Skip to content

feat: support state migrations that affects parent streams #770

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 5 commits into from
Sep 22, 2025
Merged

feat: support state migrations that affects parent streams #770

Show file tree
Hide file tree
Changes from 2 commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
568c50a
support state migrations that affects parent streams
maxi297 Sep 22, 2025
197a071
Auto-fix lint and format issues
Sep 22, 2025
54f7b0f
mypy
maxi297 Sep 22, 2025
9f9ac58
mypy again
maxi297 Sep 22, 2025
0b159f7
fix coderabbitai stuff
maxi297 Sep 22, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
63 changes: 23 additions & 40 deletions airbyte_cdk/sources/declarative/parsers/model_to_component_factory.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1264,22 +1264,12 @@ def create_concurrent_cursor_from_datetime_based_cursor(
component_definition: ComponentDefinition,
stream_name: str,
stream_namespace: Optional[str],
stream_state: MutableMapping[str, Any],
config: Config,
message_repository: Optional[MessageRepository] = None,
runtime_lookback_window: Optional[datetime.timedelta] = None,
stream_state_migrations: Optional[List[Any]] = None,
**kwargs: Any,
) -> ConcurrentCursor:
# Per-partition incremental streams can dynamically create child cursors which will pass their current
# state via the stream_state keyword argument. Incremental syncs without parent streams use the
# incoming state and connector_state_manager that is initialized when the component factory is created
stream_state = (
self._connector_state_manager.get_stream_state(stream_name, stream_namespace)
if "stream_state" not in kwargs
else kwargs["stream_state"]
)
stream_state = self.apply_stream_state_migrations(stream_state_migrations, stream_state)

component_type = component_definition.get("type")
if component_definition.get("type") != model_type.__name__:
raise ValueError(
Expand Down Expand Up @@ -1498,21 +1488,11 @@ def create_concurrent_cursor_from_incrementing_count_cursor(
component_definition: ComponentDefinition,
stream_name: str,
stream_namespace: Optional[str],
stream_state: MutableMapping[str, Any],
config: Config,
message_repository: Optional[MessageRepository] = None,
stream_state_migrations: Optional[List[Any]] = None,
**kwargs: Any,
) -> ConcurrentCursor:
# Per-partition incremental streams can dynamically create child cursors which will pass their current
# state via the stream_state keyword argument. Incremental syncs without parent streams use the
# incoming state and connector_state_manager that is initialized when the component factory is created
stream_state = (
self._connector_state_manager.get_stream_state(stream_name, stream_namespace)
if "stream_state" not in kwargs
else kwargs["stream_state"]
)
stream_state = self.apply_stream_state_migrations(stream_state_migrations, stream_state)

component_type = component_definition.get("type")
if component_definition.get("type") != model_type.__name__:
raise ValueError(
Expand Down Expand Up @@ -1587,7 +1567,6 @@ def create_concurrent_cursor_from_perpartition_cursor(
config: Config,
stream_state: MutableMapping[str, Any],
partition_router: PartitionRouter,
stream_state_migrations: Optional[List[Any]] = None,
attempt_to_create_cursor_if_not_provided: bool = False,
**kwargs: Any,
) -> ConcurrentPerPartitionCursor:
Expand Down Expand Up @@ -1647,11 +1626,9 @@ def create_concurrent_cursor_from_perpartition_cursor(
stream_namespace=stream_namespace,
config=config,
message_repository=NoopMessageRepository(),
# stream_state_migrations=stream_state_migrations, # FIXME is it expected to run migration on per partition state too?
)
)

stream_state = self.apply_stream_state_migrations(stream_state_migrations, stream_state)
# Per-partition state doesn't make sense for GroupingPartitionRouter, so force the global state
use_global_cursor = isinstance(
partition_router, GroupingPartitionRouter
Expand Down Expand Up @@ -1974,6 +1951,7 @@ def create_default_stream(
self, model: DeclarativeStreamModel, config: Config, is_parent: bool = False, **kwargs: Any
) -> AbstractStream:
primary_key = model.primary_key.__root__ if model.primary_key else None
self._migrate_state(config, model)

partition_router = self._build_stream_slicer_from_partition_router(
model.retriever,
Expand Down Expand Up @@ -2135,6 +2113,22 @@ def create_default_stream(
supports_file_transfer=hasattr(model, "file_uploader") and bool(model.file_uploader),
)

def _migrate_state(self, config, model):
stream_state = self._connector_state_manager.get_stream_state(
stream_name=model.name, namespace=None
)
if model.state_migrations:
state_transformations = [
self._create_component_from_model(state_migration, config, declarative_stream=model)
for state_migration in model.state_migrations
]
else:
state_transformations = []
stream_state = self.apply_stream_state_migrations(state_transformations, stream_state)
self._connector_state_manager.update_state_for_stream(
stream_name=model.name, namespace=None, value=stream_state
)

def _is_stop_condition_on_cursor(self, model: DeclarativeStreamModel) -> bool:
return bool(
model.incremental_sync
Expand Down Expand Up @@ -2206,17 +2200,7 @@ def _build_concurrent_cursor(
config: Config,
) -> Cursor:
stream_name = model.name or ""
stream_state = self._connector_state_manager.get_stream_state(
stream_name=stream_name, namespace=None
)

if model.state_migrations:
state_transformations = [
self._create_component_from_model(state_migration, config, declarative_stream=model)
for state_migration in model.state_migrations
]
else:
state_transformations = []
stream_state = self._connector_state_manager.get_stream_state(model.name, None)

if (
model.incremental_sync
Expand All @@ -2228,10 +2212,9 @@ def _build_concurrent_cursor(
model_type=DatetimeBasedCursorModel,
component_definition=model.incremental_sync.__dict__,
stream_name=stream_name,
stream_state=stream_state,
stream_namespace=None,
config=config or {},
stream_state=stream_state,
stream_state_migrations=state_transformations,
partition_router=stream_slicer,
attempt_to_create_cursor_if_not_provided=True, # FIXME can we remove that now?
)
Expand All @@ -2242,17 +2225,17 @@ def _build_concurrent_cursor(
component_definition=model.incremental_sync.__dict__,
stream_name=stream_name,
stream_namespace=None,
stream_state=stream_state,
config=config or {},
stream_state_migrations=state_transformations,
)
elif type(model.incremental_sync) == DatetimeBasedCursorModel:
return self.create_concurrent_cursor_from_datetime_based_cursor( # type: ignore # This is a known issue that we are creating and returning a ConcurrentCursor which does not technically implement the (low-code) StreamSlicer. However, (low-code) StreamSlicer and ConcurrentCursor both implement StreamSlicer.stream_slices() which is the primary method needed for checkpointing
model_type=type(model.incremental_sync),
component_definition=model.incremental_sync.__dict__,
stream_name=stream_name,
stream_namespace=None,
stream_state=stream_state,
config=config or {},
stream_state_migrations=state_transformations,
attempt_to_create_cursor_if_not_provided=True,
)
else:
Expand Down
Loading
Loading