chore: Add docstrings (feast-dev#3128)

* Fix base feature view docstring Signed-off-by: Felix Wang <wangfelix98@gmail.com> * Add docstring to trino source Signed-off-by: Felix Wang <wangfelix98@gmail.com> * Fix docstrings for bfv and sfv Signed-off-by: Felix Wang <wangfelix98@gmail.com> * Temp work Signed-off-by: Felix Wang <wangfelix98@gmail.com> * Update data source docstrings Signed-off-by: Felix Wang <wangfelix98@gmail.com> Signed-off-by: Felix Wang <wangfelix98@gmail.com>
srausser · Aug 24, 2022 · e610f62 · e610f62
1 parent f0594e1
commit e610f62
Show file tree

Hide file tree

Showing 9 changed files with 100 additions and 30 deletions.
diff --git a/sdk/python/feast/base_feature_view.py b/sdk/python/feast/base_feature_view.py
@@ -35,8 +35,8 @@ class BaseFeatureView(ABC):
             maintainer.
         projection: The feature view projection storing modifications to be applied to
             this base feature view at retrieval time.
-        created_timestamp (optional): The time when the base feature view was created.
-        last_updated_timestamp (optional): The time when the base feature view was last
+        created_timestamp: The time when the base feature view was created.
+        last_updated_timestamp: The time when the base feature view was last
             updated.
     """
 

diff --git a/sdk/python/feast/batch_feature_view.py b/sdk/python/feast/batch_feature_view.py
@@ -1,12 +1,16 @@
-from datetime import timedelta
-from typing import Dict, List, Optional, Union
+import warnings
+from datetime import datetime, timedelta
+from typing import Dict, List, Optional, Tuple, Union
 
+from feast import flags_helper
 from feast.data_source import DataSource
 from feast.entity import Entity
 from feast.feature_view import FeatureView
 from feast.field import Field
 from feast.protos.feast.core.DataSource_pb2 import DataSource as DataSourceProto
 
+warnings.simplefilter("once", RuntimeWarning)
+
 SUPPORTED_BATCH_SOURCES = {
     "BigQuerySource",
     "FileSource",
@@ -19,22 +23,58 @@
 
 
 class BatchFeatureView(FeatureView):
+    """
+    A batch feature view defines a logical group of features that has only a batch data source.
+
+    Attributes:
+        name: The unique name of the batch feature view.
+        entities: List of entities or entity join keys.
+        ttl: The amount of time this group of features lives. A ttl of 0 indicates that
+            this group of features lives forever. Note that large ttl's or a ttl of 0
+            can result in extremely computationally intensive queries.
+        schema: The schema of the feature view, including feature, timestamp, and entity
+            columns. If not specified, can be inferred from the underlying data source.
+        source: The batch source of data where this group of features is stored.
+        online: A boolean indicating whether online retrieval is enabled for this feature view.
+        description: A human-readable description.
+        tags: A dictionary of key-value pairs to store arbitrary metadata.
+        owner: The owner of the batch feature view, typically the email of the primary maintainer.
+    """
+
+    name: str
+    entities: List[str]
+    ttl: Optional[timedelta]
+    source: DataSource
+    schema: List[Field]
+    entity_columns: List[Field]
+    features: List[Field]
+    online: bool
+    description: str
+    tags: Dict[str, str]
+    owner: str
+    timestamp_field: str
+    materialization_intervals: List[Tuple[datetime, datetime]]
+
     def __init__(
         self,
         *,
-        name: Optional[str] = None,
+        name: str,
+        source: DataSource,
         entities: Optional[Union[List[Entity], List[str]]] = None,
         ttl: Optional[timedelta] = None,
         tags: Optional[Dict[str, str]] = None,
         online: bool = True,
         description: str = "",
         owner: str = "",
         schema: Optional[List[Field]] = None,
-        source: Optional[DataSource] = None,
     ):
+        if not flags_helper.is_test():
+            warnings.warn(
+                "Batch feature views are experimental features in alpha development. "
+                "Some functionality may still be unstable so functionality can change in the future.",
+                RuntimeWarning,
+            )
 
-        if source is None:
-            raise ValueError("Feature views need a source specified")
         if (
             type(source).__name__ not in SUPPORTED_BATCH_SOURCES
             and source.to_proto().type != DataSourceProto.SourceType.CUSTOM_SOURCE

diff --git a/sdk/python/feast/infra/offline_stores/bigquery_source.py b/sdk/python/feast/infra/offline_stores/bigquery_source.py
@@ -36,14 +36,18 @@ def __init__(
         """Create a BigQuerySource from an existing table or query.
 
