fix: Fix push sources and add docs / tests pushing via the python feature server (#2561)

* bug: Fixing push API endpoint to include a way to specify whether the registry should be looked up from a cache. Adding docs for feature server usage

Signed-off-by: Danny Chiao <[email protected]>

* fix

Signed-off-by: Danny Chiao <[email protected]>

* prune out unneeded fields in push source

Signed-off-by: Danny Chiao <[email protected]>

* prune out unneeded fields in push source

Signed-off-by: Danny Chiao <[email protected]>

* prune out unneeded fields in push source

Signed-off-by: Danny Chiao <[email protected]>

* fix comment

Signed-off-by: Danny Chiao <[email protected]>

* fix comment

Signed-off-by: Danny Chiao <[email protected]>

* fix comment

Signed-off-by: Danny Chiao <[email protected]>

* fix comment

Signed-off-by: Danny Chiao <[email protected]>

* fix generator

Signed-off-by: Danny Chiao <[email protected]>

* add data source creator teardown

Signed-off-by: Danny Chiao <[email protected]>

* add data source creator teardown

Signed-off-by: Danny Chiao <[email protected]>

* update push source to alpha

Signed-off-by: Danny Chiao <[email protected]>

* lint

Signed-off-by: Danny Chiao <[email protected]>

* lint

Signed-off-by: Danny Chiao <[email protected]>

* lint

Signed-off-by: Danny Chiao <[email protected]>

* lint

Signed-off-by: Danny Chiao <[email protected]>

* lint

Signed-off-by: Danny Chiao <[email protected]>

Author: adchia

docs/SUMMARY.md

@@ -81,7 +81,7 @@
   * [feature\_store.yaml](reference/feature-repository/feature-store-yaml.md)
   * [.feastignore](reference/feature-repository/feast-ignore.md)
 * [Feature servers](reference/feature-servers/README.md)
-  * [Local feature server](reference/feature-servers/local-feature-server.md)
+  * [Python feature server](reference/feature-servers/python-feature-server.md)
   * [Go-based feature retrieval](reference/feature-servers/go-feature-retrieval.md)
 * [\[Alpha\] Data quality monitoring](reference/dqm.md)
 * [\[Alpha\] On demand feature view](reference/alpha-on-demand-feature-view.md)

docs/reference/data-sources/push.md

@@ -1,5 +1,7 @@
 # Push source
 
+**Warning**: This is an _experimental_ feature. It's intended for early testing and feedback, and could change without warnings in future releases.
+
 ## Description
 
 Push sources allow feature values to be pushed to the online store in real time. This allows fresh feature values to be made available to applications. Push sources supercede the 
@@ -31,18 +33,14 @@ from feast.types import Int64
 
 push_source = PushSource(
     name="push_source",
-    schema=[
-        Field(name="user_id", dtype=Int64),
-        Field(name="life_time_value", dtype=Int64)
-    ],
     batch_source=BigQuerySource(table="test.test"),
 )
 
 fv = FeatureView(
     name="feature view",
     entities=["user_id"],
     schema=[Field(name="life_time_value", dtype=Int64)],
-    stream_source=push_source,
+    source=push_source,
 )
 ```
 
@@ -53,6 +51,8 @@ import pandas as pd
 
 fs = FeatureStore(...)
 feature_data_frame = pd.DataFrame()
-fs.push("push_source", feature_data_frame)
+fs.push("push_source_name", feature_data_frame)
 ```
 
+See also [Python feature server](../feature-servers/python-feature-server.md) for instructions on how to push data to a deployed feature server. 
+

docs/reference/feature-servers/README.md

@@ -2,4 +2,4 @@
 
 Feast users can choose to retrieve features from a feature server, as opposed to through the Python SDK.
 
-{% page-ref page="local-feature-server.md" %}
+{% page-ref page="python-feature-server.md" %}

docs/reference/feature-servers/go-feature-retrieval.md

@@ -2,7 +2,7 @@
 
 ## Overview
 
-The Go Feature Retrieval component is a Go implementation of the core feature serving logic, embedded in the Python SDK. It supports retrieval of feature references, feature services, and on demand feature views, and can be used either through the Python SDK or the [Python feature server](local-feature-server.md).
+The Go Feature Retrieval component is a Go implementation of the core feature serving logic, embedded in the Python SDK. It supports retrieval of feature references, feature services, and on demand feature views, and can be used either through the Python SDK or the [Python feature server](python-feature-server.md).
 
 Currently, this component only supports online serving and does not have an offline component including APIs to create feast feature repositories or apply configuration to the registry to facilitate online materialization. It also does not expose its own dedicated cli to perform feast actions. Furthermore, this component is only meant to expose an online serving API that can be called through the python SDK to facilitate faster online feature retrieval.
 

docs/reference/feature-servers/local-feature-server.md renamed to docs/reference/feature-servers/python-feature-server.md

@@ -1,15 +1,23 @@
-# Local feature server
+# Python feature server
 
 ## Overview
 
-The local feature server is an HTTP endpoint that serves features with JSON I/O. This enables users to get features from Feast using any programming language that can make HTTP requests. A [remote feature server](../alpha-aws-lambda-feature-server.md) on AWS Lambda is also available. A remote feature server on GCP Cloud Run is currently being developed.
+The feature server is an HTTP endpoint that serves features with JSON I/O. This enables users to write + read features from Feast online stores using any programming language that can make HTTP requests. 
 
