unstructured-ingest 1.2.9__py3-none-any.whl → 1.2.11__py3-none-any.whl

unstructured_ingest/processes/connectors/github.py

@@ -12,7 +12,12 @@ from unstructured_ingest.data_types.file_data import (
     FileDataSourceMetadata,
     SourceIdentifiers,
 )
-from unstructured_ingest.errors_v2 import ProviderError, UserAuthError, UserError
+from unstructured_ingest.error import (
+    ProviderError,
+    UnstructuredIngestError,
+    UserAuthError,
+    UserError,
+)
 from unstructured_ingest.interfaces import (
     AccessConfig,
     ConnectionConfig,
@@ -85,7 +90,7 @@ class GithubConnectionConfig(ConnectionConfig):
         if status_code > 500:
             return ProviderError(e.response.text)
         logger.debug(f"unhandled http error: {e}")
-        return e
+        return UnstructuredIngestError(str(e))
 
     @requires_dependencies(["requests"], extras="github")
     def wrap_error(self, e: Exception) -> Exception:
@@ -97,7 +102,7 @@ class GithubConnectionConfig(ConnectionConfig):
         if isinstance(e, HTTPError):
             return self.wrap_http_error(e=e)
         logger.debug(f"unhandled error: {e}")
-        return e
+        return UnstructuredIngestError(str(e))
 
 
 class GithubIndexerConfig(IndexerConfig):

unstructured_ingest/processes/connectors/gitlab.py

@@ -13,7 +13,7 @@ from unstructured_ingest.data_types.file_data import (
     FileDataSourceMetadata,
     SourceIdentifiers,
 )
-from unstructured_ingest.error import SourceConnectionError
+from unstructured_ingest.error import SourceConnectionError, ValueError
 from unstructured_ingest.interfaces import (
     AccessConfig,
     ConnectionConfig,

unstructured_ingest/processes/connectors/google_drive.py

@@ -13,9 +13,7 @@ from unstructured_ingest.data_types.file_data import (
     FileDataSourceMetadata,
     SourceIdentifiers,
 )
-from unstructured_ingest.error import (
-    SourceConnectionError,
-)
+from unstructured_ingest.error import SourceConnectionError, UserAuthError, ValueError
 from unstructured_ingest.interfaces import (
     AccessConfig,
     ConnectionConfig,
@@ -113,7 +111,7 @@ class GoogleDriveConnectionConfig(ConnectionConfig):
         except HttpError as exc:
             raise ValueError(f"{exc.reason}")
         except exceptions.DefaultCredentialsError:
-            raise ValueError("The provided API key is invalid.")
+            raise UserAuthError("The provided API key is invalid.")
 
 
 class GoogleDriveIndexerConfig(IndexerConfig):

unstructured_ingest/processes/connectors/ibm_watsonx/ibm_watsonx_s3.py

@@ -8,7 +8,13 @@ from typing import TYPE_CHECKING, Any, Generator, Optional, Tuple
 from pydantic import Field, Secret
 
 from unstructured_ingest.data_types.file_data import FileData
-from unstructured_ingest.errors_v2 import ProviderError, UserAuthError, UserError
+from unstructured_ingest.error import (
+    DestinationConnectionError,
+    IcebergCommitFailedException,
+    ProviderError,
+    UserAuthError,
+    UserError,
+)
 from unstructured_ingest.interfaces import (
     AccessConfig,
     ConnectionConfig,
@@ -40,10 +46,6 @@ DEFAULT_ICEBERG_URI_PATH = "/mds/iceberg"
 DEFAULT_ICEBERG_CATALOG_TYPE = "rest"
 
