unstructured-ingest 0.0.0__py3-none-any.whl → 0.0.1__py3-none-any.whl

unstructured_ingest/__version__.py

@@ -1 +1 @@
-__version__ = "0.0.0"  # pragma: no cover
+__version__ = "0.0.1"  # pragma: no cover

unstructured_ingest/connector/notion/helpers.py

@@ -5,7 +5,6 @@ from typing import List, Optional, Tuple
 from urllib.parse import urlparse
 from uuid import UUID
 
-import unstructured.ingest.connector.notion.types.blocks as notion_blocks
 from htmlBuilder.attributes import Style, Type
 from htmlBuilder.tags import (
     Body,
@@ -23,6 +22,7 @@ from htmlBuilder.tags import (
 )
 from notion_client.errors import APIResponseError
 
+import unstructured_ingest.connector.notion.types.blocks as notion_blocks
 from unstructured_ingest.connector.notion.client import Client
 from unstructured_ingest.connector.notion.interfaces import BlockBase
 from unstructured_ingest.connector.notion.types.block import Block

unstructured_ingest/logger.py

@@ -3,7 +3,7 @@ import json
 import logging
 import typing as t
 
-logger = logging.getLogger("unstructured.ingest")
+logger = logging.getLogger("unstructured_ingest")
 
 
 def default_is_data_sensitive(k: str, v: t.Any) -> bool:
@@ -119,7 +119,7 @@ def ingest_log_streaming_init(level: int) -> None:
 
 def make_default_logger(level: int) -> logging.Logger:
     """Return a custom logger."""
-    logger = logging.getLogger("unstructured.ingest")
+    logger = logging.getLogger("unstructured_ingest")
     handler = logging.StreamHandler()
     handler.name = "ingest_log_handler"
     formatter = SensitiveFormatter("%(asctime)s %(processName)-10s %(levelname)-8s %(message)s")

unstructured_ingest/v2/cli/cmds/__init__.py

@@ -15,6 +15,7 @@ from .fsspec.s3 import s3_dest_cmd, s3_src_cmd
 from .fsspec.sftp import sftp_dest_cmd, sftp_src_cmd
 from .google_drive import google_drive_src_cmd
 from .local import local_dest_cmd, local_src_cmd
+from .milvus import milvus_dest_cmd
 from .mongodb import mongodb_dest_cmd
 from .onedrive import onedrive_drive_src_cmd
 from .opensearch import opensearch_dest_cmd, opensearch_src_cmd
@@ -60,6 +61,7 @@ dest_cmds = [
     elasticsearch_dest_cmd,
     gcs_dest_cmd,
     local_dest_cmd,
+    milvus_dest_cmd,
     opensearch_dest_cmd,
     pinecone_dest_cmd,
     s3_dest_cmd,

unstructured_ingest/v2/cli/cmds/milvus.py

@@ -0,0 +1,72 @@
+from dataclasses import dataclass
+
+import click
+
+from unstructured_ingest.v2.cli.base import DestCmd
+from unstructured_ingest.v2.cli.interfaces import CliConfig
+from unstructured_ingest.v2.processes.connectors.milvus import CONNECTOR_TYPE
+
+
+@dataclass
+class MilvusCliConnectionConfig(CliConfig):
+    @staticmethod
+    def get_cli_options() -> list[click.Option]:
+        options = [
+            click.Option(
+                ["--uri"],
+                required=False,
+                type=str,
+                default=None,
+                help="Milvus uri, eg 'http://localhost:19530",
+            ),
+            click.Option(
+                ["--user"],
+                required=False,
+                type=str,
+                default=None,
+                help="Milvus user",
+            ),
+            click.Option(
+                ["--password"],
+                required=False,
+                type=str,
+                default=None,
+                help="Milvus password",
+            ),
+            click.Option(
+                ["--db-name"],
+                required=False,
+                type=str,
+                default=None,
+                help="Milvus database name",
+            ),
+        ]
+        return options
+
+
+@dataclass
+class MilvusCliUploaderConfig(CliConfig):
+    @staticmethod
+    def get_cli_options() -> list[click.Option]:
+        options = [
+            click.Option(
+                ["--collection-name"],
+                required=True,
+                type=str,
+                help="Milvus collections to write to",
+            ),
+            click.Option(
+                ["--num-of-processes"],
+                type=click.IntRange(min=1),
+                default=4,
+                help="number of processes to use when writing to support parallel writes",
+            ),
+        ]
+        return options
+
+
+milvus_dest_cmd = DestCmd(
+    cmd_name=CONNECTOR_TYPE,
+    connection_config=MilvusCliConnectionConfig,
+    uploader_config=MilvusCliUploaderConfig,
+)

