airbyte-cdk 6.7.2__tar.gz → 6.7.2.dev1__tar.gz

{airbyte_cdk-6.7.2 → airbyte_cdk-6.7.2.dev1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: airbyte-cdk
-Version: 6.7.2
+Version: 6.7.2.dev1
 Summary: A framework for writing Airbyte Connectors.
 Home-page: https://airbyte.com
 License: MIT

{airbyte_cdk-6.7.2 → airbyte_cdk-6.7.2.dev1}/airbyte_cdk/sources/declarative/concurrent_declarative_source.py

@@ -49,6 +49,7 @@ from airbyte_cdk.sources.streams.concurrent.abstract_stream import AbstractStrea
 from airbyte_cdk.sources.streams.concurrent.availability_strategy import (
     AlwaysAvailableAvailabilityStrategy,
 )
+from airbyte_cdk.sources.streams.concurrent.cursor import FinalStateCursor
 from airbyte_cdk.sources.streams.concurrent.default_stream import DefaultStream
 from airbyte_cdk.sources.streams.concurrent.helpers import get_primary_key_from_stream
 from airbyte_cdk.sources.types import Config, StreamState
@@ -69,6 +70,15 @@ class ConcurrentDeclarativeSource(ManifestDeclarativeSource, Generic[TState]):
         component_factory: Optional[ModelToComponentFactory] = None,
         **kwargs: Any,
     ) -> None:
+        # To reduce the complexity of the concurrent framework, we are not enabling RFR with synthetic
+        # cursors. We do this by no longer automatically instantiating RFR cursors when converting
+        # the declarative models into runtime components. Concurrent sources will continue to checkpoint
+        # incremental streams running in full refresh.
+        component_factory = component_factory or ModelToComponentFactory(
+            emit_connector_builder_messages=emit_connector_builder_messages,
+            disable_resumable_full_refresh=True,
+        )
+
         super().__init__(
             source_config=source_config,
             debug=debug,
@@ -191,13 +201,24 @@ class ConcurrentDeclarativeSource(ManifestDeclarativeSource, Generic[TState]):
             # these legacy Python streams the way we do low-code streams to determine if they are concurrent compatible,
             # so we need to treat them as synchronous
             if isinstance(declarative_stream, DeclarativeStream):
-                datetime_based_cursor_component_definition = name_to_stream_mapping[
+                incremental_sync_component_definition = name_to_stream_mapping[
                     declarative_stream.name
                 ].get("incremental_sync")
 
+                partition_router_component_definition = (
+                    name_to_stream_mapping[declarative_stream.name]
+                    .get("retriever")
+                    .get("partition_router")
+                )
+
+                is_substream_without_incremental = (
+                    partition_router_component_definition
+                    and not incremental_sync_component_definition
+                )
+
                 if (
-                    datetime_based_cursor_component_definition
-                    and datetime_based_cursor_component_definition.get("type", "")
+                    incremental_sync_component_definition
+                    and incremental_sync_component_definition.get("type", "")
                     == DatetimeBasedCursorModel.__name__
                     and self._stream_supports_concurrent_partition_processing(
                         declarative_stream=declarative_stream
@@ -213,7 +234,7 @@ class ConcurrentDeclarativeSource(ManifestDeclarativeSource, Generic[TState]):
                         self._constructor.create_concurrent_cursor_from_datetime_based_cursor(
                             state_manager=state_manager,
                             model_type=DatetimeBasedCursorModel,
-                            component_definition=datetime_based_cursor_component_definition,
+                            component_definition=incremental_sync_component_definition,
                             stream_name=declarative_stream.name,
                             stream_namespace=declarative_stream.namespace,
                             config=config or {},
@@ -247,6 +268,41 @@ class ConcurrentDeclarativeSource(ManifestDeclarativeSource, Generic[TState]):
                             cursor=cursor,
                         )
                     )
+                elif is_substream_without_incremental and hasattr(
+                    declarative_stream.retriever, "stream_slicer"
+                ):
+                    partition_generator = StreamSlicerPartitionGenerator(
+                        DeclarativePartitionFactory(
+                            declarative_stream.name,
+                            declarative_stream.get_json_schema(),
+                            self._retriever_factory(
+                                name_to_stream_mapping[declarative_stream.name],
+                                config,
+                                {},
+                            ),
+                            self.message_repository,
+                        ),
+                        declarative_stream.retriever.stream_slicer,
+                    )
+
+                    final_state_cursor = FinalStateCursor(
+                        stream_name=declarative_stream.name,
+                        stream_namespace=declarative_stream.namespace,
+                        message_repository=self.message_repository,
+                    )
+
+                    concurrent_streams.append(
+                        DefaultStream(
+                            partition_generator=partition_generator,
+                            name=declarative_stream.name,
+                            json_schema=declarative_stream.get_json_schema(),
+                            availability_strategy=AlwaysAvailableAvailabilityStrategy(),
+                            primary_key=get_primary_key_from_stream(declarative_stream.primary_key),
+                            cursor_field=None,
+                            logger=self.logger,
+                            cursor=final_state_cursor,
+                        )
+                    )
                 else:
                     synchronous_streams.append(declarative_stream)
             else:

{airbyte_cdk-6.7.2 → airbyte_cdk-6.7.2.dev1}/airbyte_cdk/sources/declarative/parsers/model_to_component_factory.py

