feat: Adding vector_search parameter to fields (feast-dev#4855)

* feat: Adding vector_search parameter to fields

Signed-off-by: Francisco Javier Arceo <farceo@redhat.com>

* updated field to handle linter and updated sphinx docs

Signed-off-by: Francisco Javier Arceo <farceo@redhat.com>

* Have to remove the equality test for the new fields...for now we're going to ignore them so it is backwards compatible

Signed-off-by: Francisco Javier Arceo <farceo@redhat.com>

* linter

Signed-off-by: Francisco Javier Arceo <farceo@redhat.com>

---------

Signed-off-by: Francisco Javier Arceo <farceo@redhat.com>

Author: franciscojavierarceo

protos/feast/core/Feature.proto

@@ -35,6 +35,11 @@ message FeatureSpecV2 {
     map<string,string> tags = 3;
 
     // Description of the feature.
-
     string description = 4;
+
+    // Field indicating the vector will be indexed for vector similarity search
+    bool vector_index = 5;
+
+    // Metric used for vector similarity search.
+    string vector_search_metric = 6;
 }

sdk/python/docs/source/feast.infra.online_stores.rst

@@ -7,7 +7,16 @@ Subpackages
 .. toctree::
    :maxdepth: 4
 
-   feast.infra.online_stores
+   feast.infra.online_stores.cassandra_online_store
+   feast.infra.online_stores.couchbase_online_store
+   feast.infra.online_stores.elasticsearch_online_store
+   feast.infra.online_stores.hazelcast_online_store
+   feast.infra.online_stores.hbase_online_store
+   feast.infra.online_stores.ikv_online_store
+   feast.infra.online_stores.milvus_online_store
+   feast.infra.online_stores.mysql_online_store
+   feast.infra.online_stores.postgres_online_store
+   feast.infra.online_stores.qdrant_online_store
 
 Submodules
 ----------
@@ -36,6 +45,14 @@ feast.infra.online\_stores.dynamodb module
    :undoc-members:
    :show-inheritance:
 
+feast.infra.online\_stores.faiss\_online\_store module
+------------------------------------------------------
+
+.. automodule:: feast.infra.online_stores.faiss_online_store
+   :members:
+   :undoc-members:
+   :show-inheritance:
+
 feast.infra.online\_stores.helpers module
 -----------------------------------------
 

sdk/python/docs/source/feast.infra.rst

@@ -51,6 +51,14 @@ feast.infra.provider module
    :undoc-members:
    :show-inheritance:
 
+feast.infra.supported\_async\_methods module
+--------------------------------------------
+
+.. automodule:: feast.infra.supported_async_methods
+   :members:
+   :undoc-members:
+   :show-inheritance:
+
 Module contents
 ---------------
 

sdk/python/feast/field.py

@@ -32,12 +32,16 @@ class Field:
         dtype: The type of the field, such as string or float.
         description: A human-readable description.
         tags: User-defined metadata in dictionary form.
+        vector_index: If set to True the field will be indexed for vector similarity search.
+        vector_search_metric: The metric used for vector similarity search.
     """
 
     name: str
     dtype: FeastType
     description: str
     tags: Dict[str, str]
+    vector_index: bool
+    vector_search_metric: Optional[str]
 
     def __init__(
         self,
@@ -46,6 +50,8 @@ def __init__(
         dtype: FeastType,
         description: str = "",
         tags: Optional[Dict[str, str]] = None,
+        vector_index: bool = False,
+        vector_search_metric: Optional[str] = None,
     ):
         """
         Creates a Field object.
@@ -55,11 +61,15 @@ def __init__(
             dtype: The type of the field, such as string or float.
             description (optional): A human-readable description.
             tags (optional): User-defined metadata in dictionary form.
+            vector_index (optional): If set to True the field will be indexed for vector similarity search.
+            vector_search_metric (optional): The metric used for vector similarity search.
         """
         self.name = name
         self.dtype = dtype
         self.description = description
         self.tags = tags or {}
+        self.vector_index = vector_index
+        self.vector_search_metric = vector_search_metric
 
     def __eq__(self, other):
         if type(self) != type(other):
@@ -70,6 +80,8 @@ def __eq__(self, other):
             or self.dtype != other.dtype
             or self.description != other.description
             or self.tags != other.tags
+            # or self.vector_index != other.vector_index
+            # or self.vector_search_metric != other.vector_search_metric
         ):
             return False
         return True
@@ -87,6 +99,8 @@ def __repr__(self):
             f"    dtype={self.dtype!r},\n"
             f"    description={self.description!r},\n"
             f"    tags={self.tags!r}\n"
+            f"    vector_index={self.vector_index!r}\n"
+            f"    vector_search_metric={self.vector_search_metric!r}\n"
             f")"
         )
 
@@ -96,11 +110,14 @@ def __str__(self):
     def to_proto(self) -> FieldProto:
         """Converts a Field object to its protobuf representation."""
         value_type = self.dtype.to_value_type()
+        vector_search_metric = self.vector_search_metric or ""
         return FieldProto(
             name=self.name,
             value_type=value_type.value,
             description=self.description,
             tags=self.tags,
+            vector_index=self.vector_index,
+            vector_search_metric=vector_search_metric,
         )
 
     @classmethod
@@ -112,11 +129,15 @@ def from_proto(cls, field_proto: FieldProto):
             field_proto: FieldProto protobuf object
         """
         value_type = ValueType(field_proto.value_type)
+        vector_search_metric = getattr(field_proto, "vector_search_metric", "")
+        vector_index = getattr(field_proto, "vector_index", False)
         return cls(
             name=field_proto.name,
             dtype=from_value_type(value_type=value_type),
             tags=dict(field_proto.tags),
             description=field_proto.description,
+            vector_index=vector_index,
+            vector_search_metric=vector_search_metric,
         )
 
     @classmethod

sdk/python/feast/protos/feast/core/DataSource_pb2.py

@@ -557,3 +557,18 @@ class DataSource(google.protobuf.message.Message):
     def WhichOneof(self, oneof_group: typing_extensions.Literal["options", b"options"]) -> typing_extensions.Literal["file_options", "bigquery_options", "kafka_options", "kinesis_options", "redshift_options", "request_data_options", "custom_options", "snowflake_options", "push_options", "spark_options", "trino_options", "athena_options"] | None: ...
 
 global___DataSource = DataSource
+
+class DataSourceList(google.protobuf.message.Message):
+    DESCRIPTOR: google.protobuf.descriptor.Descriptor
+
+    DATASOURCES_FIELD_NUMBER: builtins.int
+    @property
+    def datasources(self) -> google.protobuf.internal.containers.RepeatedCompositeFieldContainer[global___DataSource]: ...
+    def __init__(
+        self,
+        *,
+        datasources: collections.abc.Iterable[global___DataSource] | None = ...,
+    ) -> None: ...
+    def ClearField(self, field_name: typing_extensions.Literal["datasources", b"datasources"]) -> None: ...
+
+global___DataSourceList = DataSourceList