airbyte-cdk 6.36.0.dev0__py3-none-any.whl → 6.37.0.dev0__py3-none-any.whl

airbyte_cdk/connector_builder/models.py

@@ -21,20 +21,6 @@ class HttpRequest:
     body: Optional[str] = None
 
 
-@dataclass
-class StreamReadPages:
-    records: List[object]
-    request: Optional[HttpRequest] = None
-    response: Optional[HttpResponse] = None
-
-
-@dataclass
-class StreamReadSlices:
-    pages: List[StreamReadPages]
-    slice_descriptor: Optional[Dict[str, Any]]
-    state: Optional[List[Dict[str, Any]]] = None
-
-
 @dataclass
 class LogMessage:
     message: str
@@ -46,11 +32,27 @@ class LogMessage:
 @dataclass
 class AuxiliaryRequest:
     title: str
+    type: str
     description: str
     request: HttpRequest
     response: HttpResponse
 
 
+@dataclass
+class StreamReadPages:
+    records: List[object]
+    request: Optional[HttpRequest] = None
+    response: Optional[HttpResponse] = None
+
+
+@dataclass
+class StreamReadSlices:
+    pages: List[StreamReadPages]
+    slice_descriptor: Optional[Dict[str, Any]]
+    state: Optional[List[Dict[str, Any]]] = None
+    auxiliary_requests: Optional[List[AuxiliaryRequest]] = None
+
+
 @dataclass
 class StreamRead(object):
     logs: List[LogMessage]

airbyte_cdk/connector_builder/test_reader/helpers.py

@@ -28,7 +28,7 @@ from airbyte_cdk.utils.schema_inferrer import (
     SchemaInferrer,
 )
 
-from .types import LOG_MESSAGES_OUTPUT_TYPE
+from .types import ASYNC_AUXILIARY_REQUEST_TYPES, LOG_MESSAGES_OUTPUT_TYPE
 
 # -------
 # Parsers
@@ -226,7 +226,8 @@ def should_close_page(
         at_least_one_page_in_group
         and is_log_message(message)
         and (
-            is_page_http_request(json_message) or message.log.message.startswith("slice:")  # type: ignore[union-attr] # AirbyteMessage with MessageType.LOG has log.message
+            is_page_http_request(json_message)
+            or message.log.message.startswith(SliceLogger.SLICE_LOG_PREFIX)  # type: ignore[union-attr] # AirbyteMessage with MessageType.LOG has log.message
         )
     )
 
@@ -330,6 +331,10 @@ def is_auxiliary_http_request(message: Optional[Dict[str, Any]]) -> bool:
     return is_http_log(message) and message.get("http", {}).get("is_auxiliary", False)
 
 
+def is_async_auxiliary_request(message: AuxiliaryRequest) -> bool:
+    return message.type in ASYNC_AUXILIARY_REQUEST_TYPES
+
+
 def is_log_message(message: AirbyteMessage) -> bool:
     """
     Determines whether the provided message is of type LOG.
@@ -413,6 +418,7 @@ def handle_current_slice(
     current_slice_pages: List[StreamReadPages],
     current_slice_descriptor: Optional[Dict[str, Any]] = None,
     latest_state_message: Optional[Dict[str, Any]] = None,
+    auxiliary_requests: Optional[List[AuxiliaryRequest]] = None,
 ) -> StreamReadSlices:
     """
     Handles the current slice by packaging its pages, descriptor, and state into a StreamReadSlices instance.
@@ -421,6 +427,7 @@ def handle_current_slice(
         current_slice_pages (List[StreamReadPages]): The pages to be included in the slice.
         current_slice_descriptor (Optional[Dict[str, Any]]): Descriptor for the current slice, optional.
         latest_state_message (Optional[Dict[str, Any]]): The latest state message, optional.
+        auxiliary_requests (Optional[List[AuxiliaryRequest]]): The auxiliary requests to include, optional.
 