 
-class IcebergCommitFailedException(Exception):
-    """Failed to commit changes to the iceberg table."""
-
-
 class IbmWatsonxAccessConfig(AccessConfig):
     iam_api_key: str = Field(description="IBM IAM API Key")
     access_key_id: str = Field(description="Cloud Object Storage HMAC Access Key ID")
@@ -292,16 +294,16 @@ class IbmWatsonxUploader(SQLUploader):
             except CommitFailedException as e:
                 table.refresh()
                 logger.debug(e)
-                raise IcebergCommitFailedException(e)
-            except RESTError:
-                raise
+                raise IcebergCommitFailedException(str(e))
+            except RESTError as e:
+                raise DestinationConnectionError(str(e))
             except Exception as e:
                 raise ProviderError(f"Failed to upload data to table: {e}")
 
         try:
             return _upload_data_table(table, data_table, file_data)
-        except RESTError:
-            raise
+        except RESTError as e:
+            raise DestinationConnectionError(str(e))
         except ProviderError:
             raise
         except Exception as e:

unstructured_ingest/processes/connectors/jira.py

@@ -12,7 +12,7 @@ from unstructured_ingest.data_types.file_data import (
     FileDataSourceMetadata,
     SourceIdentifiers,
 )
-from unstructured_ingest.error import SourceConnectionError
+from unstructured_ingest.error import SourceConnectionError, ValueError
 from unstructured_ingest.interfaces import (
     AccessConfig,
     ConnectionConfig,

unstructured_ingest/processes/connectors/kafka/kafka.py

@@ -17,6 +17,8 @@ from unstructured_ingest.error import (
     DestinationConnectionError,
     SourceConnectionError,
     SourceConnectionNetworkError,
+    UnstructuredIngestError,
+    ValueError,
 )
 from unstructured_ingest.interfaces import (
     AccessConfig,
@@ -101,7 +103,7 @@ class KafkaIndexer(Indexer, ABC):
 
     @requires_dependencies(["confluent_kafka"], extras="kafka")
     def generate_messages(self) -> Generator[Any, None, None]:
-        from confluent_kafka import KafkaError, KafkaException
+        from confluent_kafka import KafkaError
 
         messages_consumed = 0
         max_empty_polls = 10
@@ -122,7 +124,7 @@ class KafkaIndexer(Indexer, ABC):
                         )
                         break
                     else:
-                        raise KafkaException(msg.error())
+                        raise UnstructuredIngestError(msg.error())
                 try:
                     empty_polls = 0
                     messages_consumed += 1
@@ -246,8 +248,6 @@ class KafkaUploader(Uploader, ABC):
             raise DestinationConnectionError(f"failed to validate connection: {e}")
 
     def produce_batch(self, elements: list[dict]) -> None:
-        from confluent_kafka.error import KafkaException
-
         producer = self.connection_config.get_producer()
         failed_producer = False
 
@@ -268,7 +268,7 @@ class KafkaUploader(Uploader, ABC):
             logger.debug(f"another iteration of kafka producer flush. Queue length: {producer_len}")
             producer.flush(timeout=self.upload_config.timeout)
         if failed_producer:
-            raise KafkaException("failed to produce all messages in batch")
+            raise UnstructuredIngestError("failed to produce all kafka messages in batch")
 
     def run_data(self, data: list[dict], file_data: FileData, **kwargs: Any) -> None:
         for element_batch in batch_generator(data, batch_size=self.upload_config.batch_size):

unstructured_ingest/processes/connectors/local.py

@@ -13,6 +13,7 @@ from unstructured_ingest.data_types.file_data import (
     FileDataSourceMetadata,
     SourceIdentifiers,
 )
+from unstructured_ingest.error import FileExistsError
 from unstructured_ingest.interfaces import (
     AccessConfig,
     ConnectionConfig,
@@ -168,7 +169,7 @@ class LocalUploaderConfig(UploaderConfig):
 
     def __post_init__(self):
         if self.output_path.exists() and self.output_path.is_file():
-            raise ValueError("output path already exists as a file")
+            raise FileExistsError(f"output path {self.output_path} already exists as a file")
 
 
 @dataclass

unstructured_ingest/processes/connectors/milvus.py

@@ -7,7 +7,12 @@ from dateutil import parser
 from pydantic import Field, Secret
 
 from unstructured_ingest.data_types.file_data import FileData
-from unstructured_ingest.error import DestinationConnectionError, WriteError
+from unstructured_ingest.error import (
+    DestinationConnectionError,
+    KeyError,
+    ValueError,
+    WriteError,
+)
 from unstructured_ingest.interfaces import (
     AccessConfig,
     ConnectionConfig,

