airbyte-cdk 6.61.6__tar.gz → 6.62.0.dev1__tar.gz

{airbyte_cdk-6.61.6 → airbyte_cdk-6.62.0.dev1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: airbyte-cdk
-Version: 6.61.6
+Version: 6.62.0.dev1
 Summary: A framework for writing Airbyte Connectors.
 Home-page: https://airbyte.com
 License: MIT
@@ -35,7 +35,6 @@ Requires-Dist: click (>=8.1.8,<9.0.0)
 Requires-Dist: cohere (>=4.21,<6.0.0) ; extra == "vector-db-based"
 Requires-Dist: cryptography (>=44.0.0,<45.0.0)
 Requires-Dist: dateparser (>=1.2.2,<2.0.0)
-Requires-Dist: ddtrace (>=3.12.3,<4.0.0) ; extra == "manifest-server"
 Requires-Dist: dpath (>=2.1.6,<3.0.0)
 Requires-Dist: dunamai (>=1.22.0,<2.0.0)
 Requires-Dist: fastapi (>=0.116.1) ; extra == "manifest-server"
@@ -44,7 +43,7 @@ Requires-Dist: genson (==1.3.0)
 Requires-Dist: google-cloud-secret-manager (>=2.17.0,<3.0.0)
 Requires-Dist: isodate (>=0.6.1,<0.7.0)
 Requires-Dist: jsonref (>=0.2,<0.3)
-Requires-Dist: jsonschema (>=4.17.3,<5.0)
+Requires-Dist: jsonschema (>=4.17.3,<4.18.0)
 Requires-Dist: langchain (==0.1.16) ; extra == "vector-db-based"
 Requires-Dist: langchain_core (==0.1.42)
 Requires-Dist: markdown ; extra == "file-based"
@@ -69,7 +68,6 @@ Requires-Dist: python-snappy (==0.7.3) ; extra == "file-based"
 Requires-Dist: python-ulid (>=3.0.0,<4.0.0)
 Requires-Dist: pytz (==2024.2)
 Requires-Dist: rapidfuzz (>=3.10.1,<4.0.0)
-Requires-Dist: referencing (>=0.36.2)
 Requires-Dist: requests
 Requires-Dist: requests_cache
 Requires-Dist: rich

{airbyte_cdk-6.61.6 → airbyte_cdk-6.62.0.dev1}/airbyte_cdk/connector_builder/connector_builder_handler.py

@@ -62,10 +62,10 @@ def should_normalize_manifest(config: Mapping[str, Any]) -> bool:
 
 def create_source(
     config: Mapping[str, Any],
-    limits: TestLimits | None = None,
-    catalog: ConfiguredAirbyteCatalog | None = None,
-    state: List[AirbyteStateMessage] | None = None,
-) -> ConcurrentDeclarativeSource:
+    limits: TestLimits,
+    catalog: Optional[ConfiguredAirbyteCatalog],
+    state: Optional[List[AirbyteStateMessage]],
+) -> ConcurrentDeclarativeSource[Optional[List[AirbyteStateMessage]]]:
     manifest = config["__injected_declarative_manifest"]
 
     # We enforce a concurrency level of 1 so that the stream is processed on a single thread
@@ -88,7 +88,7 @@ def create_source(
 
 
 def read_stream(
-    source: ConcurrentDeclarativeSource,
+    source: ConcurrentDeclarativeSource[Optional[List[AirbyteStateMessage]]],
     config: Mapping[str, Any],
     configured_catalog: ConfiguredAirbyteCatalog,
     state: List[AirbyteStateMessage],
@@ -127,7 +127,7 @@ def read_stream(
 
 
 def resolve_manifest(
-    source: ConcurrentDeclarativeSource,
+    source: ConcurrentDeclarativeSource[Optional[List[AirbyteStateMessage]]],
 ) -> AirbyteMessage:
     try:
         return AirbyteMessage(
@@ -146,7 +146,7 @@ def resolve_manifest(
 
 
 def full_resolve_manifest(
-    source: ConcurrentDeclarativeSource, limits: TestLimits
+    source: ConcurrentDeclarativeSource[Optional[List[AirbyteStateMessage]]], limits: TestLimits
 ) -> AirbyteMessage:
     try:
         manifest = {**source.resolved_manifest}

