unstructured-ingest 0.3.10__py3-none-any.whl → 0.3.11__py3-none-any.whl

test/integration/connectors/{databricks_tests → databricks}/test_volumes_native.py

@@ -1,10 +1,10 @@
 import json
 import os
-import tempfile
 import uuid
 from contextlib import contextmanager
 from dataclasses import dataclass
 from pathlib import Path
+from unittest import mock
 
 import pytest
 from databricks.sdk import WorkspaceClient
@@ -31,11 +31,15 @@ from unstructured_ingest.v2.processes.connectors.databricks.volumes_native impor
 
 
 @dataclass
-class EnvData:
+class BaseEnvData:
     host: str
+    catalog: str
+
+
+@dataclass
+class BasicAuthEnvData(BaseEnvData):
     client_id: str
     client_secret: str
-    catalog: str
 
     def get_connection_config(self) -> DatabricksNativeVolumesConnectionConfig:
         return DatabricksNativeVolumesConnectionConfig(
@@ -47,8 +51,21 @@ class EnvData:
         )
 
 
-def get_env_data() -> EnvData:
-    return EnvData(
+@dataclass
+class PATEnvData(BaseEnvData):
+    token: str
+
+    def get_connection_config(self) -> DatabricksNativeVolumesConnectionConfig:
+        return DatabricksNativeVolumesConnectionConfig(
+            host=self.host,
+            access_config=DatabricksNativeVolumesAccessConfig(
+                token=self.token,
+            ),
+        )
+
+
+def get_basic_auth_env_data() -> BasicAuthEnvData:
+    return BasicAuthEnvData(
         host=os.environ["DATABRICKS_HOST"],
         client_id=os.environ["DATABRICKS_CLIENT_ID"],
         client_secret=os.environ["DATABRICKS_CLIENT_SECRET"],
@@ -56,23 +73,30 @@ def get_env_data() -> EnvData:
     )
 
 
+def get_pat_env_data() -> PATEnvData:
+    return PATEnvData(
+        host=os.environ["DATABRICKS_HOST"],
+        catalog=os.environ["DATABRICKS_CATALOG"],
+        token=os.environ["DATABRICKS_PAT"],
+    )
+
+
 @pytest.mark.asyncio
 @pytest.mark.tags(CONNECTOR_TYPE, SOURCE_TAG)
 @requires_env(
     "DATABRICKS_HOST", "DATABRICKS_CLIENT_ID", "DATABRICKS_CLIENT_SECRET", "DATABRICKS_CATALOG"
 )
-async def test_volumes_native_source():
-    env_data = get_env_data()
-    indexer_config = DatabricksNativeVolumesIndexerConfig(
-        recursive=True,
-        volume="test-platform",
-        volume_path="databricks-volumes-test-input",
-        catalog=env_data.catalog,
-    )
-    connection_config = env_data.get_connection_config()
-    with tempfile.TemporaryDirectory() as tempdir:
-        tempdir_path = Path(tempdir)
-        download_config = DatabricksNativeVolumesDownloaderConfig(download_dir=tempdir_path)
+async def test_volumes_native_source(tmp_path: Path):
+    env_data = get_basic_auth_env_data()
+    with mock.patch.dict(os.environ, clear=True):
+        indexer_config = DatabricksNativeVolumesIndexerConfig(
+            recursive=True,
+            volume="test-platform",
+            volume_path="databricks-volumes-test-input",
+            catalog=env_data.catalog,
+        )
+        connection_config = env_data.get_connection_config()
+        download_config = DatabricksNativeVolumesDownloaderConfig(download_dir=tmp_path)
         indexer = DatabricksNativeVolumesIndexer(
             connection_config=connection_config, index_config=indexer_config
         )
@@ -89,12 +113,44 @@ async def test_volumes_native_source():
         )
 
 
+@pytest.mark.asyncio
+@pytest.mark.tags(CONNECTOR_TYPE, SOURCE_TAG)
+@requires_env("DATABRICKS_HOST", "DATABRICKS_PAT", "DATABRICKS_CATALOG")
+async def test_volumes_native_source_pat(tmp_path: Path):
+    env_data = get_pat_env_data()
+    with mock.patch.dict(os.environ, clear=True):
+        indexer_config = DatabricksNativeVolumesIndexerConfig(
+            recursive=True,
+            volume="test-platform",
+            volume_path="databricks-volumes-test-input",
+            catalog=env_data.catalog,
+        )
+        connection_config = env_data.get_connection_config()
+        download_config = DatabricksNativeVolumesDownloaderConfig(download_dir=tmp_path)
+        indexer = DatabricksNativeVolumesIndexer(
+            connection_config=connection_config, index_config=indexer_config
+        )
+        downloader = DatabricksNativeVolumesDownloader(
+            connection_config=connection_config, download_config=download_config
+        )
+        await source_connector_validation(
+            indexer=indexer,
+            downloader=downloader,
+            configs=SourceValidationConfigs(
+                test_id="databricks_volumes_native_pat",
+                expected_num_files=1,
+            ),
+        )
+
+
 def _get_volume_path(catalog: str, volume: str, volume_path: str):
     return f"/Volumes/{catalog}/default/{volume}/{volume_path}"
 