unstructured_ingest/processes/connectors/mongodb.py

@@ -14,7 +14,12 @@ from unstructured_ingest.data_types.file_data import (
     FileDataSourceMetadata,
     SourceIdentifiers,
 )
-from unstructured_ingest.error import DestinationConnectionError, SourceConnectionError
+from unstructured_ingest.error import (
+    ConnectionError,
+    DestinationConnectionError,
+    SourceConnectionError,
+    ValueError,
+)
 from unstructured_ingest.interfaces import (
     AccessConfig,
     ConnectionConfig,

unstructured_ingest/processes/connectors/neo4j.py

@@ -13,7 +13,11 @@ from pydantic import BaseModel, ConfigDict, Field, Secret, ValidationError, fiel
 
 from unstructured_ingest.data_types.entities import EntitiesData, Entity, EntityRelationship
 from unstructured_ingest.data_types.file_data import FileData
-from unstructured_ingest.error import DestinationConnectionError
+from unstructured_ingest.error import (
+    DestinationConnectionError,
+    UnstructuredIngestError,
+    ValueError,
+)
 from unstructured_ingest.interfaces import (
     AccessConfig,
     ConnectionConfig,
@@ -374,7 +378,7 @@ class Neo4jUploader(Uploader):
             if e.code == "Neo.ClientError.Schema.EquivalentSchemaRuleAlreadyExists":
                 logger.info(f"Index on nodes labeled '{label.value}' already exists.")
             else:
-                raise
+                raise UnstructuredIngestError(str(e))
 
     async def _delete_old_data_if_exists(self, file_data: FileData, client: AsyncDriver) -> None:
         logger.info(f"Deleting old data for the record '{file_data.identifier}' (if present).")

unstructured_ingest/processes/connectors/notion/client.py

@@ -8,8 +8,8 @@ from notion_client.api_endpoints import BlocksEndpoint as NotionBlocksEndpoint
 from notion_client.api_endpoints import DatabasesEndpoint as NotionDatabasesEndpoint
 from notion_client.api_endpoints import Endpoint
 from notion_client.api_endpoints import PagesEndpoint as NotionPagesEndpoint
-from notion_client.errors import HTTPResponseError, RequestTimeoutError
 
+from unstructured_ingest.error import SourceConnectionError, TimeoutError
 from unstructured_ingest.processes.connectors.notion.ingest_backoff import RetryHandler
 from unstructured_ingest.processes.connectors.notion.ingest_backoff.types import RetryStrategyConfig
 from unstructured_ingest.processes.connectors.notion.types.block import Block
@@ -132,8 +132,8 @@ class DatabasesEndpoint(NotionDatabasesEndpoint):
                 else (self.parent.client.send(request))
             )  # type: ignore
             return response.status_code
-        except httpx.TimeoutException:
-            raise RequestTimeoutError()
+        except httpx.TimeoutException as e:
+            raise TimeoutError(str(e))
 
     def query(self, database_id: str, **kwargs: Any) -> Tuple[List[Page], dict]:
         """Get a list of [Pages](https://developers.notion.com/reference/page) contained in the database.
@@ -236,8 +236,8 @@ class PagesEndpoint(NotionPagesEndpoint):
                 else (self.parent.client.send(request))
             )  # type: ignore
             return response.status_code
-        except httpx.TimeoutException:
-            raise RequestTimeoutError()
+        except httpx.TimeoutException as e:
+            raise TimeoutError(str(e))
 
 
 class Client(NotionClient):
@@ -266,9 +266,9 @@ class AsyncBlocksChildrenEndpoint(NotionBlocksChildrenEndpoint):
             )
             response.raise_for_status()
         except httpx.HTTPStatusError as e:
-            raise HTTPResponseError(f"Failed to list blocks: {str(e)}")
-        except httpx.TimeoutException:
-            raise RequestTimeoutError()
+            raise SourceConnectionError(f"Failed to list blocks: {str(e)}")
+        except httpx.TimeoutException as e:
+            raise TimeoutError(str(e))
 
         resp = response.json()
         child_blocks = [Block.from_dict(data=b) for b in resp.pop("results", [])]
@@ -307,9 +307,9 @@ class AsyncDatabasesEndpoint(NotionDatabasesEndpoint):
             )
             response.raise_for_status()
         except httpx.HTTPStatusError as e:
-            raise HTTPResponseError(f"Failed to retrieve database: {str(e)}")
-        except httpx.TimeoutException:
-            raise RequestTimeoutError()
+            raise SourceConnectionError(f"Failed to retrieve database: {str(e)}")
+        except httpx.TimeoutException as e:
+            raise TimeoutError(str(e))
 
         return Database.from_dict(data=response.json())
 
