unstructured-ingest 0.4.7__py3-none-any.whl → 0.5.0__py3-none-any.whl

test/integration/connectors/test_onedrive.py

@@ -5,13 +5,25 @@ from pathlib import Path
 import pytest
 from office365.graph_client import GraphClient
 
-from test.integration.connectors.utils.constants import BLOB_STORAGE_TAG, DESTINATION_TAG
+from test.integration.connectors.utils.constants import (
+    BLOB_STORAGE_TAG,
+    DESTINATION_TAG,
+    SOURCE_TAG,
+)
+from test.integration.connectors.utils.validation.source import (
+    SourceValidationConfigs,
+    source_connector_validation,
+)
 from test.integration.utils import requires_env
 from unstructured_ingest.v2.interfaces import FileData, SourceIdentifiers
 from unstructured_ingest.v2.processes.connectors.onedrive import (
     CONNECTOR_TYPE,
     OnedriveAccessConfig,
     OnedriveConnectionConfig,
+    OnedriveDownloader,
+    OnedriveDownloaderConfig,
+    OnedriveIndexer,
+    OnedriveIndexerConfig,
     OnedriveUploader,
     OnedriveUploaderConfig,
 )
@@ -62,9 +74,46 @@ def get_connection_config():
     return connection_config
 
 
+@pytest.mark.asyncio
+@pytest.mark.tags(CONNECTOR_TYPE, SOURCE_TAG, BLOB_STORAGE_TAG)
+@requires_env("MS_CLIENT_CRED", "MS_CLIENT_ID", "MS_TENANT_ID", "MS_USER_PNAME")
+async def test_onedrive_source(temp_dir):
+    connection_config = get_connection_config()
+    index_config = OnedriveIndexerConfig(recursive=True, path="eml")
+
+    download_config = OnedriveDownloaderConfig(download_dir=temp_dir)
+
+    # Instantiate indexer and downloader
+    indexer = OnedriveIndexer(
+        connection_config=connection_config,
+        index_config=index_config,
+    )
+    downloader = OnedriveDownloader(
+        connection_config=connection_config,
+        download_config=download_config,
+    )
+
+    # Run the source connector validation
+    await source_connector_validation(
+        indexer=indexer,
+        downloader=downloader,
+        configs=SourceValidationConfigs(
+            test_id="onedrive",
+            expected_num_files=1,
+            validate_downloaded_files=True,
+            exclude_fields_extend=[
+                "metadata.date_created",
+                "metadata.date_modified",
+                "additional_metadata.LastModified",
+                "additional_metadata.@microsoft.graph.downloadUrl",
+            ],
+        ),
+    )
+
+
 @pytest.mark.tags(CONNECTOR_TYPE, DESTINATION_TAG, BLOB_STORAGE_TAG)
 @requires_env("MS_CLIENT_CRED", "MS_CLIENT_ID", "MS_TENANT_ID", "MS_USER_PNAME")
-def test_onedrive_destination(upload_file: Path, onedrive_test_folder: str):
+def xtest_onedrive_destination(upload_file: Path, onedrive_test_folder: str):
     """
     Integration test for the OneDrive destination connector.
 

test/integration/connectors/test_sharepoint.py

@@ -0,0 +1,71 @@
+import os
+
+import pytest
+
+from test.integration.connectors.utils.constants import BLOB_STORAGE_TAG, SOURCE_TAG
+from test.integration.connectors.utils.validation.source import (
+    SourceValidationConfigs,
+    source_connector_validation,
+)
+from test.integration.utils import requires_env
+from unstructured_ingest.v2.processes.connectors.sharepoint import (
+    CONNECTOR_TYPE,
+    SharepointAccessConfig,
+    SharepointConnectionConfig,
+    SharepointDownloader,
+    SharepointDownloaderConfig,
+    SharepointIndexer,
+    SharepointIndexerConfig,
+)
+
+
+@pytest.mark.asyncio
+@pytest.mark.tags(CONNECTOR_TYPE, SOURCE_TAG, BLOB_STORAGE_TAG)
+@requires_env("SHAREPOINT_CLIENT_ID", "SHAREPOINT_CRED", "MS_TENANT_ID", "MS_USER_PNAME")
+async def test_sharepoint_source(temp_dir):
+    # Retrieve environment variables
+    site = "https://unstructuredio.sharepoint.com/sites/utic-platform-test-source"
+    client_id = os.environ["SHAREPOINT_CLIENT_ID"]
+    client_cred = os.environ["SHAREPOINT_CRED"]
+    user_pname = os.environ["MS_USER_PNAME"]
+    tenant = os.environ["MS_TENANT_ID"]
+
+    # Create connection and indexer configurations
+    access_config = SharepointAccessConfig(client_cred=client_cred)
+    connection_config = SharepointConnectionConfig(
+        client_id=client_id,
+        site=site,
+        tenant=tenant,
+        user_pname=user_pname,
+        access_config=access_config,
+    )
+    index_config = SharepointIndexerConfig(recursive=True)
+
+    download_config = SharepointDownloaderConfig(download_dir=temp_dir)
+
+    # Instantiate indexer and downloader
+    indexer = SharepointIndexer(
+        connection_config=connection_config,
+        index_config=index_config,
+    )
+    downloader = SharepointDownloader(
+        connection_config=connection_config,
+        download_config=download_config,
+    )
+
+    # Run the source connector validation
+    await source_connector_validation(
+        indexer=indexer,
+        downloader=downloader,
+        configs=SourceValidationConfigs(
+            test_id="sharepoint",
+            expected_num_files=4,
+            validate_downloaded_files=True,
+            exclude_fields_extend=[
+                "metadata.date_created",
+                "metadata.date_modified",
+                "additional_metadata.LastModified",
+                "additional_metadata.@microsoft.graph.downloadUrl",
+            ],
+        ),
+    )

test/integration/connectors/utils/validation/source.py

@@ -10,6 +10,13 @@ from pydantic import Field
 from test.integration.connectors.utils.validation.utils import ValidationConfig
 from unstructured_ingest.v2.interfaces import Downloader, FileData, Indexer
 