     Returns:
         StreamReadSlices: An object containing the current slice's pages, descriptor, and state.
@@ -429,6 +436,7 @@ def handle_current_slice(
         pages=current_slice_pages,
         slice_descriptor=current_slice_descriptor,
         state=[latest_state_message] if latest_state_message else [],
+        auxiliary_requests=auxiliary_requests if auxiliary_requests else [],
     )
 
 
@@ -486,29 +494,24 @@ def handle_auxiliary_request(json_message: Dict[str, JsonType]) -> AuxiliaryRequ
     Raises:
         ValueError: If any of the "airbyte_cdk", "stream", or "http" fields is not a dictionary.
     """
-    airbyte_cdk = json_message.get("airbyte_cdk", {})
-
-    if not isinstance(airbyte_cdk, dict):
-        raise ValueError(
-            f"Expected airbyte_cdk to be a dict, got {airbyte_cdk} of type {type(airbyte_cdk)}"
-        )
-
-    stream = airbyte_cdk.get("stream", {})
 
-    if not isinstance(stream, dict):
-        raise ValueError(f"Expected stream to be a dict, got {stream} of type {type(stream)}")
+    airbyte_cdk = get_airbyte_cdk_from_message(json_message)
+    stream = get_stream_from_airbyte_cdk(airbyte_cdk)
+    title_prefix = get_auxiliary_request_title_prefix(stream)
+    http = get_http_property_from_message(json_message)
+    request_type = get_auxiliary_request_type(stream, http)
 
-    title_prefix = "Parent stream: " if stream.get("is_substream", False) else ""
-    http = json_message.get("http", {})
-
-    if not isinstance(http, dict):
-        raise ValueError(f"Expected http to be a dict, got {http} of type {type(http)}")
+    title = title_prefix + str(http.get("title", None))
+    description = str(http.get("description", None))
+    request = create_request_from_log_message(json_message)
+    response = create_response_from_log_message(json_message)
 
     return AuxiliaryRequest(
-        title=title_prefix + str(http.get("title", None)),
-        description=str(http.get("description", None)),
-        request=create_request_from_log_message(json_message),
-        response=create_response_from_log_message(json_message),
+        title=title,
+        type=request_type,
+        description=description,
+        request=request,
+        response=response,
     )
 
 
@@ -558,7 +561,8 @@ def handle_log_message(
         at_least_one_page_in_group,
         current_page_request,
         current_page_response,
-        auxiliary_request or log_message,
+        auxiliary_request,
+        log_message,
     )
 
 
@@ -589,3 +593,97 @@ def handle_record_message(
         datetime_format_inferrer.accumulate(message.record)  # type: ignore
 