@@ -322,9 +322,9 @@ class AsyncDatabasesEndpoint(NotionDatabasesEndpoint):
             )
             response.raise_for_status()
         except httpx.HTTPStatusError as e:
-            raise HTTPResponseError(f"Failed to query database: {str(e)}")
-        except httpx.TimeoutException:
-            raise RequestTimeoutError()
+            raise SourceConnectionError(f"Failed to query database: {str(e)}")
+        except httpx.TimeoutException as e:
+            raise TimeoutError(str(e))
 
         resp = response.json()
         pages = [Page.from_dict(data=p) for p in resp.pop("results", [])]

unstructured_ingest/processes/connectors/notion/connector.py

@@ -9,7 +9,7 @@ from unstructured_ingest.data_types.file_data import (
     FileDataSourceMetadata,
     SourceIdentifiers,
 )
-from unstructured_ingest.error import SourceConnectionError
+from unstructured_ingest.error import SourceConnectionError, ValueError
 from unstructured_ingest.interfaces import (
     AccessConfig,
     ConnectionConfig,

unstructured_ingest/processes/connectors/onedrive.py

@@ -19,8 +19,9 @@ from unstructured_ingest.error import (
     DestinationConnectionError,
     SourceConnectionError,
     SourceConnectionNetworkError,
+    UserAuthError,
+    ValueError,
 )
-from unstructured_ingest.errors_v2 import UserAuthError
 from unstructured_ingest.interfaces import (
     AccessConfig,
     ConnectionConfig,

unstructured_ingest/processes/connectors/outlook.py

@@ -12,7 +12,7 @@ from unstructured_ingest.data_types.file_data import (
     FileDataSourceMetadata,
     SourceIdentifiers,
 )
-from unstructured_ingest.error import SourceConnectionError
+from unstructured_ingest.error import SourceConnectionError, ValueError
 from unstructured_ingest.interfaces import (
     AccessConfig,
     ConnectionConfig,

unstructured_ingest/processes/connectors/pinecone.py

@@ -7,8 +7,12 @@ from typing import TYPE_CHECKING, Any, Literal, Optional
 from pydantic import Field, Secret
 
 from unstructured_ingest.data_types.file_data import FileData
-from unstructured_ingest.error import DestinationConnectionError
-from unstructured_ingest.errors_v2 import UserError
+from unstructured_ingest.error import (
+    DestinationConnectionError,
+    NotFoundError,
+    UnstructuredIngestError,
+    UserError,
+)
 from unstructured_ingest.interfaces import (
     AccessConfig,
     ConnectionConfig,
@@ -220,9 +224,7 @@ class PineconeUploader(VectorDBUploader):
             if self.connection_config.index_name and not self.index_exists(
                 self.connection_config.index_name
             ):
-                raise DestinationConnectionError(
-                    f"index {self.connection_config.index_name} does not exist"
-                )
+                raise NotFoundError(f"index {self.connection_config.index_name} does not exist")
         except Exception as e:
             logger.error(f"failed to validate connection: {e}", exc_info=True)
             raise DestinationConnectionError(f"failed to validate connection: {e}")
@@ -364,7 +366,7 @@ class PineconeUploader(VectorDBUploader):
             try:
                 results = [async_result.get() for async_result in async_results]
             except PineconeApiException as api_error:
-                raise DestinationConnectionError(f"http error: {api_error}") from api_error
+                raise UnstructuredIngestError(f"http error: {api_error}") from api_error
             logger.debug(f"results: {results}")
 
     def run_data(self, data: list[dict], file_data: FileData, **kwargs: Any) -> None:

unstructured_ingest/processes/connectors/redisdb.py

@@ -6,7 +6,7 @@ from typing import TYPE_CHECKING, Any, AsyncGenerator, Generator, Optional
 from pydantic import Field, Secret, model_validator
 
 from unstructured_ingest.data_types.file_data import FileData
-from unstructured_ingest.error import DestinationConnectionError
+from unstructured_ingest.error import DestinationConnectionError, ResponseError, ValueError
 from unstructured_ingest.interfaces import (
     AccessConfig,
     ConnectionConfig,
@@ -203,7 +203,7 @@ class RedisUploader(Uploader):
                     await pipe.set(key_with_prefix, json.dumps(element)).execute()
                     redis_stack = False
                 else:
-                    raise redis_exceptions.ResponseError(message) from e
+                    raise ResponseError(message) from e
         return redis_stack
 
