better meta inferrence, to prevent indexing error while parquetting

tsfresh/convenience/bindings.py

@@ -14,6 +14,7 @@ def _feature_extraction_on_chunk_helper(
     column_value,
     default_fc_parameters,
     kind_to_fc_parameters,
+    show_warnings= True
 ):
     """
     Helper function wrapped around _do_extraction_on_chunk to use the correct format
@@ -28,18 +29,23 @@ def _feature_extraction_on_chunk_helper(
         default_fc_parameters = {}
 
     if column_sort is not None:
-        df = df.sort_values(column_sort)
+
+        data = df[[column_sort, column_value]].set_index(column_sort)
+        data = data.sort_index(level=column_sort)
+    else:
+        data = df[column_value]
 
-    chunk = df[column_id].iloc[0], df[column_kind].iloc[0], df[column_value]
+    chunk = df[column_id].iloc[0], df[column_kind].iloc[0], data
     features = _do_extraction_on_chunk(
         chunk,
         default_fc_parameters=default_fc_parameters,
         kind_to_fc_parameters=kind_to_fc_parameters,
+        show_warnings=show_warnings
     )
     features = pd.DataFrame(features, columns=[column_id, "variable", "value"])
     features["value"] = features["value"].astype("double")
 
-    return features[[column_id, "variable", "value"]]
+    return features
 
 
 def dask_feature_extraction_on_chunk(
@@ -50,6 +56,8 @@ def dask_feature_extraction_on_chunk(
     column_sort=None,
     default_fc_parameters=None,
     kind_to_fc_parameters=None,
+    taste_of_pandas_df = False,
+    show_warnings=True
 ):
     """
     Extract features on a grouped dask dataframe given the column names and the extraction settings.
@@ -124,6 +132,14 @@ def dask_feature_extraction_on_chunk(
 
     :param column_value: The name for the column keeping the value itself.
     :type column_value: str
+
+    :param taste_of_pandas_df: pandas  Dataframe with same preprocessing steps as the Dask Dataframe, (e.g. df.groupby(), etc.)
+    :type taste_of_pandas_df: pandas.DataFrame
+
+    :param show_warnings: Wether to show warings in tsfresh.
+    :type show_warnings: bool
+
+
 
     :return: A dask dataframe with the columns ``column_id``, "variable" and "value". The index is taken
             from the grouped dataframe.
@@ -138,11 +154,14 @@ def dask_feature_extraction_on_chunk(
         column_value=column_value,
         default_fc_parameters=default_fc_parameters,
         kind_to_fc_parameters=kind_to_fc_parameters,
+        show_warnings=show_warnings
     )
-    return df.apply(
-        feature_extraction,
-        meta=[(column_id, "int64"), ("variable", "object"), ("value", "float64")],
-    )
+    if taste_of_pandas_df:
+        meta = taste_of_pandas_df.apply(feature_extraction)
+    else:
+        meta = [(column_id, 'int64'), ('variable', 'object'), ('value', 'float64')]
+
+    return df.apply(feature_extraction, meta= meta)
 
 
 def spark_feature_extraction_on_chunk(
@@ -152,8 +171,9 @@ def spark_feature_extraction_on_chunk(
     column_value,
     column_sort=None,
     default_fc_parameters=None,
-    kind_to_fc_parameters=None,
+    kind_to_fc_parameters=None
 ):
+
     """
     Extract features on a grouped spark dataframe given the column names and the extraction settings.
     This wrapper function should only be used if you have a spark dataframe as input.

tsfresh/feature_extraction/extraction.py

@@ -358,7 +358,7 @@ def _f():
                         continue
                 x = data
             else:
-                x = data.values
+                x = data.values.flatten()
 
             if getattr(func, "fctype", None) == "combiner":
                 result = func(x, param=parameter_list)

tsfresh/feature_extraction/feature_calculators.py

@@ -2282,7 +2282,7 @@ def linear_trend_timewise(x, param):
     times_seconds = (ix - ix[0]).total_seconds()
     times_hours = np.asarray(times_seconds / float(3600))
 
-    linReg = linregress(times_hours, x.values)
+    linReg = linregress(times_hours, x.values.flatten())
 
     return [
         ('attr_"{}"'.format(config["attr"]), getattr(linReg, config["attr"]))