     return records_count
+
+
+# -------
+# Reusable Getters
+# -------
+
+
+def get_airbyte_cdk_from_message(json_message: Dict[str, JsonType]) -> dict:  # type: ignore
+    """
+    Retrieves the "airbyte_cdk" dictionary from the provided JSON message.
+
+    This function validates that the extracted "airbyte_cdk" is of type dict,
+    raising a ValueError if the validation fails.
+
+    Parameters:
+        json_message (Dict[str, JsonType]): A dictionary representing the JSON message.
+
+    Returns:
+        dict: The "airbyte_cdk" dictionary extracted from the JSON message.
+
+    Raises:
+        ValueError: If the "airbyte_cdk" field is not a dictionary.
+    """
+    airbyte_cdk = json_message.get("airbyte_cdk", {})
+
+    if not isinstance(airbyte_cdk, dict):
+        raise ValueError(
+            f"Expected airbyte_cdk to be a dict, got {airbyte_cdk} of type {type(airbyte_cdk)}"
+        )
+
+    return airbyte_cdk
+
+
+def get_stream_from_airbyte_cdk(airbyte_cdk: dict) -> dict:  # type: ignore
+    """
+    Retrieves the "stream" dictionary from the provided "airbyte_cdk" dictionary.
+
+    This function ensures that the extracted "stream" is of type dict,
+    raising a ValueError if the validation fails.
+
+    Parameters:
+        airbyte_cdk (dict): The dictionary representing the Airbyte CDK data.
+
+    Returns:
+        dict: The "stream" dictionary extracted from the Airbyte CDK data.
+
+    Raises:
+        ValueError: If the "stream" field is not a dictionary.
+    """
+
+    stream = airbyte_cdk.get("stream", {})
+
+    if not isinstance(stream, dict):
+        raise ValueError(f"Expected stream to be a dict, got {stream} of type {type(stream)}")
+
+    return stream
+
+
+def get_auxiliary_request_title_prefix(stream: dict) -> str:  # type: ignore
+    """
+    Generates a title prefix based on the stream type.
+    """
+    return "Parent stream: " if stream.get("is_substream", False) else ""
+
+
+def get_http_property_from_message(json_message: Dict[str, JsonType]) -> dict:  # type: ignore
+    """
+    Retrieves the "http" dictionary from the provided JSON message.
+
+    This function validates that the extracted "http" is of type dict,
+    raising a ValueError if the validation fails.
+
+    Parameters:
+        json_message (Dict[str, JsonType]): A dictionary representing the JSON message.
+
+    Returns:
+        dict: The "http" dictionary extracted from the JSON message.
+
+    Raises:
+        ValueError: If the "http" field is not a dictionary.
+    """
+    http = json_message.get("http", {})
+
+    if not isinstance(http, dict):
+        raise ValueError(f"Expected http to be a dict, got {http} of type {type(http)}")
+
+    return http
+
+
+def get_auxiliary_request_type(stream: dict, http: dict) -> str:  # type: ignore
+    """
+    Determines the type of the auxiliary request based on the stream and HTTP properties.
+    """
+    return "PARENT_STREAM" if stream.get("is_substream", False) else str(http.get("type", None))

airbyte_cdk/connector_builder/test_reader/message_grouper.py

@@ -6,6 +6,7 @@
 from typing import Any, Dict, Iterator, List, Mapping, Optional
 
 from airbyte_cdk.connector_builder.models import (
+    AuxiliaryRequest,
     HttpRequest,
     HttpResponse,
     StreamReadPages,
@@ -24,6 +25,7 @@ from .helpers import (
     handle_current_slice,
     handle_log_message,
     handle_record_message,
+    is_async_auxiliary_request,
     is_config_update_message,
     is_log_message,
     is_record_message,
@@ -89,6 +91,7 @@ def get_message_groups(
     current_page_request: Optional[HttpRequest] = None
     current_page_response: Optional[HttpResponse] = None
     latest_state_message: Optional[Dict[str, Any]] = None
+    slice_auxiliary_requests: List[AuxiliaryRequest] = []
 
     while records_count < limit and (message := next(messages, None)):
         json_message = airbyte_message_to_json(message)
@@ -106,6 +109,7 @@ def get_message_groups(
                 current_slice_pages,
                 current_slice_descriptor,
                 latest_state_message,
+                slice_auxiliary_requests,
             )
             current_slice_descriptor = parse_slice_description(message.log.message)  # type: ignore
             current_slice_pages = []
@@ -118,7 +122,8 @@ def get_message_groups(
                 at_least_one_page_in_group,
                 current_page_request,
                 current_page_response,
-                log_or_auxiliary_request,
+                auxiliary_request,
+                log_message,
             ) = handle_log_message(
                 message,
                 json_message,
@@ -126,8 +131,15 @@ def get_message_groups(
                 current_page_request,
                 current_page_response,
             )
-            if log_or_auxiliary_request:
-                yield log_or_auxiliary_request
+
+            if auxiliary_request:
+                if is_async_auxiliary_request(auxiliary_request):
+                    slice_auxiliary_requests.append(auxiliary_request)
+                else:
+                    yield auxiliary_request
+
+            if log_message:
+                yield log_message
         elif is_trace_with_error(message):
             if message.trace is not None:
                 yield message.trace
@@ -157,4 +169,5 @@ def get_message_groups(
                 current_slice_pages,
                 current_slice_descriptor,
                 latest_state_message,
+                slice_auxiliary_requests,
             )