 

unstructured_ingest/processes/connectors/salesforce.py

@@ -25,7 +25,12 @@ from unstructured_ingest.data_types.file_data import (
     FileDataSourceMetadata,
     SourceIdentifiers,
 )
-from unstructured_ingest.error import SourceConnectionError, SourceConnectionNetworkError
+from unstructured_ingest.error import (
+    MissingCategoryError,
+    SourceConnectionError,
+    SourceConnectionNetworkError,
+    ValueError,
+)
 from unstructured_ingest.interfaces import (
     AccessConfig,
     ConnectionConfig,
@@ -41,11 +46,6 @@ from unstructured_ingest.processes.connector_registry import (
 )
 from unstructured_ingest.utils.dep_check import requires_dependencies
 
-
-class MissingCategoryError(Exception):
-    """There are no categories with that name."""
-
-
 CONNECTOR_TYPE = "salesforce"
 
 if TYPE_CHECKING:

unstructured_ingest/processes/connectors/sharepoint.py

@@ -10,10 +10,13 @@ from unstructured_ingest.data_types.file_data import (
     FileData,
 )
 from unstructured_ingest.error import (
+    NotFoundError,
     SourceConnectionError,
     SourceConnectionNetworkError,
+    UserAuthError,
+    UserError,
+    ValueError,
 )
-from unstructured_ingest.errors_v2 import UserAuthError, UserError
 from unstructured_ingest.logger import logger
 from unstructured_ingest.processes.connector_registry import (
     SourceRegistryEntry,
@@ -239,7 +242,7 @@ class SharepointDownloader(OnedriveDownloader):
         file = site_drive_item.get_by_path(server_relative_path).get().execute_query()
 
         if not file:
-            raise FileNotFoundError(f"file not found: {server_relative_path}")
+            raise NotFoundError(f"file not found: {server_relative_path}")
         return file
 
 

unstructured_ingest/processes/connectors/slack.py

@@ -13,7 +13,7 @@ from unstructured_ingest.data_types.file_data import (
     FileDataSourceMetadata,
     SourceIdentifiers,
 )
-from unstructured_ingest.error import SourceConnectionError
+from unstructured_ingest.error import SourceConnectionError, ValueError
 from unstructured_ingest.interfaces import (
     AccessConfig,
     ConnectionConfig,

unstructured_ingest/processes/connectors/sql/databricks_delta_tables.py

@@ -8,6 +8,7 @@ from typing import TYPE_CHECKING, Any, Generator, Optional
 from pydantic import Field, Secret
 
 from unstructured_ingest.data_types.file_data import FileData
+from unstructured_ingest.error import ValueError
 from unstructured_ingest.logger import logger
 from unstructured_ingest.processes.connector_registry import (
     DestinationRegistryEntry,

unstructured_ingest/processes/connectors/sql/sqlite.py

@@ -7,6 +7,7 @@ from typing import TYPE_CHECKING, Any, Generator
 from pydantic import Field, Secret, model_validator
 
 from unstructured_ingest.data_types.file_data import FileData
+from unstructured_ingest.error import ValueError
 from unstructured_ingest.logger import logger
 from unstructured_ingest.processes.connector_registry import (
     DestinationRegistryEntry,

unstructured_ingest/processes/connectors/vectara.py

@@ -9,7 +9,7 @@ from typing import Any, Dict, Mapping, Optional
 from pydantic import Field, Secret
 
 from unstructured_ingest.data_types.file_data import FileData
-from unstructured_ingest.error import DestinationConnectionError
+from unstructured_ingest.error import DestinationConnectionError, ValueError
 from unstructured_ingest.interfaces import (
     AccessConfig,
     ConnectionConfig,

