unstructured-ingest 0.3.2__py3-none-any.whl → 0.3.3__py3-none-any.whl

test/integration/connectors/test_lancedb.py

@@ -14,9 +14,9 @@ from upath import UPath
 from test.integration.connectors.utils.constants import DESTINATION_TAG
 from unstructured_ingest.v2.interfaces.file_data import FileData, SourceIdentifiers
 from unstructured_ingest.v2.processes.connectors.lancedb.aws import (
-    LanceDBS3AccessConfig,
-    LanceDBS3ConnectionConfig,
-    LanceDBS3Uploader,
+    LanceDBAwsAccessConfig,
+    LanceDBAwsConnectionConfig,
+    LanceDBAwsUploader,
 )
 from unstructured_ingest.v2.processes.connectors.lancedb.azure import (
     LanceDBAzureAccessConfig,
@@ -156,7 +156,7 @@ def _get_uri(target: Literal["local", "s3", "gcs", "az"], local_base_path: Path)
 
 def _get_uploader(
     uri: str,
-) -> Union[LanceDBAzureUploader, LanceDBAzureUploader, LanceDBS3Uploader, LanceDBGSPUploader]:
+) -> Union[LanceDBAzureUploader, LanceDBAzureUploader, LanceDBAwsUploader, LanceDBGSPUploader]:
     target = uri.split("://", maxsplit=1)[0] if uri.startswith(("s3", "az", "gs")) else "local"
     if target == "az":
         azure_connection_string = os.getenv("AZURE_DEST_CONNECTION_STR")
@@ -170,10 +170,10 @@ def _get_uploader(
         )
 
     elif target == "s3":
-        return LanceDBS3Uploader(
+        return LanceDBAwsUploader(
             upload_config=LanceDBUploaderConfig(table_name=TABLE_NAME),
-            connection_config=LanceDBS3ConnectionConfig(
-                access_config=LanceDBS3AccessConfig(
+            connection_config=LanceDBAwsConnectionConfig(
+                access_config=LanceDBAwsAccessConfig(
                     aws_access_key_id=os.getenv("S3_INGEST_TEST_ACCESS_KEY"),
                     aws_secret_access_key=os.getenv("S3_INGEST_TEST_SECRET_KEY"),
                 ),

test/integration/connectors/test_milvus.py

@@ -15,6 +15,7 @@ from pymilvus.milvus_client import IndexParams
 from test.integration.connectors.utils.constants import DESTINATION_TAG, env_setup_path
 from test.integration.connectors.utils.docker import healthcheck_wait
 from test.integration.connectors.utils.docker_compose import docker_compose_context
+from unstructured_ingest.error import DestinationConnectionError
 from unstructured_ingest.v2.interfaces import FileData, SourceIdentifiers
 from unstructured_ingest.v2.processes.connectors.milvus import (
     CONNECTOR_TYPE,
@@ -24,9 +25,10 @@ from unstructured_ingest.v2.processes.connectors.milvus import (
     MilvusUploadStager,
 )
 
-DB_URI = "http://localhost:19530"
 DB_NAME = "test_database"
-COLLECTION_NAME = "test_collection"
+EXISTENT_COLLECTION_NAME = "test_collection"
+NONEXISTENT_COLLECTION_NAME = "nonexistent_collection"
+DB_URI = "http://localhost:19530"
 
 
 def get_schema() -> CollectionSchema:
@@ -55,7 +57,9 @@ def get_index_params() -> IndexParams:
     return index_params
 
 
-@pytest.fixture
+# NOTE: Precheck tests are read-only so they don't interfere with destination test,
+# using scope="module" we can limit number of times the docker-compose has to be run
+@pytest.fixture(scope="module")
 def collection():
     docker_client = docker.from_env()
     with docker_compose_context(docker_compose_path=env_setup_path / "milvus"):
@@ -73,10 +77,10 @@ def collection():
             schema = get_schema()
             index_params = get_index_params()
             collection_resp = milvus_client.create_collection(
-                collection_name=COLLECTION_NAME, schema=schema, index_params=index_params
+                collection_name=EXISTENT_COLLECTION_NAME, schema=schema, index_params=index_params
             )
-            print(f"Created collection {COLLECTION_NAME}: {collection_resp}")
-            yield COLLECTION_NAME
+            print(f"Created collection {EXISTENT_COLLECTION_NAME}: {collection_resp}")
+            yield EXISTENT_COLLECTION_NAME
         finally:
             milvus_client.close()
 