unstructured_ingest/v2/logger.py

@@ -5,7 +5,7 @@ from logging import Formatter, Logger, StreamHandler, getLevelName, getLogger
 from typing import Any, Callable
 
 log_level = os.getenv("INGEST_LOG_LEVEL", "INFO")
-LOGGER_NAME = "unstructured.ingest.v2"
+LOGGER_NAME = "unstructured_ingest.v2"
 
 
 def default_is_data_sensitive(k: str, v: Any) -> bool:

unstructured_ingest/v2/pipeline/interfaces.py

@@ -146,6 +146,8 @@ class PipelineStep(ABC):
             logger.error(f"Exception raised while running {self.identifier}", exc_info=e)
             if "file_data_path" in kwargs:
                 self.context.status[kwargs["file_data_path"]] = {self.identifier: str(e)}
+            else:
+                self.context.status[self.identifier] = {"step_error": str(e)}
             if self.context.raise_on_error:
                 raise e
             return None
@@ -158,6 +160,8 @@ class PipelineStep(ABC):
             logger.error(f"Exception raised while running {self.identifier}", exc_info=e)
             if "file_data_path" in kwargs:
                 self.context.status[kwargs["file_data_path"]] = {self.identifier: str(e)}
+            else:
+                self.context.status[self.identifier] = {"step_error": str(e)}
             if self.context.raise_on_error:
                 raise e
             return None

unstructured_ingest/v2/processes/connectors/__init__.py

@@ -1,7 +1,6 @@
 from __future__ import annotations
 
-import unstructured.ingest.v2.processes.connectors.fsspec  # noqa: F401
-
+import unstructured_ingest.v2.processes.connectors.fsspec  # noqa: F401
 from unstructured_ingest.v2.processes.connector_registry import (
     add_destination_entry,
     add_source_entry,
@@ -19,6 +18,8 @@ from .google_drive import CONNECTOR_TYPE as GOOGLE_DRIVE_CONNECTOR_TYPE
 from .google_drive import google_drive_source_entry
 from .local import CONNECTOR_TYPE as LOCAL_CONNECTOR_TYPE
 from .local import local_destination_entry, local_source_entry
+from .milvus import CONNECTOR_TYPE as MILVUS_CONNECTOR_TYPE
+from .milvus import milvus_destination_entry
 from .mongodb import CONNECTOR_TYPE as MONGODB_CONNECTOR_TYPE
 from .mongodb import mongodb_destination_entry
 from .onedrive import CONNECTOR_TYPE as ONEDRIVE_CONNECTOR_TYPE
@@ -75,3 +76,4 @@ add_source_entry(source_type=SHAREPOINT_CONNECTOR_TYPE, entry=sharepoint_source_
 add_destination_entry(
     destination_type=SINGLESTORE_CONNECTOR_TYPE, entry=singlestore_destination_entry
 )
+add_destination_entry(destination_type=MILVUS_CONNECTOR_TYPE, entry=milvus_destination_entry)