unstructured_ingest/processes/connectors/weaviate/cloud.py

@@ -4,6 +4,7 @@ from typing import TYPE_CHECKING, Any, Generator, Optional
 
 from pydantic import Field, Secret
 
+from unstructured_ingest.error import ValueError
 from unstructured_ingest.processes.connector_registry import DestinationRegistryEntry
 from unstructured_ingest.processes.connectors.weaviate.weaviate import (
     WeaviateAccessConfig,

unstructured_ingest/processes/connectors/weaviate/weaviate.py

@@ -11,7 +11,7 @@ from dateutil import parser
 from pydantic import Field, Secret
 
 from unstructured_ingest.data_types.file_data import FileData
-from unstructured_ingest.error import DestinationConnectionError, WriteError
+from unstructured_ingest.error import DestinationConnectionError, ValueError, WriteError
 from unstructured_ingest.interfaces import (
     AccessConfig,
     ConnectionConfig,

unstructured_ingest/processes/connectors/zendesk/client.py

@@ -4,7 +4,13 @@ from typing import TYPE_CHECKING, Any, AsyncGenerator, Literal, Optional, Union
 
 from pydantic import BaseModel, Field, HttpUrl
 
-from unstructured_ingest.errors_v2 import ProviderError, RateLimitError, UserAuthError, UserError
+from unstructured_ingest.error import (
+    ProviderError,
+    RateLimitError,
+    UnstructuredIngestError,
+    UserAuthError,
+    UserError,
+)
 from unstructured_ingest.logger import logger
 from unstructured_ingest.utils.dep_check import requires_dependencies
 from unstructured_ingest.utils.string_and_date_utils import fix_unescaped_unicode
@@ -211,7 +217,7 @@ class ZendeskClient:
 
         if not isinstance(e, httpx.HTTPStatusError):
             logger.error(f"unhandled exception from Zendesk client: {e}", exc_info=True)
-            return e
+            return UnstructuredIngestError(str(e))
         url = e.request.url
         response_code = e.response.status_code
         if response_code == 401:

unstructured_ingest/processes/connectors/zendesk/zendesk.py

@@ -13,6 +13,7 @@ from unstructured_ingest.data_types.file_data import (
     FileDataSourceMetadata,
     SourceIdentifiers,
 )
+from unstructured_ingest.error import SourceConnectionError, ValueError
 from unstructured_ingest.interfaces import (
     AccessConfig,
     ConnectionConfig,
@@ -222,7 +223,9 @@ class ZendeskDownloader(Downloader):
             ticket = ZendeskTicket.model_validate(zendesk_filedata.additional_metadata.content)
             await self.download_ticket(ticket=ticket, download_path=download_path)
         else:
-            raise RuntimeError(f"Item type {item_type} cannot be handled by the downloader")
+            raise SourceConnectionError(
+                f"Item type {item_type} cannot be handled by the downloader"
+            )
         return super().generate_download_response(
             file_data=zendesk_filedata, download_path=download_path
         )

unstructured_ingest/processes/partitioner.py

@@ -5,7 +5,7 @@ from typing import Any, Optional
 
 from pydantic import BaseModel, Field, SecretStr
 
-from unstructured_ingest.errors_v2 import UserError
+from unstructured_ingest.error import UserError
 from unstructured_ingest.interfaces.process import BaseProcess
 from unstructured_ingest.logger import logger
 from unstructured_ingest.unstructured_api import call_api_async

unstructured_ingest/unstructured_api.py

@@ -2,7 +2,7 @@ from dataclasses import fields
 from pathlib import Path
 from typing import TYPE_CHECKING, Optional
 
-from unstructured_ingest.errors_v2 import ProviderError, QuotaError, UserAuthError, UserError
+from unstructured_ingest.error import ProviderError, QuotaError, UserAuthError, UserError
 from unstructured_ingest.logger import logger
 
 if TYPE_CHECKING:

{unstructured_ingest-1.2.9.dist-info → unstructured_ingest-1.2.11.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: unstructured_ingest
-Version: 1.2.9
+Version: 1.2.11
 Summary: Local ETL data pipeline to get data RAG ready
 Author-email: Unstructured Technologies <devops@unstructuredai.io>
 License-Expression: Apache-2.0