@@ -139,3 +143,27 @@ async def test_milvus_destination(
     uploader.run(path=staged_filepath, file_data=file_data)
     with uploader.get_client() as client:
         validate_count(client=client, expected_count=expected_count)
+
+
+@pytest.mark.tags(CONNECTOR_TYPE, DESTINATION_TAG)
+def test_precheck_succeeds(collection: str):
+    uploader = MilvusUploader(
+        connection_config=MilvusConnectionConfig(uri=DB_URI),
+        upload_config=MilvusUploaderConfig(db_name=DB_NAME, collection_name=collection),
+    )
+    uploader.precheck()
+
+
+@pytest.mark.tags(CONNECTOR_TYPE, DESTINATION_TAG)
+def test_precheck_fails_on_nonexistent_collection(collection: str):
+    uploader = MilvusUploader(
+        connection_config=MilvusConnectionConfig(uri=DB_URI),
+        upload_config=MilvusUploaderConfig(
+            db_name=DB_NAME, collection_name=NONEXISTENT_COLLECTION_NAME
+        ),
+    )
+    with pytest.raises(
+        DestinationConnectionError,
+        match=f"Collection '{NONEXISTENT_COLLECTION_NAME}' does not exist",
+    ):
+        uploader.precheck()

test/integration/connectors/weaviate/test_cloud.py

@@ -0,0 +1,34 @@
+import pytest
+from pydantic import ValidationError
+
+from unstructured_ingest.v2.processes.connectors.weaviate.cloud import (
+    CloudWeaviateAccessConfig,
+    CloudWeaviateConnectionConfig,
+)
+
+
+def test_weaviate_failing_connection_config():
+    with pytest.raises(ValidationError):
+        CloudWeaviateConnectionConfig(
+            access_config=CloudWeaviateAccessConfig(api_key="my key", password="password"),
+            username="username",
+            cluster_url="clusterurl",
+        )
+
+
+def test_weaviate_connection_config_happy_path():
+    CloudWeaviateConnectionConfig(
+        access_config=CloudWeaviateAccessConfig(
+            api_key="my key",
+        ),
+        cluster_url="clusterurl",
+    )
+
+
+def test_weaviate_connection_config_anonymous():
+    CloudWeaviateConnectionConfig(
+        access_config=CloudWeaviateAccessConfig(api_key="my key", password="password"),
+        username="username",
+        anonymous=True,
+        cluster_url="clusterurl",
+    )

test/unit/test_utils.py

@@ -8,7 +8,11 @@ import pytz
 
 from unstructured_ingest.cli.utils import extract_config
 from unstructured_ingest.interfaces import BaseConfig
-from unstructured_ingest.utils.string_and_date_utils import ensure_isoformat_datetime, json_to_dict
+from unstructured_ingest.utils.string_and_date_utils import (
+    ensure_isoformat_datetime,
+    json_to_dict,
+    truncate_string_bytes,
+)
 
 
 @dataclass
@@ -162,3 +166,19 @@ def test_ensure_isoformat_datetime_fails_on_string():
 def test_ensure_isoformat_datetime_fails_on_int():
     with pytest.raises(TypeError):
         ensure_isoformat_datetime(1111)
+
+
+def test_truncate_string_bytes_return_truncated_string():
+    test_string = "abcdef안녕하세요ghijklmn방갑습니opqrstu 더 길어지면 안되는 문자열vwxyz"
+    max_bytes = 11
+    result = truncate_string_bytes(test_string, max_bytes)
+    assert result == "abcdef안"
+    assert len(result.encode("utf-8")) <= max_bytes
+
+
+def test_truncate_string_bytes_return_untouched_string():
+    test_string = "abcdef"
+    max_bytes = 11
+    result = truncate_string_bytes(test_string, max_bytes)
+    assert result == "abcdef"
+    assert len(result.encode("utf-8")) <= max_bytes

unstructured_ingest/__version__.py

@@ -1 +1 @@
-__version__ = "0.3.2"  # pragma: no cover
+__version__ = "0.3.3"  # pragma: no cover

unstructured_ingest/utils/string_and_date_utils.py

@@ -37,3 +37,13 @@ def ensure_isoformat_datetime(timestamp: t.Union[datetime, str]) -> str:
             raise ValueError(f"String '{timestamp}' could not be parsed as a datetime.") from e
     else:
         raise TypeError(f"Expected input type datetime or str, but got {type(timestamp)}.")
+
+
+def truncate_string_bytes(string: str, max_bytes: int, encoding: str = "utf-8") -> str:
+    """
+    Truncates a string to a specified maximum number of bytes.
+    """
+    encoded_string = str(string).encode(encoding)
+    if len(encoded_string) <= max_bytes:
+        return string
+    return encoded_string[:max_bytes].decode(encoding, errors="ignore")

unstructured_ingest/v2/processes/connectors/astradb.py

@@ -19,6 +19,7 @@ from unstructured_ingest.error import (
 )
 from unstructured_ingest.utils.data_prep import batch_generator
 from unstructured_ingest.utils.dep_check import requires_dependencies
+from unstructured_ingest.utils.string_and_date_utils import truncate_string_bytes
 from unstructured_ingest.v2.constants import RECORD_ID_LABEL
 from unstructured_ingest.v2.interfaces import (
     AccessConfig,
@@ -50,6 +51,8 @@ if TYPE_CHECKING:
 