{airbyte_cdk-6.61.6 → airbyte_cdk-6.62.0.dev1}/airbyte_cdk/connector_builder/main.py

@@ -34,7 +34,7 @@ from airbyte_cdk.utils.traced_exception import AirbyteTracedException
 
 def get_config_and_catalog_from_args(
     args: List[str],
-) -> Tuple[str, Mapping[str, Any], Optional[ConfiguredAirbyteCatalog], List[AirbyteStateMessage]]:
+) -> Tuple[str, Mapping[str, Any], Optional[ConfiguredAirbyteCatalog], Any]:
     # TODO: Add functionality for the `debug` logger.
     #  Currently, no one `debug` level log will be displayed during `read` a stream for a connector created through `connector-builder`.
     parsed_args = AirbyteEntrypoint.parse_args(args)
@@ -70,7 +70,7 @@ def get_config_and_catalog_from_args(
 
 
 def handle_connector_builder_request(
-    source: ConcurrentDeclarativeSource,
+    source: ConcurrentDeclarativeSource[Optional[List[AirbyteStateMessage]]],
     command: str,
     config: Mapping[str, Any],
     catalog: Optional[ConfiguredAirbyteCatalog],

{airbyte_cdk-6.61.6 → airbyte_cdk-6.62.0.dev1}/airbyte_cdk/connector_builder/test_reader/reader.py

@@ -85,7 +85,7 @@ class TestReader:
 
     def run_test_read(
         self,
-        source: ConcurrentDeclarativeSource,
+        source: ConcurrentDeclarativeSource[Optional[List[AirbyteStateMessage]]],
         config: Mapping[str, Any],
         configured_catalog: ConfiguredAirbyteCatalog,
         stream_name: str,
@@ -383,7 +383,7 @@ class TestReader:
 
     def _read_stream(
         self,
-        source: ConcurrentDeclarativeSource,
+        source: ConcurrentDeclarativeSource[Optional[List[AirbyteStateMessage]]],
         config: Mapping[str, Any],
         configured_catalog: ConfiguredAirbyteCatalog,
         state: List[AirbyteStateMessage],

{airbyte_cdk-6.61.6 → airbyte_cdk-6.62.0.dev1}/airbyte_cdk/legacy/sources/declarative/incremental/per_partition_cursor.py

@@ -146,8 +146,10 @@ class PerPartitionCursor(DeclarativeCursor):
             if "state" in stream_state:
                 self._state_to_migrate_from = stream_state["state"]
 
-        # Set parent state for partition routers based on parent streams
-        self._partition_router.set_initial_state(stream_state)
+        # We used to set the parent state through this method but since moving the SubstreamPartitionRouter to the
+        # Concurrent CDK/AbstractStream, the state is passed at the __init__ stage and this does not need to be called.
+        # We are still keeping this line as a comment to be explicit about the past behavior.
+        # self._partition_router.set_initial_state(stream_state)
 
     def observe(self, stream_slice: StreamSlice, record: Record) -> None:
         self._cursor_per_partition[self._to_partition_key(stream_slice.partition)].observe(

{airbyte_cdk-6.61.6 → airbyte_cdk-6.62.0.dev1}/airbyte_cdk/manifest_server/Dockerfile

@@ -11,7 +11,7 @@ FROM python:3.12-slim-bookworm
 RUN apt-get update && \
     apt-get install -y git && \
     rm -rf /var/lib/apt/lists/* && \
-    pip install poetry==2.0.1
+    pip install poetry==1.8.3
 
 # Configure poetry to not create virtual environments and disable interactive mode
 ENV POETRY_NO_INTERACTION=1 \
@@ -42,4 +42,4 @@ USER airbyte:airbyte
 
 EXPOSE 8080
 
-CMD ["uvicorn", "airbyte_cdk.manifest_server.app:app", "--host", "0.0.0.0", "--port", "8080"]
+CMD ["uvicorn", "airbyte_cdk.manifest_server.app:app", "--host", "0.0.0.0", "--port", "8080"]

{airbyte_cdk-6.61.6 → airbyte_cdk-6.62.0.dev1}/airbyte_cdk/manifest_server/README.md