+NONSTANDARD_METADATA_FIELDS = {
+    "additional_metadata.@microsoft.graph.downloadUrl": [
+        "additional_metadata",
+        "@microsoft.graph.downloadUrl",
+    ]
+}
+
 
 class SourceValidationConfigs(ValidationConfig):
     expected_number_indexed_file_data: Optional[int] = None
@@ -26,7 +33,7 @@ class SourceValidationConfigs(ValidationConfig):
     def get_exclude_fields(self) -> list[str]:
         exclude_fields = self.exclude_fields
         exclude_fields.extend(self.exclude_fields_extend)
-        return exclude_fields
+        return list(set(exclude_fields))
 
     def run_file_data_validation(
         self, predownload_file_data: FileData, postdownload_file_data: FileData
@@ -45,8 +52,13 @@ class SourceValidationConfigs(ValidationConfig):
         exclude_fields = self.get_exclude_fields()
         # Ignore fields that dynamically change every time the tests run
         copied_data = data.copy()
+
         for exclude_field in exclude_fields:
-            exclude_field_vals = exclude_field.split(".")
+            exclude_field_vals = (
+                NONSTANDARD_METADATA_FIELDS[exclude_field]
+                if exclude_field in NONSTANDARD_METADATA_FIELDS
+                else exclude_field.split(".")
+            )
             if len(exclude_field_vals) == 1:
                 current_val = copied_data
                 drop_field = exclude_field_vals[0]
@@ -261,21 +273,38 @@ async def source_connector_validation(
     indexer.precheck()
     download_dir = downloader.download_config.download_dir
     test_output_dir = configs.test_output_dir()
-    for file_data in indexer.run():
-        assert file_data
-        predownload_file_data = file_data.model_copy(deep=True)
-        all_predownload_file_data.append(predownload_file_data)
-        if downloader.is_async():
-            resp = await downloader.run_async(file_data=file_data)
-        else:
-            resp = downloader.run(file_data=file_data)
-        if isinstance(resp, list):
-            for r in resp:
-                postdownload_file_data = r["file_data"].model_copy(deep=True)
+    if indexer.is_async():
+        async for file_data in indexer.run_async():
+            assert file_data
+            predownload_file_data = file_data.model_copy(deep=True)
+            all_predownload_file_data.append(predownload_file_data)
+            if downloader.is_async():
+                resp = await downloader.run_async(file_data=file_data)
+            else:
+                resp = downloader.run(file_data=file_data)
+            if isinstance(resp, list):
+                for r in resp:
+                    postdownload_file_data = r["file_data"].model_copy(deep=True)
+                    all_postdownload_file_data.append(postdownload_file_data)
+            else:
+                postdownload_file_data = resp["file_data"].model_copy(deep=True)
+                all_postdownload_file_data.append(postdownload_file_data)
+    else:
+        for file_data in indexer.run():
+            assert file_data
+            predownload_file_data = file_data.model_copy(deep=True)
+            all_predownload_file_data.append(predownload_file_data)
+            if downloader.is_async():
+                resp = await downloader.run_async(file_data=file_data)
+            else:
+                resp = downloader.run(file_data=file_data)
+            if isinstance(resp, list):
+                for r in resp:
+                    postdownload_file_data = r["file_data"].model_copy(deep=True)
+                    all_postdownload_file_data.append(postdownload_file_data)
+            else:
+                postdownload_file_data = resp["file_data"].model_copy(deep=True)
                 all_postdownload_file_data.append(postdownload_file_data)
-        else:
-            postdownload_file_data = resp["file_data"].model_copy(deep=True)
-            all_postdownload_file_data.append(postdownload_file_data)
     if not overwrite_fixtures:
         print("Running validation")
         run_all_validations(

test/integration/embedders/test_bedrock.py

@@ -31,7 +31,7 @@ def get_aws_credentials() -> dict:
 def test_bedrock_embedder(embedder_file: Path):
     aws_credentials = get_aws_credentials()
     embedder_config = EmbedderConfig(
-        embedding_provider="aws-bedrock",
+        embedding_provider="bedrock",
         embedding_aws_access_key_id=aws_credentials["aws_access_key_id"],
         embedding_aws_secret_access_key=aws_credentials["aws_secret_access_key"],
     )

test/integration/partitioners/test_partitioner.py

@@ -1,4 +1,3 @@
-import json
 import os
 from pathlib import Path
 
@@ -15,6 +14,9 @@ all_partition_files = [path for path in assets_dir.iterdir() if path.is_file()]
 non_image_partition_files = [
     path for path in all_partition_files if path.suffix not in [".jpg", ".png", ".tif"]
 ]
+supported_fast_partition_files = [
+    path for path in non_image_partition_files if path.suffix != ".eml"
+]
 image_partition_files = [
     path for path in all_partition_files if path not in non_image_partition_files
 ]
@@ -33,18 +35,13 @@ async def test_partitioner_api_hi_res(partition_file: Path):
     )
     partitioner = Partitioner(config=partitioner_config)
     results = await partitioner.run_async(filename=partition_file)
-    results_dir = int_test_dir / "results"
-    results_dir.mkdir(exist_ok=True)
-    results_path = results_dir / f"{partition_file.name}.json"
-    with results_path.open("w") as f:
-        json.dump(results, f, indent=2)
     assert results
 