 CONNECTOR_TYPE = "astradb"
 
+MAX_CONTENT_PARAM_BYTE_SIZE = 8000
+
 
 class AstraDBAccessConfig(AccessConfig):
     token: str = Field(description="Astra DB Token with access to the database.")
@@ -301,7 +304,20 @@ class AstraDBUploadStager(UploadStager):
         default_factory=lambda: AstraDBUploadStagerConfig()
     )
 
+    def truncate_dict_elements(self, element_dict: dict) -> None:
+        text = element_dict.pop("text", None)
+        if text is not None:
+            element_dict["text"] = truncate_string_bytes(text, MAX_CONTENT_PARAM_BYTE_SIZE)
+        metadata = element_dict.get("metadata")
+        if metadata is not None and isinstance(metadata, dict):
+            text_as_html = element_dict["metadata"].pop("text_as_html", None)
+            if text_as_html is not None:
+                element_dict["metadata"]["text_as_html"] = truncate_string_bytes(
+                    text_as_html, MAX_CONTENT_PARAM_BYTE_SIZE
+                )
+
     def conform_dict(self, element_dict: dict, file_data: FileData) -> dict:
+        self.truncate_dict_elements(element_dict)
         return {
             "$vector": element_dict.pop("embeddings", None),
             "content": element_dict.pop("text", None),

unstructured_ingest/v2/processes/connectors/lancedb/__init__.py

@@ -6,12 +6,25 @@ from .aws import CONNECTOR_TYPE as LANCEDB_S3_CONNECTOR_TYPE
 from .aws import lancedb_aws_destination_entry
 from .azure import CONNECTOR_TYPE as LANCEDB_AZURE_CONNECTOR_TYPE
 from .azure import lancedb_azure_destination_entry
+from .cloud import CONNECTOR_TYPE as LANCEDB_CLOUD_CONNECTOR_TYPE
+from .cloud import lancedb_cloud_destination_entry
 from .gcp import CONNECTOR_TYPE as LANCEDB_GCS_CONNECTOR_TYPE
 from .gcp import lancedb_gcp_destination_entry
 from .local import CONNECTOR_TYPE as LANCEDB_LOCAL_CONNECTOR_TYPE
 from .local import lancedb_local_destination_entry
 
-add_destination_entry(LANCEDB_S3_CONNECTOR_TYPE, lancedb_aws_destination_entry)
-add_destination_entry(LANCEDB_AZURE_CONNECTOR_TYPE, lancedb_azure_destination_entry)
-add_destination_entry(LANCEDB_GCS_CONNECTOR_TYPE, lancedb_gcp_destination_entry)
-add_destination_entry(LANCEDB_LOCAL_CONNECTOR_TYPE, lancedb_local_destination_entry)
+add_destination_entry(
+    destination_type=LANCEDB_S3_CONNECTOR_TYPE, entry=lancedb_aws_destination_entry
+)
+add_destination_entry(
+    destination_type=LANCEDB_AZURE_CONNECTOR_TYPE, entry=lancedb_azure_destination_entry
+)
+add_destination_entry(
+    destination_type=LANCEDB_GCS_CONNECTOR_TYPE, entry=lancedb_gcp_destination_entry
+)
+add_destination_entry(
+    destination_type=LANCEDB_LOCAL_CONNECTOR_TYPE, entry=lancedb_local_destination_entry
+)
+add_destination_entry(
+    destination_type=LANCEDB_CLOUD_CONNECTOR_TYPE, entry=lancedb_cloud_destination_entry
+)

unstructured_ingest/v2/processes/connectors/lancedb/aws.py

@@ -15,28 +15,28 @@ from unstructured_ingest.v2.processes.connectors.lancedb.lancedb import (
 CONNECTOR_TYPE = "lancedb_aws"
 
 
-class LanceDBS3AccessConfig(AccessConfig):
+class LanceDBAwsAccessConfig(AccessConfig):
     aws_access_key_id: str = Field(description="The AWS access key ID to use.")
     aws_secret_access_key: str = Field(description="The AWS secret access key to use.")
 
