airbyte-source-github 1.8.29.dev202506231149__py3-none-any.whl → 1.8.29.dev202506231532__py3-none-any.whl

{airbyte_source_github-1.8.29.dev202506231149.dist-info → airbyte_source_github-1.8.29.dev202506231532.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: airbyte-source-github
-Version: 1.8.29.dev202506231149
+Version: 1.8.29.dev202506231532
 Summary: Source implementation for GitHub.
 License: MIT
 Author: Airbyte

{airbyte_source_github-1.8.29.dev202506231149.dist-info → airbyte_source_github-1.8.29.dev202506231532.dist-info}/RECORD

@@ -2,7 +2,7 @@ source_github/__init__.py,sha256=punPc3v0mXEYOun7cbkfM5KUhgjv72B9DgDhI4VtzcQ,113
 source_github/backoff_strategies.py,sha256=ChkWEmqZL7Qrr1zidOfjZWlMWIVRVO2yqZ7QE_HaZP8,2287
 source_github/config_migrations.py,sha256=H58hHqAnuvb0B8IXHW4aEDZ3HotEg7HdA2rXDG9XW7A,3832
 source_github/constants.py,sha256=Hj3Q4y7OoU-Iff4m9gEC2CjwmWJYXhNbHVNjg8EBLmQ,238
-source_github/errors_handlers.py,sha256=aKMf12iVtf4bnQhsST1lfgOFuauTZM7jvF8o8NzJHO8,7116
+source_github/errors_handlers.py,sha256=wAaihlPrYqPXhso2oCubje1p8yW79ol8t2g0zgQ30Wo,7052
 source_github/github_schema.py,sha256=2AXmTN_s_VqvZAR1GPqus5HhSXWokS4N7HMQQqefaYw,1600315
 source_github/graphql.py,sha256=edd7EoXwD5rRDHrjm4ZBy5VGkruTScL_Xipy4CBmTFc,11625
 source_github/run.py,sha256=onA-rP2aVhWHvDquKZdR1381CU66rnzqJ7EFMS5dd4Q,407
@@ -56,9 +56,9 @@ source_github/schemas/workflow_runs.json,sha256=XDmIsjtzka-ItEonImD3ZATZjxRNkbFo
 source_github/schemas/workflows.json,sha256=gSNw8WZaVKbX4AL97PbjZHzvxcOltXqv9Ao1RNQOFXM,1470
 source_github/source.py,sha256=1o8eayigi4xSUeNHdCd-mhNswGUq_XQrVk2eihTjm1o,14246
 source_github/spec.json,sha256=7LOQm01fP_RvPF-HifhNPJ7i0AxT2LTNPaLAA3uOfNY,7443
-source_github/streams.py,sha256=Dr1zWvKtkX4yq3THjBOx5Rkixa6Hnyso3kJ8t886gGE,76956
+source_github/streams.py,sha256=oL91L04kf0NANfdAJbXghvg2G-InAfjOFFqKF7rDb2s,77308
 source_github/utils.py,sha256=Ztd8VWwzTNUg_A96_8R9XSKorIcBa8wJ6aYUqygRkyk,5492
-airbyte_source_github-1.8.29.dev202506231149.dist-info/METADATA,sha256=y-a-yBgngPTljU56yBbbyaMM5dSGZ6-GYDZK22WMyFQ,5218
-airbyte_source_github-1.8.29.dev202506231149.dist-info/WHEEL,sha256=b4K_helf-jlQoXBBETfwnf4B04YC67LOev0jo4fX5m8,88
-airbyte_source_github-1.8.29.dev202506231149.dist-info/entry_points.txt,sha256=gYhqVrTAZvMwuYByg0b_-o115yUFLLcfNxMrLZmiW9k,55
-airbyte_source_github-1.8.29.dev202506231149.dist-info/RECORD,,
+airbyte_source_github-1.8.29.dev202506231532.dist-info/METADATA,sha256=RA_9qGWOu4OBbHUkqhr67po9hEFHCvUPZ-m7QIIiKNA,5218
+airbyte_source_github-1.8.29.dev202506231532.dist-info/WHEEL,sha256=b4K_helf-jlQoXBBETfwnf4B04YC67LOev0jo4fX5m8,88
+airbyte_source_github-1.8.29.dev202506231532.dist-info/entry_points.txt,sha256=gYhqVrTAZvMwuYByg0b_-o115yUFLLcfNxMrLZmiW9k,55
+airbyte_source_github-1.8.29.dev202506231532.dist-info/RECORD,,