 
 @pytest.mark.parametrize(
     "partition_file",
-    non_image_partition_files,
-    ids=[path.name for path in non_image_partition_files],
+    supported_fast_partition_files,
+    ids=[path.name for path in supported_fast_partition_files],
 )
 @requires_env("UNSTRUCTURED_API_KEY", "UNSTRUCTURED_API_URL")
 @pytest.mark.asyncio
@@ -68,7 +65,11 @@ async def test_partitioner_api_fast_error(partition_file: Path):
     api_key = os.getenv("UNSTRUCTURED_API_KEY")
     api_url = os.getenv("UNSTRUCTURED_API_URL")
     partitioner_config = PartitionerConfig(
-        strategy="fast", partition_by_api=True, api_key=api_key, partition_endpoint=api_url
+        strategy="fast",
+        partition_by_api=True,
+        api_key=api_key,
+        partition_endpoint=api_url,
+        raise_unsupported_filetype=True,
     )
     partitioner = Partitioner(config=partitioner_config)
     with pytest.raises(UserError):

unstructured_ingest/__version__.py

@@ -1 +1 @@
-__version__ = "0.4.7"  # pragma: no cover
+__version__ = "0.5.0"  # pragma: no cover

unstructured_ingest/cli/interfaces.py

@@ -417,7 +417,7 @@ class CliEmbeddingConfig(EmbeddingConfig, CliMixin):
         embed_providers = [
             "openai",
             "huggingface",
-            "aws-bedrock",
+            "bedrock",
             "vertexai",
             "voyageai",
             "octoai",

unstructured_ingest/interfaces.py

@@ -226,7 +226,7 @@ class EmbeddingConfig(BaseConfig):
             )
 
             return OctoAIEmbeddingEncoder(config=OctoAiEmbeddingConfig(**kwargs))
-        elif self.provider == "aws-bedrock":
+        elif self.provider == "bedrock":
             from unstructured_ingest.embed.bedrock import (
                 BedrockEmbeddingConfig,
                 BedrockEmbeddingEncoder,

unstructured_ingest/v2/pipeline/pipeline.py

@@ -268,6 +268,7 @@ class Pipeline:
 
         # Partition content
         elements = self.partitioner_step(downloaded_data)
+        elements = self.clean_results(results=elements)
         # Download data non longer needed, delete if possible
         self.downloader_step.delete_cache()
         elements = self.clean_results(results=elements)

unstructured_ingest/v2/processes/connectors/onedrive.py

@@ -105,6 +105,7 @@ class OnedriveIndexerConfig(IndexerConfig):
 class OnedriveIndexer(Indexer):
     connection_config: OnedriveConnectionConfig
     index_config: OnedriveIndexerConfig
+    connector_type: str = CONNECTOR_TYPE
 
     def precheck(self) -> None:
         try:
@@ -172,7 +173,7 @@ class OnedriveIndexer(Indexer):
         )
         return FileData(
             identifier=drive_item.id,
-            connector_type=CONNECTOR_TYPE,
+            connector_type=self.connector_type,
             source_identifiers=SourceIdentifiers(
                 fullpath=server_path, filename=drive_item.name, rel_path=rel_path
             ),
@@ -201,7 +202,8 @@ class OnedriveIndexer(Indexer):
         token_resp = await asyncio.to_thread(self.connection_config.get_token)
         if "error" in token_resp:
             raise SourceConnectionError(
-                f"[{CONNECTOR_TYPE}]: {token_resp['error']} ({token_resp.get('error_description')})"
+                f"[{self.connector_type}]: {token_resp['error']} "
+                f"({token_resp.get('error_description')})"
             )
 
         client = await asyncio.to_thread(self.connection_config.get_client)
@@ -221,6 +223,7 @@ class OnedriveDownloaderConfig(DownloaderConfig):
 class OnedriveDownloader(Downloader):
     connection_config: OnedriveConnectionConfig
     download_config: OnedriveDownloaderConfig
+    connector_type: str = CONNECTOR_TYPE
 
     @SourceConnectionNetworkError.wrap
     def _fetch_file(self, file_data: FileData) -> DriveItem:
@@ -260,7 +263,9 @@ class OnedriveDownloader(Downloader):
                     file.download_session(f).execute_query()
             return self.generate_download_response(file_data=file_data, download_path=download_path)
         except Exception as e:
-            logger.error(f"[{CONNECTOR_TYPE}] Exception during downloading: {e}", exc_info=True)
+            logger.error(
+                f"[{self.connector_type}] Exception during downloading: {e}", exc_info=True
+            )
             # Re-raise to see full stack trace locally
             raise
 

unstructured_ingest/v2/processes/connectors/sharepoint.py

@@ -56,6 +56,7 @@ class SharepointIndexerConfig(OnedriveIndexerConfig):
 class SharepointIndexer(OnedriveIndexer):
     connection_config: SharepointConnectionConfig
     index_config: SharepointIndexerConfig
+    connector_type: str = CONNECTOR_TYPE
 
     @requires_dependencies(["office365"], extras="sharepoint")
     async def run_async(self, **kwargs: Any) -> AsyncIterator[FileData]:
@@ -64,7 +65,8 @@ class SharepointIndexer(OnedriveIndexer):
         token_resp = await asyncio.to_thread(self.connection_config.get_token)
         if "error" in token_resp:
             raise SourceConnectionError(
-                f"[{CONNECTOR_TYPE}]: {token_resp['error']} ({token_resp.get('error_description')})"
+                f"[{self.connector_type}]: {token_resp['error']} "
+                f"({token_resp.get('error_description')})"
             )
 
         client = await asyncio.to_thread(self.connection_config.get_client)
@@ -90,6 +92,7 @@ class SharepointDownloaderConfig(OnedriveDownloaderConfig):
 class SharepointDownloader(OnedriveDownloader):
     connection_config: SharepointConnectionConfig
     download_config: SharepointDownloaderConfig
+    connector_type: str = CONNECTOR_TYPE
 
