unstructured-ingest 1.0.33__py3-none-any.whl → 1.0.34__py3-none-any.whl

unstructured_ingest/__version__.py

@@ -1 +1 @@
-__version__ = "1.0.33"  # pragma: no cover
+__version__ = "1.0.34"  # pragma: no cover

unstructured_ingest/processes/connectors/confluence.py

@@ -12,6 +12,7 @@ from unstructured_ingest.data_types.file_data import (
     SourceIdentifiers,
 )
 from unstructured_ingest.error import SourceConnectionError
+from unstructured_ingest.errors_v2 import UserAuthError, UserError
 from unstructured_ingest.interfaces import (
     AccessConfig,
     ConnectionConfig,
@@ -96,7 +97,7 @@ class ConfluenceConnectionConfig(ConnectionConfig):
 
     @requires_dependencies(["atlassian"], extras="confluence")
     @contextmanager
-    def get_client(self) -> "Confluence":
+    def get_client(self) -> Generator["Confluence", None, None]:
         from atlassian import Confluence
 
         access_configs = self.access_config.get_secret_value()
@@ -126,15 +127,36 @@ class ConfluenceIndexer(Indexer):
 
     def precheck(self) -> bool:
         try:
-            # Attempt to retrieve a list of spaces with limit=1.
-            # This should only succeed if all creds are valid
-            with self.connection_config.get_client() as client:
+            self.connection_config.get_client()
+        except Exception as e:
+            logger.exception(f"Failed to connect to Confluence: {e}")
+            raise UserAuthError(f"Failed to connect to Confluence: {e}")
+
+        with self.connection_config.get_client() as client:
+            # opportunistically check the first space in list of all spaces
+            try:
                 client.get_all_spaces(limit=1)
+            except Exception as e:
+                logger.exception(f"Failed to connect to find any Confluence space: {e}")
+                raise UserError(f"Failed to connect to find any Confluence space: {e}")
+
             logger.info("Connection to Confluence successful.")
-            return True
-        except Exception as e:
-            logger.error(f"Failed to connect to Confluence: {e}", exc_info=True)
-            raise SourceConnectionError(f"Failed to connect to Confluence: {e}")
+
+            # If specific spaces are provided, check if we can access them
+            errors = []
+
+            if self.index_config.spaces:
+                for space_key in self.index_config.spaces:
+                    try:
+                        client.get_space(space_key)
+                    except Exception as e:
+                        logger.exception(f"Failed to connect to Confluence: {e}")
+                        errors.append(f"Failed to connect to '{space_key}' space, cause: '{e}'")
+
+            if errors:
+                raise UserError("\n".join(errors))
+
+        return True
 
     def _get_space_ids_and_keys(self) -> List[Tuple[str, int]]:
         """
@@ -406,7 +428,7 @@ class ConfluenceDownloader(Downloader):
                     expand="history.lastUpdated,version,body.view",
                 )
         except Exception as e:
-            logger.error(f"Failed to retrieve page with ID {doc_id}: {e}", exc_info=True)
+            logger.exception(f"Failed to retrieve page with ID {doc_id}: {e}")
             raise SourceConnectionError(f"Failed to retrieve page with ID {doc_id}: {e}")
 
         if not page:

{unstructured_ingest-1.0.33.dist-info → unstructured_ingest-1.0.34.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: unstructured_ingest
-Version: 1.0.33
+Version: 1.0.34
 Summary: Local ETL data pipeline to get data RAG ready
 Author-email: Unstructured Technologies <devops@unstructuredai.io>
 License-Expression: Apache-2.0

{unstructured_ingest-1.0.33.dist-info → unstructured_ingest-1.0.34.dist-info}/RECORD