airbyte_cdk/connector_builder/test_reader/types.py

@@ -71,5 +71,13 @@ LOG_MESSAGES_OUTPUT_TYPE = tuple[
     bool,
     HttpRequest | None,
     HttpResponse | None,
-    AuxiliaryRequest | AirbyteLogMessage | None,
+    AuxiliaryRequest | None,
+    AirbyteLogMessage | None,
+]
+
+ASYNC_AUXILIARY_REQUEST_TYPES = [
+    "ASYNC_CREATE",
+    "ASYNC_POLL",
+    "ASYNC_ABORT",
+    "ASYNC_DELETE",
 ]

airbyte_cdk/entrypoint.py

@@ -37,8 +37,8 @@ from airbyte_cdk.sources import Source
 from airbyte_cdk.sources.connector_state_manager import HashableStreamDescriptor
 from airbyte_cdk.sources.utils.schema_helpers import check_config_against_spec_or_exit, split_config
 
-# from airbyte_cdk.utils import PrintBuffer, is_cloud_environment, message_utils  # add PrintBuffer back once fixed
-from airbyte_cdk.utils import is_cloud_environment, message_utils
+from airbyte_cdk.utils import PrintBuffer, is_cloud_environment, message_utils  # add PrintBuffer back once fixed
+# from airbyte_cdk.utils import is_cloud_environment, message_utils
 from airbyte_cdk.utils.airbyte_secrets_utils import get_secrets, update_secrets
 from airbyte_cdk.utils.constants import ENV_REQUEST_CACHE_PATH
 from airbyte_cdk.utils.traced_exception import AirbyteTracedException
@@ -337,11 +337,11 @@ def launch(source: Source, args: List[str]) -> None:
     parsed_args = source_entrypoint.parse_args(args)
     # temporarily removes the PrintBuffer because we're seeing weird print behavior for concurrent syncs
     # Refer to: https://github.com/airbytehq/oncall/issues/6235
-    # with PrintBuffer():
-    for message in source_entrypoint.run(parsed_args):
-        # simply printing is creating issues for concurrent CDK as Python uses different two instructions to print: one for the message and
-        # the other for the break line. Adding `\n` to the message ensure that both are printed at the same time
-        print(f"{message}\n", end="", flush=True)
+    with PrintBuffer():
+        for message in source_entrypoint.run(parsed_args):
+            # simply printing is creating issues for concurrent CDK as Python uses different two instructions to print: one for the message and
+            # the other for the break line. Adding `\n` to the message ensure that both are printed at the same time
+            print(f"{message}\n", end="", flush=True)
 
 
 def _init_internal_request_filter() -> None:

airbyte_cdk/sources/declarative/auth/token_provider.py

@@ -58,6 +58,7 @@ class SessionTokenProvider(TokenProvider):
                 "Obtains session token",
                 None,
                 is_auxiliary=True,
+                type="AUTH",
             ),
         )
         if response is None:

airbyte_cdk/sources/declarative/concurrent_declarative_source.py

@@ -24,7 +24,6 @@ from airbyte_cdk.sources.declarative.incremental.datetime_based_cursor import Da
 from airbyte_cdk.sources.declarative.incremental.per_partition_with_global import (
     PerPartitionWithGlobalCursor,
 )
-from airbyte_cdk.sources.declarative.interpolation import InterpolatedString
 from airbyte_cdk.sources.declarative.manifest_declarative_source import ManifestDeclarativeSource
 from airbyte_cdk.sources.declarative.models.declarative_component_schema import (
     ConcurrencyLevel as ConcurrencyLevelModel,
@@ -36,17 +35,16 @@ from airbyte_cdk.sources.declarative.parsers.model_to_component_factory import (
     ModelToComponentFactory,
 )
 from airbyte_cdk.sources.declarative.partition_routers import AsyncJobPartitionRouter
-from airbyte_cdk.sources.declarative.requesters import HttpRequester
 from airbyte_cdk.sources.declarative.retrievers import AsyncRetriever, Retriever, SimpleRetriever
 from airbyte_cdk.sources.declarative.stream_slicers.declarative_partition_generator import (
     DeclarativePartitionFactory,
     StreamSlicerPartitionGenerator,
 )
-from airbyte_cdk.sources.declarative.transformations.add_fields import AddFields
 from airbyte_cdk.sources.declarative.types import ConnectionDefinition
 from airbyte_cdk.sources.source import TState
 from airbyte_cdk.sources.streams import Stream
 from airbyte_cdk.sources.streams.concurrent.abstract_stream import AbstractStream
+from airbyte_cdk.sources.streams.concurrent.abstract_stream_facade import AbstractStreamFacade
 from airbyte_cdk.sources.streams.concurrent.availability_strategy import (
     AlwaysAvailableAvailabilityStrategy,
 )
@@ -121,6 +119,12 @@ class ConcurrentDeclarativeSource(ManifestDeclarativeSource, Generic[TState]):
             message_repository=self.message_repository,
         )
 