 
 @contextmanager
-def databricks_destination_context(env_data: EnvData, volume: str, volume_path) -> WorkspaceClient:
+def databricks_destination_context(
+    env_data: BasicAuthEnvData, volume: str, volume_path
+) -> WorkspaceClient:
     client = WorkspaceClient(
         host=env_data.host, client_id=env_data.client_id, client_secret=env_data.client_secret
     )
@@ -137,7 +193,7 @@ def validate_upload(client: WorkspaceClient, catalog: str, volume: str, volume_p
     "DATABRICKS_HOST", "DATABRICKS_CLIENT_ID", "DATABRICKS_CLIENT_SECRET", "DATABRICKS_CATALOG"
 )
 async def test_volumes_native_destination(upload_file: Path):
-    env_data = get_env_data()
+    env_data = get_basic_auth_env_data()
     volume_path = f"databricks-volumes-test-output-{uuid.uuid4()}"
     file_data = FileData(
         source_identifiers=SourceIdentifiers(fullpath=upload_file.name, filename=upload_file.name),

test/integration/connectors/sql/test_postgres.py

@@ -15,7 +15,7 @@ from test.integration.connectors.utils.validation.source import (
     SourceValidationConfigs,
     source_connector_validation,
 )
-from unstructured_ingest.v2.interfaces import FileData
+from unstructured_ingest.v2.interfaces import FileData, SourceIdentifiers
 from unstructured_ingest.v2.processes.connectors.sql.postgres import (
     CONNECTOR_TYPE,
     PostgresAccessConfig,
@@ -119,7 +119,11 @@ def validate_destination(
 async def test_postgres_destination(upload_file: Path, temp_dir: Path):
     # the postgres destination connector doesn't leverage the file data but is required as an input,
     # mocking it with arbitrary values to meet the base requirements:
-    mock_file_data = FileData(identifier="mock file data", connector_type=CONNECTOR_TYPE)
+    mock_file_data = FileData(
+        identifier="mock file data",
+        connector_type=CONNECTOR_TYPE,
+        source_identifiers=SourceIdentifiers(filename=upload_file.name, fullpath=upload_file.name),
+    )
     with docker_compose_context(
         docker_compose_path=env_setup_path / "sql" / "postgres" / "destination"
     ):

test/integration/connectors/sql/test_singlestore.py

@@ -15,7 +15,7 @@ from test.integration.connectors.utils.validation.source import (
     SourceValidationConfigs,
     source_connector_validation,
 )
-from unstructured_ingest.v2.interfaces import FileData
+from unstructured_ingest.v2.interfaces import FileData, SourceIdentifiers
 from unstructured_ingest.v2.processes.connectors.sql.singlestore import (
     CONNECTOR_TYPE,
     SingleStoreAccessConfig,
@@ -103,7 +103,11 @@ def validate_destination(
 @pytest.mark.asyncio
 @pytest.mark.tags(CONNECTOR_TYPE, DESTINATION_TAG, "sql")
 async def test_singlestore_destination(upload_file: Path, temp_dir: Path):
-    mock_file_data = FileData(identifier="mock file data", connector_type=CONNECTOR_TYPE)
+    mock_file_data = FileData(
+        identifier="mock file data",
+        connector_type=CONNECTOR_TYPE,
+        source_identifiers=SourceIdentifiers(filename=upload_file.name, fullpath=upload_file.name),
+    )
     with docker_compose_context(
         docker_compose_path=env_setup_path / "sql" / "singlestore" / "destination"
     ):

test/integration/connectors/sql/test_snowflake.py

@@ -17,7 +17,7 @@ from test.integration.connectors.utils.validation.source import (
     source_connector_validation,
 )
 from test.integration.utils import requires_env
-from unstructured_ingest.v2.interfaces import FileData
+from unstructured_ingest.v2.interfaces import FileData, SourceIdentifiers
 from unstructured_ingest.v2.processes.connectors.sql.snowflake import (
     CONNECTOR_TYPE,
     SnowflakeAccessConfig,
@@ -170,7 +170,11 @@ async def test_snowflake_destination(
 ):
     # the postgres destination connector doesn't leverage the file data but is required as an input,
     # mocking it with arbitrary values to meet the base requirements:
-    mock_file_data = FileData(identifier="mock file data", connector_type=CONNECTOR_TYPE)
+    mock_file_data = FileData(
+        identifier="mock file data",
+        connector_type=CONNECTOR_TYPE,
+        source_identifiers=SourceIdentifiers(filename=upload_file.name, fullpath=upload_file.name),
+    )
     init_db_destination()
     stager = SnowflakeUploadStager()
     staged_path = stager.run(

test/integration/connectors/sql/test_sqlite.py

@@ -15,7 +15,7 @@ from test.integration.connectors.utils.validation.source import (
     SourceValidationConfigs,
     source_connector_validation,
 )
-from unstructured_ingest.v2.interfaces import FileData
+from unstructured_ingest.v2.interfaces import FileData, SourceIdentifiers
 from unstructured_ingest.v2.processes.connectors.sql.sqlite import (
     CONNECTOR_TYPE,
     SQLiteConnectionConfig,
@@ -116,7 +116,11 @@ async def test_sqlite_destination(
 ):
     # the sqlite destination connector doesn't leverage the file data but is required as an input,
     # mocking it with arbitrary values to meet the base requirements:
-    mock_file_data = FileData(identifier="mock file data", connector_type=CONNECTOR_TYPE)
+    mock_file_data = FileData(
+        identifier="mock file data",
+        connector_type=CONNECTOR_TYPE,
+        source_identifiers=SourceIdentifiers(filename=upload_file.name, fullpath=upload_file.name),
+    )
     stager = SQLiteUploadStager()
     staged_path = stager.run(
         elements_filepath=upload_file,

unstructured_ingest/__version__.py

@@ -1 +1 @@
-__version__ = "0.3.10"  # pragma: no cover
+__version__ = "0.3.11"  # pragma: no cover

unstructured_ingest/v2/interfaces/file_data.py

@@ -36,7 +36,7 @@ class FileDataSourceMetadata(BaseModel):
 class FileData(BaseModel):
     identifier: str
     connector_type: str
-    source_identifiers: Optional[SourceIdentifiers] = None
+    source_identifiers: SourceIdentifiers
     metadata: FileDataSourceMetadata = Field(default_factory=lambda: FileDataSourceMetadata())
     additional_metadata: dict[str, Any] = Field(default_factory=dict)
     reprocess: bool = False
@@ -73,6 +73,7 @@ class BatchItem(BaseModel):
 class BatchFileData(FileData):
     identifier: str = Field(init=False)
     batch_items: list[BatchItem]
+    source_identifiers: Optional[SourceIdentifiers] = None
 