@@ -154,25 +154,3 @@ docker run -p 8080:8080 manifest-server
 ```
 
 Note: The container runs on port 8080 by default.
-
-## Datadog APM
-
-The manifest server supports Datadog APM tracing for monitoring and observability:
-
-### Configuration
-
-To enable Datadog tracing, set the environment variable:
-
-```bash
-export DD_ENABLED=true
-```
-
-This requires the `ddtrace` dependency, which is included in the `manifest-server` extra. For additional configuration options via environment variables, see [ddtrace configuration](https://ddtrace.readthedocs.io/en/stable/configuration.html).
-
-### Usage
-
-```bash
-# Run with Datadog tracing enabled
-DD_ENABLED=true manifest-server start
-```
-

{airbyte_cdk-6.61.6 → airbyte_cdk-6.62.0.dev1}/airbyte_cdk/manifest_server/app.py

@@ -1,9 +1,3 @@
-import os
-
-if os.getenv("DD_ENABLED", "false").lower() == "true":
-    # Auto-instrumentation should be imported as early as possible.
-    import ddtrace.auto  # noqa: F401
-
 from fastapi import FastAPI
 
 from .routers import capabilities, health, manifest

{airbyte_cdk-6.61.6 → airbyte_cdk-6.62.0.dev1}/airbyte_cdk/manifest_server/cli/_common.py

@@ -7,7 +7,6 @@ import rich_click as click
 
 # Import server dependencies with graceful fallback
 try:
-    import ddtrace  # noqa: F401
     import fastapi  # noqa: F401
     import uvicorn  # noqa: F401
 

{airbyte_cdk-6.61.6 → airbyte_cdk-6.62.0.dev1}/airbyte_cdk/manifest_server/command_processor/processor.py

@@ -21,10 +21,12 @@ from airbyte_cdk.test.entrypoint_wrapper import AirbyteEntrypointException, Entr
 
 
 class ManifestCommandProcessor:
-    _source: ConcurrentDeclarativeSource
+    _source: ConcurrentDeclarativeSource[Optional[List[AirbyteStateMessage]]]
     _logger = logging.getLogger("airbyte.manifest-server")
 
-    def __init__(self, source: ConcurrentDeclarativeSource) -> None:
+    def __init__(
+        self, source: ConcurrentDeclarativeSource[Optional[List[AirbyteStateMessage]]]
+    ) -> None:
         self._source = source
 
     def test_read(
@@ -39,6 +41,7 @@ class ManifestCommandProcessor:
         """
         Test the read method of the source.
         """
+
         test_read_handler = TestReader(
             max_pages_per_slice=page_limit,
             max_slices=slice_limit,

{airbyte_cdk-6.61.6 → airbyte_cdk-6.62.0.dev1}/airbyte_cdk/manifest_server/command_processor/utils.py

@@ -63,7 +63,7 @@ def build_source(
     record_limit: Optional[int] = None,
     page_limit: Optional[int] = None,
     slice_limit: Optional[int] = None,
-) -> ConcurrentDeclarativeSource:
+) -> ConcurrentDeclarativeSource[Optional[List[AirbyteStateMessage]]]:
     # We enforce a concurrency level of 1 so that the stream is processed on a single thread
     # to retain ordering for the grouping of the builder message responses.
     definition = copy.deepcopy(manifest)

{airbyte_cdk-6.61.6 → airbyte_cdk-6.62.0.dev1}/airbyte_cdk/manifest_server/routers/manifest.py

@@ -40,7 +40,7 @@ def safe_build_source(
     page_limit: Optional[int] = None,
     slice_limit: Optional[int] = None,
     record_limit: Optional[int] = None,
-) -> ConcurrentDeclarativeSource:
+) -> ConcurrentDeclarativeSource[Optional[List[AirbyteStateMessage]]]:
     """Wrapper around build_source that converts ValidationError to HTTPException."""
     try:
         return build_source(

{airbyte_cdk-6.61.6 → airbyte_cdk-6.62.0.dev1}/airbyte_cdk/sources/declarative/concurrent_declarative_source.py

@@ -162,17 +162,16 @@ def _get_declarative_component_schema() -> Dict[str, Any]:
 #  is no longer inherited from since the only external dependency is from that class.
 #
 # todo: It is worth investigating removal of the Generic[TState] since it will always be Optional[List[AirbyteStateMessage]]
-class ConcurrentDeclarativeSource(AbstractSource):
+class ConcurrentDeclarativeSource(AbstractSource, Generic[TState]):
     # By default, we defer to a value of 2. A value lower than could cause a PartitionEnqueuer to be stuck in a state of deadlock
     # because it has hit the limit of futures but not partition reader is consuming them.
     _LOWEST_SAFE_CONCURRENCY_LEVEL = 2
 
     def __init__(
         self,
-        catalog: Optional[ConfiguredAirbyteCatalog] = None,
-        config: Optional[Mapping[str, Any]] = None,
-        state: Optional[List[AirbyteStateMessage]] = None,
-        *,