         Args:
-            name (optional): Name for the source. Defaults to the table if not specified.
+            name (optional): Name for the source. Defaults to the table if not specified, in which
+                case the table must be specified.
             timestamp_field (optional): Event timestamp field used for point in time
                 joins of feature values.
+            table (optional): BigQuery table where the features are stored. Exactly one of 'table'
+                and 'query' must be specified.
             table (optional): The BigQuery table where features can be found.
             created_timestamp_column (optional): Timestamp column when row was created, used for deduplicating rows.
             field_mapping (optional): A dictionary mapping of column names in this data source to feature names in a feature table
                 or view. Only used for feature columns, not entities or timestamp columns.
-            query (optional): SQL query to execute to generate data for this data source.
+            query (optional): The query to be executed to obtain the features. Exactly one of 'table'
+                and 'query' must be specified.
             description (optional): A human-readable description.
             tags (optional): A dictionary of key-value pairs to store arbitrary metadata.
             owner (optional): The owner of the bigquery source, typically the email of the primary

diff --git a/sdk/python/feast/infra/offline_stores/contrib/athena_offline_store/athena_source.py b/sdk/python/feast/infra/offline_stores/contrib/athena_offline_store/athena_source.py
@@ -38,24 +38,24 @@ def __init__(
 
         Args:
             timestamp_field : event timestamp column.
-            table (optional): Athena table where the features are stored.
+            table (optional): Athena table where the features are stored. Exactly one of 'table'
+                and 'query' must be specified.
             database: Athena Database Name
             data_source (optional): Athena data source
             created_timestamp_column (optional): Timestamp column indicating when the
                 row was created, used for deduplicating rows.
             field_mapping (optional): A dictionary mapping of column names in this data
                 source to column names in a feature table or view.
             date_partition_column : Timestamp column used for partitioning.
-            query (optional): The query to be executed to obtain the features.
-            name (optional): Name for the source. Defaults to the table_ref if not specified.
+            query (optional): The query to be executed to obtain the features. Exactly one of 'table'
+                and 'query' must be specified.
+            name (optional): Name for the source. Defaults to the table if not specified, in which
+                case the table must be specified.
             description (optional): A human-readable description.
             tags (optional): A dictionary of key-value pairs to store arbitrary metadata.
             owner (optional): The owner of the athena source, typically the email of the primary
                 maintainer.
-
-
         """
-
         _database = "default" if table and not database else database
         self.athena_options = AthenaOptions(
             table=table, query=query, database=_database, data_source=data_source

diff --git a/sdk/python/feast/infra/offline_stores/contrib/trino_offline_store/trino_source.py b/sdk/python/feast/infra/offline_stores/contrib/trino_offline_store/trino_source.py
@@ -97,6 +97,27 @@ def __init__(
         tags: Optional[Dict[str, str]] = None,
         owner: Optional[str] = "",
     ):
+        """
+        Creates a TrinoSource object.
+
+        Args:
+            name (optional): Name for the source. Defaults to the table if not specified, in which
+                case the table must be specified.
+            timestamp_field (optional): Event timestamp field used for point in time
+                joins of feature values.
+            table (optional): Trino table where the features are stored. Exactly one of 'table' and
+                'query' must be specified.
+            created_timestamp_column (optional): Timestamp column indicating when the
+                row was created, used for deduplicating rows.
+            field_mapping (optional): A dictionary mapping of column names in this data
+                source to column names in a feature table or view.
+            query (optional): The query to be executed to obtain the features. Exactly one of 'table'
+                and 'query' must be specified.
+            description (optional): A human-readable description.
+            tags (optional): A dictionary of key-value pairs to store arbitrary metadata.
+            owner (optional): The owner of the snowflake source, typically the email of the primary
+                maintainer.
+        """
         # If no name, use the table as the default name.
         if name is None and table is None:
             raise DataSourceNoNameException()

diff --git a/sdk/python/feast/infra/offline_stores/redshift_source.py b/sdk/python/feast/infra/offline_stores/redshift_source.py
@@ -43,16 +43,19 @@ def __init__(
         Creates a RedshiftSource object.
 