@@ -387,6 +387,7 @@ class ModelToComponentFactory:
         emit_connector_builder_messages: bool = False,
         disable_retries: bool = False,
         disable_cache: bool = False,
+        disable_resumable_full_refresh: bool = False,
         message_repository: Optional[MessageRepository] = None,
     ):
         self._init_mappings()
@@ -395,6 +396,7 @@ class ModelToComponentFactory:
         self._emit_connector_builder_messages = emit_connector_builder_messages
         self._disable_retries = disable_retries
         self._disable_cache = disable_cache
+        self._disable_resumable_full_refresh = disable_resumable_full_refresh
         self._message_repository = message_repository or InMemoryMessageRepository(  # type: ignore
             self._evaluate_log_level(emit_connector_builder_messages)
         )
@@ -1339,6 +1341,8 @@ class ModelToComponentFactory:
                 if model.incremental_sync
                 else None
             )
+        elif self._disable_resumable_full_refresh:
+            return stream_slicer
         elif stream_slicer:
             # For the Full-Refresh sub-streams, we use the nested `ChildPartitionResumableFullRefreshCursor`
             return PerPartitionCursor(

{airbyte_cdk-6.7.2 → airbyte_cdk-6.7.2.dev1}/airbyte_cdk/sources/streams/concurrent/cursor.py

@@ -5,7 +5,18 @@
 import functools
 import logging
 from abc import ABC, abstractmethod
-from typing import Any, Callable, Iterable, List, Mapping, MutableMapping, Optional, Protocol, Tuple
+from typing import (
+    Any,
+    Callable,
+    Iterable,
+    List,
+    Mapping,
+    MutableMapping,
+    Optional,
+    Protocol,
+    Tuple,
+    Union,
+)
 
 from airbyte_cdk.sources.connector_state_manager import ConnectorStateManager
 from airbyte_cdk.sources.message import MessageRepository
@@ -175,7 +186,9 @@ class ConcurrentCursor(Cursor):
         self.start, self._concurrent_state = self._get_concurrent_state(stream_state)
         self._lookback_window = lookback_window
         self._slice_range = slice_range
-        self._most_recent_cursor_value_per_partition: MutableMapping[StreamSlice, Any] = {}
+        self._most_recent_cursor_value_per_partition: MutableMapping[
+            Union[StreamSlice, Mapping[str, Any], None], Any
+        ] = {}
         self._has_closed_at_least_one_slice = False
         self._cursor_granularity = cursor_granularity
         # Flag to track if the logger has been triggered (per stream)
@@ -216,10 +229,13 @@ class ConcurrentCursor(Cursor):
         most_recent_cursor_value = self._most_recent_cursor_value_per_partition.get(
             record.associated_slice
         )
-        cursor_value = self._extract_cursor_value(record)
+        try:
+            cursor_value = self._extract_cursor_value(record)
 
-        if most_recent_cursor_value is None or most_recent_cursor_value < cursor_value:
-            self._most_recent_cursor_value_per_partition[record.associated_slice] = cursor_value
+            if most_recent_cursor_value is None or most_recent_cursor_value < cursor_value:
+                self._most_recent_cursor_value_per_partition[record.associated_slice] = cursor_value
+        except ValueError:
+            self._log_for_record_without_cursor_value()
 
     def _extract_cursor_value(self, record: Record) -> Any:
         return self._connector_state_converter.parse_value(self._cursor_field.extract_value(record))
@@ -459,10 +475,13 @@ class ConcurrentCursor(Cursor):
         try:
             record_cursor_value: CursorValueType = self._extract_cursor_value(record)  # type: ignore  # cursor_field is converted to an InterpolatedString in __post_init__
         except ValueError:
-            if not self._should_be_synced_logger_triggered:
-                LOGGER.warning(
-                    f"Could not find cursor field `{self.cursor_field.cursor_field_key}` in record. The incremental sync will assume it needs to be synced"
-                )
-                self._should_be_synced_logger_triggered = True
+            self._log_for_record_without_cursor_value()
             return True
         return self.start <= record_cursor_value <= self._end_provider()
+
+    def _log_for_record_without_cursor_value(self) -> None:
+        if not self._should_be_synced_logger_triggered:
+            LOGGER.warning(
+                f"Could not find cursor field `{self.cursor_field.cursor_field_key}` in record for stream {self._stream_name}. The incremental sync will assume it needs to be synced"
+            )
+            self._should_be_synced_logger_triggered = True

{airbyte_cdk-6.7.2 → airbyte_cdk-6.7.2.dev1}/airbyte_cdk/sources/streams/concurrent/default_stream.py

@@ -67,6 +67,7 @@ class DefaultStream(AbstractStream):
             name=self.name,
             json_schema=dict(self._json_schema),
             supported_sync_modes=[SyncMode.full_refresh],
+            is_resumable=False,
         )
 
         if self._namespace:

{airbyte_cdk-6.7.2 → airbyte_cdk-6.7.2.dev1}/pyproject.toml

@@ -23,7 +23,7 @@ classifiers = [
 keywords = ["airbyte", "connector-development-kit", "cdk"]
 
 # Python CDK uses dynamic versioning: https://github.com/mtkennerly/poetry-dynamic-versioning
-version = "6.7.2"  # Version will be calculated dynamically.
+version = "6.7.2dev1"  # Version will be calculated dynamically.
 
 [tool.poetry-dynamic-versioning]
 enable = false