+    # TODO: Remove this. This property is necessary to safely migrate Stripe during the transition state.
+    @property
+    def is_partially_declarative(self) -> bool:
+        """This flag used to avoid unexpected AbstractStreamFacade processing as concurrent streams."""
+        return False
+
     def read(
         self,
         logger: logging.Logger,
@@ -321,9 +325,6 @@ class ConcurrentDeclarativeSource(ManifestDeclarativeSource, Generic[TState]):
                     incremental_sync_component_definition
                     and incremental_sync_component_definition.get("type", "")
                     == DatetimeBasedCursorModel.__name__
-                    and self._stream_supports_concurrent_partition_processing(
-                        declarative_stream=declarative_stream
-                    )
                     and hasattr(declarative_stream.retriever, "stream_slicer")
                     and isinstance(
                         declarative_stream.retriever.stream_slicer, PerPartitionWithGlobalCursor
@@ -375,6 +376,14 @@ class ConcurrentDeclarativeSource(ManifestDeclarativeSource, Generic[TState]):
                     )
                 else:
                     synchronous_streams.append(declarative_stream)
+            # TODO: Remove this. This check is necessary to safely migrate Stripe during the transition state.
+            # Condition below needs to ensure that concurrent support is not lost for sources that already support
+            # it before migration, but now are only partially migrated to declarative implementation (e.g., Stripe).
+            elif (
+                isinstance(declarative_stream, AbstractStreamFacade)
+                and self.is_partially_declarative
+            ):
+                concurrent_streams.append(declarative_stream.get_underlying_stream())
             else:
                 synchronous_streams.append(declarative_stream)
 
@@ -390,9 +399,6 @@ class ConcurrentDeclarativeSource(ManifestDeclarativeSource, Generic[TState]):
             and bool(incremental_sync_component_definition)
             and incremental_sync_component_definition.get("type", "")
             == DatetimeBasedCursorModel.__name__
-            and self._stream_supports_concurrent_partition_processing(
-                declarative_stream=declarative_stream
-            )
             and hasattr(declarative_stream.retriever, "stream_slicer")
             and (
                 isinstance(declarative_stream.retriever.stream_slicer, DatetimeBasedCursor)
@@ -400,72 +406,6 @@ class ConcurrentDeclarativeSource(ManifestDeclarativeSource, Generic[TState]):
             )
         )
 