     @field_validator("batch_items")
     @classmethod
@@ -104,3 +105,12 @@ def file_data_from_file(path: str) -> FileData:
         logger.debug(f"{path} not valid for batch file data")
 
     return FileData.from_file(path=path)
+
+
+def file_data_from_dict(data: dict) -> FileData:
+    try:
+        return BatchFileData.model_validate(data)
+    except ValidationError:
+        logger.debug(f"{data} not valid for batch file data")
+
+    return FileData.model_validate(data)

unstructured_ingest/v2/processes/connectors/astradb.py

@@ -30,6 +30,7 @@ from unstructured_ingest.v2.interfaces import (
     FileDataSourceMetadata,
     Indexer,
     IndexerConfig,
+    SourceIdentifiers,
     Uploader,
     UploaderConfig,
     UploadStager,
@@ -267,6 +268,7 @@ class AstraDBDownloader(Downloader):
             raise SourceConnectionNetworkError(f"failed to download file {file_data.identifier}")
 
         # modify input file_data for download_response
+        file_data.source_identifiers = SourceIdentifiers(filename=filename, fullpath=filename)
         cast_file_data = FileData.cast(file_data=file_data)
         cast_file_data.identifier = filename
         cast_file_data.metadata.date_processed = str(time())

unstructured_ingest/v2/processes/connectors/couchbase.py

@@ -27,6 +27,7 @@ from unstructured_ingest.v2.interfaces import (
     FileDataSourceMetadata,
     Indexer,
     IndexerConfig,
+    SourceIdentifiers,
     Uploader,
     UploaderConfig,
     UploadStager,
@@ -261,6 +262,7 @@ class CouchbaseDownloader(Downloader):
                 exc_info=True,
             )
             raise SourceConnectionNetworkError(f"failed to download file {file_data.identifier}")
+        file_data.source_identifiers = SourceIdentifiers(filename=filename, fullpath=filename)
         cast_file_data = FileData.cast(file_data=file_data)
         cast_file_data.identifier = filename_id
         cast_file_data.metadata.date_processed = str(time.time())

unstructured_ingest/v2/processes/connectors/databricks/volumes.py

@@ -14,6 +14,7 @@ from unstructured_ingest.error import (
 )
 from unstructured_ingest.utils.dep_check import requires_dependencies
 from unstructured_ingest.v2.interfaces import (
+    AccessConfig,
     ConnectionConfig,
     Downloader,
     DownloaderConfig,
@@ -52,6 +53,10 @@ class DatabricksPathMixin(BaseModel):
         return path
 
 
+class DatabricksVolumesAccessConfig(AccessConfig):
+    token: Optional[str] = Field(default=None, description="Databricks Personal Access Token")
+
+
 class DatabricksVolumesConnectionConfig(ConnectionConfig, ABC):
     host: Optional[str] = Field(
         default=None,

unstructured_ingest/v2/processes/connectors/databricks/volumes_aws.py

@@ -3,12 +3,12 @@ from typing import Optional
 
 from pydantic import Field, Secret
 
-from unstructured_ingest.v2.interfaces import AccessConfig
 from unstructured_ingest.v2.processes.connector_registry import (
     DestinationRegistryEntry,
     SourceRegistryEntry,
 )
 from unstructured_ingest.v2.processes.connectors.databricks.volumes import (
+    DatabricksVolumesAccessConfig,
     DatabricksVolumesConnectionConfig,
     DatabricksVolumesDownloader,
     DatabricksVolumesDownloaderConfig,
@@ -21,7 +21,7 @@ from unstructured_ingest.v2.processes.connectors.databricks.volumes import (
 CONNECTOR_TYPE = "databricks_volumes_aws"
 
 
-class DatabricksAWSVolumesAccessConfig(AccessConfig):
+class DatabricksAWSVolumesAccessConfig(DatabricksVolumesAccessConfig):
     account_id: Optional[str] = Field(
         default=None,
         description="The Databricks account ID for the Databricks " "accounts endpoint",

unstructured_ingest/v2/processes/connectors/databricks/volumes_azure.py

@@ -3,12 +3,12 @@ from typing import Optional
 
 from pydantic import Field, Secret
 
-from unstructured_ingest.v2.interfaces import AccessConfig
 from unstructured_ingest.v2.processes.connector_registry import (
     DestinationRegistryEntry,
     SourceRegistryEntry,
 )
 from unstructured_ingest.v2.processes.connectors.databricks.volumes import (
+    DatabricksVolumesAccessConfig,
     DatabricksVolumesConnectionConfig,
     DatabricksVolumesDownloader,
     DatabricksVolumesDownloaderConfig,
@@ -21,7 +21,7 @@ from unstructured_ingest.v2.processes.connectors.databricks.volumes import (
 CONNECTOR_TYPE = "databricks_volumes_azure"
 
 
-class DatabricksAzureVolumesAccessConfig(AccessConfig):
+class DatabricksAzureVolumesAccessConfig(DatabricksVolumesAccessConfig):
     account_id: Optional[str] = Field(
         default=None,
         description="The Databricks account ID for the Databricks " "accounts endpoint.",

unstructured_ingest/v2/processes/connectors/databricks/volumes_gcp.py

@@ -3,12 +3,12 @@ from typing import Optional
 
 from pydantic import Field, Secret
 
-from unstructured_ingest.v2.interfaces import AccessConfig
 from unstructured_ingest.v2.processes.connector_registry import (
     DestinationRegistryEntry,
     SourceRegistryEntry,
 )
 from unstructured_ingest.v2.processes.connectors.databricks.volumes import (
+    DatabricksVolumesAccessConfig,
     DatabricksVolumesConnectionConfig,
     DatabricksVolumesDownloader,
     DatabricksVolumesDownloaderConfig,
@@ -21,7 +21,7 @@ from unstructured_ingest.v2.processes.connectors.databricks.volumes import (
 CONNECTOR_TYPE = "databricks_volumes_gcp"
 