 
-class LanceDBS3ConnectionConfig(LanceDBRemoteConnectionConfig):
-    access_config: Secret[LanceDBS3AccessConfig]
+class LanceDBAwsConnectionConfig(LanceDBRemoteConnectionConfig):
+    access_config: Secret[LanceDBAwsAccessConfig]
 
     def get_storage_options(self) -> dict:
         return {**self.access_config.get_secret_value().model_dump(), "timeout": self.timeout}
 
 
 @dataclass
-class LanceDBS3Uploader(LanceDBUploader):
+class LanceDBAwsUploader(LanceDBUploader):
     upload_config: LanceDBUploaderConfig
-    connection_config: LanceDBS3ConnectionConfig
+    connection_config: LanceDBAwsConnectionConfig
     connector_type: str = CONNECTOR_TYPE
 
 
 lancedb_aws_destination_entry = DestinationRegistryEntry(
-    connection_config=LanceDBS3ConnectionConfig,
-    uploader=LanceDBS3Uploader,
+    connection_config=LanceDBAwsConnectionConfig,
+    uploader=LanceDBAwsUploader,
     uploader_config=LanceDBUploaderConfig,
     upload_stager_config=LanceDBUploadStagerConfig,
     upload_stager=LanceDBUploadStager,

unstructured_ingest/v2/processes/connectors/lancedb/cloud.py

@@ -0,0 +1,42 @@
+from dataclasses import dataclass
+
+from pydantic import Field, Secret
+
+from unstructured_ingest.v2.interfaces.connector import AccessConfig
+from unstructured_ingest.v2.processes.connector_registry import DestinationRegistryEntry
+from unstructured_ingest.v2.processes.connectors.lancedb.lancedb import (
+    LanceDBRemoteConnectionConfig,
+    LanceDBUploader,
+    LanceDBUploaderConfig,
+    LanceDBUploadStager,
+    LanceDBUploadStagerConfig,
+)
+
+CONNECTOR_TYPE = "lancedb_cloud"
+
+
+class LanceDBCloudAccessConfig(AccessConfig):
+    api_key: str = Field(description="Api key associated with LanceDb cloud")
+
+
+class LanceDBCloudConnectionConfig(LanceDBRemoteConnectionConfig):
+    access_config: Secret[LanceDBCloudAccessConfig]
+
+    def get_storage_options(self) -> dict:
+        return {**self.access_config.get_secret_value().model_dump(), "timeout": self.timeout}
+
+
+@dataclass
+class LanceDBCloudUploader(LanceDBUploader):
+    upload_config: LanceDBUploaderConfig
+    connection_config: LanceDBCloudConnectionConfig
+    connector_type: str = CONNECTOR_TYPE
+
+
+lancedb_cloud_destination_entry = DestinationRegistryEntry(
+    connection_config=LanceDBCloudConnectionConfig,
+    uploader=LanceDBCloudUploader,
+    uploader_config=LanceDBUploaderConfig,
+    upload_stager_config=LanceDBUploadStagerConfig,
+    upload_stager=LanceDBUploadStager,
+)

unstructured_ingest/v2/processes/connectors/milvus.py

@@ -8,7 +8,7 @@ import pandas as pd
 from dateutil import parser
 from pydantic import Field, Secret
 
-from unstructured_ingest.error import WriteError
+from unstructured_ingest.error import DestinationConnectionError, WriteError
 from unstructured_ingest.utils.data_prep import flatten_dict
 from unstructured_ingest.utils.dep_check import requires_dependencies
 from unstructured_ingest.v2.constants import RECORD_ID_LABEL
@@ -66,7 +66,6 @@ class MilvusConnectionConfig(ConnectionConfig):
 
 
 class MilvusUploadStagerConfig(UploadStagerConfig):
-
     fields_to_include: Optional[list[str]] = None
     """If set - list of fields to include in the output.
     Unspecified fields are removed from the elements.
@@ -174,6 +173,14 @@ class MilvusUploader(Uploader):
     upload_config: MilvusUploaderConfig
     connector_type: str = CONNECTOR_TYPE
 
+    @DestinationConnectionError.wrap
+    def precheck(self):
+        with self.get_client() as client:
+            if not client.has_collection(self.upload_config.collection_name):
+                raise DestinationConnectionError(
+                    f"Collection '{self.upload_config.collection_name}' does not exist"
+                )
+
     @contextmanager
     def get_client(self) -> Generator["MilvusClient", None, None]:
         client = self.connection_config.get_client()
@@ -218,7 +225,6 @@ class MilvusUploader(Uploader):
             f"db in collection {self.upload_config.collection_name}"
         )
         with self.get_client() as client:
-
             try:
                 res = client.insert(collection_name=self.upload_config.collection_name, data=data)
             except MilvusException as milvus_exception:

unstructured_ingest/v2/processes/connectors/weaviate/cloud.py

@@ -55,10 +55,11 @@ class CloudWeaviateConnectionConfig(WeaviateConnectionConfig):
             "client_secret": access_config.client_secret is not None,
             "client_password": access_config.password is not None and self.username is not None,
         }
-        if len(auths) == 0:
+        existing_auths = [auth_method for auth_method, flag in auths.items() if flag]
+
+        if len(existing_auths) == 0:
             raise ValueError("No auth values provided and anonymous is False")
-        if len(auths) > 1:
-            existing_auths = [auth_method for auth_method, flag in auths.items() if flag]
+        if len(existing_auths) > 1:
             raise ValueError(
                 "Multiple auth values provided, only one approach can be used: {}".format(
                     ", ".join(existing_auths)