-    def _stream_supports_concurrent_partition_processing(
-        self, declarative_stream: DeclarativeStream
-    ) -> bool:
-        """
-        Many connectors make use of stream_state during interpolation on a per-partition basis under the assumption that
-        state is updated sequentially. Because the concurrent CDK engine processes different partitions in parallel,
-        stream_state is no longer a thread-safe interpolation context. It would be a race condition because a cursor's
-        stream_state can be updated in any order depending on which stream partition's finish first.
-
-        We should start to move away from depending on the value of stream_state for low-code components that operate
-        per-partition, but we need to gate this otherwise some connectors will be blocked from publishing. See the
-        cdk-migrations.md for the full list of connectors.
-        """
-
-        if isinstance(declarative_stream.retriever, SimpleRetriever) and isinstance(
-            declarative_stream.retriever.requester, HttpRequester
-        ):
-            http_requester = declarative_stream.retriever.requester
-            if "stream_state" in http_requester._path.string:
-                self.logger.warning(
-                    f"Low-code stream '{declarative_stream.name}' uses interpolation of stream_state in the HttpRequester which is not thread-safe. Defaulting to synchronous processing"
-                )
-                return False
-
-            request_options_provider = http_requester._request_options_provider
-            if request_options_provider.request_options_contain_stream_state():
-                self.logger.warning(
-                    f"Low-code stream '{declarative_stream.name}' uses interpolation of stream_state in the HttpRequester which is not thread-safe. Defaulting to synchronous processing"
-                )
-                return False
-
-            record_selector = declarative_stream.retriever.record_selector
-            if isinstance(record_selector, RecordSelector):
-                if (
-                    record_selector.record_filter
-                    and not isinstance(
-                        record_selector.record_filter, ClientSideIncrementalRecordFilterDecorator
-                    )
-                    and "stream_state" in record_selector.record_filter.condition
-                ):
-                    self.logger.warning(
-                        f"Low-code stream '{declarative_stream.name}' uses interpolation of stream_state in the RecordFilter which is not thread-safe. Defaulting to synchronous processing"
-                    )
-                    return False
-
-                for add_fields in [
-                    transformation
-                    for transformation in record_selector.transformations
-                    if isinstance(transformation, AddFields)
-                ]:
-                    for field in add_fields.fields:
-                        if isinstance(field.value, str) and "stream_state" in field.value:
-                            self.logger.warning(
-                                f"Low-code stream '{declarative_stream.name}' uses interpolation of stream_state in the AddFields which is not thread-safe. Defaulting to synchronous processing"
-                            )
-                            return False
-                        if (
-                            isinstance(field.value, InterpolatedString)
-                            and "stream_state" in field.value.string
-                        ):
-                            self.logger.warning(
-                                f"Low-code stream '{declarative_stream.name}' uses interpolation of stream_state in the AddFields which is not thread-safe. Defaulting to synchronous processing"
-                            )
-                            return False
-        return True
-
     @staticmethod
     def _get_retriever(
         declarative_stream: DeclarativeStream, stream_state: Mapping[str, Any]