     @SourceConnectionNetworkError.wrap
     @requires_dependencies(["office365"], extras="onedrive")

unstructured_ingest/v2/processes/embedder.py

@@ -18,7 +18,7 @@ class EmbedderConfig(BaseModel):
             "openai",
             "azure-openai",
             "huggingface",
-            "aws-bedrock",
+            "bedrock",
             "vertexai",
             "voyageai",
             "octoai",
@@ -162,7 +162,7 @@ class EmbedderConfig(BaseModel):
         if self.embedding_provider == "octoai":
             return self.get_octoai_embedder(embedding_kwargs=kwargs)
 
-        if self.embedding_provider == "aws-bedrock":
+        if self.embedding_provider == "bedrock":
             return self.get_bedrock_embedder()
 
         if self.embedding_provider == "vertexai":

unstructured_ingest/v2/processes/partitioner.py

@@ -1,3 +1,4 @@
+import json
 from abc import ABC
 from dataclasses import dataclass
 from pathlib import Path
@@ -7,6 +8,7 @@ from pydantic import BaseModel, Field, SecretStr
 
 from unstructured_ingest.utils.data_prep import flatten_dict
 from unstructured_ingest.utils.dep_check import requires_dependencies
+from unstructured_ingest.v2.errors import UserError
 from unstructured_ingest.v2.interfaces.process import BaseProcess
 from unstructured_ingest.v2.logger import logger
 from unstructured_ingest.v2.unstructured_api import call_api_async
@@ -73,6 +75,9 @@ class PartitionerConfig(BaseModel):
     hi_res_model_name: Optional[str] = Field(
         default=None, description="Model name for hi-res strategy."
     )
+    raise_unsupported_filetype: bool = Field(
+        default=False, description="Raise an error if the file type is not supported"
+    )
 
     def model_post_init(self, __context: Any) -> None:
         if self.metadata_exclude and self.metadata_include:
@@ -151,13 +156,25 @@ class Partitioner(BaseProcess, ABC):
         class FileDataSourceMetadata(DataSourceMetadata):
             filesize_bytes: Optional[int] = None
 
+        metadata = metadata or {}
         logger.debug(f"using local partition with kwargs: {self.config.to_partition_kwargs()}")
         logger.debug(f"partitioning file {filename} with metadata {metadata}")
-        elements = partition(
-            filename=str(filename.resolve()),
-            data_source_metadata=FileDataSourceMetadata.from_dict(metadata),
-            **self.config.to_partition_kwargs(),
-        )
+        try:
+            elements = partition(
+                filename=str(filename.resolve()),
+                data_source_metadata=FileDataSourceMetadata.from_dict(metadata),
+                **self.config.to_partition_kwargs(),
+            )
+        except ValueError as sdk_error:
+            if (
+                self.is_unstructured_error_unsupported_filetype(sdk_error=sdk_error)
+                and not self.config.raise_unsupported_filetype
+            ):
+                logger.warning(
+                    f"Unsupported file type for strategy {self.config.strategy}: {filename}"
+                )
+                return []
+            raise sdk_error
         return self.postprocess(elements=elements_to_dicts(elements))
 
     @requires_dependencies(dependencies=["unstructured_client"], extras="remote")
@@ -179,10 +196,37 @@ class Partitioner(BaseProcess, ABC):
             element["metadata"]["data_source"] = metadata
         return self.postprocess(elements=elements)
 
+    def is_unstructured_error_unsupported_filetype(self, sdk_error: ValueError) -> bool:
+        error_msg = sdk_error.args[0]
+        return (
+            "Invalid file" in error_msg
+            or "Unstructured schema" in error_msg
+            or "fast strategy is not available for image files" in error_msg
+        )
+
+    def is_client_error_unsupported_filetype(self, error: UserError) -> bool:
+        error_msg = error.args[0]
+        error_dict = json.loads(error_msg)
+        details = error_dict["detail"]
+        return "fast strategy is not available for image files" in details or (
+            "file type" in details.lower() and "is not supported" in details.lower()
+        )
+
     def run(self, filename: Path, metadata: Optional[dict] = None, **kwargs) -> list[dict]:
         return self.partition_locally(filename, metadata=metadata, **kwargs)
 
     async def run_async(
         self, filename: Path, metadata: Optional[dict] = None, **kwargs
     ) -> list[dict]:
-        return await self.partition_via_api(filename, metadata=metadata, **kwargs)
+        try:
+            return await self.partition_via_api(filename, metadata=metadata, **kwargs)
+        except UserError as user_error:
+            if (
+                self.is_client_error_unsupported_filetype(error=user_error)
+                and not self.config.raise_unsupported_filetype
+            ):
+                logger.warning(
+                    f"Unsupported file type for strategy {self.config.strategy}: {filename}"
+                )
+                return []
+            raise user_error

