Merge pull request #240 from antoinecarme/numpy_20_and_pandas_30_impact

Numpy 20 and pandas 30 impact
antoinecarme · Jun 16, 2024 · 942b221 · 942b221
2 parents 9d90dba + 7191dfb
commit 942b221
Show file tree

Hide file tree

Showing 155 changed files with 10,392 additions and 10,955 deletions.
diff --git a/pyaf/TS/Intermittent_Models.py b/pyaf/TS/Intermittent_Models.py
@@ -113,7 +113,7 @@ def compute_forecast(self, df, alpha, method, horizon_index = 1):
         for h in range(horizon_index):
             lCounts_df.loc[-(h+1), self.mCycleResidueName] = None
         df3 = df1.merge(df2 , how='left', on=('index' , 'index'))
-        df4 = df3.fillna(method='ffill')
+        df4 = df3.ffill()
         # fill first empty fit data with zero counts (when signal starts with zeros)
         i = 0
         while(np.isnan(df4.loc[i , 'forecast'])):

diff --git a/pyaf/TS/MissingData.py b/pyaf/TS/MissingData.py
@@ -84,9 +84,9 @@ def apply_signal_imputation_method(self, iInputDS, iSignal):
             iInputDS[iSignal] = lSignal
 
         elif(lMethod == "PreviousValue"):
-            lSignal = iInputDS[iSignal].fillna(method='ffill')
+            lSignal = iInputDS[iSignal].ffill()
             # replace the first empty values with the first known value
-            lSignal = lSignal.fillna(method='bfill')
+            lSignal = lSignal.bfill()
             iInputDS[iSignal] = lSignal
 
         return iInputDS

diff --git a/pyaf/TS/Perf.py b/pyaf/TS/Perf.py
@@ -157,7 +157,7 @@ def compute_R2(self, signal , estimator):
     def compute_LnQ(self, signal , estimator):
         min_signal , min_estimator = signal.min() , estimator.min()
         # return +inf if the signals are not strictly positive (discard the model)
-        self.mLnQ = np.Inf
+        self.mLnQ = np.inf
         if(min_signal > 0.0 and min_estimator > 0.0):
             log_diff = np.log(estimator) - np.log(signal)
             self.mLnQ = np.sum(log_diff * log_diff)

diff --git a/pyaf/TS/SignalDecomposition_Cycle.py b/pyaf/TS/SignalDecomposition_Cycle.py
@@ -220,7 +220,6 @@ def fit(self):
         self.mEncodedValueDict = self.compute_target_means_by_cycle_value(self.mCycleFrame, self.getCycleName())
 
         self.mCycleFrame[lName + '_enc'] = self.mCycleFrame[lName].map(self.mEncodedValueDict).fillna(self.mDefaultValue)
-        self.mCycleFrame[lName + '_enc'].fillna(self.mDefaultValue, inplace=True);
         self.mCycleFrame[lName + '_NotEncoded'] = self.mCycleFrame[lName];
         self.mCycleFrame[lName] = self.mCycleFrame[lName + '_enc'];
 

diff --git a/pyaf/TS/SignalDecomposition_Trend.py b/pyaf/TS/SignalDecomposition_Trend.py
@@ -145,7 +145,7 @@ def fit_specific(self):
 
 
     def compute(self, df):
-        Y_pred = df[self.mSignal].shift(1).rolling(self.mWindow).mean().fillna(method='bfill');
+        Y_pred = df[self.mSignal].shift(1).rolling(self.mWindow).mean().ffill();
         Y_pred.fillna(self.mMean , inplace=True)
         return Y_pred
 
@@ -168,7 +168,7 @@ def fit_specific(self):
 
 
     def compute(self, df):
-        Y_pred = df[self.mSignal].shift(1).rolling(self.mWindow).median().fillna(method='bfill');
+        Y_pred = df[self.mSignal].shift(1).rolling(self.mWindow).median().ffill();
         Y_pred.fillna(self.mMean , inplace=True)
         return Y_pred
 

diff --git a/tests/basic_checks/test_temporal_demo_hourly_D_H_ErrorMessage.py b/tests/basic_checks/test_temporal_demo_hourly_D_H_ErrorMessage.py
@@ -4,8 +4,8 @@
 import numpy as np
 import pandas as pd
 
-DATA_FREQ = 'H'
-PERIODS = ["D" , "H"]
+DATA_FREQ = 'h'
+PERIODS = ["D" , "h"]
 H = 365
 N = H * 10
 lDateColumn = "Date"

diff --git a/tests/basic_checks/test_temporal_demo_hourly_D_H_ErrorMessage_2.py b/tests/basic_checks/test_temporal_demo_hourly_D_H_ErrorMessage_2.py
@@ -4,8 +4,8 @@
 import numpy as np
 import pandas as pd
 