 
-class DatabricksGoogleVolumesAccessConfig(AccessConfig):
+class DatabricksGoogleVolumesAccessConfig(DatabricksVolumesAccessConfig):
     account_id: Optional[str] = Field(
         default=None,
         description="The Databricks account ID for the Databricks " "accounts endpoint.",

unstructured_ingest/v2/processes/connectors/databricks/volumes_native.py

@@ -3,12 +3,12 @@ from typing import Optional
 
 from pydantic import Field, Secret
 
-from unstructured_ingest.v2.interfaces import AccessConfig
 from unstructured_ingest.v2.processes.connector_registry import (
     DestinationRegistryEntry,
     SourceRegistryEntry,
 )
 from unstructured_ingest.v2.processes.connectors.databricks.volumes import (
+    DatabricksVolumesAccessConfig,
     DatabricksVolumesConnectionConfig,
     DatabricksVolumesDownloader,
     DatabricksVolumesDownloaderConfig,
@@ -21,7 +21,7 @@ from unstructured_ingest.v2.processes.connectors.databricks.volumes import (
 CONNECTOR_TYPE = "databricks_volumes"
 
 
-class DatabricksNativeVolumesAccessConfig(AccessConfig):
+class DatabricksNativeVolumesAccessConfig(DatabricksVolumesAccessConfig):
     client_id: Optional[str] = Field(default=None, description="Client ID of the OAuth app.")
     client_secret: Optional[str] = Field(
         default=None, description="Client Secret of the OAuth app."

unstructured_ingest/v2/processes/connectors/elasticsearch/elasticsearch.py

@@ -255,6 +255,7 @@ class ElasticsearchDownloader(Downloader):
                 exc_info=True,
             )
             raise SourceConnectionNetworkError(f"failed to download file {file_data.identifier}")
+        file_data.source_identifiers = SourceIdentifiers(filename=filename, fullpath=filename)
         cast_file_data = FileData.cast(file_data=file_data)
         cast_file_data.identifier = filename_id
         cast_file_data.metadata.date_processed = str(time())
@@ -264,7 +265,6 @@ class ElasticsearchDownloader(Downloader):
             "index_name": index_name,
             "document_id": record_id,
         }
-        cast_file_data.source_identifiers = SourceIdentifiers(filename=filename, fullpath=filename)
         return super().generate_download_response(
             file_data=cast_file_data,
             download_path=download_path,

unstructured_ingest/v2/processes/connectors/mongodb.py

@@ -198,14 +198,13 @@ class MongoDBDownloader(Downloader):
         concatenated_values = "\n".join(str(value) for value in flattened_dict.values())
 
         # Create a FileData object for each document with source_identifiers
-        cast_file_data = FileData.cast(file_data=file_data)
-        cast_file_data.identifier = str(doc_id)
         filename = f"{doc_id}.txt"
-        cast_file_data.source_identifiers = SourceIdentifiers(
+        file_data.source_identifiers = SourceIdentifiers(
             filename=filename,
             fullpath=filename,
-            rel_path=filename,
         )
+        cast_file_data = FileData.cast(file_data=file_data)
+        cast_file_data.identifier = str(doc_id)
 
         # Determine the download path
         download_path = self.get_download_path(file_data=cast_file_data)

unstructured_ingest/v2/processes/connectors/sql/sql.py

@@ -28,6 +28,7 @@ from unstructured_ingest.v2.interfaces import (
     FileDataSourceMetadata,
     Indexer,
     IndexerConfig,
+    SourceIdentifiers,
     Uploader,
     UploaderConfig,
     UploadStager,
@@ -218,6 +219,10 @@ class SQLDownloader(Downloader, ABC):
         )
         download_path.parent.mkdir(parents=True, exist_ok=True)
         result.to_csv(download_path, index=False)
+        file_data.source_identifiers = SourceIdentifiers(
+            filename=filename,
+            fullpath=filename,
+        )
         cast_file_data = FileData.cast(file_data=file_data)
         cast_file_data.identifier = filename_id
         return super().generate_download_response(

