_compile(os.path.join(out_dir,"A2B_unit_test.c"), os.path.join(out_dir,"A2B_unit_test"))
end = time.time()
print("Finished in "+str(end-start)+" seconds.\n\n")
print("Now running...\n")
start = time.time()
!./Validation/A2B_unit_... | Now compiling, should take ~2 seconds...
Compiling executable...
Executing `gcc -Ofast -fopenmp -march=native -funroll-loops Validation/A2B_unit_test.c -o Validation/A2B_unit_test -lm`...
Finished executing in 0.6135389804840088 seconds.
Finished compilation.
Finished in 0.6216833591461182 seconds.
Now running...
d... | BSD-2-Clause | in_progress/Tutorial-Start_to_Finish_UnitTest-GiRaFFE_NRPy-A2B.ipynb | Steve-Hawk/nrpytutorial |
Step 3: Code validation: Verify that relative error in numerical solution converges to zero at the expected order \[Back to [top](toc)\]$$\label{convergence}$$Now that we have shown that when we use a quadratic vector potential, we get roundoff-level agreement (which is to be expected, since the finite-differencing us... | import numpy as np
import matplotlib.pyplot as plt
Data1 = np.loadtxt("out1-numer.txt")
Data2 = np.loadtxt("out7-numer.txt")
# print("Convergence test: All should be approximately 2\n")
# convergence = np.log(np.divide(np.abs(Data1),np.abs(Data2)))/np.log(2)
# for i in range(len(convergence[:,0])):
# print(converg... | Face | Res | L2 norm | Conv. Order
Int | Dx | 0.0000005 | --
-- | Dx/2 | 0.0000001 | 2.03057
-x | Dx | 0.0000008 | --
-- | Dx/2 | 0.0000002 | 2.08857
+x | Dx | 0.0000008 | --
-- | Dx/2 | 0.0000002 | 2.08857
-y | Dx | 0.0000008 | --
-- | Dx/2 | 0.0000002 | 1.64224
+y | Dx | 0.0000016 | ... | BSD-2-Clause | in_progress/Tutorial-Start_to_Finish_UnitTest-GiRaFFE_NRPy-A2B.ipynb | Steve-Hawk/nrpytutorial |
Step 4: Output this notebook to $\LaTeX$-formatted PDF file \[Back to [top](toc)\]$$\label{latex_pdf_output}$$The following code cell converts this Jupyter notebook into a proper, clickable $\LaTeX$-formatted PDF file. After the cell is successfully run, the generated PDF may be found in the root NRPy+ tutorial direct... | !jupyter nbconvert --to latex --template latex_nrpy_style.tplx --log-level='WARN' Tutorial-Start_to_Finish_UnitTest-GiRaFFE_NRPy-A2B.ipynb
!pdflatex -interaction=batchmode Tutorial-Start_to_Finish_UnitTest-GiRaFFE_NRPy-A2B.tex
!pdflatex -interaction=batchmode Tutorial-Start_to_Finish_UnitTest-GiRaFFE_NRPy-A2B.tex
!pdfl... | This is pdfTeX, Version 3.14159265-2.6-1.40.18 (TeX Live 2017/Debian) (preloaded format=pdflatex)
restricted \write18 enabled.
entering extended mode
This is pdfTeX, Version 3.14159265-2.6-1.40.18 (TeX Live 2017/Debian) (preloaded format=pdflatex)
restricted \write18 enabled.
entering extended mode
This is pdfTeX, Ve... | BSD-2-Clause | in_progress/Tutorial-Start_to_Finish_UnitTest-GiRaFFE_NRPy-A2B.ipynb | Steve-Hawk/nrpytutorial |
http://www.yr.no/place/Norway/Telemark/Vinje/Haukeliseter/climate.month12.html | import matplotlib.pyplot as plt
import matplotlib.dates as dates
import numpy as np
import csv
import pandas as pd
import datetime
from datetime import date
import calendar
%matplotlib inline
year = np.arange(2000,2017, 1)
T_av = [-4.1,\
-8.2,\
-10.7,\
-4.3,\
-4.1,\
-5.5,\
-0... | _____no_output_____ | MIT | yr_Dec_clim_2000_2016.ipynb | franzihe/Python_Masterthesis |
DB2 Jupyter Notebook ExtensionsVersion: 2021-08-23 This code is imported as a Jupyter notebook extension in any notebooks you create with DB2 code in it. Place the following line of code in any notebook that you want to use these commands with:&37;run db2.ipynbThis code defines a Jupyter/Python magic command called `%... | #