{unstructured_ingest-0.4.7.dist-info → unstructured_ingest-0.5.0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: unstructured-ingest
-Version: 0.4.7
+Version: 0.5.0
 Summary: A library that prepares raw documents for downstream ML tasks.
 Home-page: https://github.com/Unstructured-IO/unstructured-ingest
 Author: Unstructured Technologies
@@ -23,30 +23,30 @@ Requires-Python: >=3.9.0,<3.14
 Description-Content-Type: text/markdown
 License-File: LICENSE.md
 Requires-Dist: pandas
-Requires-Dist: pydantic>=2.7
 Requires-Dist: dataclasses-json
+Requires-Dist: pydantic>=2.7
+Requires-Dist: click
+Requires-Dist: tqdm
 Requires-Dist: python-dateutil
 Requires-Dist: opentelemetry-sdk
-Requires-Dist: tqdm
-Requires-Dist: click
 Provides-Extra: airtable
 Requires-Dist: pyairtable; extra == "airtable"
 Provides-Extra: astradb
 Requires-Dist: astrapy; extra == "astradb"
 Provides-Extra: azure
-Requires-Dist: fsspec; extra == "azure"
 Requires-Dist: adlfs; extra == "azure"
+Requires-Dist: fsspec; extra == "azure"
 Provides-Extra: azure-ai-search
 Requires-Dist: azure-search-documents; extra == "azure-ai-search"
 Provides-Extra: bedrock
-Requires-Dist: boto3; extra == "bedrock"
 Requires-Dist: aioboto3; extra == "bedrock"
+Requires-Dist: boto3; extra == "bedrock"
 Provides-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"
+Requires-Dist: fsspec; extra == "box"
 Provides-Extra: chroma
 Requires-Dist: chromadb; extra == "chroma"
 Provides-Extra: clarifai
@@ -63,8 +63,8 @@ Requires-Dist: databricks-sql-connector; extra == "databricks-delta-tables"
 Provides-Extra: databricks-volumes
 Requires-Dist: databricks-sdk; extra == "databricks-volumes"
 Provides-Extra: delta-table
-Requires-Dist: boto3; extra == "delta-table"
 Requires-Dist: deltalake; extra == "delta-table"
+Requires-Dist: boto3; extra == "delta-table"
 Provides-Extra: discord
 Requires-Dist: discord.py; extra == "discord"
 Provides-Extra: doc
@@ -92,19 +92,19 @@ Requires-Dist: voyageai; extra == "embed-voyageai"
 Provides-Extra: epub
 Requires-Dist: unstructured[epub]; extra == "epub"
 Provides-Extra: gcs
-Requires-Dist: fsspec; extra == "gcs"
-Requires-Dist: bs4; extra == "gcs"
 Requires-Dist: gcsfs; extra == "gcs"
+Requires-Dist: bs4; extra == "gcs"
+Requires-Dist: fsspec; extra == "gcs"
 Provides-Extra: github
-Requires-Dist: requests; extra == "github"
 Requires-Dist: pygithub>1.58.0; extra == "github"
+Requires-Dist: requests; 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: urllib3; extra == "hubspot"
 Requires-Dist: hubspot-api-client; extra == "hubspot"
+Requires-Dist: urllib3; extra == "hubspot"
 Provides-Extra: jira
 Requires-Dist: atlassian-python-api; extra == "jira"
 Provides-Extra: kafka
@@ -127,15 +127,15 @@ Requires-Dist: cymple; extra == "neo4j"
 Requires-Dist: neo4j; extra == "neo4j"
 Provides-Extra: notion
 Requires-Dist: httpx; extra == "notion"
-Requires-Dist: htmlBuilder; extra == "notion"
-Requires-Dist: notion-client; extra == "notion"
 Requires-Dist: backoff; extra == "notion"
+Requires-Dist: notion-client; extra == "notion"
+Requires-Dist: htmlBuilder; extra == "notion"
 Provides-Extra: odt
 Requires-Dist: unstructured[odt]; extra == "odt"
 Provides-Extra: onedrive
-Requires-Dist: msal; extra == "onedrive"
 Requires-Dist: Office365-REST-Python-Client; extra == "onedrive"
 Requires-Dist: bs4; extra == "onedrive"
+Requires-Dist: msal; extra == "onedrive"
 Provides-Extra: openai
 Requires-Dist: tiktoken; extra == "openai"
 Requires-Dist: openai; extra == "openai"
@@ -144,8 +144,8 @@ Requires-Dist: opensearch-py; extra == "opensearch"
 Provides-Extra: org
 Requires-Dist: unstructured[org]; extra == "org"
 Provides-Extra: outlook
-Requires-Dist: msal; extra == "outlook"
 Requires-Dist: Office365-REST-Python-Client; extra == "outlook"
+Requires-Dist: msal; extra == "outlook"
 Provides-Extra: pdf
 Requires-Dist: unstructured[pdf]; extra == "pdf"
 Provides-Extra: pinecone
@@ -169,35 +169,35 @@ Requires-Dist: unstructured[rst]; extra == "rst"
 Provides-Extra: rtf
 Requires-Dist: unstructured[rtf]; extra == "rtf"
 Provides-Extra: s3
-Requires-Dist: fsspec; extra == "s3"
 Requires-Dist: s3fs; extra == "s3"
+Requires-Dist: fsspec; extra == "s3"
 Provides-Extra: salesforce
 Requires-Dist: simple-salesforce; extra == "salesforce"
 Provides-Extra: sftp
 Requires-Dist: paramiko; extra == "sftp"
 Requires-Dist: fsspec; extra == "sftp"
 Provides-Extra: sharepoint
-Requires-Dist: msal; extra == "sharepoint"
 Requires-Dist: Office365-REST-Python-Client; extra == "sharepoint"
+Requires-Dist: msal; extra == "sharepoint"
 Provides-Extra: singlestore
 Requires-Dist: singlestoredb; extra == "singlestore"
 Provides-Extra: slack
 Requires-Dist: slack-sdk[optional]; extra == "slack"
 Provides-Extra: snowflake
-Requires-Dist: snowflake-connector-python; extra == "snowflake"
 Requires-Dist: psycopg2-binary; extra == "snowflake"
+Requires-Dist: snowflake-connector-python; extra == "snowflake"
 Provides-Extra: togetherai
 Requires-Dist: together; extra == "togetherai"
 Provides-Extra: tsv
 Requires-Dist: unstructured[tsv]; extra == "tsv"
 Provides-Extra: vastdb
 Requires-Dist: vastdb; extra == "vastdb"
-Requires-Dist: pyarrow; extra == "vastdb"
 Requires-Dist: ibis; extra == "vastdb"
+Requires-Dist: pyarrow; extra == "vastdb"
 Provides-Extra: vectara
-Requires-Dist: aiofiles; extra == "vectara"
 Requires-Dist: httpx; extra == "vectara"
 Requires-Dist: requests; extra == "vectara"
+Requires-Dist: aiofiles; extra == "vectara"
 Provides-Extra: weaviate
 Requires-Dist: weaviate-client; extra == "weaviate"
 Provides-Extra: wikipedia