-DATA_FREQ = 'H'
-PERIODS = ["D" , "H"]
+DATA_FREQ = 'h'
+PERIODS = ["D" , "h"]
 H = 365
 N = H * 10
 lDateColumn = "Date"

diff --git a/tests/basic_checks/test_temporal_demo_hourly_D_H_ErrorMessage_3.py b/tests/basic_checks/test_temporal_demo_hourly_D_H_ErrorMessage_3.py
@@ -4,8 +4,8 @@
 import numpy as np
 import pandas as pd
 
-DATA_FREQ = 'H'
-PERIODS = ["T" , "H"]
+DATA_FREQ = 'h'
+PERIODS = ["min" , "h"]
 H = 365
 N = H * 10
 lDateColumn = "Date"

diff --git a/tests/bugs/issue_179/issue_179_cumprod_overflow.py b/tests/bugs/issue_179/issue_179_cumprod_overflow.py
@@ -17,7 +17,7 @@
 N = 1000
 lTimeVar = 'Time'
 lSignalVar = 'Signal'
-df[lTimeVar] = pd.date_range("2018-01-01", periods=N, freq="H")
+df[lTimeVar] = pd.date_range("2018-01-01", periods=N, freq="h")
 df[lSignalVar] = np.random.random(df.shape[0]) 
 df.info()
 print(df.head())

diff --git a/tests/bugs/issue_216/issue_216_prediction_interval_shaded_area.py b/tests/bugs/issue_216/issue_216_prediction_interval_shaded_area.py
@@ -5,7 +5,7 @@
     # generate a daily signal covering one year 2016 in a pandas dataframe
     N = 365
     np.random.seed(seed=1960)
-    df_train = pd.DataFrame({"Date" : pd.date_range(start="2016-01-25", periods=N, freq='H'),
+    df_train = pd.DataFrame({"Date" : pd.date_range(start="2016-01-25", periods=N, freq='h'),
                              "Signal" : (np.arange(N)//40 + np.arange(N) % 21 + np.random.randn(N))})
     # print(df_train.head(N))
 

diff --git a/tests/bugs/issue_70/test_artificial_filter_seasonals_day.py b/tests/bugs/issue_70/test_artificial_filter_seasonals_day.py
@@ -7,7 +7,7 @@
 
 b1 = tsds.generate_random_TS(N = 4000 , FREQ = 'D', seed = 0, trendtype = "constant", cycle_length = 24, transform = "None", sigma = 0.1, exog_count = 0);
 
-df = b1.mPastData
+df = b1.mPastData.copy()
 df[b1.mSignalVar] = df[b1.mName]
 
 lEngine = autof.cForecastEngine()

diff --git a/tests/bugs/issue_70/test_artificial_filter_seasonals_hour.py b/tests/bugs/issue_70/test_artificial_filter_seasonals_hour.py
@@ -7,7 +7,7 @@
 
 b1 = tsds.generate_random_TS(N = 4000 , FREQ = 'H', seed = 0, trendtype = "constant", cycle_length = 24, transform = "None", sigma = 0.1, exog_count = 0);
 
-df = b1.mPastData
+df = b1.mPastData.copy()
 df[b1.mSignalVar] = df[b1.mName]
 
 lEngine = autof.cForecastEngine()

diff --git a/tests/bugs/issue_70/test_artificial_filter_seasonals_min.py b/tests/bugs/issue_70/test_artificial_filter_seasonals_min.py
@@ -7,7 +7,7 @@
 
 b1 = tsds.generate_random_TS(N = 40000 , FREQ = 'min', seed = 0, trendtype = "constant", cycle_length = 24, transform = "None", sigma = 0.1, exog_count = 0);
 
-df = b1.mPastData
+df = b1.mPastData.copy()
 df[b1.mSignalVar] = df[b1.mName]
 
 lEngine = autof.cForecastEngine()

diff --git a/tests/bugs/issue_70/test_artificial_filter_seasonals_second.py b/tests/bugs/issue_70/test_artificial_filter_seasonals_second.py
@@ -7,7 +7,7 @@
 
 b1 = tsds.generate_random_TS(N = 40000 , FREQ = 'S', seed = 0, trendtype = "constant", cycle_length = 24, transform = "None", sigma = 0.1, exog_count = 0);
 
