Do not attempt to compute ODFVs when there are no ODFVs (#2090)

* Do not attempt to compute ODFVs when there are no ODFVs Signed-off-by: Felix Wang <[email protected]> * Make on_demand_feature_views an optional parameter for retrieval jobs Signed-off-by: Felix Wang <[email protected]>
feast-dev · Dec 2, 2021 · ce84d38 · ce84d38
1 parent fc9b767
commit ce84d38
Show file tree

Hide file tree

Showing 4 changed files with 17 additions and 19 deletions.
diff --git a/sdk/python/feast/infra/offline_stores/bigquery.py b/sdk/python/feast/infra/offline_stores/bigquery.py
@@ -107,11 +107,7 @@ def pull_latest_from_table_or_query(
 
         # When materializing a single feature view, we don't need full feature names. On demand transforms aren't materialized
         return BigQueryRetrievalJob(
-            query=query,
-            client=client,
-            config=config,
-            full_feature_names=False,
-            on_demand_feature_views=None,
+            query=query, client=client, config=config, full_feature_names=False,
         )
 
     @staticmethod
@@ -200,7 +196,7 @@ def __init__(
         client: bigquery.Client,
         config: RepoConfig,
         full_feature_names: bool,
-        on_demand_feature_views: Optional[List[OnDemandFeatureView]],
+        on_demand_feature_views: Optional[List[OnDemandFeatureView]] = None,
     ):
         if not isinstance(query, str):
             self._query_generator = query
@@ -215,7 +211,9 @@ def query_generator() -> Iterator[str]:
         self.client = client
         self.config = config
         self._full_feature_names = full_feature_names
-        self._on_demand_feature_views = on_demand_feature_views
+        self._on_demand_feature_views = (
+            on_demand_feature_views if on_demand_feature_views else []
+        )
 
     @property
     def full_feature_names(self) -> bool:

diff --git a/sdk/python/feast/infra/offline_stores/file.py b/sdk/python/feast/infra/offline_stores/file.py
@@ -35,14 +35,16 @@ def __init__(
         self,
         evaluation_function: Callable,
         full_feature_names: bool,
-        on_demand_feature_views: Optional[List[OnDemandFeatureView]],
+        on_demand_feature_views: Optional[List[OnDemandFeatureView]] = None,
     ):
         """Initialize a lazy historical retrieval job"""
 
         # The evaluation function executes a stored procedure to compute a historical retrieval.
         self.evaluation_function = evaluation_function
         self._full_feature_names = full_feature_names
-        self._on_demand_feature_views = on_demand_feature_views
+        self._on_demand_feature_views = (
+            on_demand_feature_views if on_demand_feature_views else []
+        )
 
     @property
     def full_feature_names(self) -> bool:
@@ -333,7 +335,5 @@ def evaluate_offline_job():
 
         # When materializing a single feature view, we don't need full feature names. On demand transforms aren't materialized
         return FileRetrievalJob(
-            evaluation_function=evaluate_offline_job,
-            full_feature_names=False,
-            on_demand_feature_views=None,
+            evaluation_function=evaluate_offline_job, full_feature_names=False,
         )
diff --git a/sdk/python/feast/infra/offline_stores/offline_store.py b/sdk/python/feast/infra/offline_stores/offline_store.py
@@ -41,7 +41,7 @@ def on_demand_feature_views(self) -> Optional[List[OnDemandFeatureView]]:
     def to_df(self) -> pd.DataFrame:
         """Return dataset as Pandas DataFrame synchronously including on demand transforms"""
         features_df = self._to_df_internal()
-        if self.on_demand_feature_views is None:
+        if not self.on_demand_feature_views:
             return features_df
 
         # TODO(adchia): Fix requirement to specify dependent feature views in feature_refs
@@ -63,7 +63,7 @@ def _to_arrow_internal(self) -> pyarrow.Table:
 
     def to_arrow(self) -> pyarrow.Table:
         """Return dataset as pyarrow Table synchronously"""
-        if self.on_demand_feature_views is None:
+        if not self.on_demand_feature_views:
             return self._to_arrow_internal()
 
         features_df = self._to_df_internal()

diff --git a/sdk/python/feast/infra/offline_stores/redshift.py b/sdk/python/feast/infra/offline_stores/redshift.py
@@ -101,7 +101,6 @@ def pull_latest_from_table_or_query(
             s3_resource=s3_resource,
             config=config,
             full_feature_names=False,
-            on_demand_feature_views=None,
         )
 
     @staticmethod
@@ -189,7 +188,7 @@ def __init__(
         s3_resource,
         config: RepoConfig,
         full_feature_names: bool,
-        on_demand_feature_views: Optional[List[OnDemandFeatureView]],
+        on_demand_feature_views: Optional[List[OnDemandFeatureView]] = None,
     ):
         """Initialize RedshiftRetrievalJob object.
 
@@ -199,7 +198,7 @@ def __init__(
             s3_resource: boto3 s3 resource object
             config: Feast repo config
             full_feature_names: Whether to add the feature view prefixes to the feature names
-            on_demand_feature_views: A list of on demand transforms to apply at retrieval time
+            on_demand_feature_views (optional): A list of on demand transforms to apply at retrieval time
         """
         if not isinstance(query, str):
             self._query_generator = query
@@ -220,7 +219,9 @@ def query_generator() -> Iterator[str]:
             + str(uuid.uuid4())
         )
         self._full_feature_names = full_feature_names
-        self._on_demand_feature_views = on_demand_feature_views
+        self._on_demand_feature_views = (
+            on_demand_feature_views if on_demand_feature_views else []
+        )
 
     @property
     def full_feature_names(self) -> bool:
@@ -346,7 +347,6 @@ def _upload_entity_df_and_get_entity_schema(
             s3_resource,
             config,
             full_feature_names=False,
-            on_demand_feature_views=None,
         ).to_df()
         return dict(zip(limited_entity_df.columns, limited_entity_df.dtypes))
     else: