airbyte-cdk 6.13.0__tar.gz → 6.13.0.dev0__tar.gz

{airbyte_cdk-6.13.0 → airbyte_cdk-6.13.0.dev0}/PKG-INFO

@@ -1,6 +1,6 @@
-Metadata-Version: 2.3
+Metadata-Version: 2.1
 Name: airbyte-cdk
-Version: 6.13.0
+Version: 6.13.0.dev0
 Summary: A framework for writing Airbyte Connectors.
 Home-page: https://airbyte.com
 License: MIT

{airbyte_cdk-6.13.0 → airbyte_cdk-6.13.0.dev0}/airbyte_cdk/sources/declarative/concurrent_declarative_source.py

@@ -194,10 +194,11 @@ class ConcurrentDeclarativeSource(ManifestDeclarativeSource, Generic[TState]):
             # Some low-code sources use a combination of DeclarativeStream and regular Python streams. We can't inspect
             # 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)
-                and name_to_stream_mapping[declarative_stream.name]["retriever"]["type"]
+            if isinstance(declarative_stream, DeclarativeStream) and (
+                name_to_stream_mapping[declarative_stream.name]["retriever"]["type"]
                 == "SimpleRetriever"
+                or name_to_stream_mapping[declarative_stream.name]["retriever"]["type"]
+                == "AsyncRetriever"
             ):
                 incremental_sync_component_definition = name_to_stream_mapping[
                     declarative_stream.name
@@ -217,6 +218,11 @@ class ConcurrentDeclarativeSource(ManifestDeclarativeSource, Generic[TState]):
                     and not incremental_sync_component_definition
                 )
 
+                is_async_job_stream = (
+                    name_to_stream_mapping[declarative_stream.name].get("retriever", {}).get("type")
+                    == "AsyncRetriever"
+                )
+
                 if self._is_datetime_incremental_without_partition_routing(
                     declarative_stream, incremental_sync_component_definition
                 ):
@@ -268,15 +274,26 @@ class ConcurrentDeclarativeSource(ManifestDeclarativeSource, Generic[TState]):
                 elif (
                     is_substream_without_incremental or is_without_partition_router_or_cursor
                 ) and hasattr(declarative_stream.retriever, "stream_slicer"):
+                    if is_async_job_stream:
+                        # A stream's AsyncRetriever must be shared across all partitions because it uses a
+                        # shared JobRepository to manage the state of jobs requests and when they are ready
+                        async_retriever = declarative_stream.retriever
+
+                        def async_retriever_factory_method() -> Retriever:
+                            return async_retriever
+
+                        retriever_factory = async_retriever_factory_method
+                    else:
+                        retriever_factory = self._retriever_factory(
+                            name_to_stream_mapping[declarative_stream.name],
+                            config,
+                            {},
+                        )
                     partition_generator = StreamSlicerPartitionGenerator(
                         DeclarativePartitionFactory(
                             declarative_stream.name,
                             declarative_stream.get_json_schema(),
-                            self._retriever_factory(
-                                name_to_stream_mapping[declarative_stream.name],
-                                config,
-                                {},
-                            ),
+                            retriever_factory,
                             self.message_repository,
                         ),
                         declarative_stream.retriever.stream_slicer,

{airbyte_cdk-6.13.0 → airbyte_cdk-6.13.0.dev0}/airbyte_cdk/sources/declarative/requesters/paginators/default_paginator.py

@@ -112,27 +112,39 @@ class DefaultPaginator(Paginator):
             )
         if isinstance(self.url_base, str):
             self.url_base = InterpolatedString(string=self.url_base, parameters=parameters)
-        self._token: Optional[Any] = self.pagination_strategy.initial_token
+
+    def get_initial_token(self) -> Optional[Any]:
+        """
+        Return the page token that should be used for the first request of a stream
+
+        WARNING: get_initial_token() should not be used by streams that use RFR that perform checkpointing
+        of state using page numbers. Because paginators are stateless
+        """
+        return self.pagination_strategy.initial_token
 
     def next_page_token(
-        self, response: requests.Response, last_page_size: int, last_record: Optional[Record]
+        self,
+        response: requests.Response,
+        last_page_size: int,
+        last_record: Optional[Record],
+        last_page_token_value: Optional[Any] = None,
     ) -> Optional[Mapping[str, Any]]:
-        self._token = self.pagination_strategy.next_page_token(
-            response, last_page_size, last_record
+        next_page_token = self.pagination_strategy.next_page_token(
+            response=response,
+            last_page_size=last_page_size,
+            last_record=last_record,
+            last_page_token_value=last_page_token_value,
         )
-        if self._token:
-            return {"next_page_token": self._token}
+        if next_page_token:
+            return {"next_page_token": next_page_token}
         else:
             return None
 
-    def path(self) -> Optional[str]:
-        if (
-            self._token
-            and self.page_token_option
-            and isinstance(self.page_token_option, RequestPath)
-        ):
+    def path(self, next_page_token: Optional[Mapping[str, Any]]) -> Optional[str]:
+        token = next_page_token.get("next_page_token") if next_page_token else None
+        if token and self.page_token_option and isinstance(self.page_token_option, RequestPath):
             # Replace url base to only return the path
-            return str(self._token).replace(self.url_base.eval(self.config), "")  # type: ignore # url_base is casted to a InterpolatedString in __post_init__
+            return str(token).replace(self.url_base.eval(self.config), "")  # type: ignore # url_base is casted to a InterpolatedString in __post_init__
         else:
             return None
 
@@ -143,7 +155,7 @@ class DefaultPaginator(Paginator):
         stream_slice: Optional[StreamSlice] = None,
         next_page_token: Optional[Mapping[str, Any]] = None,
     ) -> MutableMapping[str, Any]:
-        return self._get_request_options(RequestOptionType.request_parameter)
+        return self._get_request_options(RequestOptionType.request_parameter, next_page_token)
 
     def get_request_headers(
         self,
@@ -152,7 +164,7 @@ class DefaultPaginator(Paginator):
         stream_slice: Optional[StreamSlice] = None,
         next_page_token: Optional[Mapping[str, Any]] = None,
     ) -> Mapping[str, str]:
-        return self._get_request_options(RequestOptionType.header)
+        return self._get_request_options(RequestOptionType.header, next_page_token)
 
     def get_request_body_data(
         self,
@@ -161,7 +173,7 @@ class DefaultPaginator(Paginator):
         stream_slice: Optional[StreamSlice] = None,
         next_page_token: Optional[Mapping[str, Any]] = None,
     ) -> Mapping[str, Any]:
-        return self._get_request_options(RequestOptionType.body_data)
+        return self._get_request_options(RequestOptionType.body_data, next_page_token)
 
     def get_request_body_json(
         self,
@@ -170,25 +182,21 @@ class DefaultPaginator(Paginator):
         stream_slice: Optional[StreamSlice] = None,
         next_page_token: Optional[Mapping[str, Any]] = None,
     ) -> Mapping[str, Any]:
-        return self._get_request_options(RequestOptionType.body_json)
-
-    def reset(self, reset_value: Optional[Any] = None) -> None:
-        if reset_value:
-            self.pagination_strategy.reset(reset_value=reset_value)
-        else:
-            self.pagination_strategy.reset()
-        self._token = self.pagination_strategy.initial_token
+        return self._get_request_options(RequestOptionType.body_json, next_page_token)
 
-    def _get_request_options(self, option_type: RequestOptionType) -> MutableMapping[str, Any]:
+    def _get_request_options(
+        self, option_type: RequestOptionType, next_page_token: Optional[Mapping[str, Any]]
+    ) -> MutableMapping[str, Any]:
         options = {}
 
+        token = next_page_token.get("next_page_token") if next_page_token else None
         if (
             self.page_token_option
-            and self._token is not None
+            and token is not None
             and isinstance(self.page_token_option, RequestOption)
             and self.page_token_option.inject_into == option_type
         ):
-            options[self.page_token_option.field_name.eval(config=self.config)] = self._token  # type: ignore # field_name is always cast to an interpolated string
+            options[self.page_token_option.field_name.eval(config=self.config)] = token  # type: ignore # field_name is always cast to an interpolated string
         if (
             self.page_size_option
             and self.pagination_strategy.get_page_size()
@@ -204,6 +212,9 @@ class PaginatorTestReadDecorator(Paginator):
     """
     In some cases, we want to limit the number of requests that are made to the backend source. This class allows for limiting the number of
     pages that are queried throughout a read command.
+
+    WARNING: This decorator is not currently thread-safe like the rest of the low-code framework because it has
+    an internal state to track the current number of pages counted so that it can exit early during a test read
     """
 
     _PAGE_COUNT_BEFORE_FIRST_NEXT_CALL = 1
@@ -217,17 +228,27 @@ class PaginatorTestReadDecorator(Paginator):
         self._decorated = decorated
         self._page_count = self._PAGE_COUNT_BEFORE_FIRST_NEXT_CALL
 
+    def get_initial_token(self) -> Optional[Any]:
+        self._page_count = self._PAGE_COUNT_BEFORE_FIRST_NEXT_CALL
+        return self._decorated.get_initial_token()
+
     def next_page_token(
-        self, response: requests.Response, last_page_size: int, last_record: Optional[Record]
+        self,
+        response: requests.Response,
+        last_page_size: int,
+        last_record: Optional[Record],
+        last_page_token_value: Optional[Any] = None,
     ) -> Optional[Mapping[str, Any]]:
         if self._page_count >= self._maximum_number_of_pages:
             return None
 
         self._page_count += 1
-        return self._decorated.next_page_token(response, last_page_size, last_record)
+        return self._decorated.next_page_token(
+            response, last_page_size, last_record, last_page_token_value
+        )
 
-    def path(self) -> Optional[str]:
-        return self._decorated.path()
+    def path(self, next_page_token: Optional[Mapping[str, Any]]) -> Optional[str]:
+        return self._decorated.path(next_page_token)
 
     def get_request_params(
         self,
@@ -272,7 +293,3 @@ class PaginatorTestReadDecorator(Paginator):
         return self._decorated.get_request_body_json(
             stream_state=stream_state, stream_slice=stream_slice, next_page_token=next_page_token
         )
-
-    def reset(self, reset_value: Optional[Any] = None) -> None:
-        self._decorated.reset()
-        self._page_count = self._PAGE_COUNT_BEFORE_FIRST_NEXT_CALL

{airbyte_cdk-6.13.0 → airbyte_cdk-6.13.0.dev0}/airbyte_cdk/sources/declarative/requesters/paginators/no_pagination.py

@@ -19,7 +19,7 @@ class NoPagination(Paginator):
 
     parameters: InitVar[Mapping[str, Any]]
 
-    def path(self) -> Optional[str]:
+    def path(self, next_page_token: Optional[Mapping[str, Any]]) -> Optional[str]:
         return None
 
     def get_request_params(
@@ -58,11 +58,14 @@ class NoPagination(Paginator):
     ) -> Mapping[str, Any]:
         return {}
 
+    def get_initial_token(self) -> Optional[Any]:
+        return None
+
     def next_page_token(
-        self, response: requests.Response, last_page_size: int, last_record: Optional[Record]
-    ) -> Mapping[str, Any]:
+        self,
+        response: requests.Response,
+        last_page_size: int,
+        last_record: Optional[Record],
+        last_page_token_value: Optional[Any],
+    ) -> Optional[Mapping[str, Any]]:
         return {}
-
-    def reset(self, reset_value: Optional[Any] = None) -> None:
-        # No state to reset
-        pass

{airbyte_cdk-6.13.0 → airbyte_cdk-6.13.0.dev0}/airbyte_cdk/sources/declarative/requesters/paginators/paginator.py

@@ -24,14 +24,18 @@ class Paginator(ABC, RequestOptionsProvider):
     """
 
     @abstractmethod
-    def reset(self, reset_value: Optional[Any] = None) -> None:
+    def get_initial_token(self) -> Optional[Any]:
         """
-        Reset the pagination's inner state
+        Get the page token that should be included in the request to get the first page of records
         """
 
     @abstractmethod
     def next_page_token(
-        self, response: requests.Response, last_page_size: int, last_record: Optional[Record]
+        self,
+        response: requests.Response,
+        last_page_size: int,
+        last_record: Optional[Record],
+        last_page_token_value: Optional[Any],
     ) -> Optional[Mapping[str, Any]]:
         """
         Returns the next_page_token to use to fetch the next page of records.
@@ -39,12 +43,13 @@ class Paginator(ABC, RequestOptionsProvider):
         :param response: the response to process
         :param last_page_size: the number of records read from the response
         :param last_record: the last record extracted from the response
+        :param last_page_token_value: The current value of the page token made on the last request
         :return: A mapping {"next_page_token": <token>} for the next page from the input response object. Returning None means there are no more pages to read in this response.
         """
         pass
 
     @abstractmethod
-    def path(self) -> Optional[str]:
+    def path(self, next_page_token: Optional[Mapping[str, Any]]) -> Optional[str]:
         """
         Returns the URL path to hit to fetch the next page of records
 