{unstructured_ingest-0.3.2.dist-info → unstructured_ingest-0.3.3.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: unstructured-ingest
-Version: 0.3.2
+Version: 0.3.3
 Summary: A library that prepares raw documents for downstream ML tasks.
 Home-page: https://github.com/Unstructured-IO/unstructured-ingest
 Author: Unstructured Technologies
@@ -22,20 +22,20 @@ Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
 Requires-Python: >=3.9.0,<3.13
 Description-Content-Type: text/markdown
 License-File: LICENSE.md
-Requires-Dist: python-dateutil
+Requires-Dist: dataclasses-json
 Requires-Dist: pydantic>=2.7
-Requires-Dist: opentelemetry-sdk
-Requires-Dist: click
-Requires-Dist: tqdm
 Requires-Dist: pandas
-Requires-Dist: dataclasses-json
+Requires-Dist: tqdm
+Requires-Dist: python-dateutil
+Requires-Dist: click
+Requires-Dist: opentelemetry-sdk
 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
@@ -78,8 +78,8 @@ Requires-Dist: sentence-transformers; extra == "embed-huggingface"
 Provides-Extra: embed-mixedbreadai
 Requires-Dist: mixedbread-ai; extra == "embed-mixedbreadai"
 Provides-Extra: embed-octoai
-Requires-Dist: tiktoken; extra == "embed-octoai"
 Requires-Dist: openai; extra == "embed-octoai"
+Requires-Dist: tiktoken; extra == "embed-octoai"
 Provides-Extra: embed-vertexai
 Requires-Dist: vertexai; extra == "embed-vertexai"
 Provides-Extra: embed-voyageai
@@ -91,8 +91,8 @@ Requires-Dist: bs4; extra == "gcs"
 Requires-Dist: gcsfs; 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
@@ -106,6 +106,8 @@ Provides-Extra: kafka
 Requires-Dist: confluent-kafka; extra == "kafka"
 Provides-Extra: kdbai
 Requires-Dist: kdbai-client>=1.4.0; extra == "kdbai"
+Provides-Extra: lancedb
+Requires-Dist: lancedb; extra == "lancedb"
 Provides-Extra: md
 Requires-Dist: unstructured[md]; extra == "md"
 Provides-Extra: milvus
@@ -122,12 +124,12 @@ Requires-Dist: httpx; extra == "notion"
 Provides-Extra: odt
 Requires-Dist: unstructured[odt]; extra == "odt"
 Provides-Extra: onedrive
-Requires-Dist: bs4; extra == "onedrive"
 Requires-Dist: msal; extra == "onedrive"
 Requires-Dist: Office365-REST-Python-Client; extra == "onedrive"
+Requires-Dist: bs4; extra == "onedrive"
 Provides-Extra: openai
-Requires-Dist: tiktoken; extra == "openai"
 Requires-Dist: openai; extra == "openai"
+Requires-Dist: tiktoken; extra == "openai"
 Provides-Extra: opensearch
 Requires-Dist: opensearch-py; extra == "opensearch"
 Provides-Extra: org
@@ -171,8 +173,8 @@ Requires-Dist: singlestoredb; extra == "singlestore"
 Provides-Extra: slack
 Requires-Dist: slack-sdk[optional]; extra == "slack"
 Provides-Extra: snowflake
-Requires-Dist: psycopg2-binary; extra == "snowflake"
 Requires-Dist: snowflake-connector-python; extra == "snowflake"
+Requires-Dist: psycopg2-binary; extra == "snowflake"
 Provides-Extra: togetherai
 Requires-Dist: together; extra == "togetherai"
 Provides-Extra: tsv

{unstructured_ingest-0.3.2.dist-info → unstructured_ingest-0.3.3.dist-info}/RECORD