{unstructured_ingest-0.3.10.dist-info → unstructured_ingest-0.3.11.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: unstructured-ingest
-Version: 0.3.10
+Version: 0.3.11
 Summary: A library that prepares raw documents for downstream ML tasks.
 Home-page: https://github.com/Unstructured-IO/unstructured-ingest
 Author: Unstructured Technologies
@@ -22,14 +22,14 @@ Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
 Requires-Python: >=3.9.0,<3.13
 Description-Content-Type: text/markdown
 License-File: LICENSE.md
+Requires-Dist: ndjson
+Requires-Dist: pandas
+Requires-Dist: dataclasses-json
+Requires-Dist: pydantic>=2.7
 Requires-Dist: opentelemetry-sdk
 Requires-Dist: python-dateutil
 Requires-Dist: click
-Requires-Dist: ndjson
-Requires-Dist: pydantic>=2.7
-Requires-Dist: pandas
 Requires-Dist: tqdm
-Requires-Dist: dataclasses-json
 Provides-Extra: airtable
 Requires-Dist: pyairtable; extra == "airtable"
 Provides-Extra: astradb
@@ -42,8 +42,8 @@ Requires-Dist: azure-search-documents; extra == "azure-ai-search"
 Provides-Extra: bedrock
 Requires-Dist: boto3; extra == "bedrock"
 Provides-Extra: biomed
-Requires-Dist: bs4; extra == "biomed"
 Requires-Dist: requests; extra == "biomed"
+Requires-Dist: bs4; extra == "biomed"
 Provides-Extra: box
 Requires-Dist: fsspec; extra == "box"
 Requires-Dist: boxfs; extra == "box"
@@ -61,8 +61,8 @@ Requires-Dist: unstructured[tsv]; extra == "csv"
 Provides-Extra: databricks-volumes
 Requires-Dist: databricks-sdk; extra == "databricks-volumes"
 Provides-Extra: delta-table
-Requires-Dist: deltalake; extra == "delta-table"
 Requires-Dist: boto3; extra == "delta-table"
+Requires-Dist: deltalake; extra == "delta-table"
 Provides-Extra: discord
 Requires-Dist: discord-py; extra == "discord"
 Provides-Extra: doc
@@ -70,8 +70,8 @@ Requires-Dist: unstructured[docx]; extra == "doc"
 Provides-Extra: docx
 Requires-Dist: unstructured[docx]; extra == "docx"
 Provides-Extra: dropbox
-Requires-Dist: dropboxdrivefs; extra == "dropbox"
 Requires-Dist: fsspec; extra == "dropbox"
+Requires-Dist: dropboxdrivefs; extra == "dropbox"
 Provides-Extra: duckdb
 Requires-Dist: duckdb; extra == "duckdb"
 Provides-Extra: elasticsearch
@@ -90,19 +90,19 @@ Requires-Dist: voyageai; extra == "embed-voyageai"
 Provides-Extra: epub
 Requires-Dist: unstructured[epub]; extra == "epub"
 Provides-Extra: gcs
-Requires-Dist: bs4; extra == "gcs"
 Requires-Dist: fsspec; extra == "gcs"
 Requires-Dist: gcsfs; extra == "gcs"
+Requires-Dist: bs4; extra == "gcs"
 Provides-Extra: github
-Requires-Dist: pygithub>1.58.0; extra == "github"
 Requires-Dist: requests; extra == "github"
+Requires-Dist: pygithub>1.58.0; extra == "github"
 Provides-Extra: gitlab
 Requires-Dist: python-gitlab; extra == "gitlab"
 Provides-Extra: google-drive
 Requires-Dist: google-api-python-client; extra == "google-drive"
 Provides-Extra: hubspot
-Requires-Dist: hubspot-api-client; extra == "hubspot"
 Requires-Dist: urllib3; extra == "hubspot"
+Requires-Dist: hubspot-api-client; extra == "hubspot"
 Provides-Extra: jira
 Requires-Dist: atlassian-python-api; extra == "jira"
 Provides-Extra: kafka
@@ -120,20 +120,20 @@ Requires-Dist: pymongo; extra == "mongodb"
 Provides-Extra: msg
 Requires-Dist: unstructured[msg]; extra == "msg"
 Provides-Extra: neo4j
-Requires-Dist: cymple; extra == "neo4j"
 Requires-Dist: neo4j; extra == "neo4j"
+Requires-Dist: cymple; extra == "neo4j"
 Requires-Dist: networkx; extra == "neo4j"
 Provides-Extra: notion
-Requires-Dist: htmlBuilder; extra == "notion"
 Requires-Dist: backoff; extra == "notion"
 Requires-Dist: notion-client; extra == "notion"
 Requires-Dist: httpx; extra == "notion"
+Requires-Dist: htmlBuilder; extra == "notion"
 Provides-Extra: odt
 Requires-Dist: unstructured[odt]; extra == "odt"
 Provides-Extra: onedrive
 Requires-Dist: Office365-REST-Python-Client; extra == "onedrive"
-Requires-Dist: bs4; extra == "onedrive"
 Requires-Dist: msal; extra == "onedrive"
+Requires-Dist: bs4; extra == "onedrive"
 Provides-Extra: openai
 Requires-Dist: openai; extra == "openai"
 Requires-Dist: tiktoken; extra == "openai"
@@ -165,8 +165,8 @@ Requires-Dist: unstructured[rst]; extra == "rst"
 Provides-Extra: rtf
 Requires-Dist: unstructured[rtf]; extra == "rtf"
 Provides-Extra: s3
-Requires-Dist: s3fs; extra == "s3"
 Requires-Dist: fsspec; extra == "s3"
+Requires-Dist: s3fs; extra == "s3"
 Provides-Extra: salesforce
 Requires-Dist: simple-salesforce; extra == "salesforce"
 Provides-Extra: sftp

{unstructured_ingest-0.3.10.dist-info → unstructured_ingest-0.3.11.dist-info}/RECORD