+        catalog: Optional[ConfiguredAirbyteCatalog],
+        config: Optional[Mapping[str, Any]],
+        state: TState,
         source_config: ConnectionDefinition,
         debug: bool = False,
         emit_connector_builder_messages: bool = False,
@@ -704,7 +703,7 @@ class ConcurrentDeclarativeSource(AbstractSource):
                             stream_slicer=declarative_stream.retriever.stream_slicer,
                             slice_limit=self._limits.max_slices
                             if self._limits
-                            else None,  # technically not needed because create_declarative_stream() -> create_simple_retriever() will apply the decorator. But for consistency and depending how we build create_default_stream, this may be needed later
+                            else None,  # technically not needed because create_default_stream() -> create_simple_retriever() will apply the decorator. But for consistency and depending how we build create_default_stream, this may be needed later
                         )
                     else:
                         if (
@@ -773,7 +772,7 @@ class ConcurrentDeclarativeSource(AbstractSource):
                         declarative_stream.retriever.stream_slicer,
                         slice_limit=self._limits.max_slices
                         if self._limits
-                        else None,  # technically not needed because create_declarative_stream() -> create_simple_retriever() will apply the decorator. But for consistency and depending how we build create_default_stream, this may be needed later
+                        else None,  # technically not needed because create_default_stream() -> create_simple_retriever() will apply the decorator. But for consistency and depending how we build create_default_stream, this may be needed later
                     )
 
                     final_state_cursor = FinalStateCursor(

{airbyte_cdk-6.61.6 → airbyte_cdk-6.62.0.dev1}/airbyte_cdk/sources/declarative/incremental/concurrent_partition_cursor.py

@@ -11,6 +11,13 @@ from copy import deepcopy
 from datetime import timedelta
 from typing import Any, Callable, Iterable, List, Mapping, MutableMapping, Optional
 
+from airbyte_cdk.models import (
+    AirbyteStateBlob,
+    AirbyteStateMessage,
+    AirbyteStateType,
+    AirbyteStreamState,
+    StreamDescriptor,
+)
 from airbyte_cdk.sources.connector_state_manager import ConnectorStateManager
 from airbyte_cdk.sources.declarative.incremental.global_substream_cursor import (
     Timer,
@@ -48,7 +55,7 @@ class ConcurrentPerPartitionCursor(Cursor):
     Manages state per partition when a stream has many partitions, preventing data loss or duplication.
 
     Attributes:
-        DEFAULT_MAX_PARTITIONS_NUMBER (int): Maximum number of partitions to retain in memory (default is 10,000).
+        DEFAULT_MAX_PARTITIONS_NUMBER (int): Maximum number of partitions to retain in memory (default is 10,000). This limit needs to be higher than the number of threads we might enqueue (which is represented by ThreadPoolManager.DEFAULT_MAX_QUEUE_SIZE). If not, we could have partitions that have been generated and submitted to the ThreadPool but got deleted from the ConcurrentPerPartitionCursor and when closing them, it will generate KeyError.
 
     - **Partition Limitation Logic**
       Ensures the number of tracked partitions does not exceed the specified limit to prevent memory overuse. Oldest partitions are removed when the limit is reached.
@@ -128,6 +135,7 @@ class ConcurrentPerPartitionCursor(Cursor):
 
         # FIXME this is a temporary field the time of the migration from declarative cursors to concurrent ones
         self._attempt_to_create_cursor_if_not_provided = attempt_to_create_cursor_if_not_provided
+        self._synced_some_data = False
 
     @property
     def cursor_field(self) -> CursorField:
@@ -168,8 +176,8 @@ class ConcurrentPerPartitionCursor(Cursor):
         with self._lock:
             self._semaphore_per_partition[partition_key].acquire()
             if not self._use_global_cursor:
-                self._cursor_per_partition[partition_key].close_partition(partition=partition)
                 cursor = self._cursor_per_partition[partition_key]
+                cursor.close_partition(partition=partition)
                 if (
                     partition_key in self._partitions_done_generating_stream_slices
