docs/changes/newsfragments/236.change

@@ -0,0 +1 @@
+Add :meth:`.BaseFeatureStorage.read` method for storage-like objects by `Synchon Mandal`_

docs/changes/newsfragments/236.enh

@@ -0,0 +1 @@
+Adapt :meth:`.HDF5FeatureStorage.read_df` due to addition of :meth:`.HDF5FeatureStorage.read` by `Synchon Mandal`_

junifer/storage/base.py

@@ -104,6 +104,34 @@ class BaseFeatureStorage(ABC):
             klass=NotImplementedError,
         )
 
+    @abstractmethod
+    def read(
+        self,
+        feature_name: Optional[str] = None,
+        feature_md5: Optional[str] = None,
+    ) -> Dict[
+        str, Union[str, List[Union[int, str, Dict[str, str]]], np.ndarray]
+    ]:
+        """Read stored feature.
+
+        Parameters
+        ----------
+        feature_name : str, optional
+            Name of the feature to read (default None).
+        feature_md5 : str, optional
+            MD5 hash of the feature to read (default None).
+
+        Returns
+        -------
+        dict
+            The stored feature as a dictionary.
+
+        """
+        raise_error(
+            msg="Concrete classes need to implement read().",
+            klass=NotImplementedError,
+        )
+
     @abstractmethod
     def read_df(
         self,

junifer/storage/hdf5.py

@@ -328,15 +328,14 @@ class HDF5FeatureStorage(BaseFeatureStorage):
 
         return data
 
-    def read_df(
+    def read(
         self,
         feature_name: Optional[str] = None,
-        feature_md5: Optional[bool] = None,
-    ) -> pd.DataFrame:
-        """Read feature into a pandas.DataFrame.
-
-        Either one of ``feature_name`` or ``feature_md5`` needs to be
-        specified.
+        feature_md5: Optional[str] = None,
+    ) -> Dict[
+        str, Union[str, List[Union[int, str, Dict[str, str]]], np.ndarray]
+    ]:
+        """Read stored feature.
 
         Parameters
         ----------
@@ -347,8 +346,8 @@ class HDF5FeatureStorage(BaseFeatureStorage):
 
         Returns
         -------
-        pandas.DataFrame
-            The features as a dataframe.
+        dict
+            The stored feature as a dictionary.
 
         Raises
         ------
@@ -427,10 +426,46 @@ class HDF5FeatureStorage(BaseFeatureStorage):
             title=md5,
             slash="ignore",
         )
+        return hdf_data
+
+    def read_df(
+        self,
+        feature_name: Optional[str] = None,
+        feature_md5: Optional[str] = None,
+    ) -> pd.DataFrame:
+        """Read feature into a pandas.DataFrame.
+
+        Either one of ``feature_name`` or ``feature_md5`` needs to be
+        specified.
+
+        Parameters
+        ----------
+        feature_name : str, optional
+            Name of the feature to read (default None).
+        feature_md5 : str, optional
+            MD5 hash of the feature to read (default None).
+
+        Returns
+        -------
+        pandas.DataFrame
+            The features as a dataframe.
+
+        Raises
+        ------
+        IOError
+            If HDF5 file does not exist.
+
+        """
+        hdf_data = self.read(
+            feature_name=feature_name, feature_md5=feature_md5
+        )
         reshaped_data = hdf_data["data"]
 
         # Generate index for the data
-        logger.debug(f"Generating pandas.MultiIndex for {md5} ...")
+        logger.debug(
+            "Generating pandas.MultiIndex for feature "
+            f"{feature_name or feature_md5} ..."
+        )
 
         if hdf_data["kind"] == "matrix":
             # Set index for element
@@ -477,17 +512,26 @@ class HDF5FeatureStorage(BaseFeatureStorage):
         idx_df = pd.DataFrame(data=element_idx)  # type: ignore
         # Create multiindex from dataframe
         hdf_data_idx = pd.MultiIndex.from_frame(df=idx_df)
-        logger.debug(f"Generated pandas.MultiIndex for {md5} ...")
+        logger.debug(
+            "Generated pandas.MultiIndex for feature "
+            f"{feature_name or feature_md5} ..."
+        )
 