@@ -19,8 +19,8 @@ test/integration/connectors/test_onedrive.py,sha256=KIkBwKh1hnv203VCL2UABnDkS_bP
 test/integration/connectors/test_pinecone.py,sha256=suPFi40d6rHXurQQLIpCzW5XRTdgzlP-f-KLPhGCUHo,10208
 test/integration/connectors/test_qdrant.py,sha256=hyuqSJDaylkQVxWh7byD8jo8bwPuBxSa8MWRD3sBu-Y,7906
 test/integration/connectors/test_s3.py,sha256=PJaAwFRF2lXMQlkbv9JHpngPc6706ML7zowOlXT3TcY,7033
-test/integration/connectors/databricks_tests/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-test/integration/connectors/databricks_tests/test_volumes_native.py,sha256=9Ndo0Q8uzBGGOYYjmTV6EdZtaTRy97lHAXvJczTOJe4,5859
+test/integration/connectors/databricks/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+test/integration/connectors/databricks/test_volumes_native.py,sha256=ig60-nCdLF0GsgJowG9eRaG28iuoYHtuf12HdK6OE1I,7764
 test/integration/connectors/duckdb/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 test/integration/connectors/duckdb/conftest.py,sha256=rlBHMJTiJ2a5xbvIxTOyhhcuTBc9DO-yTzD6Kf8X3hY,301
 test/integration/connectors/duckdb/test_duckdb.py,sha256=tZfHJYNILVqwT20XD-aJUFZ67TnJvHLpfAxNvNiE51o,2891
@@ -30,10 +30,10 @@ test/integration/connectors/elasticsearch/conftest.py,sha256=-i4_7MkIxSQENz7nuD2
 test/integration/connectors/elasticsearch/test_elasticsearch.py,sha256=Lm8topVssTsqcI8H2Tzohuxb9j-CFHv9orM6WfAqCZw,11933
 test/integration/connectors/elasticsearch/test_opensearch.py,sha256=fWpZrhzRiVpm9AOlZvgZRCjyXSYvWG7-8j06x-HR3PY,11311
 test/integration/connectors/sql/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-test/integration/connectors/sql/test_postgres.py,sha256=BNoH1Ha3hdpxk--DqE5cg8zwnx14rN-1LzQ_YJ1U5us,6734
-test/integration/connectors/sql/test_singlestore.py,sha256=W4bNc7HoVrtfHfDTdQF8rcwKdX-AdIDHV-phJeViSmU,5917
-test/integration/connectors/sql/test_snowflake.py,sha256=-fzpaB1LlFOW2bZmJHQMbH6T0zU30dMypAHLRAbT4d0,7258
-test/integration/connectors/sql/test_sqlite.py,sha256=NjIcAvujSiuqjTaitjjyLEsS3oazPyQzTnzCZ6QjIkM,5720
+test/integration/connectors/sql/test_postgres.py,sha256=DXyHMZBQgrV2HyVflkoBpT1mewSnvw3ugoHtGR5o8OM,6876
+test/integration/connectors/sql/test_singlestore.py,sha256=pzCPo8IW3c9VH-f3UdJS5MjPjkHarJPSepAxV0ZVajo,6059
+test/integration/connectors/sql/test_snowflake.py,sha256=MiTzepeeJlv147CyzCGyd16MRk5QeUw4g4L3TTi5gVY,7400
+test/integration/connectors/sql/test_sqlite.py,sha256=rSkjv3KpslAvt_8LQecJUT0lOLtuZSvhtlW2deJovLI,5862
 test/integration/connectors/utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 test/integration/connectors/utils/constants.py,sha256=0zSPnsZVqJuNhXduXvdXFQLZTRIQa5Fo_1qjBYVCfb8,209
 test/integration/connectors/utils/docker.py,sha256=8uOTJ3AVG1dxK4OiLvOLfRxL_TsYQX2KKCID9TZ7-Ac,4995
@@ -94,7 +94,7 @@ test/unit/v2/partitioners/test_partitioner.py,sha256=iIYg7IpftV3LusoO4H8tr1IHY1U
 test/unit/v2/utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 test/unit/v2/utils/data_generator.py,sha256=UoYVNjG4S4wlaA9gceQ82HIpF9_6I1UTHD1_GrQBHp0,973
 unstructured_ingest/__init__.py,sha256=U4S_2y3zgLZVfMenHRaJFBW8yqh2mUBuI291LGQVOJ8,35
-unstructured_ingest/__version__.py,sha256=agVWTd6LRRiz8p3CAvwS6-Qb_jT60tZHXAK83ewMwL4,43
+unstructured_ingest/__version__.py,sha256=CJalz6YpEm8DAhzCP5dryU5ddzKQaSQOzTObKxfOVHs,43
 unstructured_ingest/error.py,sha256=qDncnJgbf5ils956RcO2CGlAKYDT5OaEM9Clv1JVTNc,1448
 unstructured_ingest/interfaces.py,sha256=OYVUP0bzBJpT-Lz92BDyz_hLBvyfxkuSwWHhUdnUayA,31493
 unstructured_ingest/logger.py,sha256=S5nSqGcABoQyeicgRnBQFjDScCaTvFVivOCvbo-laL0,4479
@@ -379,7 +379,7 @@ unstructured_ingest/v2/cli/utils/model_conversion.py,sha256=7eEIkk1KU51-ZNiIfI1K
 unstructured_ingest/v2/interfaces/__init__.py,sha256=9VO09XuTvyOcFF8ZDKN169fNb_uA5TAYzPsiPHOyxhQ,963
 unstructured_ingest/v2/interfaces/connector.py,sha256=qUFFJ3qgDMenTCZMtVRjq1DIwsVak6pxNjQOH2eVkMw,1623
 unstructured_ingest/v2/interfaces/downloader.py,sha256=Lj3nTY1hPA71GfNeedFVCdHdZsHLle8qrx5RtXAy9GY,2940
-unstructured_ingest/v2/interfaces/file_data.py,sha256=weeCEp14QGEGjCBtJBByrAcAqA87Jx62qw3Ed7x-crU,3550
+unstructured_ingest/v2/interfaces/file_data.py,sha256=7MyRlj5dijQsCR6W18wQ8fEgJigGKwoOYc10g9A6PSo,3834