unstructured_ingest/v2/processes/connectors/milvus.py

@@ -0,0 +1,200 @@
+import json
+import multiprocessing as mp
+from dataclasses import dataclass, field
+from pathlib import Path
+from typing import TYPE_CHECKING, Any, Optional, Union
+
+import pandas as pd
+from dateutil import parser
+
+from unstructured_ingest.enhanced_dataclass import enhanced_field
+from unstructured_ingest.error import WriteError
+from unstructured_ingest.utils.data_prep import flatten_dict
+from unstructured_ingest.utils.dep_check import requires_dependencies
+from unstructured_ingest.v2.interfaces import (
+    AccessConfig,
+    ConnectionConfig,
+    FileData,
+    UploadContent,
+    Uploader,
+    UploaderConfig,
+    UploadStager,
+    UploadStagerConfig,
+)
+from unstructured_ingest.v2.logger import logger
+from unstructured_ingest.v2.processes.connector_registry import (
+    DestinationRegistryEntry,
+)
+
+if TYPE_CHECKING:
+    from pymilvus import MilvusClient
+
+CONNECTOR_TYPE = "milvus"
+
+
+@dataclass
+class MilvusAccessConfig(AccessConfig):
+    password: Optional[str] = None
+    token: Optional[str] = None
+
+
+@dataclass
+class MilvusConnectionConfig(ConnectionConfig):
+    access_config: MilvusAccessConfig = enhanced_field(
+        sensitive=True, default_factory=lambda: MilvusAccessConfig()
+    )
+    uri: Optional[str] = None
+    user: Optional[str] = None
+    db_name: Optional[str] = None
+
+    def get_connection_kwargs(self) -> dict[str, Any]:
+        access_config_dict = self.access_config.to_dict()
+        connection_config_dict = self.to_dict()
+        connection_config_dict.pop("access_config", None)
+        connection_config_dict.update(access_config_dict)
+        # Drop any that were not set explicitly
+        connection_config_dict = {k: v for k, v in connection_config_dict.items() if v is not None}
+        return connection_config_dict
+
+    @requires_dependencies(["pymilvus"], extras="milvus")
+    def get_client(self) -> "MilvusClient":
+        from pymilvus import MilvusClient
+
+        return MilvusClient(**self.get_connection_kwargs())
+
+
+@dataclass
+class MilvusUploadStagerConfig(UploadStagerConfig):
+    pass
+
+
+@dataclass
+class MilvusUploadStager(UploadStager):
+    upload_stager_config: MilvusUploadStagerConfig = field(
+        default_factory=lambda: MilvusUploadStagerConfig()
+    )
+
+    @staticmethod
+    def parse_date_string(date_string: str) -> float:
+        try:
+            timestamp = float(date_string)
+            return timestamp
+        except ValueError:
+            pass
+        return parser.parse(date_string).timestamp()
+
+    @classmethod
+    def conform_dict(cls, data: dict) -> None:
+        datetime_columns = [
+            "data_source_date_created",
+            "data_source_date_modified",
+            "data_source_date_processed",
+            "last_modified",
+        ]
+
+        json_dumps_fields = ["languages", "data_source_permissions_data"]
+
+        # TODO: milvus sdk doesn't seem to support defaults via the schema yet,
+        #  remove once that gets updated
+        defaults = {"is_continuation": False}
+
+        if metadata := data.pop("metadata", None):
+            data.update(flatten_dict(metadata, keys_to_omit=["data_source_record_locator"]))
+        for datetime_column in datetime_columns:
+            if datetime_column in data:
+                data[datetime_column] = cls.parse_date_string(data[datetime_column])
+        for json_dumps_field in json_dumps_fields:
+            if json_dumps_field in data:
+                data[json_dumps_field] = json.dumps(data[json_dumps_field])
+        for default in defaults:
+            if default not in data:
+                data[default] = defaults[default]
+
+    def run(
+        self,
+        elements_filepath: Path,
+        file_data: FileData,
+        output_dir: Path,
+        output_filename: str,
+        **kwargs: Any,
+    ) -> Path:
+        with open(elements_filepath) as elements_file:
+            elements_contents: list[dict[str, Any]] = json.load(elements_file)
+        for element in elements_contents:
+            self.conform_dict(data=element)
+
+        output_path = Path(output_dir) / Path(f"{output_filename}.json")
+        output_path.parent.mkdir(parents=True, exist_ok=True)
+        with output_path.open("w") as output_file:
+            json.dump(elements_contents, output_file, indent=2)
+        return output_path
+
+
+@dataclass
+class MilvusUploaderConfig(UploaderConfig):
+    collection_name: str
+    num_of_processes: int = 4
+
+
+@dataclass
+class MilvusUploader(Uploader):
+    connection_config: MilvusConnectionConfig
+    upload_config: MilvusUploaderConfig
+    connector_type: str = CONNECTOR_TYPE
+
+    def upload(self, content: UploadContent) -> None:
+        file_extension = content.path.suffix
+        if file_extension == ".json":
+            self.upload_json(content=content)
+        elif file_extension == ".csv":
+            self.upload_csv(content=content)
+        else:
+            raise ValueError(f"Unsupported file extension: {file_extension}")
+
+    @requires_dependencies(["pymilvus"], extras="milvus")
+    def insert_results(self, data: Union[dict, list[dict]]):
+        from pymilvus import MilvusException
+
+        logger.debug(
+            f"uploading {len(data)} entries to {self.connection_config.db_name} "
+            f"db in collection {self.upload_config.collection_name}"
+        )
+        client = self.connection_config.get_client()
+
+        try:
+            res = client.insert(collection_name=self.upload_config.collection_name, data=data)
+        except MilvusException as milvus_exception:
+            raise WriteError("failed to upload records to milvus") from milvus_exception
+        if "err_count" in res and isinstance(res["err_count"], int) and res["err_count"] > 0:
+            err_count = res["err_count"]
+            raise WriteError(f"failed to upload {err_count} docs")
+
+    def upload_csv(self, content: UploadContent) -> None:
+        df = pd.read_csv(content.path)
+        data = df.to_dict(orient="records")
+        self.insert_results(data=data)
+
+    def upload_json(self, content: UploadContent) -> None:
+        with content.path.open("r") as file:
+            data: list[dict] = json.load(file)
+        self.insert_results(data=data)
+
+    def run(self, contents: list[UploadContent], **kwargs: Any) -> None:
+        if self.upload_config.num_of_processes == 1:
+            for content in contents:
+                self.upload(content=content)
+
+        else:
+            with mp.Pool(
+                processes=self.upload_config.num_of_processes,
+            ) as pool:
+                pool.map(self.upload, contents)
+
+
+milvus_destination_entry = DestinationRegistryEntry(
+    connection_config=MilvusConnectionConfig,
+    uploader=MilvusUploader,
+    uploader_config=MilvusUploaderConfig,
+    upload_stager=MilvusUploadStager,
+    upload_stager_config=MilvusUploadStagerConfig,
+)