         Args:
-            name (optional): Name for the source. Defaults to the table_ref if not specified.
+            name (optional): Name for the source. Defaults to the table if not specified, in which
+                case the table must be specified.
             timestamp_field (optional): Event timestamp field used for point in time
                 joins of feature values.
-            table (optional): Redshift table where the features are stored.
+            table (optional): Redshift table where the features are stored. Exactly one of 'table'
+                and 'query' must be specified.
             schema (optional): Redshift schema in which the table is located.
             created_timestamp_column (optional): Timestamp column indicating when the
                 row was created, used for deduplicating rows.
             field_mapping (optional): A dictionary mapping of column names in this data
                 source to column names in a feature table or view.
-            query (optional): The query to be executed to obtain the features.
+            query (optional): The query to be executed to obtain the features. Exactly one of 'table'
+                and 'query' must be specified.
             description (optional): A human-readable description.
             tags (optional): A dictionary of key-value pairs to store arbitrary metadata.
             owner (optional): The owner of the redshift source, typically the email of the primary

diff --git a/sdk/python/feast/infra/offline_stores/snowflake_source.py b/sdk/python/feast/infra/offline_stores/snowflake_source.py
@@ -40,14 +40,17 @@ def __init__(
         Creates a SnowflakeSource object.
 
         Args:
-            name (optional): Name for the source. Defaults to the table if not specified.
+            name (optional): Name for the source. Defaults to the table if not specified, in which
+                case the table must be specified.
             timestamp_field (optional): Event timestamp field used for point in time
                 joins of feature values.
             database (optional): Snowflake database where the features are stored.
             warehouse (optional): Snowflake warehouse where the database is stored.
             schema (optional): Snowflake schema in which the table is located.
-            table (optional): Snowflake table where the features are stored.
-            query (optional): The query to be executed to obtain the features.
+            table (optional): Snowflake table where the features are stored. Exactly one of 'table'
+                and 'query' must be specified.
+            query (optional): The query to be executed to obtain the features. Exactly one of 'table'
+                and 'query' must be specified.
             created_timestamp_column (optional): Timestamp column indicating when the
                 row was created, used for deduplicating rows.
             field_mapping (optional): A dictionary mapping of column names in this data

diff --git a/sdk/python/feast/stream_feature_view.py b/sdk/python/feast/stream_feature_view.py
@@ -33,8 +33,8 @@
 @typechecked
 class StreamFeatureView(FeatureView):
     """
-    NOTE: Stream Feature Views are not yet fully implemented and exist to allow users to register their stream sources and
-    schemas with Feast.
+    A stream feature view defines a logical group of features that has both a stream data source and
+    a batch data source.
 
     Attributes:
         name: The unique name of the stream feature view.
@@ -44,15 +44,14 @@ class StreamFeatureView(FeatureView):
             can result in extremely computationally intensive queries.
         schema: The schema of the feature view, including feature, timestamp, and entity
             columns. If not specified, can be inferred from the underlying data source.
-        source: DataSource. The stream source of data where this group of features is stored.
+        source: The stream source of data where this group of features is stored.
         aggregations: List of aggregations registered with the stream feature view.
         mode: The mode of execution.
         timestamp_field: Must be specified if aggregations are specified. Defines the timestamp column on which to aggregate windows.
-        online: Defines whether this stream feature view is used in online feature retrieval.
+        online: A boolean indicating whether online retrieval is enabled for this feature view.
         description: A human-readable description.
         tags: A dictionary of key-value pairs to store arbitrary metadata.
-        owner: The owner of the on demand feature view, typically the email of the primary
-            maintainer.
+        owner: The owner of the stream feature view, typically the email of the primary maintainer.
         udf: The user defined transformation function. This transformation function should have all of the corresponding imports imported within the function.
     """
 
@@ -92,7 +91,7 @@ def __init__(
     ):
         if not flags_helper.is_test():
             warnings.warn(
-                "Stream Feature Views are experimental features in alpha development. "
+                "Stream feature views are experimental features in alpha development. "
                 "Some functionality may still be unstable so functionality can change in the future.",
                 RuntimeWarning,
             )

diff --git a/sdk/python/tests/unit/test_feature_views.py b/sdk/python/tests/unit/test_feature_views.py
@@ -24,7 +24,7 @@ def test_create_batch_feature_view():
         source=batch_source,
     )
 
-    with pytest.raises(ValueError):
+    with pytest.raises(TypeError):
         BatchFeatureView(
             name="test batch feature view", entities=[], ttl=timedelta(days=30)
         )