 ## CLI
 
-There is a new CLI command that starts the server: `feast serve`. By default Feast uses port 6566; the port be overridden by a `--port` flag.
+There is a CLI command that starts the server: `feast serve`. By default, Feast uses port 6566; the port be overridden by a `--port` flag.
+
+## Deploying as a service
+
+One can also deploy a feature server by building a docker image that bundles in the project's `feature_store.yaml`. See [helm chart](https://github.com/feast-dev/feast/blob/master/infra/charts/feast-python-server) for example.
+
+A [remote feature server](../alpha-aws-lambda-feature-server.md) on AWS Lambda is available. A remote feature server on GCP Cloud Run is currently being developed.
+
 
 ## Example
 
+### Initializing a feature server
 Here's the local feature server usage example with the local template:
 
 ```bash
@@ -41,6 +49,7 @@ INFO:     Uvicorn running on http://127.0.0.1:6566 (Press CTRL+C to quit)
 09/10/2021 10:42:11 AM INFO:Uvicorn running on http://127.0.0.1:6566 (Press CTRL+C to quit)
 ```
 
+### Retrieving features from the online store
 After the server starts, we can execute cURL commands from another terminal tab:
 
 ```bash
@@ -142,3 +151,45 @@ curl -X POST \
     }
   }' | jq
 ```
+
+### Pushing features to the online store
+You can push data corresponding to a push source to the online store (note that timestamps need to be strings):
+
+```text
+curl -X POST "http://localhost:6566/push" -d '{
+    "push_source_name": "driver_hourly_stats_push_source",
+    "df": {
+            "driver_id": [1001],
+            "event_timestamp": ["2022-05-13 10:59:42"],
+            "created": ["2022-05-13 10:59:42"],
+            "conv_rate": [1.0],
+            "acc_rate": [1.0],
+            "avg_daily_trips": [1000]
+    }
+  }' | jq
+```
+
+or equivalently from Python:
+```python
+import json
+import requests
+import pandas as pd
+from datetime import datetime
+
+event_dict = {
+    "driver_id": [1001],
+    "event_timestamp": [str(datetime(2021, 5, 13, 10, 59, 42))],
+    "created": [str(datetime(2021, 5, 13, 10, 59, 42))],
+    "conv_rate": [1.0],
+    "acc_rate": [1.0],
+    "avg_daily_trips": [1000],
+    "string_feature": "test2",
+}
+push_data = {
+    "push_source_name":"driver_stats_push_source",
+    "df":event_dict
+}
+requests.post(
+    "http://localhost:6566/push", 
+    data=json.dumps(push_data))
+```

protos/feast/core/DataSource.proto

@@ -222,8 +222,7 @@ message DataSource {
   // Defines options for DataSource that supports pushing data to it. This allows data to be pushed to
   // the online store on-demand, such as by stream consumers.
   message PushOptions {
-    // Mapping of feature name to type
-    map<string, feast.types.ValueType.Enum> schema = 1;
+    reserved 1;
   }
 
 

sdk/python/feast/data_source.py

@@ -21,7 +21,7 @@
 
 from feast import type_map
 from feast.data_format import StreamFormat
-from feast.field import Field, from_value_type
+from feast.field import Field
 from feast.protos.feast.core.DataSource_pb2 import DataSource as DataSourceProto
 from feast.repo_config import RepoConfig, get_data_source_class_from_type
 from feast.types import VALUE_TYPES_TO_FEAST_TYPES
@@ -714,45 +714,35 @@ class PushSource(DataSource):
     A source that can be used to ingest features on request
     """
 
-    name: str
-    schema: List[Field]
+    # TODO(adchia): consider adding schema here in case where Feast manages pushing events to the offline store
+    # TODO(adchia): consider a "mode" to support pushing raw vs transformed events
     batch_source: DataSource
-    timestamp_field: str
 
     def __init__(
         self,
         *,
         name: str,
-        schema: List[Field],
         batch_source: DataSource,
         description: Optional[str] = "",
         tags: Optional[Dict[str, str]] = None,
         owner: Optional[str] = "",
-        timestamp_field: Optional[str] = "",
     ):
         """
         Creates a PushSource object.
         Args:
             name: Name of the push source
-            schema: Schema mapping from the input feature name to a ValueType
             batch_source: The batch source that backs this push source. It's used when materializing from the offline
                 store to the online store, and when retrieving historical features.
             description (optional): A human-readable description.
             tags (optional): A dictionary of key-value pairs to store arbitrary metadata.
             owner (optional): The owner of the data source, typically the email of the primary
                 maintainer.
-            timestamp_field (optional): Event timestamp foe;d used for point in time
-                joins of feature values.
 