{unstructured_ingest-0.4.7.dist-info → unstructured_ingest-0.5.0.dist-info}/RECORD

@@ -15,11 +15,12 @@ test/integration/connectors/test_milvus.py,sha256=7mI6zznN0PTxDL9DLogH1k3dxx6R8D
 test/integration/connectors/test_mongodb.py,sha256=0A6DvF-iTCSZzOefisd_i20j9li8uNWTF2wyLGwlhco,12446
 test/integration/connectors/test_neo4j.py,sha256=r4TRYtTXeeOdcRcfa_gvslhSKvoIWrwN1FRJ5XRoH4k,8456
 test/integration/connectors/test_notion.py,sha256=ueXyVqYWzP4LuZYe6PauptkXNG6qkoV3srltFOSSKTA,5403
-test/integration/connectors/test_onedrive.py,sha256=rjgN2LhaW1htEMBJPxmlP_kcRB7p_oOeZcogFlHyJH4,3721
+test/integration/connectors/test_onedrive.py,sha256=iwiDK0kWCfQbIEPnWUzzAA5PiCsHcmFZSxEcIZy_6cc,5229
 test/integration/connectors/test_pinecone.py,sha256=acKEu1vnAk0Ht3FhCnGtOEKaj_YlgCzZB7wRU17ehQ0,12407
 test/integration/connectors/test_qdrant.py,sha256=Yme3ZZ5zIbaZ-yYLUqN2oy0hsrcAfvlleRLYWMSYeSE,8049
 test/integration/connectors/test_redis.py,sha256=1aKwOb-K4zCxZwHmgW_WzGJwqLntbWTbpGQ-rtUwN9o,4360
 test/integration/connectors/test_s3.py,sha256=E1dypeag_E3OIfpQWIz3jb7ctRHRD63UtyTrzyvJzpc,7473
+test/integration/connectors/test_sharepoint.py,sha256=8HlcnrP4K8oPUzef6AA11P2cMlxSp7tiddTkT4JOeRU,2378
 test/integration/connectors/test_vectara.py,sha256=4kKOOTGUjeZw2jKRcgVDI7ifbRPRZfjjVO4d_7H5C6I,8710
 test/integration/connectors/databricks/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 test/integration/connectors/databricks/test_volumes_native.py,sha256=KqiapQAV0s_Zv0CO8BwYoiCk30dwrSZzuigUWNRIem0,9559
@@ -46,7 +47,7 @@ test/integration/connectors/utils/docker_compose.py,sha256=GVTB6Cel05c0VQ2n4AwkQ
 test/integration/connectors/utils/validation/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 test/integration/connectors/utils/validation/destination.py,sha256=ZvMSvqz9in35xaoUJGx9rG8oWCU3FYlfLLQ6sfdI0pw,2649
 test/integration/connectors/utils/validation/equality.py,sha256=R6d_1c-Si5518WJcBcshF_wBRnywnZ0ORQ-NL0xNmGo,2602
-test/integration/connectors/utils/validation/source.py,sha256=VALU5ms_JBu_eFkp2WQ7oZtJKozJ8MZSJ7h7ZA3Fz_Q,12296
+test/integration/connectors/utils/validation/source.py,sha256=xnAZI26ILdeMhgrWAGrU2N2fqK58YNGkfyUhJekZ0Ho,13541
 test/integration/connectors/utils/validation/utils.py,sha256=xYYvAbqP6_lZyH09_JjB4w2Sf8aQPvDVT5vZTs05ILs,1428
 test/integration/connectors/weaviate/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 test/integration/connectors/weaviate/conftest.py,sha256=6Q6QdrLJmGHowRFSmoVSzup2EX6qASfS2Z5tqlpTm9M,387
@@ -55,7 +56,7 @@ test/integration/connectors/weaviate/test_local.py,sha256=gXMpnzVcrNQdptDjx0haPW
 test/integration/embedders/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 test/integration/embedders/conftest.py,sha256=B2W771RbijR7G_GybsCzRyIvOzXqzbKZdRIlNDd5AGY,334
 test/integration/embedders/test_azure_openai.py,sha256=YQ3uq2-NuxtTyGsSgMNa10pcITLKMJ4E1scTGFgwujw,1790
-test/integration/embedders/test_bedrock.py,sha256=ZehreheLgY9Bqdjk-3MQOaou9IP-H3Pcz7WWiOWAxTU,3557
+test/integration/embedders/test_bedrock.py,sha256=vmjoi1uUk-LX4Yz0ZPn6Ry1JdVEsyIhLhPbSPmkeT9o,3553
 test/integration/embedders/test_huggingface.py,sha256=qFblyXounVNRaNkk3gbKoBqU5E2dNecgKU2Bz2LyOa8,989
 test/integration/embedders/test_mixedbread.py,sha256=lLz_cooyC38VSo-FMHbhKpHvYs3QzA20NOIvM5oooaw,1998
 test/integration/embedders/test_octoai.py,sha256=qs-bqZ7iGWO_BzUZvKJmOHBT3cmFSkEYbleWhj3snJc,2197
@@ -65,7 +66,7 @@ test/integration/embedders/test_vertexai.py,sha256=4-E4plJXFf1b02RhOqOCBHR2GA4gT
 test/integration/embedders/test_voyageai.py,sha256=Gm3sVjhsym1ASIDfr-sZoCbpsNMaAk_l4E3-dtjRCQ4,1832