-df = b1.mPastData
+df = b1.mPastData.copy()
 df[b1.mSignalVar] = df[b1.mName]
 
 lEngine = autof.cForecastEngine()

diff --git a/tests/bugs/issue_70/test_artificial_keep_all_seasonals_day.py b/tests/bugs/issue_70/test_artificial_keep_all_seasonals_day.py
@@ -7,7 +7,7 @@
 
 b1 = tsds.generate_random_TS(N = 4000 , FREQ = 'D', seed = 0, trendtype = "constant", cycle_length = 24, transform = "None", sigma = 0.1, exog_count = 0);
 
-df = b1.mPastData
+df = b1.mPastData.copy()
 df[b1.mSignalVar] = df[b1.mName]
 
 lEngine = autof.cForecastEngine()

diff --git a/tests/bugs/issue_70/test_artificial_keep_all_seasonals_hour.py b/tests/bugs/issue_70/test_artificial_keep_all_seasonals_hour.py
@@ -7,7 +7,7 @@
 
 b1 = tsds.generate_random_TS(N = 4000 , FREQ = 'H', seed = 0, trendtype = "constant", cycle_length = 24, transform = "None", sigma = 0.1, exog_count = 0);
 
-df = b1.mPastData
+df = b1.mPastData.copy()
 df[b1.mSignalVar] = df[b1.mName]
 
 lEngine = autof.cForecastEngine()

diff --git a/tests/bugs/issue_70/test_artificial_keep_all_seasonals_minute.py b/tests/bugs/issue_70/test_artificial_keep_all_seasonals_minute.py
@@ -7,7 +7,7 @@
 
 b1 = tsds.generate_random_TS(N = 40000 , FREQ = 'min', seed = 0, trendtype = "constant", cycle_length = 24, transform = "None", sigma = 0.1, exog_count = 0);
 
-df = b1.mPastData
+df = b1.mPastData.copy()
 df[b1.mSignalVar] = df[b1.mName]
 
 lEngine = autof.cForecastEngine()

diff --git a/tests/bugs/issue_70/test_artificial_keep_all_seasonals_second.py b/tests/bugs/issue_70/test_artificial_keep_all_seasonals_second.py
@@ -7,7 +7,7 @@
 
 b1 = tsds.generate_random_TS(N = 40000 , FREQ = 'S', seed = 0, trendtype = "constant", cycle_length = 24, transform = "None", sigma = 0.1, exog_count = 0);
 
-df = b1.mPastData
+df = b1.mPastData.copy()
 df[b1.mSignalVar] = df[b1.mName]
 
 lEngine = autof.cForecastEngine()

diff --git a/tests/bugs/issue_82/issue_82_long_cycles.py b/tests/bugs/issue_82/issue_82_long_cycles.py
@@ -14,7 +14,7 @@
 for cyc in lValues:
     print("TEST_CYCLES_START", cyc)
     b1 = tsds.generate_random_TS(N = 3200 , FREQ = 'H', seed = 0, trendtype = "constant", cycle_length = cyc, transform = "None", sigma = 0.1, exog_count = 0, ar_order=0);
-    df = b1.mPastData
+    df = b1.mPastData.copy()
     df[b1.mSignalVar] = df[b1.mName]
 
     # df.tail(10)

diff --git a/tests/bugs/test_random_exogenous.py b/tests/bugs/test_random_exogenous.py
@@ -9,7 +9,7 @@
     # warnings.simplefilter("error")
 
     b1 = tsds.generate_random_TS(N = 160 , FREQ = 'D', seed = 0, trendtype = "constant", cycle_length = 12, transform = "None", sigma = 0.10, exog_count = 1280);
-    df = b1.mPastData
+    df = b1.mPastData.copy()
     df[b1.mSignalVar] = df[b1.mName]
 
     df.info()

diff --git a/tests/croston/test_croston_fpp2_counts_example_optimize.py b/tests/croston/test_croston_fpp2_counts_example_optimize.py
@@ -4,7 +4,7 @@ def create_dataset():
     lCounts = "0 2 0 1 0 11 0 0 0 0 2 0 6 3 0 0 0 0 0 7 0 0 0 0 0 0 0 3 1 0 0 1 0 1 0 0".split()
     lCounts = [float(c) for c in lCounts]
     N = len(lCounts)
-    lDates = pd.date_range(start="2000-01-01", periods=N, freq='m')
+    lDates = pd.date_range(start="2000-01-01", periods=N, freq='ME')
 
     df = pd.DataFrame({"Date" : lDates, "Signal" : lCounts})
     return df

diff --git a/tests/func/test_ar.py b/tests/func/test_ar.py
@@ -7,7 +7,7 @@
 #get_ipython().magic('matplotlib inline')
 
 b1 = tsds.generate_random_TS(N = 320 , FREQ = 'D', seed = 0, trendtype = "constant", cycle_length = 0, transform = "None", sigma = 0.1, ar_order = 12);
-df = b1.mPastData
+df = b1.mPastData.copy()
 