{airbyte_cdk-6.13.0 → airbyte_cdk-6.13.0.dev0}/airbyte_cdk/sources/declarative/requesters/paginators/strategies/cursor_pagination_strategy.py

@@ -43,7 +43,6 @@ class CursorPaginationStrategy(PaginationStrategy):
     )
 
     def __post_init__(self, parameters: Mapping[str, Any]) -> None:
-        self._initial_cursor = None
         if isinstance(self.cursor_value, str):
             self._cursor_value = InterpolatedString.create(self.cursor_value, parameters=parameters)
         else:
@@ -57,10 +56,19 @@ class CursorPaginationStrategy(PaginationStrategy):
 
     @property
     def initial_token(self) -> Optional[Any]:
-        return self._initial_cursor
+        """
+        CursorPaginationStrategy does not have an initial value because the next cursor is typically included
+        in the response of the first request. For Resumable Full Refresh streams that checkpoint the page
+        cursor, the next cursor should be read from the state or stream slice object.
+        """
+        return None
 
     def next_page_token(
-        self, response: requests.Response, last_page_size: int, last_record: Optional[Record]
+        self,
+        response: requests.Response,
+        last_page_size: int,
+        last_record: Optional[Record],
+        last_page_token_value: Optional[Any] = None,
     ) -> Optional[Any]:
         decoded_response = next(self.decoder.decode(response))
 
@@ -87,8 +95,5 @@ class CursorPaginationStrategy(PaginationStrategy):
         )
         return token if token else None
 
-    def reset(self, reset_value: Optional[Any] = None) -> None:
-        self._initial_cursor = reset_value
-
     def get_page_size(self) -> Optional[int]:
         return self.page_size

{airbyte_cdk-6.13.0 → airbyte_cdk-6.13.0.dev0}/airbyte_cdk/sources/declarative/requesters/paginators/strategies/offset_increment.py

@@ -52,7 +52,6 @@ class OffsetIncrement(PaginationStrategy):
     inject_on_first_request: bool = False
 
     def __post_init__(self, parameters: Mapping[str, Any]) -> None:
-        self._offset = 0
         page_size = str(self.page_size) if isinstance(self.page_size, int) else self.page_size
         if page_size:
             self._page_size: Optional[InterpolatedString] = InterpolatedString(
@@ -64,11 +63,15 @@ class OffsetIncrement(PaginationStrategy):
     @property
     def initial_token(self) -> Optional[Any]:
         if self.inject_on_first_request:
-            return self._offset
+            return 0
         return None
 
     def next_page_token(
-        self, response: requests.Response, last_page_size: int, last_record: Optional[Record]
+        self,
+        response: requests.Response,
+        last_page_size: int,
+        last_record: Optional[Record],
+        last_page_token_value: Optional[Any] = None,
     ) -> Optional[Any]:
         decoded_response = next(self.decoder.decode(response))
 
@@ -78,14 +81,20 @@ class OffsetIncrement(PaginationStrategy):
             and last_page_size < self._page_size.eval(self.config, response=decoded_response)
         ) or last_page_size == 0:
             return None
+        elif last_page_token_value is None:
+            # If the OffsetIncrement strategy does not inject on the first request, the incoming last_page_token_value
+            # will be None. For this case, we assume that None was the first page and progress to the next offset
+            return 0 + last_page_size
+        elif not isinstance(last_page_token_value, int):
+            raise ValueError(
+                "The page token for a OffsetIncrement pagination strategy must be an integer"
+            )
         else:
-            self._offset += last_page_size
-            return self._offset
+            next_page_token_value = last_page_token_value + last_page_size
+            return next_page_token_value
 
     def reset(self, reset_value: Optional[Any] = 0) -> None:
-        if reset_value is None:
-            self._offset = 0
-        elif not isinstance(reset_value, int):
+        if not isinstance(reset_value, int):
             raise ValueError(
                 f"Reset value {reset_value} for OffsetIncrement pagination strategy was not an integer"
             )

{airbyte_cdk-6.13.0 → airbyte_cdk-6.13.0.dev0}/airbyte_cdk/sources/declarative/requesters/paginators/strategies/page_increment.py