 unstructured_ingest/v2/interfaces/indexer.py,sha256=gsa1MLhFa82BzD2h4Yb7ons0VxRwKINZOrzvHAahwVU,846
 unstructured_ingest/v2/interfaces/process.py,sha256=BgglTu5K93FnDDopZKKr_rkK2LTZOguR6kcQjKHjF40,392
 unstructured_ingest/v2/interfaces/processor.py,sha256=VX7JqXlbG1plxMK8THWhWINPbTICaaUEk4XUXhnOixY,3303
@@ -408,18 +408,18 @@ unstructured_ingest/v2/processes/partitioner.py,sha256=agpHwB9FR8OZVQqE7zFEb0IcD
 unstructured_ingest/v2/processes/uncompress.py,sha256=Z_XfsITGdyaRwhtNUc7bMj5Y2jLuBge8KoK4nxhqKag,2425
 unstructured_ingest/v2/processes/connectors/__init__.py,sha256=CTWLEmaKLTjbqeUQGI0fxJobsqDOc1d2ZKJoXh98Lww,5432
 unstructured_ingest/v2/processes/connectors/airtable.py,sha256=eeZJe-bBNxt5Sa-XEFCdcGeJCguJU5WN2Mv9kLp5dVQ,8917
-unstructured_ingest/v2/processes/connectors/astradb.py,sha256=Y4rgEZZRMOLCVeC5jHbPWxnVtl1A-fPeU8BEibMZIlQ,14696
+unstructured_ingest/v2/processes/connectors/astradb.py,sha256=xhUMoUdnrfAY1isZGqsV4lZUsnZNpbvgLyQWQbR4hVo,14814
 unstructured_ingest/v2/processes/connectors/azure_ai_search.py,sha256=ngPDpU0oZ6m5sxIlB6u5ebQpqCS_SJ-_amCC1KQ03EQ,11529
 unstructured_ingest/v2/processes/connectors/chroma.py,sha256=G1DQHhhFQCS2RLF0cVvoUH9QO8KkVjIyNZ9nKh__aHw,7220
 unstructured_ingest/v2/processes/connectors/confluence.py,sha256=-Y1OU_ZXhZQNj5NH3EN01CP8QKKZJaJ9xkXoAlSgnIk,7604
-unstructured_ingest/v2/processes/connectors/couchbase.py,sha256=5Da-JWHXHcVcmJ_rqj4hZwURh_z7sJRcYpyx83Bc7zM,12132
+unstructured_ingest/v2/processes/connectors/couchbase.py,sha256=i7vuNKsUkN93JRVmg4--MO0ZgbjvhIqt46oYqk9zFSQ,12250
 unstructured_ingest/v2/processes/connectors/delta_table.py,sha256=SotSXZQ85_6TO906YvFi3yTml8jE9A_zV6nBJ4oTx8A,7075
 unstructured_ingest/v2/processes/connectors/gitlab.py,sha256=ufE65Z8q_tC4oppGg5BsGXwSaL7RbEXcaagJQYsylNo,9984
 unstructured_ingest/v2/processes/connectors/google_drive.py,sha256=5k7pdAzJGXSdyPCzW9vu2OaAjGVTo2JevDyGaXM1Hvk,13370
 unstructured_ingest/v2/processes/connectors/kdbai.py,sha256=VRDAiou_7oWOIAgQTdOGQWxudzQEDopXM8XkfkQ2j6g,5004
 unstructured_ingest/v2/processes/connectors/local.py,sha256=ZvWTj6ZYkwnvQMNFsZWoaQyp9zp0WVqAywMaHJ2kcAc,7153
 unstructured_ingest/v2/processes/connectors/milvus.py,sha256=I57hyH5nz_p7utmUOkvt_6vCPxNIVQMoukplUgIyYi8,8503
-unstructured_ingest/v2/processes/connectors/mongodb.py,sha256=yzcsHTJXVyuASeLr56OSratDQRiw5v68JSHPYpMEzzY,14309
+unstructured_ingest/v2/processes/connectors/mongodb.py,sha256=cL0QUQZF_s2brh3nNNeAywXVpaIiND4b5JTAFlYjLjw,14273
 unstructured_ingest/v2/processes/connectors/neo4j.py,sha256=lRtWe6xWYogT-y_r_o7HWvlFMf_OIPGQq_Z-5v7IOq0,14163
 unstructured_ingest/v2/processes/connectors/onedrive.py,sha256=H8qk53YJXAPrPyISze0dybZdDFv5B7dVO3fIr10dVU8,15982
 unstructured_ingest/v2/processes/connectors/outlook.py,sha256=KgNGM8hImRhy6_SpswRP2VwRD4VOrqqJoySgxf2oduI,9290
@@ -429,17 +429,17 @@ unstructured_ingest/v2/processes/connectors/sharepoint.py,sha256=Ndn2Wm7RupfjAtl
 unstructured_ingest/v2/processes/connectors/slack.py,sha256=Z73VmQ3oUY09KoLEi5OBdQeDt4ONEY_02SglWQc6HXE,9252
 unstructured_ingest/v2/processes/connectors/utils.py,sha256=8kd0g7lo9NqnpaIkjeO-Ut6erhwUNH_gS9koevpe3WE,878
 unstructured_ingest/v2/processes/connectors/databricks/__init__.py,sha256=jO71UTC7bLA_N12CrLWJzh_yZML5gfT7VohxzCpUGWg,1848
-unstructured_ingest/v2/processes/connectors/databricks/volumes.py,sha256=p7sjCYZb7JmY3v3Xy1gm-q0O7oamLTsSFf2EWXYfXYQ,6447
-unstructured_ingest/v2/processes/connectors/databricks/volumes_aws.py,sha256=I1MJwe5LOxoPLjwo00H0XbXO6u_SJHWYgsj4s6ePoyI,2754
-unstructured_ingest/v2/processes/connectors/databricks/volumes_azure.py,sha256=P4rfcE3td7WyuuguRgUnGQytCMDpfeYrrpshBZuVynY,3539
-unstructured_ingest/v2/processes/connectors/databricks/volumes_gcp.py,sha256=UUotY_-HpgSEJkvdQfZTlbxY7CRLZ4ctL8TlryeFvxk,2790
-unstructured_ingest/v2/processes/connectors/databricks/volumes_native.py,sha256=Wk7s2_u5G0BOV5slvGc8IlUf7ivznY9PrgPqe6nlJKM,2897
+unstructured_ingest/v2/processes/connectors/databricks/volumes.py,sha256=kI_ThB5e-DS8-GiQP5TQ8cP3fiGRm-V2AuNlGoSjH6I,6613
+unstructured_ingest/v2/processes/connectors/databricks/volumes_aws.py,sha256=TA2e_1SIr4VaEI62873eyReCNfgmQ51_2Pko2I04pPM,2747
