extractor: add explicit numerical columns tconvert_columns_to_category

extractors.py

@@ -251,7 +251,7 @@ def apply_tags(data, tags, base_tags=None, additional_tags=[], minimal=True):
 
 
     @staticmethod
-    def convert_columns_to_category(data, additional_columns:list = [], excluded_columns:set = {}):
+    def convert_columns_to_category(data, additional_columns:list = [], excluded_columns:set = {}, numerical_columns:set = {}):
         excluded_columns = set(excluded_columns).union(DEFAULT_CATEGORICALS_COLUMN_EXCLUSION_SET)
 
         col_list = []
@@ -265,11 +265,16 @@ def convert_columns_to_category(data, additional_columns:list = [], excluded_col
             if s < threshold:
                 col_list.append(col)
 
+        logd(f"{excluded_columns}=")
+        logd(f"{col_list}=")
         # convert selected columns to Categorical
         for col in col_list:
             data[col] = data[col].astype('category')
             data[col] = data[col].cat.as_ordered()
 
+        for col in numerical_columns:
+            data[col] = data[col].astype('float')
+
         return data
 
 

plots.py

@@ -81,7 +81,7 @@ def read_data(self):
         data_list = list(map(dask.delayed(functools.partial(read_from_file, sample=self.sample, sample_seed=self.sample_seed, filter_query=self.filter_query))
                              , data_set.get_file_list()))
         concat_result = dask.delayed(pd.concat)(data_list)
-        convert_columns_result = dask.delayed(RawExtractor.convert_columns_to_category)(concat_result, excluded_columns=self.numerical_columns)
+        convert_columns_result = dask.delayed(RawExtractor.convert_columns_to_category)(concat_result, numerical_columns=self.numerical_columns)
         logd(f'PlottingReaderFeather::read_data: {data_list=}')
         logd(f'PlottingReaderFeather::read_data: {convert_columns_result=}')
         # d = dask.compute(convert_columns_result)