 test/integration/embedders/utils.py,sha256=Sqqg-X31ZV1hojqPQBaZgM2lb2u8cG6s6OnH9JRsFjs,2717
 test/integration/partitioners/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-test/integration/partitioners/test_partitioner.py,sha256=MEQJbRoc01uPLT6O8CkXeQF_DXK21nz3KVJkzkBtsgM,2835
+test/integration/partitioners/test_partitioner.py,sha256=6sdZhhtqEICBPqEgpKrCQIfJ-7hKcwuTFqjWs1mbQf8,2787
 test/unit/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 test/unit/test_error.py,sha256=RflmngCdFNKOLXVfLnUdNfY3Mfg3k7DTEzfIl0B-syU,840
 test/unit/test_html.py,sha256=LKGi_QaH4U4gktrbd2NcURL-d-0Rm1UnG5Y6r9EvTG0,4489
@@ -102,9 +103,9 @@ 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=i2QrUEuUnVPQuTv5hg_JWbhbwm5k6KU4hPIFq0SIgdc,42
+unstructured_ingest/__version__.py,sha256=A09Ks7MDqP-QtYP9TIQMxydOZeCTtu9i7xoq5wdy4As,42
 unstructured_ingest/error.py,sha256=qDncnJgbf5ils956RcO2CGlAKYDT5OaEM9Clv1JVTNc,1448
-unstructured_ingest/interfaces.py,sha256=OYVUP0bzBJpT-Lz92BDyz_hLBvyfxkuSwWHhUdnUayA,31493
+unstructured_ingest/interfaces.py,sha256=7DOnDpGvUNlCoFR7UPRGmOarqH5sFtuUOO5vf8X3oTM,31489
 unstructured_ingest/logger.py,sha256=S5nSqGcABoQyeicgRnBQFjDScCaTvFVivOCvbo-laL0,4479
 unstructured_ingest/main.py,sha256=82G_7eG4PNhc_xIqj4Y_sFbDV9VI-nwSfsfJQMzovMk,169
 unstructured_ingest/processor.py,sha256=XKKrvbxsb--5cDzz4hB3-GfWZYyIjJ2ah8FpzQKF_DM,2760
@@ -112,7 +113,7 @@ unstructured_ingest/cli/__init__.py,sha256=9kNcBOHuXON5lB1MJU9QewEhwPmId56vXqB29
 unstructured_ingest/cli/cli.py,sha256=LutBTBYMqboKw8cputHVszpenyfnySzcUC15ifwuYyg,1049
 unstructured_ingest/cli/cmd_factory.py,sha256=UdHm1KacTombpF6DxyTSwTCuApsKHUYw_kVu5Nhcy3Y,364
 unstructured_ingest/cli/common.py,sha256=I0El08FHz5kxw7iz0VWOWPrvcJD1rBgXJSwVIpVmmwU,204
-unstructured_ingest/cli/interfaces.py,sha256=lpaaOdAQ4NMsawVaHSk5lXCcZ0Mw85kRzfElu1ODCB0,24090
+unstructured_ingest/cli/interfaces.py,sha256=pvEwNfYwINx3-TQ0LPudjpYNR3PnanUiXpEePPEtRSw,24086
 unstructured_ingest/cli/utils.py,sha256=KNhkFNKOeEihc8HlvMz_MTbYVQNFklrBKbC8xg9h1xE,7982
 unstructured_ingest/cli/base/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 unstructured_ingest/cli/base/cmd.py,sha256=BbfjA2v203Jh-7DL6bzxQ7fOeNixd5BsBMuzXz6h5IQ,583
@@ -399,7 +400,7 @@ unstructured_ingest/v2/interfaces/uploader.py,sha256=rrZLTjmTcrDL-amQIKzIP6j2fW-
 unstructured_ingest/v2/pipeline/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 unstructured_ingest/v2/pipeline/interfaces.py,sha256=-Y6gPnl-SbNxIx5-dQCmiYSPKUMjivrRlBLIKIUWVeM,8658
 unstructured_ingest/v2/pipeline/otel.py,sha256=K3pQvWVgWzyOWMKCBUofsH7wTZPJ0Ysw5sLjMBLW41I,1088
-unstructured_ingest/v2/pipeline/pipeline.py,sha256=4IwCWMlBrMpZI6V82q5nzrbyQNDVM62AQsWt6MUBWa8,16508
+unstructured_ingest/v2/pipeline/pipeline.py,sha256=b37fQGm_lGutQ3Jc0qePB15lkBiFavH9tCso3inm-3I,16564
 unstructured_ingest/v2/pipeline/steps/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 unstructured_ingest/v2/pipeline/steps/chunk.py,sha256=LK2ldM24TE4ukX_Z6Z81LpF53orMaRkddM3uhLtT5EQ,3221
 unstructured_ingest/v2/pipeline/steps/download.py,sha256=nZ4B0d9p-6TgWqrBoKUQPlr8m6dz1RGNr_3OjUhRpWg,8259
@@ -413,9 +414,9 @@ unstructured_ingest/v2/pipeline/steps/upload.py,sha256=We4OAtStuZwWKKBCOPhfeAz_v
 unstructured_ingest/v2/processes/__init__.py,sha256=FaHWSCGyc7GWVnAsNEUUj7L8hT8gCVY3_hUE2VzWtUg,462
 unstructured_ingest/v2/processes/chunker.py,sha256=31-7ojsM2coIt2rMR0KOb82IxLVJfNHbqYUOsDkhxN8,5491
 unstructured_ingest/v2/processes/connector_registry.py,sha256=vkEe6jpgdYtZCxMj59s5atWGgmPuxAEXRUoTt-MJ7wc,2198
-unstructured_ingest/v2/processes/embedder.py,sha256=uiuCOSwwasHp4eqtewMvgnM86WVch7HDFiWqpGLahvo,7812