 df['Signal'] = df[b1.mName]
 

diff --git a/tests/func/test_const_signal.py b/tests/func/test_const_signal.py
@@ -7,7 +7,7 @@
 #get_ipython().magic('matplotlib inline')
 
 b1 = tsds.generate_random_TS(N = 320 , FREQ = 'D', seed = 0, trendtype = "constant", cycle_length = 0, transform = "None", sigma = 0.0, ar_order = 0);
-df = b1.mPastData
+df = b1.mPastData.copy()
 
 df['Signal'] = df[b1.mName]
 

diff --git a/tests/func/test_cycle.py b/tests/func/test_cycle.py
@@ -7,7 +7,7 @@
 #get_ipython().magic('matplotlib inline')
 
 b1 = tsds.generate_random_TS(N = 320 , FREQ = 'D', seed = 0, trendtype = "constant", cycle_length = 12, transform = "None", sigma = 0.0);
-df = b1.mPastData
+df = b1.mPastData.copy()
 
 df['Signal'] = df[b1.mName]
 

diff --git a/tests/func/test_cycles_full.py b/tests/func/test_cycles_full.py
@@ -12,7 +12,7 @@
 for cyc in lValues:
     print("TEST_CYCLES_START", cyc)
     b1 = tsds.generate_random_TS(N = 320 , FREQ = 'D', seed = 0, trendtype = "constant", cycle_length = cyc, transform = "None", sigma = 0.1, exog_count = 0);
-    df = b1.mPastData
+    df = b1.mPastData.copy()
 
     df['Signal'] = df[b1.mName]
 

diff --git a/tests/perf_MASE_RMSSE/user_test_MAPE.py b/tests/perf_MASE_RMSSE/user_test_MAPE.py
@@ -9,8 +9,8 @@
 def prepare_dataset():
     df = pd.read_csv("data/real-life/veturilo.csv", parse_dates=True, usecols=["ts","qnty"], index_col="ts" )
     df.qnty.unique()
-    df.qnty = df.qnty.replace("?", np.NaN).fillna(method='ffill').astype('uint8')
-    df = df.qnty.resample("H").mean().to_frame()
+    df.qnty = df.qnty.replace("?", np.nan).ffill().astype('uint8')
+    df = df.qnty.resample("h").mean().to_frame()
 
     df_predict = df.iloc[:]
 

diff --git a/tests/perf_MASE_RMSSE/user_test_MASE.py b/tests/perf_MASE_RMSSE/user_test_MASE.py
@@ -9,8 +9,8 @@
 def prepare_dataset():
     df = pd.read_csv("data/real-life/veturilo.csv", parse_dates=True, usecols=["ts","qnty"], index_col="ts" )
     df.qnty.unique()
-    df.qnty = df.qnty.replace("?", np.NaN).fillna(method='ffill').astype('uint8')
-    df = df.qnty.resample("H").mean().to_frame()
+    df.qnty = df.qnty.replace("?", np.nan).ffill().astype('uint8')
+    df = df.qnty.resample("h").mean().to_frame()
 
     df_predict = df.iloc[:]
 

diff --git a/tests/perf_MASE_RMSSE/user_test_RMSE.py b/tests/perf_MASE_RMSSE/user_test_RMSE.py
@@ -9,8 +9,8 @@
 def prepare_dataset():
     df = pd.read_csv("data/real-life/veturilo.csv", parse_dates=True, usecols=["ts","qnty"], index_col="ts" )
     df.qnty.unique()
-    df.qnty = df.qnty.replace("?", np.NaN).fillna(method='ffill').astype('uint8')
-    df = df.qnty.resample("H").mean().to_frame()
+    df.qnty = df.qnty.replace("?", np.nan).ffill().astype('uint8')
+    df = df.qnty.resample("h").mean().to_frame()
 
     df_predict = df.iloc[:]
 

diff --git a/tests/perf_MASE_RMSSE/user_test_RMSSE.py b/tests/perf_MASE_RMSSE/user_test_RMSSE.py
@@ -9,8 +9,8 @@
 def prepare_dataset():
     df = pd.read_csv("data/real-life/veturilo.csv", parse_dates=True, usecols=["ts","qnty"], index_col="ts" )
     df.qnty.unique()
-    df.qnty = df.qnty.replace("?", np.NaN).fillna(method='ffill').astype('uint8')
-    df = df.qnty.resample("H").mean().to_frame()
+    df.qnty = df.qnty.replace("?", np.nan).ffill().astype('uint8')
+    df = df.qnty.resample("h").mean().to_frame()
 
     df_predict = df.iloc[:]