hopeit.dataframes 0.25.1__tar.gz → 0.25.2__tar.gz

{hopeit_dataframes-0.25.1 → hopeit_dataframes-0.25.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: hopeit.dataframes
-Version: 0.25.1
+Version: 0.25.2
 Summary: Hopeit Engine Dataframes Toolkit
 Author-email: Leo Smerling <contact@hopeit.com.ar>, Pablo Canto <contact@hopeit.com.ar>
 License: Apache 2
@@ -24,7 +24,7 @@ Classifier: Topic :: Software Development :: Libraries :: Python Modules
 Classifier: Framework :: AsyncIO
 Requires-Python: >=3.9
 Description-Content-Type: text/plain
-Requires-Dist: hopeit.engine[fs-storage]==0.25.1
+Requires-Dist: hopeit.engine[fs-storage]==0.25.2
 Requires-Dist: pandas
 Requires-Dist: numpy
 Provides-Extra: pyarrow

{hopeit_dataframes-0.25.1 → hopeit_dataframes-0.25.2}/src/hopeit/dataframes/__init__.py

@@ -118,9 +118,9 @@ class DataFrames(Generic[DataFrameT, DataObject]):
         return datatype._from_dataobjects(dataobjects)  # type: ignore  # pylint: disable=protected-access
 
     @staticmethod
-    def to_dataobjects(obj: DataFrameT) -> List[DataObject]:
+    def to_dataobjects(obj: DataFrameT, *, normalize_null_values: bool = False) -> List[DataObject]:
         """Converts `@dataframe` object to a list of standard `@dataobject`s"""
-        return obj._to_dataobjects()  # type: ignore  # pylint: disable=protected-access
+        return obj._to_dataobjects(normalize_null_values)  # type: ignore  # pylint: disable=protected-access
 
     @staticmethod
     def from_array(datatype: Type[DataFrameT], array: np.ndarray) -> DataFrameT:

{hopeit_dataframes-0.25.1 → hopeit_dataframes-0.25.2}/src/hopeit/dataframes/dataframe.py

@@ -37,6 +37,12 @@ def _series_to_int(field_name: str, x: pd.Series) -> pd.Series:
     return x.astype(np.int64)
 
 
+def _series_to_bool(field_name: str, x: pd.Series) -> pd.Series:
+    if x.isnull().values.any():  # type: ignore[union-attr]
+        raise ValueError(f"Field `{field_name}` is not nullable")
+    return x.astype(bool)
+
+
 def _series_to_float(field_name: str, x: pd.Series) -> pd.Series:
     if x.isnull().values.any():  # type: ignore[union-attr]
         raise ValueError(f"Field `{field_name}` is not nullable")
@@ -51,15 +57,19 @@ def _series_to_str(field_name: str, x: pd.Series) -> pd.Series:
 
 # Functions to do type coercion
 def _series_to_int_nullable(_field_name: str, x: pd.Series) -> pd.Series:
-    return x[x.notna()].astype(np.int64)
+    return x.dropna().astype(np.int64)
+
+
+def _series_to_bool_nullable(_field_name: str, x: pd.Series) -> pd.Series:
+    return x.dropna().astype(bool)
 
 
 def _series_to_float_nullable(_field_name: str, x: pd.Series) -> pd.Series:
-    return x[x.notna()].astype(np.float64)
+    return x.dropna().astype(np.float64)
 
 
 def _series_to_str_nullable(_field_name: str, x: pd.Series) -> pd.Series:
-    return x[x.notna()].astype(str)
+    return x.dropna().astype(str)
 
 
 def _series_to_datetime(field_name: str, x: pd.Series) -> pd.Series:
@@ -75,11 +85,11 @@ def _series_to_utc_datetime(field_name: str, x: pd.Series) -> pd.Series:
 
 
 def _series_to_datetime_nullable(_field_name: str, x: pd.Series) -> pd.Series:
-    return pd.to_datetime(x)
+    return pd.to_datetime(x.dropna())
 
 
 def _series_to_utc_datetime_nullable(_field_name: str, x: pd.Series) -> pd.Series:
-    return pd.to_datetime(x, utc=True)
+    return pd.to_datetime(x.dropna(), utc=True)
 
 
 class DataFrameMixin(Generic[DataFrameT, DataObject]):
@@ -89,13 +99,17 @@ class DataFrameMixin(Generic[DataFrameT, DataObject]):
     Do not use this class directly, instead use `@dataframe` class decorator.
     """
 
+    DataFrameValueType = Union[int, bool, float, str, date, datetime, None]
+
     DATATYPE_MAPPING = {
         int: _series_to_int,
+        bool: _series_to_bool,
         float: _series_to_float,
         str: _series_to_str,
         date: _series_to_datetime,
         datetime: _series_to_utc_datetime,
         Union[int, None]: _series_to_int_nullable,
+        Union[bool, None]: _series_to_bool_nullable,
         Union[float, None]: _series_to_float_nullable,
         Union[str, None]: _series_to_str_nullable,
         Union[date, None]: _series_to_datetime_nullable,
@@ -145,7 +159,17 @@ class DataFrameMixin(Generic[DataFrameT, DataObject]):
     def __getitem__(self, key) -> "DataFrameT":
         return self._from_df(self.__df[key])
 
-    def _to_dataobjects(self) -> List[DataObject]:
+    def _normalize_null_values(
+        self, value: Union[DataFrameValueType, pd.Timestamp]
+    ) -> DataFrameValueType:
+        return None if pd.isnull(value) else value
+
+    def _to_dataobjects(self, normalize_null_values: bool) -> List[DataObject]:
+        if normalize_null_values:
+            return [
+                self.DataObject(**{k: self._normalize_null_values(v) for k, v in fields.items()})
+                for fields in self.__df.to_dict(orient="records")
+            ]
         return [self.DataObject(**fields) for fields in self.__df.to_dict(orient="records")]
 
     def event_id(self, *args, **kwargs) -> str:

{hopeit_dataframes-0.25.1 → hopeit_dataframes-0.25.2}/src/hopeit.dataframes.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: hopeit.dataframes
-Version: 0.25.1
+Version: 0.25.2
 Summary: Hopeit Engine Dataframes Toolkit
 Author-email: Leo Smerling <contact@hopeit.com.ar>, Pablo Canto <contact@hopeit.com.ar>
 License: Apache 2
@@ -24,7 +24,7 @@ Classifier: Topic :: Software Development :: Libraries :: Python Modules
 Classifier: Framework :: AsyncIO
 Requires-Python: >=3.9
 Description-Content-Type: text/plain
-Requires-Dist: hopeit.engine[fs-storage]==0.25.1
+Requires-Dist: hopeit.engine[fs-storage]==0.25.2
 Requires-Dist: pandas
 Requires-Dist: numpy
 Provides-Extra: pyarrow

hopeit_dataframes-0.25.2/src/hopeit.dataframes.egg-info/requires.txt

@@ -0,0 +1,6 @@
+hopeit.engine[fs-storage]==0.25.2
+pandas
+numpy
+
+[pyarrow]
+pyarrow

hopeit_dataframes-0.25.1/src/hopeit.dataframes.egg-info/requires.txt

@@ -1,6 +0,0 @@
-hopeit.engine[fs-storage]==0.25.1
-pandas
-numpy
-
-[pyarrow]
-pyarrow