+unstructured_ingest/v2/processes/embedder.py,sha256=PTBlRgNbAXkSaLg7JrZzHwAoqpHmopg8jNU1TmaXguU,7804
 unstructured_ingest/v2/processes/filter.py,sha256=kjUmMw2SDq2bme0JCAOxs6cJriIG6Ty09KOznS-xz08,2145
-unstructured_ingest/v2/processes/partitioner.py,sha256=agpHwB9FR8OZVQqE7zFEb0IcDPCOPA_BZjLzLF71nOY,8194
+unstructured_ingest/v2/processes/partitioner.py,sha256=ZC9mt85I3o_SLR4DvE7vPBGphMET994phFkTuT-L9B8,9998
 unstructured_ingest/v2/processes/uncompress.py,sha256=Z_XfsITGdyaRwhtNUc7bMj5Y2jLuBge8KoK4nxhqKag,2425
 unstructured_ingest/v2/processes/connectors/__init__.py,sha256=KO1zn-96Qa49TOSZn-gv_RUMGMCmUcdtHoeJqCpxPLY,6219
 unstructured_ingest/v2/processes/connectors/airtable.py,sha256=eeZJe-bBNxt5Sa-XEFCdcGeJCguJU5WN2Mv9kLp5dVQ,8917
@@ -433,12 +434,12 @@ unstructured_ingest/v2/processes/connectors/local.py,sha256=ZvWTj6ZYkwnvQMNFsZWo
 unstructured_ingest/v2/processes/connectors/milvus.py,sha256=wmcu9NVy3gYlQGT25inN5w_QrhFoL8-hRq0pJFSNw8g,8866
 unstructured_ingest/v2/processes/connectors/mongodb.py,sha256=cL0QUQZF_s2brh3nNNeAywXVpaIiND4b5JTAFlYjLjw,14273
 unstructured_ingest/v2/processes/connectors/neo4j.py,sha256=HU1IwchTM7Q1kkeIFVe-Lg6gInMItBpgkDkVwuTvkGY,14259
-unstructured_ingest/v2/processes/connectors/onedrive.py,sha256=b616B_-9MfU6gxvpw7IBUa2szNFURA_VP8q5j2FXxnA,17632
+unstructured_ingest/v2/processes/connectors/onedrive.py,sha256=EM9fq67RsiudZvZbi6nDXkS-i6W0xLvbkNvD0G-Ni5E,17779
 unstructured_ingest/v2/processes/connectors/outlook.py,sha256=KgNGM8hImRhy6_SpswRP2VwRD4VOrqqJoySgxf2oduI,9290
 unstructured_ingest/v2/processes/connectors/pinecone.py,sha256=bQDCch7OGiQgpWO3n3ncLuQ4XCWqDc7ZWEB-Qrqkss8,10730
 unstructured_ingest/v2/processes/connectors/redisdb.py,sha256=p0AY4ukBNpwAemV4bWzpScvVbLTVlI3DzsCNUKiBI5M,6757
 unstructured_ingest/v2/processes/connectors/salesforce.py,sha256=2CiO2ZZiZ1Y1-nB7wcDlDVcpW2B7ut9wCj66rkkqho0,11616
-unstructured_ingest/v2/processes/connectors/sharepoint.py,sha256=SdcbOEUzgi1sUZJA6doZDm-a8d4F3Qtud-OVbDKW7Ng,4456
+unstructured_ingest/v2/processes/connectors/sharepoint.py,sha256=f0F7KioXgucVc3tVASTa67ynlTa4s9_FKGPHop6Xm0A,4563
 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/vectara.py,sha256=BlI_4nkpNR99aYxDd9eusm5LQsVB9EI0r-5Kc1D7pgQ,12255
@@ -562,9 +563,9 @@ unstructured_ingest/v2/processes/connectors/weaviate/cloud.py,sha256=bXtfEYLquR-
 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=yJza_jBSEFnzZRq5L6vJ0Mm3uS1uxkOiKIimPpUyQds,12418
-unstructured_ingest-0.4.7.dist-info/LICENSE.md,sha256=SxkKP_62uIAKb9mb1eH7FH4Kn2aYT09fgjKpJt5PyTk,11360
-unstructured_ingest-0.4.7.dist-info/METADATA,sha256=yGcahQ8fZmoU_c1h02b76tRn5w0uj_931AAQKlFrqxs,8051
-unstructured_ingest-0.4.7.dist-info/WHEEL,sha256=GV9aMThwP_4oNCtvEC2ec3qUYutgWeAzklro_0m4WJQ,91
-unstructured_ingest-0.4.7.dist-info/entry_points.txt,sha256=gUAAFnjFPnBgThJSEbw0N5ZjxtaKlT1s9e05_arQrNw,70
-unstructured_ingest-0.4.7.dist-info/top_level.txt,sha256=DMuDMHZRMdeay8v8Kdi855muIv92F0OkutvBCaBEW6M,25
-unstructured_ingest-0.4.7.dist-info/RECORD,,
+unstructured_ingest-0.5.0.dist-info/LICENSE.md,sha256=SxkKP_62uIAKb9mb1eH7FH4Kn2aYT09fgjKpJt5PyTk,11360
+unstructured_ingest-0.5.0.dist-info/METADATA,sha256=dyxZ7tfjq1tkZPJgaK6ZanQwB6pteSIznmfUhAgnT64,8051
+unstructured_ingest-0.5.0.dist-info/WHEEL,sha256=GV9aMThwP_4oNCtvEC2ec3qUYutgWeAzklro_0m4WJQ,91
+unstructured_ingest-0.5.0.dist-info/entry_points.txt,sha256=gUAAFnjFPnBgThJSEbw0N5ZjxtaKlT1s9e05_arQrNw,70
+unstructured_ingest-0.5.0.dist-info/top_level.txt,sha256=DMuDMHZRMdeay8v8Kdi855muIv92F0OkutvBCaBEW6M,25
+unstructured_ingest-0.5.0.dist-info/RECORD,,