@@ -10,8 +10,8 @@ test/integration/connectors/test_azure_ai_search.py,sha256=dae4GifRiKue5YpsxworD
 test/integration/connectors/test_confluence.py,sha256=xcPmZ_vi_pkCt-tUPn10P49FH9i_9YUbrAPO6fYk5rU,3521
 test/integration/connectors/test_delta_table.py,sha256=GSzWIkbEUzOrRPt2F1uO0dabcp7kTFDj75BhhI2y-WU,6856
 test/integration/connectors/test_kafka.py,sha256=j7jsNWZumNBv9v-5Bpx8geUUXpxxad5EuA4CMRsl4R8,7104
-test/integration/connectors/test_lancedb.py,sha256=O3YF6MVBkCsCgklXCJe8Kpy8aKGfafASVH4PspmpcYs,7628
-test/integration/connectors/test_milvus.py,sha256=CVmYw9iEeKT_0OtShxye2E6i1LbWzzDA8JtwJRkYQlA,4763
+test/integration/connectors/test_lancedb.py,sha256=8hRlqw3zYOcFCu6PPlejquSvvEM_3OEBzKTQbNm_Zmg,7635
+test/integration/connectors/test_milvus.py,sha256=p4UujDr_tsRaQDmhDmDZp38t8oSFm7hrTqiq6NNuhGo,5933
 test/integration/connectors/test_mongodb.py,sha256=YeS_DUnVYN02F76j87W8RhXGHnJMzQYb3n-L1-oWGXI,12254
 test/integration/connectors/test_onedrive.py,sha256=KIkBwKh1hnv203VCL2UABnDkS_bP4NxOFm1AL8EPGLA,3554
 test/integration/connectors/test_pinecone.py,sha256=X10OWZ6IrO6YyhuR3ydMAZOQq3u2f5u_lCjKNYUUcnI,7558
@@ -35,6 +35,7 @@ test/integration/connectors/utils/docker_compose.py,sha256=GVTB6Cel05c0VQ2n4AwkQ
 test/integration/connectors/utils/validation.py,sha256=SwvPVuHjJxTo8xEUwnuL9FZNpu3sZZ8iouOz5xh_kB8,14272
 test/integration/connectors/weaviate/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 test/integration/connectors/weaviate/conftest.py,sha256=6Q6QdrLJmGHowRFSmoVSzup2EX6qASfS2Z5tqlpTm9M,387
+test/integration/connectors/weaviate/test_cloud.py,sha256=07VxNRxWWcgTstFfpoZ1FlVnEhcBnQlo5nosWKjKz_4,979
 test/integration/connectors/weaviate/test_local.py,sha256=SK6iEwQUKiCd0X99BEk8GlQoLaCcJcFPt09NN526Ct0,4508
 test/integration/embedders/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 test/integration/embedders/conftest.py,sha256=B2W771RbijR7G_GybsCzRyIvOzXqzbKZdRIlNDd5AGY,334
@@ -54,7 +55,7 @@ test/unit/test_chunking_utils.py,sha256=0iPwfnMPpyTm-yOE0BXMnEQQP4iguS6NhOqgMQU5
 test/unit/test_error.py,sha256=RflmngCdFNKOLXVfLnUdNfY3Mfg3k7DTEzfIl0B-syU,840
 test/unit/test_interfaces.py,sha256=XNj8qasc1ltaeUv-2y31rv7R9xquo0rgRrMvBZoNZLw,9623
 test/unit/test_logger.py,sha256=0SKndXE_VRd8XmUHkrj7zuBQHZscXx3ZQllMEOvtF9Y,2380
-test/unit/test_utils.py,sha256=xJ9WGpHBihWpQWvIzd6z99UIdZJba8U7c31h3q6C9To,4800
+test/unit/test_utils.py,sha256=Q6mp9YZPah8z3-2lreyRbmAc7m2Y_w26_N9vocSInoA,5421
 test/unit/embed/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 test/unit/embed/test_mixedbreadai.py,sha256=XFNJDP5pIgF3eQYwBiuEWmH3zZWx72Wpwyv-Q4m0DJg,1332
 test/unit/embed/test_octoai.py,sha256=Ha9EgAW64Q45hFj51tToe8RyKXWXwqAkdDqSFDMu37Q,831
@@ -81,7 +82,7 @@ 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=Js7MXQhyIj1akVjPNsLkmZxqoOHDGOr2opEPgFOSTZQ,42
+unstructured_ingest/__version__.py,sha256=p5uBTX3-kWJF_Qc2XUwgA0BcGSwYkkJo-kqLi89Vqo4,42
 unstructured_ingest/error.py,sha256=qDncnJgbf5ils956RcO2CGlAKYDT5OaEM9Clv1JVTNc,1448
 unstructured_ingest/interfaces.py,sha256=OYVUP0bzBJpT-Lz92BDyz_hLBvyfxkuSwWHhUdnUayA,31493
