feat: Update stream fcos to have watermark and sliding interval (#2765)

* Add sliding window to aggregations Signed-off-by: Kevin Zhang <[email protected]> * Fix Signed-off-by: Kevin Zhang <[email protected]> * update apis Signed-off-by: Kevin Zhang <[email protected]> * Lint Signed-off-by: Kevin Zhang <[email protected]> * Fix lint Signed-off-by: Kevin Zhang <[email protected]> * Fix Signed-off-by: Kevin Zhang <[email protected]> * Fix Signed-off-by: Kevin Zhang <[email protected]>
feast-dev · Jun 8, 2022 · 3256952 · 3256952
1 parent d25e8d4
commit 3256952
Show file tree

Hide file tree

Showing 6 changed files with 85 additions and 3 deletions.
diff --git a/protos/feast/core/Aggregation.proto b/protos/feast/core/Aggregation.proto
@@ -11,4 +11,5 @@ message Aggregation {
     string column = 1;
     string function = 2;
     google.protobuf.Duration time_window = 3;
+    google.protobuf.Duration slide_interval = 4;
 }
diff --git a/protos/feast/core/DataSource.proto b/protos/feast/core/DataSource.proto
@@ -22,6 +22,7 @@ option go_package = "github.com/feast-dev/feast/go/protos/feast/core";
 option java_outer_classname = "DataSourceProto";
 option java_package = "feast.proto.core";
 
+import "google/protobuf/duration.proto";
 import "feast/core/DataFormat.proto";
 import "feast/types/Value.proto";
 import "feast/core/Feature.proto";
@@ -135,6 +136,7 @@ message DataSource {
     // Defines the stream data format encoding feature/entity data in Kafka messages.
     StreamFormat message_format = 3;
 
+    google.protobuf.Duration watermark = 4;
   }
 
   // Defines options for DataSource that sources features from Kinesis records.

diff --git a/sdk/python/feast/aggregation.py b/sdk/python/feast/aggregation.py
@@ -14,30 +14,45 @@ class Aggregation:
         column: str  # Column name of the feature we are aggregating.
         function: str  # Provided built in aggregations sum, max, min, count mean
         time_window: timedelta  # The time window for this aggregation.
+        slide_interval: timedelta # The sliding window for these aggregations
     """
 
     column: str
     function: str
     time_window: Optional[timedelta]
+    slide_interval: Optional[timedelta]
 
     def __init__(
         self,
         column: Optional[str] = "",
         function: Optional[str] = "",
         time_window: Optional[timedelta] = None,
+        slide_interval: Optional[timedelta] = None,
     ):
         self.column = column or ""
         self.function = function or ""
         self.time_window = time_window
+        if not slide_interval:
+            self.slide_interval = self.time_window
+        else:
+            self.slide_interval = slide_interval
 
     def to_proto(self) -> AggregationProto:
         window_duration = None
         if self.time_window is not None:
             window_duration = Duration()
             window_duration.FromTimedelta(self.time_window)
 
+        slide_interval_duration = None
+        if self.slide_interval is not None:
+            slide_interval_duration = Duration()
+            slide_interval_duration.FromTimedelta(self.slide_interval)
+
         return AggregationProto(
-            column=self.column, function=self.function, time_window=window_duration
+            column=self.column,
+            function=self.function,
+            time_window=window_duration,
+            slide_interval=slide_interval_duration,
         )
 
     @classmethod
@@ -48,10 +63,16 @@ def from_proto(cls, agg_proto: AggregationProto):
             else agg_proto.time_window.ToTimedelta()
         )
 
+        slide_interval = (
+            timedelta(days=0)
+            if agg_proto.slide_interval.ToNanoseconds() == 0
+            else agg_proto.slide_interval.ToTimedelta()
+        )
         aggregation = cls(
             column=agg_proto.column,
             function=agg_proto.function,
             time_window=time_window,
+            slide_interval=slide_interval,
         )
         return aggregation
 
@@ -63,6 +84,7 @@ def __eq__(self, other):
             self.column != other.column
             or self.function != other.function
             or self.time_window != other.time_window
+            or self.slide_interval != other.slide_interval
         ):
             return False
 

diff --git a/sdk/python/feast/data_source.py b/sdk/python/feast/data_source.py
@@ -15,8 +15,10 @@
 import enum
 import warnings
 from abc import ABC, abstractmethod
+from datetime import timedelta
 from typing import Any, Callable, Dict, Iterable, List, Optional, Tuple, Union
 
+from google.protobuf.duration_pb2 import Duration
 from google.protobuf.json_format import MessageToJson
 
 from feast import type_map
@@ -47,11 +49,16 @@ class KafkaOptions:
     """
 
     def __init__(
-        self, bootstrap_servers: str, message_format: StreamFormat, topic: str,
+        self,
+        bootstrap_servers: str,
+        message_format: StreamFormat,
+        topic: str,
+        watermark: Optional[timedelta] = None,
     ):
         self.bootstrap_servers = bootstrap_servers
         self.message_format = message_format
         self.topic = topic
+        self.watermark = watermark or None
 