         # Convert to DataFrame
-        logger.debug(f"Converting HDF5 data for {md5} to pandas.DataFrame ...")
+        logger.debug(
+            "Converting HDF5 data for feature "
+            f"{feature_name or feature_md5} to pandas.DataFrame ..."
+        )
         df = pd.DataFrame(
             data=reshaped_data,
             index=hdf_data_idx,
             columns=columns,  # type: ignore
             dtype=reshaped_data.dtype,
         )
-        logger.debug(f"Converted HDF5 data for {md5} to pandas.DataFrame ...")
+        logger.debug(
+            "Converted HDF5 data for feature "
+            f"{feature_name or feature_md5} to pandas.DataFrame ..."
+        )
 
         return df
 

junifer/storage/sqlite.py

@@ -235,6 +235,38 @@ class SQLiteFeatureStorage(PandasBaseFeatureStorage):
                 out[md5][k] = json.loads(v)
         return out
 
+    def read(
+        self,
+        feature_name: Optional[str] = None,
+        feature_md5: Optional[str] = None,
+    ) -> Dict[
+        str, Union[str, List[Union[int, str, Dict[str, str]]], np.ndarray]
+    ]:
+        """Read stored feature.
+
+        Parameters
+        ----------
+        feature_name : str, optional
+            Name of the feature to read (default None).
+        feature_md5 : str, optional
+            MD5 hash of the feature to read (default None).
+
+        Returns
+        -------
+        dict
+            The stored feature as a dictionary.
+
+        """
+        raise_error(
+            msg=(
+                "read() for SQLiteFeatureStorage is not currently planned to "
+                "be implemented in the near future. If you need the "
+                "functionality, contact the junifer developers "
+                "(https://juaml.github.io/junifer/main/help.html)."
+            ),
+            klass=NotImplementedError,
+        )
+
     def read_df(
         self,
         feature_name: Optional[str] = None,

junifer/storage/tests/test_hdf5.py

@@ -220,6 +220,37 @@ def test_store_metadata_ignore_duplicate(tmp_path: Path) -> None:
     assert len(features) == 1
 
 
+def test_read(tmp_path: Path) -> None:
+    """Test read.
+
+    Parameters
+    ----------
+    tmp_path : pathlib.Path
+        The path to the test directory.
+
+    """
+    uri = tmp_path / "test_read.hdf5"
+    storage = HDF5FeatureStorage(uri=uri)
+    # Data to store
+    to_store_data = {
+        "meta": {
+            "element": {"subject": "test"},
+            "dependencies": ["numpy"],
+            "marker": {"name": "test"},
+            "type": "BOLD",
+        },
+        "data": np.array([[1, 10]]),
+        "col_names": ["f1", "f2"],
+    }
+    # Store data
+    storage.store(kind="vector", **to_store_data)
+    # Read and check
+    stored_data = storage.read(feature_name="BOLD_test")
+    assert ["column_headers", "data", "element", "kind"] == list(
+        stored_data.keys()
+    )
+
+
 def test_read_df_params_error(tmp_path: Path) -> None:
     """Test parameter validation errors for read_df.
 

junifer/storage/tests/test_sqlite.py

@@ -253,6 +253,22 @@ def test_upsert_invalid_option(tmp_path: Path) -> None:
         SQLiteFeatureStorage(uri=uri, upsert="wrong")
 
 
+def test_read(tmp_path: Path) -> None:
+    """Test reading of stored feature.
+
+    Parameters
+    ----------
+    tmp_path : pathlib.Path
+        The path to the test directory.
+
+    """
+    uri = tmp_path / "test_read.sqlite"
+    storage = SQLiteFeatureStorage(uri=uri, upsert="ignore")
+    # Check error
+    with pytest.raises(NotImplementedError):
+        storage.read()
+
+
 # TODO: can the tests be separated?
 def test_store_df_and_read_df(tmp_path: Path) -> None:
     """Test storing dataframe and reading of stored table into dataframe.

junifer/storage/tests/test_storage_base.py

@@ -38,6 +38,12 @@ def test_BaseFeatureStorage() -> None:
         def list_features(self):
             super().list_features()
 
+        def read(self, feature_name=None, feature_md5=None):
+            super().read(
+                feature_name=feature_name,
+                feature_md5=feature_md5,
+            )
+
         def read_df(self, feature_name=None, feature_md5=None):
             super().read_df(
                 feature_name=feature_name,
@@ -66,6 +72,9 @@ def test_BaseFeatureStorage() -> None:
     with pytest.raises(NotImplementedError):
         st.list_features()
 
+    with pytest.raises(NotImplementedError):
+        st.read(None)
+
     with pytest.raises(NotImplementedError):
         st.read_df(None)