@@ -31,7 +31,6 @@ class PageIncrement(PaginationStrategy):
     inject_on_first_request: bool = False
 
     def __post_init__(self, parameters: Mapping[str, Any]) -> None:
-        self._page = self.start_from_page
         if isinstance(self.page_size, int) or (self.page_size is None):
             self._page_size = self.page_size
         else:
@@ -43,28 +42,30 @@ class PageIncrement(PaginationStrategy):
     @property
     def initial_token(self) -> Optional[Any]:
         if self.inject_on_first_request:
-            return self._page
+            return self.start_from_page
         return None
 
     def next_page_token(
-        self, response: requests.Response, last_page_size: int, last_record: Optional[Record]
+        self,
+        response: requests.Response,
+        last_page_size: int,
+        last_record: Optional[Record],
+        last_page_token_value: Optional[Any],
     ) -> Optional[Any]:
         # Stop paginating when there are fewer records than the page size or the current page has no records
         if (self._page_size and last_page_size < self._page_size) or last_page_size == 0:
             return None
-        else:
-            self._page += 1
-            return self._page
-
-    def reset(self, reset_value: Optional[Any] = None) -> None:
-        if reset_value is None:
-            self._page = self.start_from_page
-        elif not isinstance(reset_value, int):
+        elif last_page_token_value is None:
+            # If the PageIncrement strategy does not inject on the first request, the incoming last_page_token_value
+            # may be None. When this is the case, we assume we've already requested the first page specified by
+            # start_from_page and must now get the next page
+            return self.start_from_page + 1
+        elif not isinstance(last_page_token_value, int):
             raise ValueError(
-                f"Reset value {reset_value} for PageIncrement pagination strategy was not an integer"
+                "The page token for a PageIncrement pagination strategy must be an integer"
             )
         else:
-            self._page = reset_value
+            return last_page_token_value + 1
 
     def get_page_size(self) -> Optional[int]:
         return self._page_size

{airbyte_cdk-6.13.0 → airbyte_cdk-6.13.0.dev0}/airbyte_cdk/sources/declarative/requesters/paginators/strategies/pagination_strategy.py

@@ -4,7 +4,7 @@
 
 from abc import abstractmethod
 from dataclasses import dataclass
-from typing import Any, Optional
+from typing import Any, Mapping, Optional
 
 import requests
 
@@ -26,22 +26,21 @@ class PaginationStrategy:
 
     @abstractmethod
     def next_page_token(
-        self, response: requests.Response, last_page_size: int, last_record: Optional[Record]
+        self,
+        response: requests.Response,
+        last_page_size: int,
+        last_record: Optional[Record],
+        last_page_token_value: Optional[Any],
     ) -> Optional[Any]:
         """
         :param response: response to process
         :param last_page_size: the number of records read from the response
         :param last_record: the last record extracted from the response
+        :param last_page_token_value: The current value of the page token made on the last request
         :return: next page token. Returns None if there are no more pages to fetch
         """
         pass
 
-    @abstractmethod
-    def reset(self, reset_value: Optional[Any] = None) -> None:
-        """
-        Reset the pagination's inner state
-        """
-
     @abstractmethod
     def get_page_size(self) -> Optional[int]:
         """

{airbyte_cdk-6.13.0 → airbyte_cdk-6.13.0.dev0}/airbyte_cdk/sources/declarative/requesters/paginators/strategies/stop_condition.py

@@ -44,19 +44,19 @@ class StopConditionPaginationStrategyDecorator(PaginationStrategy):
         self._stop_condition = stop_condition
 
     def next_page_token(
-        self, response: requests.Response, last_page_size: int, last_record: Optional[Record]
+        self,
+        response: requests.Response,
+        last_page_size: int,
+        last_record: Optional[Record],
+        last_page_token_value: Optional[Any] = None,
     ) -> Optional[Any]:
-        # We evaluate in reverse order because the assumption is that most of the APIs using data feed structure will return records in
-        # descending order. In terms of performance/memory, we return the records lazily
+        # We evaluate in reverse order because the assumption is that most of the APIs using data feed structure
+        # will return records in descending order. In terms of performance/memory, we return the records lazily
         if last_record and self._stop_condition.is_met(last_record):
             return None
-        return self._delegate.next_page_token(response, last_page_size, last_record)
-
-    def reset(self, reset_value: Optional[Any] = None) -> None:
-        if reset_value:
-            self._delegate.reset(reset_value)
-        else:
-            self._delegate.reset()
+        return self._delegate.next_page_token(
+            response, last_page_size, last_record, last_page_token_value
+        )
 
     def get_page_size(self) -> Optional[int]:
         return self._delegate.get_page_size()