     @classmethod
     def from_proto(cls, kafka_options_proto: DataSourceProto.KafkaOptions):
@@ -64,11 +71,18 @@ def from_proto(cls, kafka_options_proto: DataSourceProto.KafkaOptions):
         Returns:
             Returns a BigQueryOptions object based on the kafka_options protobuf
         """
-
+        watermark = None
+        if kafka_options_proto.HasField("watermark"):
+            watermark = (
+                timedelta(days=0)
+                if kafka_options_proto.watermark.ToNanoseconds() == 0
+                else kafka_options_proto.watermark.ToTimedelta()
+            )
         kafka_options = cls(
             bootstrap_servers=kafka_options_proto.bootstrap_servers,
             message_format=StreamFormat.from_proto(kafka_options_proto.message_format),
             topic=kafka_options_proto.topic,
+            watermark=watermark,
         )
 
         return kafka_options
@@ -80,11 +94,16 @@ def to_proto(self) -> DataSourceProto.KafkaOptions:
         Returns:
             KafkaOptionsProto protobuf
         """
+        watermark_duration = None
+        if self.watermark is not None:
+            watermark_duration = Duration()
+            watermark_duration.FromTimedelta(self.watermark)
 
         kafka_options_proto = DataSourceProto.KafkaOptions(
             bootstrap_servers=self.bootstrap_servers,
             message_format=self.message_format.to_proto(),
             topic=self.topic,
+            watermark=watermark_duration,
         )
 
         return kafka_options_proto
@@ -369,7 +388,32 @@ def __init__(
         owner: Optional[str] = "",
         timestamp_field: Optional[str] = "",
         batch_source: Optional[DataSource] = None,
+        watermark: Optional[timedelta] = None,
     ):
+        """
+        Creates a KafkaSource stream source object.
+        Args:
+            name: str. Name of data source, which should be unique within a project
+            event_timestamp_column (optional): str. (Deprecated) Event timestamp column used for point in time
+                joins of feature values.
+            bootstrap_servers: str. The servers of the kafka broker in the form "localhost:9092".
+            message_format: StreamFormat. StreamFormat of serialized messages.
+            topic: str. The name of the topic to read from in the kafka source.
+            created_timestamp_column (optional): str. Timestamp column indicating when the row
+                was created, used for deduplicating rows.
+            field_mapping (optional): dict(str, str). 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 entity or timestamp columns.
+            date_partition_column (optional): str. Timestamp column used for partitioning.
+            description (optional): str. A human-readable description.
+            tags (optional): dict(str, str). A dictionary of key-value pairs to store arbitrary metadata.
+            owner (optional): str. The owner of the data source, typically the email of the primary
+                maintainer.
+            timestamp_field (optional): str. Event timestamp field used for point
+                in time joins of feature values.
+            batch_source: DataSource. The datasource that acts as a batch source.
+            watermark: timedelta. The watermark for stream data. Specifically how late stream data can arrive without being discarded.
+        """
         positional_attributes = [
             "name",
             "event_timestamp_column",
@@ -425,10 +469,12 @@ def __init__(
             timestamp_field=timestamp_field,
         )
         self.batch_source = batch_source
+
         self.kafka_options = KafkaOptions(
             bootstrap_servers=_bootstrap_servers,
             message_format=_message_format,
             topic=_topic,
+            watermark=watermark,
         )
 
     def __eq__(self, other):
@@ -445,6 +491,7 @@ def __eq__(self, other):
             != other.kafka_options.bootstrap_servers
             or self.kafka_options.message_format != other.kafka_options.message_format
             or self.kafka_options.topic != other.kafka_options.topic
+            or self.kafka_options.watermark != other.kafka_options.watermark
         ):
             return False
 
@@ -455,6 +502,13 @@ def __hash__(self):
 
     @staticmethod
     def from_proto(data_source: DataSourceProto):
+        watermark = None
+        if data_source.kafka_options.HasField("watermark"):
+            watermark = (
+                timedelta(days=0)
+                if data_source.kafka_options.watermark.ToNanoseconds() == 0
+                else data_source.kafka_options.watermark.ToTimedelta()
+            )
         return KafkaSource(
             name=data_source.name,
             event_timestamp_column=data_source.timestamp_field,
@@ -463,6 +517,7 @@ def from_proto(data_source: DataSourceProto):
             message_format=StreamFormat.from_proto(
                 data_source.kafka_options.message_format
             ),
+            watermark=watermark,
             topic=data_source.kafka_options.topic,
             created_timestamp_column=data_source.created_timestamp_column,
             timestamp_field=data_source.timestamp_field,

diff --git a/sdk/python/tests/integration/registration/test_registry.py b/sdk/python/tests/integration/registration/test_registry.py
@@ -319,6 +319,7 @@ def simple_udf(x: int):
         message_format=AvroFormat(""),
         topic="topic",
         batch_source=FileSource(path="some path"),
+        watermark=timedelta(days=1),
     )
 
     sfv = StreamFeatureView(

diff --git a/sdk/python/tests/integration/registration/test_stream_feature_view_apply.py b/sdk/python/tests/integration/registration/test_stream_feature_view_apply.py
@@ -27,6 +27,7 @@ def test_apply_stream_feature_view(environment) -> None:
         message_format=AvroFormat(""),
         topic="topic",
         batch_source=FileSource(path="test_path", timestamp_field="event_timestamp"),
+        watermark=timedelta(days=1),
     )
 
     @stream_feature_view(