# Set up Jupyter MAGIC commands "sql".
# %sql will return results from a DB2 select statement or execute a DB2 command
#
# IBM 2021: George Baklarz
# Version 2021-07-13
#
from __future__ import print_function
from IPython.display import HTML as pHTML, Image as pImage, display as pdisplay, Javascript as Javascript
f... | _____no_output_____ | Apache-2.0 | db2.ipynb | Db2-DTE-POC/Db2-Openshift-11.5.4 |
OptionsThere are four options that can be set with the **`%sql`** command. These options are shown below with the default value shown in parenthesis.- **`MAXROWS n (10)`** - The maximum number of rows that will be displayed before summary information is shown. If the answer set is less than this number of rows, it wil... | def setOptions(inSQL):
global _settings, _display
cParms = inSQL.split()
cnt = 0
while cnt < len(cParms):
if cParms[cnt].upper() == 'MAXROWS':
if cnt+1 < len(cParms):
try:
_settings["maxrows"] = int(cParms[cnt+1])
ex... | _____no_output_____ | Apache-2.0 | db2.ipynb | Db2-DTE-POC/Db2-Openshift-11.5.4 |
SQL HelpThe calling format of this routine is:```sqlhelp()```This code displays help related to the %sql magic command. This help is displayed when you issue a %sql or %%sql command by itself, or use the %sql -h flag. | def sqlhelp():
global _environment
if (_environment["jupyter"] == True):
sd = '<td style="text-align:left;">'
ed1 = '</td>'
ed2 = '</td>'
sh = '<th style="text-align:left;">'
eh1 = '</th>'
eh2 = '</th>'
sr = '<tr>'
er = '</tr>'
... | _____no_output_____ | Apache-2.0 | db2.ipynb | Db2-DTE-POC/Db2-Openshift-11.5.4 |
Connection HelpThe calling format of this routine is:```connected_help()```This code displays help related to the CONNECT command. This code is displayed when you issue a %sql CONNECT command with no arguments or you are running a SQL statement and there isn't any connection to a database yet. | def connected_help():
sd = '<td style="text-align:left;">'
ed = '</td>'
sh = '<th style="text-align:left;">'
eh = '</th>'
sr = '<tr>'
er = '</tr>'
if (_environment['jupyter'] == True):
helpConnect = """
<h3>Connecting to Db2</h3>
<p>The CONNECT c... | _____no_output_____ | Apache-2.0 | db2.ipynb | Db2-DTE-POC/Db2-Openshift-11.5.4 |
Prompt for Connection InformationIf you are running an SQL statement and have not yet connected to a database, the %sql command will prompt you for connection information. In order to connect to a database, you must supply:- Database name - Host name (IP address or name)- Port number- Userid- Password- Secure socketTh... | # Prompt for Connection information
def connected_prompt():
global _settings
_database = ''
_hostname = ''
_port = ''
_uid = ''
_pwd = ''
_ssl = ''
print("Enter the database connection details (Any empty value will cancel the connection)")
_database = input("Enter the... | _____no_output_____ | Apache-2.0 | db2.ipynb | Db2-DTE-POC/Db2-Openshift-11.5.4 |
Connect Syntax ParserThe parseConnect routine is used to parse the CONNECT command that the user issued within the %sql command. The format of the command is:```parseConnect(inSQL)```The inSQL string contains the CONNECT keyword with some additional parameters. The format of the CONNECT command is one of:```CONNECT RE... | # Parse the CONNECT statement and execute if possible
def parseConnect(inSQL,local_ns):
global _settings, _connected
_connected = False
cParms = inSQL.split()
cnt = 0
_settings["ssl"] = ""
while cnt < len(cParms):
if cParms[cnt].upper() == 'TO':
if cnt+... | _____no_output_____ | Apache-2.0 | db2.ipynb | Db2-DTE-POC/Db2-Openshift-11.5.4 |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.