 unstructured_ingest/logger.py,sha256=S5nSqGcABoQyeicgRnBQFjDScCaTvFVivOCvbo-laL0,4479
@@ -342,7 +343,7 @@ unstructured_ingest/utils/compression.py,sha256=NNiY-2S2Gf3at7zC1PYxMijaEza9vVSz
 unstructured_ingest/utils/data_prep.py,sha256=IDAedOSBdgZpD9IY4tLJT-rmKGV7GHtU6KRj6VM-_tE,4666
 unstructured_ingest/utils/dep_check.py,sha256=SXXcUna2H0RtxA6j1S2NGkvQa9JP2DujWhmyBa7776Y,2400
 unstructured_ingest/utils/google_filetype.py,sha256=YVspEkiiBrRUSGVeVbsavvLvTmizdy2e6TsjigXTSRU,468
-unstructured_ingest/utils/string_and_date_utils.py,sha256=LwcbLmWpwt1zEabLlyUd5kIf9oOWcZxsRzxDglLCMeU,1375
+unstructured_ingest/utils/string_and_date_utils.py,sha256=kijtPlGAbH376vVjFSo5H_ZhW-FEcMC2sCNsSNwDOjo,1729
 unstructured_ingest/utils/table.py,sha256=aWjcowDVSClNpEAdR6PY3H7khKu4T6T3QqQE6GjmQ_M,3469
 unstructured_ingest/v2/__init__.py,sha256=U4S_2y3zgLZVfMenHRaJFBW8yqh2mUBuI291LGQVOJ8,35
 unstructured_ingest/v2/constants.py,sha256=pDspTYz-nEojHBqrZNfssGEiujmVa02pIWL63PQP9sU,103
@@ -394,7 +395,7 @@ 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=8M3aYYNbOkS2SYG2B_HLHMgX4V69-Oz1VqpQcRQMiVg,5167
 unstructured_ingest/v2/processes/connectors/airtable.py,sha256=eeZJe-bBNxt5Sa-XEFCdcGeJCguJU5WN2Mv9kLp5dVQ,8917
-unstructured_ingest/v2/processes/connectors/astradb.py,sha256=zsIElFNNqVCXcLqBw6C8bRoyPQDrGNPkTWeA0FYYO94,14703
+unstructured_ingest/v2/processes/connectors/astradb.py,sha256=QTUQ-cv_iZi9eaXRRHQNKhtgFn-Pi20AXdSVaDFg9DM,15498
 unstructured_ingest/v2/processes/connectors/azure_ai_search.py,sha256=-6IijSWGqj-85vD0c4l5wdMHp-LF371jO8j53PPRB4I,12002
 unstructured_ingest/v2/processes/connectors/chroma.py,sha256=skrxRPHZ8y3JxNa0dt5SVitHiDQ5WVxLvY_kh2-QUrQ,8029
 unstructured_ingest/v2/processes/connectors/confluence.py,sha256=qQApDcmPBGg4tHXwSOj4JPkAbrO9GQ4NRlaETjhp25U,7003
@@ -404,7 +405,7 @@ unstructured_ingest/v2/processes/connectors/gitlab.py,sha256=yBgCeLy9iCVI8bBDcHH
 unstructured_ingest/v2/processes/connectors/google_drive.py,sha256=EEwXK1Anlu-eXl5qxmdDIqPYW7eMSez6WGlTPG2vSn8,13121
 unstructured_ingest/v2/processes/connectors/kdbai.py,sha256=8bGHbZctJ_Tl1AUSMnI7CCZ7CgEtTRVcRuvlB1HPlqQ,5907
 unstructured_ingest/v2/processes/connectors/local.py,sha256=a3stgnIkhBbXPIQD0O-RaRM-Eb-szHj9Yy4Fz881-9c,6723
-unstructured_ingest/v2/processes/connectors/milvus.py,sha256=Bzv2fa852BcM4_Pr-I_DPvLmjPoXv0Z7BeEA8qSKCDc,9725
+unstructured_ingest/v2/processes/connectors/milvus.py,sha256=3sV0Yv2vYMLyxszKCqAqlMcHHJSBR-GGbaZf1nvobLE,10089
 unstructured_ingest/v2/processes/connectors/mongodb.py,sha256=XLuprTCY0D9tAh_qn81MjJrDN9YaNqMlKe7BJl3eTZc,14998
 unstructured_ingest/v2/processes/connectors/onedrive.py,sha256=heZMtOIrCySi552ldIk8iH0pSRXZ0W2LeD-CcNOwCFQ,15979
 unstructured_ingest/v2/processes/connectors/outlook.py,sha256=KgNGM8hImRhy6_SpswRP2VwRD4VOrqqJoySgxf2oduI,9290
@@ -435,9 +436,10 @@ unstructured_ingest/v2/processes/connectors/kafka/__init__.py,sha256=mQJ9Ex-QCfh
 unstructured_ingest/v2/processes/connectors/kafka/cloud.py,sha256=qprsfI8VH0mVTa1MOCpa2D4coyopinQ5ag2KXcAecXE,3296