+unstructured_ingest/v2/processes/connectors/databricks/volumes_azure.py,sha256=cb-EUW0T-linZMkbU6AcKEGWnFHQvhpO5Abtps4P2X0,3532
+unstructured_ingest/v2/processes/connectors/databricks/volumes_gcp.py,sha256=tR8NubkyHw49IpW_42g6w1Koxlm56EPiPf1lB-eoRSI,2783
+unstructured_ingest/v2/processes/connectors/databricks/volumes_native.py,sha256=dJLD1fueXf8_0AfC4cg0G7siJZVefz68iuEx2Kq7rMs,2890
 unstructured_ingest/v2/processes/connectors/duckdb/__init__.py,sha256=5sVvJCWhU-YkjHIwk4W6BZCanFYK5W4xTpWtQ8xzeB4,561
 unstructured_ingest/v2/processes/connectors/duckdb/base.py,sha256=FVblIddorGCh9D9GZ8zLVUm8n39PJA5JLoJeWd-tSy8,2610
 unstructured_ingest/v2/processes/connectors/duckdb/duckdb.py,sha256=oUHHaLpO2pWW2Lu4Mc-XFjrA0ze97205WQ_xP95ua4M,4296
 unstructured_ingest/v2/processes/connectors/duckdb/motherduck.py,sha256=mU5x6SnbFgRsVicNGh4y4gtR6ek7eQFinI0dQQmzMds,4481
 unstructured_ingest/v2/processes/connectors/elasticsearch/__init__.py,sha256=Zzc0JNPP-eFqpwWw1Gp-XC8H-s__IgkYKzoagECycZY,829
-unstructured_ingest/v2/processes/connectors/elasticsearch/elasticsearch.py,sha256=Ws6NoHvbXkxizbcvr1bKGkMMB7l_e-sfkQQGU6u7e0Y,18818
+unstructured_ingest/v2/processes/connectors/elasticsearch/elasticsearch.py,sha256=MEKU64OsiQmbLPb3ken-WWCIV6-pnFbs_6kjJweG-SY,18813
 unstructured_ingest/v2/processes/connectors/elasticsearch/opensearch.py,sha256=qRz8Fyr2RSZIPZGkhPeme6AZxM0aX-c_xOa1ZtSr2Kg,6781
 unstructured_ingest/v2/processes/connectors/fsspec/__init__.py,sha256=TtdeImM7Ypl_n6sl7I1JqX6bGSG0t_FqvCqE3Cy24og,1846
 unstructured_ingest/v2/processes/connectors/fsspec/azure.py,sha256=NWsouhaxeyxrS_WlZJ70X2YIdioFH5LSaRLhnCPYAH0,6034
@@ -470,16 +470,16 @@ unstructured_ingest/v2/processes/connectors/sql/__init__.py,sha256=E16CXRBw8fZKT
 unstructured_ingest/v2/processes/connectors/sql/postgres.py,sha256=BATfX1PQGT2kl8jAbdNKXTojYKJxh3pJV9-h3OBnHGo,5124
 unstructured_ingest/v2/processes/connectors/sql/singlestore.py,sha256=-2E9dsdNhjAiuzeSBytBbAhljOhvQ8kN8wvlUESvLo8,5465
 unstructured_ingest/v2/processes/connectors/sql/snowflake.py,sha256=TApGi5G0W0TktJFmo4QWDR3X3R-MUQTKbIxjAX_M8ZI,7402
-unstructured_ingest/v2/processes/connectors/sql/sql.py,sha256=LlhdBL_yPsPwNzDPwAblUDLBOsvorD6bB3OUrxPk_PQ,15112
+unstructured_ingest/v2/processes/connectors/sql/sql.py,sha256=HC-qnhdpuyScKoGh50pPjkQLGSac_mOAnuB2FZwSVl0,15265
 unstructured_ingest/v2/processes/connectors/sql/sqlite.py,sha256=Q5RAqn5Ccw-pbeKZLkiMn5IVw6EemCMukXzLlS7pDhc,5162
 unstructured_ingest/v2/processes/connectors/weaviate/__init__.py,sha256=NMiwnVWan69KnzVELvaqX34tMhCytIa-C8EDsXVKsEo,856
 unstructured_ingest/v2/processes/connectors/weaviate/cloud.py,sha256=bXtfEYLquR-BszZ5S_lQ4JbETNs9Vozgpfm8x9egAmE,6251
 unstructured_ingest/v2/processes/connectors/weaviate/embedded.py,sha256=S8Zg8StuZT-k7tCg1D5YShO1-vJYYk9-M1bE1fIqx64,3014
 unstructured_ingest/v2/processes/connectors/weaviate/local.py,sha256=LuTBKPseVewsz8VqxRPRLfGEm3BeI9nBZxpy7ZU5tOA,2201
 unstructured_ingest/v2/processes/connectors/weaviate/weaviate.py,sha256=X1yv1H_orDQ-J965EMXhR2XaURqe8vovSi9n1fk85B4,10499
-unstructured_ingest-0.3.10.dist-info/LICENSE.md,sha256=SxkKP_62uIAKb9mb1eH7FH4Kn2aYT09fgjKpJt5PyTk,11360
-unstructured_ingest-0.3.10.dist-info/METADATA,sha256=emdZRN3DNrVAwB6fUYQonLRpymZzh4ktHT73x2_vdxg,7623
-unstructured_ingest-0.3.10.dist-info/WHEEL,sha256=GV9aMThwP_4oNCtvEC2ec3qUYutgWeAzklro_0m4WJQ,91
-unstructured_ingest-0.3.10.dist-info/entry_points.txt,sha256=gUAAFnjFPnBgThJSEbw0N5ZjxtaKlT1s9e05_arQrNw,70
-unstructured_ingest-0.3.10.dist-info/top_level.txt,sha256=DMuDMHZRMdeay8v8Kdi855muIv92F0OkutvBCaBEW6M,25
-unstructured_ingest-0.3.10.dist-info/RECORD,,
+unstructured_ingest-0.3.11.dist-info/LICENSE.md,sha256=SxkKP_62uIAKb9mb1eH7FH4Kn2aYT09fgjKpJt5PyTk,11360
+unstructured_ingest-0.3.11.dist-info/METADATA,sha256=4-py_Sf-ahdzXF0l1evY4aI_s5Vz4oc6Gtenhegc6Vo,7623
+unstructured_ingest-0.3.11.dist-info/WHEEL,sha256=GV9aMThwP_4oNCtvEC2ec3qUYutgWeAzklro_0m4WJQ,91
+unstructured_ingest-0.3.11.dist-info/entry_points.txt,sha256=gUAAFnjFPnBgThJSEbw0N5ZjxtaKlT1s9e05_arQrNw,70
+unstructured_ingest-0.3.11.dist-info/top_level.txt,sha256=DMuDMHZRMdeay8v8Kdi855muIv92F0OkutvBCaBEW6M,25
+unstructured_ingest-0.3.11.dist-info/RECORD,,