source_github/errors_handlers.py

@@ -43,6 +43,13 @@ GITHUB_DEFAULT_ERROR_MAPPING = DEFAULT_ERROR_MAPPING | {
 }
 
 
+def is_conflict_with_empty_repository(response_or_exception: Optional[Union[requests.Response, Exception]] = None) -> bool:
+    if isinstance(response_or_exception, requests.Response) and response_or_exception.status_code == requests.codes.CONFLICT:
+        response_data = response_or_exception.json()
+        return response_data.get("message") == "Git Repository is empty."
+    return False
+
+
 class GithubStreamABCErrorHandler(HttpStatusErrorHandler):
     def __init__(self, stream: HttpStream, **kwargs):  # type: ignore # noqa
         self.stream = stream
@@ -88,6 +95,13 @@ class GithubStreamABCErrorHandler(HttpStatusErrorHandler):
                     error_message=f"Response status code: {response_or_exception.status_code}. Retrying...",
                 )
 
+            if is_conflict_with_empty_repository(response_or_exception=response_or_exception):
+                return ErrorResolution(
+                    response_action=ResponseAction.IGNORE,
+                    failure_type=FailureType.config_error,
+                    error_message="Git Repository is empty.",
+                )
+
         return super().interpret_response(response_or_exception)
 
 
@@ -135,18 +149,3 @@ class GitHubGraphQLErrorHandler(GithubStreamABCErrorHandler):
                 )
 
         return super().interpret_response(response_or_exception)
-
-
-class CommitsErrorHandler(GithubStreamABCErrorHandler):
-    def interpret_response(self, response_or_exception: Optional[Union[requests.Response, Exception]] = None) -> ErrorResolution:
-        if isinstance(response_or_exception, requests.Response):
-            if response_or_exception.status_code == requests.codes.CONFLICT:
-                response_data = response_or_exception.json()
-                if response_data.get("message") == "Git Repository is empty.":
-                    return ErrorResolution(
-                        response_action=ResponseAction.IGNORE,
-                        failure_type=FailureType.config_error,
-                        error_message="Git Repository is empty.",
-                    )
-
-        return super().interpret_response(response_or_exception)

source_github/streams.py

@@ -26,10 +26,10 @@ from . import constants
 from .backoff_strategies import ContributorActivityBackoffStrategy, GithubStreamABCBackoffStrategy
 from .errors_handlers import (
     GITHUB_DEFAULT_ERROR_MAPPING,
-    CommitsErrorHandler,
     ContributorActivityErrorHandler,
     GitHubGraphQLErrorHandler,
     GithubStreamABCErrorHandler,
+    is_conflict_with_empty_repository,
 )
 from .graphql import (
     CursorStorage,
@@ -101,10 +101,6 @@ class GithubStreamABC(HttpStream, ABC):
         stream_slice: Mapping[str, Any] = None,
         next_page_token: Mapping[str, Any] = None,
     ) -> Iterable[Mapping]:
-        if response.status_code == 409:
-            response_data = response.json()
-            if response_data.get("message") == "Git Repository is empty.":
-                return
         for record in response.json():  # GitHub puts records in an array.
             yield self.transform(record=record, stream_slice=stream_slice)
 
@@ -222,6 +218,24 @@ class GithubStream(GithubStreamABC):
         record["repository"] = stream_slice["repository"]
         return record
 
+    def parse_response(
+        self,
+        response: requests.Response,
+        stream_state: Mapping[str, Any],
+        stream_slice: Mapping[str, Any] = None,
+        next_page_token: Mapping[str, Any] = None,
+    ) -> Iterable[Mapping]:
+        if is_conflict_with_empty_repository(response):
+            # I would expect that this should be handled (skipped) by the error handler, but it seems like
+            # ignored this error but continue to processing records. This may be fixed in latest CDK versions.
+            return
+        yield from super().parse_response(
+            response=response,
+            stream_state=stream_state,
+            stream_slice=stream_slice,
+            next_page_token=next_page_token,
+        )
+
 
 class SemiIncrementalMixin(CheckpointMixin):
     """
@@ -668,11 +682,6 @@ class Commits(IncrementalMixin, GithubStream):
         self.branches_stream = Branches(**kwargs)
         self.repositories_stream = RepositoryStats(**kwargs)
 
-    def get_error_handler(self) -> Optional[ErrorHandler]:
-        return CommitsErrorHandler(
-            logger=self.logger, max_retries=self.max_retries, error_mapping=GITHUB_DEFAULT_ERROR_MAPPING, stream=self
-        )
-
     def request_params(self, stream_state: Mapping[str, Any], stream_slice: Mapping[str, Any] = None, **kwargs) -> MutableMapping[str, Any]:
         params = super(IncrementalMixin, self).request_params(stream_state=stream_state, stream_slice=stream_slice, **kwargs)
         since = self.get_starting_point(stream_state=stream_state, stream_slice=stream_slice)