@@ -1,5 +1,5 @@
 unstructured_ingest/__init__.py,sha256=U4S_2y3zgLZVfMenHRaJFBW8yqh2mUBuI291LGQVOJ8,35
-unstructured_ingest/__version__.py,sha256=moV2VeZrrB_QVKOvny6NjEoowwTiGToZWfDpKig5QOQ,43
+unstructured_ingest/__version__.py,sha256=S3Vgmk2V2EWfbef_sUbnJb_d5x0m64Z8D_xx-_9kXOM,43
 unstructured_ingest/error.py,sha256=qDncnJgbf5ils956RcO2CGlAKYDT5OaEM9Clv1JVTNc,1448
 unstructured_ingest/errors_v2.py,sha256=9RuRCi7lbDxCguDz07y5RiHoQiFIOWwOD7xqzJ2B3Yw,436
 unstructured_ingest/logger.py,sha256=7e_7UeK6hVOd5BQ6i9NzRUAPCS_DF839Y8TjUDywraY,1428
@@ -66,7 +66,7 @@ unstructured_ingest/processes/connectors/airtable.py,sha256=smx5qBSUKwM8V6Xcc7ik
 unstructured_ingest/processes/connectors/astradb.py,sha256=Ob9wQgDxa6BXDPZBOqooNKQgvjIZcMwIe4fW3VlI7h8,18929
 unstructured_ingest/processes/connectors/azure_ai_search.py,sha256=szhSRXzUHk0DE2hGFfjGc_jNFzlUwiRlCtIkuu7tmnk,11524
 unstructured_ingest/processes/connectors/chroma.py,sha256=q5_Fu4xb6_W_NyrPxVa3-jVwZLqVdlBNlR4dFvbd7l0,7235
-unstructured_ingest/processes/connectors/confluence.py,sha256=1oT4A83jSOWR8u8kldHImOBqSLxctdlsR-AZpzJfO9w,21098
+unstructured_ingest/processes/connectors/confluence.py,sha256=C62LVwZYk7H8RfiPb0mbxig2osW5u7KvHIlz4qOJU-0,21954
 unstructured_ingest/processes/connectors/couchbase.py,sha256=KCHoYDNya9B05NIB5D78zXoizFyfpJRepcYBe1nLSOs,12298
 unstructured_ingest/processes/connectors/delta_table.py,sha256=2DFox_Vzoopt_D3Jy3rCjrrTGMutG2INIrwCeoIohRY,7340
 unstructured_ingest/processes/connectors/discord.py,sha256=6yEJ_agfKUqsV43wFsbMkcd8lcLJC0uqbo4izjdZ3rU,5294
@@ -231,8 +231,8 @@ unstructured_ingest/utils/ndjson.py,sha256=nz8VUOPEgAFdhaDOpuveknvCU4x82fVwqE01q
 unstructured_ingest/utils/pydantic_models.py,sha256=BT_j15e4rX40wQbt8LUXbqfPhA3rJn1PHTI_G_A_EHY,1720
 unstructured_ingest/utils/string_and_date_utils.py,sha256=oXOI6rxXq-8ncbk7EoJK0WCcTXWj75EzKl8pfQMID3U,2522
 unstructured_ingest/utils/table.py,sha256=WZechczgVFvlodUWFcsnCGvBNh1xRm6hr0VbJTPxKAc,3669
-unstructured_ingest-1.0.33.dist-info/METADATA,sha256=d8F0hFb3s-aLloV1TGFLDHRa8CHwuTduHFS1neEHu6s,8747
-unstructured_ingest-1.0.33.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
-unstructured_ingest-1.0.33.dist-info/entry_points.txt,sha256=gUAAFnjFPnBgThJSEbw0N5ZjxtaKlT1s9e05_arQrNw,70
-unstructured_ingest-1.0.33.dist-info/licenses/LICENSE.md,sha256=SxkKP_62uIAKb9mb1eH7FH4Kn2aYT09fgjKpJt5PyTk,11360
-unstructured_ingest-1.0.33.dist-info/RECORD,,
+unstructured_ingest-1.0.34.dist-info/METADATA,sha256=Pw-KP4al9gteAFj6lqY7xkFRjWj1rTAgN960UsAUZAM,8747
+unstructured_ingest-1.0.34.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
+unstructured_ingest-1.0.34.dist-info/entry_points.txt,sha256=gUAAFnjFPnBgThJSEbw0N5ZjxtaKlT1s9e05_arQrNw,70
+unstructured_ingest-1.0.34.dist-info/licenses/LICENSE.md,sha256=SxkKP_62uIAKb9mb1eH7FH4Kn2aYT09fgjKpJt5PyTk,11360
+unstructured_ingest-1.0.34.dist-info/RECORD,,