airbyte_cdk/sources/declarative/declarative_component_schema.yaml

@@ -82,7 +82,6 @@ definitions:
           - stream_interval
           - stream_partition
           - stream_slice
-          - stream_state
         examples:
           - "{{ record['updates'] }}"
           - "{{ record['MetaData']['LastUpdatedTime'] }}"
@@ -1491,7 +1490,11 @@ definitions:
       limit:
         title: Limit
         description: The maximum number of calls allowed within the interval.
-        type: integer
+        anyOf:
+          - type: integer
+          - type: string
+        interpolation_context:
+          - config
       interval:
         title: Interval
         description: The time interval for the rate limit.
@@ -1776,7 +1779,6 @@ definitions:
           - stream_interval
           - stream_partition
           - stream_slice
-          - stream_state
         examples:
           - "/products"
           - "/quotes/{{ stream_partition['id'] }}/quote_line_groups"
@@ -1826,7 +1828,6 @@ definitions:
           - stream_interval
           - stream_partition
           - stream_slice
-          - stream_state
         examples:
           - |
             [{"clause": {"type": "timestamp", "operator": 10, "parameters":
@@ -1844,7 +1845,6 @@ definitions:
           - stream_interval
           - stream_partition
           - stream_slice
-          - stream_state
         examples:
           - sort_order: "ASC"
             sort_field: "CREATED_AT"
@@ -1865,7 +1865,6 @@ definitions:
           - stream_interval
           - stream_partition
           - stream_slice
-          - stream_state
         examples:
           - Output-Format: JSON
           - Version: "{{ config['version'] }}"
@@ -1882,7 +1881,6 @@ definitions:
           - stream_interval
           - stream_partition
           - stream_slice
-          - stream_state
         examples:
           - unit: "day"
           - query: 'last_event_time BETWEEN TIMESTAMP "{{ stream_interval.start_time }}" AND TIMESTAMP "{{ stream_interval.end_time }}"'
@@ -2237,7 +2235,6 @@ definitions:
         interpolation_context:
           - config
           - record
-          - stream_state
           - stream_slice
       new:
         type: string
@@ -2251,7 +2248,6 @@ definitions:
         interpolation_context:
           - config
           - record
-          - stream_state
           - stream_slice
       $parameters:
         type: object
@@ -2901,7 +2897,6 @@ definitions:
           - stream_interval
           - stream_partition
           - stream_slice
-          - stream_state
         examples:
           - "{{ record['created_at'] >= stream_interval['start_time'] }}"
           - "{{ record.status in ['active', 'expired'] }}"
@@ -3689,12 +3684,6 @@ interpolation:
     - title: stream_slice
       description: This variable is deprecated. Use stream_interval or stream_partition instead.
       type: object
-    - title: stream_state
-      description: The current state of the stream. The object's keys are defined by the incremental sync's cursor_field the and partition router's values.
-      type: object
-      examples:
-        - created_at: "2020-01-01 00:00:00.000+00:00"
-        - updated_at: "2020-01-02 00:00:00.000+00:00"
   macros:
     - title: now_utc
       description: Returns the current date and time in the UTC timezone.
@@ -3759,6 +3748,16 @@ interpolation:
         - "{{ format_datetime(config['start_time'], '%Y-%m-%d') }}"
         - "{{ format_datetime(config['start_date'], '%Y-%m-%dT%H:%M:%S.%fZ') }}"
         - "{{ format_datetime(config['start_date'], '%Y-%m-%dT%H:%M:%S.%fZ', '%a, %d %b %Y %H:%M:%S %z') }}"
+    - title: str_to_datetime
+      description: Converts a string to a datetime object with UTC timezone.
+      arguments:
+        s: The string to convert.
+      return_type: datetime.datetime
+      examples:
+        - "{{ str_to_datetime('2022-01-14') }}"
+        - "{{ str_to_datetime('2022-01-01 13:45:30') }}"
+        - "{{ str_to_datetime('2022-01-01T13:45:30+00:00') }}"
+        - "{{ str_to_datetime('2022-01-01T13:45:30.123456Z') }}"
   filters:
     - title: hash
       description: Convert the specified value to a hashed string.

airbyte_cdk/sources/declarative/decoders/composite_raw_decoder.py

@@ -107,6 +107,16 @@ class CsvParser(Parser):
     encoding: Optional[str] = "utf-8"
     delimiter: Optional[str] = ","
 
+    def _get_delimiter(self) -> Optional[str]:
+        """
+        Get delimiter from the configuration. Check for the escape character and decode it.
+        """
+        if self.delimiter is not None:
+            if self.delimiter.startswith("\\"):
+                self.delimiter = self.delimiter.encode("utf-8").decode("unicode_escape")
+
+        return self.delimiter
+
     def parse(
         self,
         data: BufferedIOBase,
@@ -115,8 +125,9 @@ class CsvParser(Parser):
         Parse CSV data from decompressed bytes.
         """
         text_data = TextIOWrapper(data, encoding=self.encoding)  # type: ignore
-        reader = csv.DictReader(text_data, delimiter=self.delimiter or ",")
-        yield from reader
+        reader = csv.DictReader(text_data, delimiter=self._get_delimiter() or ",")
+        for row in reader:
+            yield row
 
 
 @dataclass

airbyte_cdk/sources/declarative/incremental/concurrent_partition_cursor.py

@@ -251,7 +251,6 @@ class ConcurrentPerPartitionCursor(Cursor):
         self._message_repository.emit_message(state_message)
 
     def stream_slices(self) -> Iterable[StreamSlice]:
-        print("stream_slices")
         if self._timer.is_running():
             raise RuntimeError("stream_slices has been executed more than once.")