 unstructured_ingest/v2/processes/connectors/kafka/kafka.py,sha256=qEv_yaG94KekFtfS06KgpTTbqeJkje0hn5uOjsMMngw,9414
 unstructured_ingest/v2/processes/connectors/kafka/local.py,sha256=vwLZjvc_C17zOqcrzic0aIoPwS98sqYiwiMknw2IcK4,2586
-unstructured_ingest/v2/processes/connectors/lancedb/__init__.py,sha256=lHUPCOiyOGu1IME1QiyFBZaB8z8e3bP8Y8TkqKs32Qk,906
-unstructured_ingest/v2/processes/connectors/lancedb/aws.py,sha256=yR8V4O-oI_nUKJtHTLxhteEJpPDPn-_d2IkkXvgThJ0,1406
+unstructured_ingest/v2/processes/connectors/lancedb/__init__.py,sha256=LW37xZrn48JeHluRNulLTreUPdaF-ZU81F7MCUHcCv8,1253
+unstructured_ingest/v2/processes/connectors/lancedb/aws.py,sha256=eeXWsh8UeVm1Ur53C4MEnpLplfO8U91KYgk--0kk5pE,1413
 unstructured_ingest/v2/processes/connectors/lancedb/azure.py,sha256=Ms5vQVRIpTF1Q2qBl_bET9wbgaf4diPaH-iR8kJlr4E,1461
+unstructured_ingest/v2/processes/connectors/lancedb/cloud.py,sha256=BFy0gW2OZ_qaZJM97m-tNsFaJPi9zOKrrd2y4thcNP0,1341
 unstructured_ingest/v2/processes/connectors/lancedb/gcp.py,sha256=p5BPaFtS3y3Yh8PIr3tUqsAXrUYu4QYYAWQNh5W2ucE,1361
 unstructured_ingest/v2/processes/connectors/lancedb/lancedb.py,sha256=7WIShs2V3dpN6wUhDTt1j2rvdiPp6yopbh7XYkb9T3s,5129
 unstructured_ingest/v2/processes/connectors/lancedb/local.py,sha256=_7-6iO6B60gAWwJUUrmlsRzYMFIBeZgu_QT3mhw5L0I,1272
@@ -453,13 +455,13 @@ unstructured_ingest/v2/processes/connectors/sql/snowflake.py,sha256=jl524VudwmFK
 unstructured_ingest/v2/processes/connectors/sql/sql.py,sha256=LFzGeAUagLknK07DsXg2oSG7ZAgR6VqT9wfI_tYlHUg,14782
 unstructured_ingest/v2/processes/connectors/sql/sqlite.py,sha256=9605K36nQ5-gBxzt1daYKYotON1SE85RETusqCJrbdk,5230
 unstructured_ingest/v2/processes/connectors/weaviate/__init__.py,sha256=eXamSnQdzzMvt62z80B8nmlkwDKO-Pogln_K_zLz53A,1067
-unstructured_ingest/v2/processes/connectors/weaviate/cloud.py,sha256=2g1Fm2J0ppfy2jCw4b5YtrsWrSD3VcrAaqiE7FlpIAg,6236
+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=ln1p9ahFTaT-qsL7p4bgw_IqnU60As_l6vVAqUWyQVE,11655
-unstructured_ingest-0.3.2.dist-info/LICENSE.md,sha256=SxkKP_62uIAKb9mb1eH7FH4Kn2aYT09fgjKpJt5PyTk,11360
-unstructured_ingest-0.3.2.dist-info/METADATA,sha256=rqTWqewB8eIrgrHJ-8AsNtehy35eSHKseCsveXTwN3Y,7326
-unstructured_ingest-0.3.2.dist-info/WHEEL,sha256=GV9aMThwP_4oNCtvEC2ec3qUYutgWeAzklro_0m4WJQ,91
-unstructured_ingest-0.3.2.dist-info/entry_points.txt,sha256=gUAAFnjFPnBgThJSEbw0N5ZjxtaKlT1s9e05_arQrNw,70
-unstructured_ingest-0.3.2.dist-info/top_level.txt,sha256=DMuDMHZRMdeay8v8Kdi855muIv92F0OkutvBCaBEW6M,25
-unstructured_ingest-0.3.2.dist-info/RECORD,,
+unstructured_ingest-0.3.3.dist-info/LICENSE.md,sha256=SxkKP_62uIAKb9mb1eH7FH4Kn2aYT09fgjKpJt5PyTk,11360
+unstructured_ingest-0.3.3.dist-info/METADATA,sha256=AEumzINrBNXXeBEBQiIB8309_9OkIWhLeo7Giqzl1ew,7393
+unstructured_ingest-0.3.3.dist-info/WHEEL,sha256=GV9aMThwP_4oNCtvEC2ec3qUYutgWeAzklro_0m4WJQ,91
+unstructured_ingest-0.3.3.dist-info/entry_points.txt,sha256=gUAAFnjFPnBgThJSEbw0N5ZjxtaKlT1s9e05_arQrNw,70
+unstructured_ingest-0.3.3.dist-info/top_level.txt,sha256=DMuDMHZRMdeay8v8Kdi855muIv92F0OkutvBCaBEW6M,25
+unstructured_ingest-0.3.3.dist-info/RECORD,,