         """
         super().__init__(name=name, description=description, tags=tags, owner=owner)
-        self.schema = sorted(schema)  # TODO: add schema inference from a batch source
         self.batch_source = batch_source
         if not self.batch_source:
             raise ValueError(f"batch_source is needed for push source {self.name}")
-        if not timestamp_field:
-            raise ValueError(f"timestamp field is needed for push source {self.name}")
-        self.timestamp_field = timestamp_field
 
     def validate(self, config: RepoConfig):
         pass
@@ -764,38 +754,25 @@ def get_table_column_names_and_types(
 
     @staticmethod
     def from_proto(data_source: DataSourceProto):
-        schema_pb = data_source.push_options.schema
-        schema = []
-        for key, val in schema_pb.items():
-            schema.append(Field(name=key, dtype=from_value_type(ValueType(val))))
-
         assert data_source.HasField("batch_source")
         batch_source = DataSource.from_proto(data_source.batch_source)
 
         return PushSource(
             name=data_source.name,
-            schema=sorted(schema),
             batch_source=batch_source,
-            timestamp_field=data_source.timestamp_field,
             description=data_source.description,
             tags=dict(data_source.tags),
             owner=data_source.owner,
         )
 
     def to_proto(self) -> DataSourceProto:
-        schema_pb = {}
-        for field in self.schema:
-            schema_pb[field.name] = field.dtype.to_value_type().value
         batch_source_proto = None
         if self.batch_source:
             batch_source_proto = self.batch_source.to_proto()
 
-        options = DataSourceProto.PushOptions(schema=schema_pb,)
         data_source_proto = DataSourceProto(
             name=self.name,
             type=DataSourceProto.PUSH_SOURCE,
-            push_options=options,
-            timestamp_field=self.timestamp_field,
             description=self.description,
             tags=self.tags,
             owner=self.owner,

sdk/python/feast/feature_server.py

@@ -94,9 +94,7 @@ def push(body=Depends(get_body)):
     @app.post("/write-to-online-store")
     def write_to_online_store(body=Depends(get_body)):
         warnings.warn(
-            "write_to_online_store is an experimental feature. "
-            "This API is unstable and it could be changed in the future. "
-            "We do not guarantee that future changes will maintain backward compatibility.",
+            "write_to_online_store is deprecated. Please consider using /push instead",
             RuntimeWarning,
         )
         try:

sdk/python/feast/feature_store.py

@@ -92,7 +92,6 @@
 
 warnings.simplefilter("once", DeprecationWarning)
 
-
 if TYPE_CHECKING:
     from feast.embedded_go.online_features_service import EmbeddedOnlineFeatureServer
 
@@ -1186,16 +1185,25 @@ def tqdm_builder(length):
             )
 
     @log_exceptions_and_usage
-    def push(self, push_source_name: str, df: pd.DataFrame):
+    def push(
+        self, push_source_name: str, df: pd.DataFrame, allow_registry_cache: bool = True
+    ):
         """
         Push features to a push source. This updates all the feature views that have the push source as stream source.
         Args:
             push_source_name: The name of the push source we want to push data to.
             df: the data being pushed.
+            allow_registry_cache: whether to allow cached versions of the registry.
         """
+        warnings.warn(
+            "Push source is an experimental feature. "
+            "This API is unstable and it could and might change in the future. "
+            "We do not guarantee that future changes will maintain backward compatibility.",
+            RuntimeWarning,
+        )
         from feast.data_source import PushSource
 
-        all_fvs = self.list_feature_views(allow_cache=True)
+        all_fvs = self.list_feature_views(allow_cache=allow_registry_cache)
 
         fvs_with_push_sources = {
             fv
@@ -1208,7 +1216,9 @@ def push(self, push_source_name: str, df: pd.DataFrame):
         }
 
         for fv in fvs_with_push_sources:
-            self.write_to_online_store(fv.name, df, allow_registry_cache=True)
+            self.write_to_online_store(
+                fv.name, df, allow_registry_cache=allow_registry_cache
+            )
 
     @log_exceptions_and_usage
     def write_to_online_store(

sdk/python/feast/inference.py

@@ -2,7 +2,7 @@
 from typing import List
 
 from feast import BigQuerySource, Entity, FileSource, RedshiftSource, SnowflakeSource
-from feast.data_source import DataSource, RequestSource
+from feast.data_source import DataSource, PushSource, RequestSource
 from feast.errors import RegistryInferenceFailure
 from feast.feature_view import FeatureView
 from feast.field import Field, from_value_type
@@ -74,6 +74,8 @@ def update_data_sources_with_inferred_event_timestamp_col(
     for data_source in data_sources:
         if isinstance(data_source, RequestSource):
             continue
+        if isinstance(data_source, PushSource):
+            data_source = data_source.batch_source
         if data_source.timestamp_field is None or data_source.timestamp_field == "":
             # prepare right match pattern for data source
             ts_column_type_regex_pattern = ""