                     and self._semaphore_per_partition[partition_key]._value == 0
@@ -213,8 +221,10 @@ class ConcurrentPerPartitionCursor(Cursor):
         if not any(
             semaphore_item[1]._value for semaphore_item in self._semaphore_per_partition.items()
         ):
-            self._global_cursor = self._new_global_cursor
-            self._lookback_window = self._timer.finish()
+            if self._synced_some_data:
+                # we only update those if we actually synced some data
+                self._global_cursor = self._new_global_cursor
+                self._lookback_window = self._timer.finish()
             self._parent_state = self._partition_router.get_stream_state()
         self._emit_state_message(throttle=False)
 
@@ -422,9 +432,6 @@ class ConcurrentPerPartitionCursor(Cursor):
         if stream_state.get("parent_state"):
             self._parent_state = stream_state["parent_state"]
 
-        # Set parent state for partition routers based on parent streams
-        self._partition_router.set_initial_state(stream_state)
-
     def _set_global_state(self, stream_state: Mapping[str, Any]) -> None:
         """
         Initializes the global cursor state from the provided stream state.
@@ -458,6 +465,7 @@ class ConcurrentPerPartitionCursor(Cursor):
         except ValueError:
             return
 
+        self._synced_some_data = True
         record_cursor = self._connector_state_converter.output_format(
             self._connector_state_converter.parse_value(record_cursor_value)
         )
@@ -541,3 +549,45 @@ class ConcurrentPerPartitionCursor(Cursor):
 
     def limit_reached(self) -> bool:
         return self._number_of_partitions > self.SWITCH_TO_GLOBAL_LIMIT
+
+    @staticmethod
+    def get_parent_state(
+        stream_state: Optional[StreamState], parent_stream_name: str
+    ) -> Optional[AirbyteStateMessage]:
+        if not stream_state:
+            return None
+
+        if "parent_state" not in stream_state:
+            logger.warning(
+                f"Trying to get_parent_state for stream `{parent_stream_name}` when there are not parent state in the state"
+            )
+            return None
+        elif parent_stream_name not in stream_state["parent_state"]:
+            logger.info(
+                f"Could not find parent state for stream `{parent_stream_name}`. On parents available are {list(stream_state['parent_state'].keys())}"
+            )
+            return None
+
+        return AirbyteStateMessage(
+            type=AirbyteStateType.STREAM,
+            stream=AirbyteStreamState(
+                stream_descriptor=StreamDescriptor(parent_stream_name, None),
+                stream_state=AirbyteStateBlob(stream_state["parent_state"][parent_stream_name]),
+            ),
+        )
+
+    @staticmethod
+    def get_global_state(
+        stream_state: Optional[StreamState], parent_stream_name: str
+    ) -> Optional[AirbyteStateMessage]:
+        return (
+            AirbyteStateMessage(
+                type=AirbyteStateType.STREAM,
+                stream=AirbyteStreamState(
+                    stream_descriptor=StreamDescriptor(parent_stream_name, None),
+                    stream_state=AirbyteStateBlob(stream_state["state"]),
+                ),
+            )
+            if stream_state and "state" in stream_state
+            else None
+        )

{airbyte_cdk-6.61.6 → airbyte_cdk-6.62.0.dev1}/airbyte_cdk/sources/declarative/incremental/global_substream_cursor.py

@@ -192,8 +192,10 @@ class GlobalSubstreamCursor(DeclarativeCursor):
             # Example: {"global_state_format_key": "global_state_format_value"}
             self._stream_cursor.set_initial_state(stream_state)
 
-        # Set parent state for partition routers based on parent streams
-        self._partition_router.set_initial_state(stream_state)
+        # We used to set the parent state through this method but since moving the SubstreamPartitionRouter to the
+        # Concurrent CDK/AbstractStream, the state is passed at the __init__ stage and this does not need to be called.
+        # We are still keeping this line as a comment to be explicit about the past behavior.
+        # self._partition_router.set_initial_state(stream_state)
 
     def _inject_lookback_into_stream_cursor(self, lookback_window: int) -> None:
         """