unstructured_ingest/v2/processes/connectors/pinecone.py

@@ -5,10 +5,6 @@ from dataclasses import dataclass, field
 from pathlib import Path
 from typing import TYPE_CHECKING, Any, Optional
 
-from unstructured.ingest.v2.logger import logger
-from unstructured.ingest.v2.processes.connector_registry import (
-    DestinationRegistryEntry,
-)
 from unstructured.staging.base import flatten_dict
 from unstructured.utils import requires_dependencies
 
@@ -24,6 +20,10 @@ from unstructured_ingest.v2.interfaces import (
     UploadStager,
     UploadStagerConfig,
 )
+from unstructured_ingest.v2.logger import logger
+from unstructured_ingest.v2.processes.connector_registry import (
+    DestinationRegistryEntry,
+)
 
 if TYPE_CHECKING:
     from pinecone import Index as PineconeIndex

{unstructured_ingest-0.0.0.dist-info → unstructured_ingest-0.0.1.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: unstructured-ingest
-Version: 0.0.0
+Version: 0.0.1
 Summary: A library that prepares raw documents for downstream ML tasks.
 Home-page: https://github.com/Unstructured-IO/unstructured-ingest
 Author: Unstructured Technologies
@@ -29,22 +29,22 @@ Requires-Dist: pyairtable ; extra == 'airtable'
 Provides-Extra: astra
 Requires-Dist: astrapy ; extra == 'astra'
 Provides-Extra: azure
-Requires-Dist: fsspec ; extra == 'azure'
 Requires-Dist: adlfs ; extra == 'azure'
+Requires-Dist: fsspec ; extra == 'azure'
 Provides-Extra: azure-cognitive-search
 Requires-Dist: azure-search-documents ; extra == 'azure-cognitive-search'
 Provides-Extra: bedrock
-Requires-Dist: boto3 ; extra == 'bedrock'
 Requires-Dist: langchain-community ; extra == 'bedrock'
+Requires-Dist: boto3 ; extra == 'bedrock'
 Provides-Extra: biomed
 Requires-Dist: bs4 ; extra == 'biomed'
 Provides-Extra: box
-Requires-Dist: boxfs ; extra == 'box'
 Requires-Dist: fsspec ; extra == 'box'
+Requires-Dist: boxfs ; extra == 'box'
 Provides-Extra: chroma
-Requires-Dist: importlib-metadata >=7.1.0 ; extra == 'chroma'
 Requires-Dist: typer <=0.9.0 ; extra == 'chroma'
 Requires-Dist: chromadb ; extra == 'chroma'
+Requires-Dist: importlib-metadata >=7.1.0 ; extra == 'chroma'
 Provides-Extra: clarifai
 Requires-Dist: clarifai ; extra == 'clarifai'
 Provides-Extra: confluence
@@ -72,21 +72,21 @@ Requires-Dist: huggingface ; extra == 'embed-huggingface'
 Requires-Dist: langchain-community ; extra == 'embed-huggingface'
 Requires-Dist: sentence-transformers ; extra == 'embed-huggingface'
 Provides-Extra: embed-octoai
-Requires-Dist: openai ; extra == 'embed-octoai'
 Requires-Dist: tiktoken ; extra == 'embed-octoai'
+Requires-Dist: openai ; extra == 'embed-octoai'
 Provides-Extra: embed-vertexai
-Requires-Dist: langchain-google-vertexai ; extra == 'embed-vertexai'
 Requires-Dist: langchain-community ; extra == 'embed-vertexai'
 Requires-Dist: langchain ; extra == 'embed-vertexai'
+Requires-Dist: langchain-google-vertexai ; extra == 'embed-vertexai'
 Provides-Extra: embed-voyageai
 Requires-Dist: langchain-voyageai ; extra == 'embed-voyageai'
 Requires-Dist: langchain ; 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: fsspec ; extra == 'gcs'
 Provides-Extra: github
 Requires-Dist: pygithub >1.58.0 ; extra == 'github'
 Provides-Extra: gitlab
@@ -94,38 +94,40 @@ 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
 Requires-Dist: confluent-kafka ; extra == 'kafka'
 Provides-Extra: md
 Requires-Dist: unstructured[md] ; extra == 'md'
+Provides-Extra: milvus
+Requires-Dist: pymilvus ; extra == 'milvus'
 Provides-Extra: mongodb
 Requires-Dist: pymongo ; extra == 'mongodb'
 Provides-Extra: msg
 Requires-Dist: unstructured[msg] ; extra == 'msg'
 Provides-Extra: notion
-Requires-Dist: notion-client ; extra == 'notion'
 Requires-Dist: htmlBuilder ; extra == 'notion'
+Requires-Dist: notion-client ; extra == 'notion'
 Provides-Extra: odt
 Requires-Dist: unstructured[odt] ; extra == 'odt'
 Provides-Extra: onedrive
-Requires-Dist: msal ; extra == 'onedrive'
 Requires-Dist: bs4 ; extra == 'onedrive'
 Requires-Dist: Office365-REST-Python-Client ; extra == 'onedrive'
+Requires-Dist: msal ; extra == 'onedrive'
 Provides-Extra: openai
-Requires-Dist: openai ; extra == 'openai'
 Requires-Dist: tiktoken ; extra == 'openai'
 Requires-Dist: langchain-community ; extra == 'openai'
+Requires-Dist: openai ; extra == 'openai'
 Provides-Extra: opensearch
 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
@@ -150,11 +152,11 @@ Requires-Dist: s3fs ; extra == 's3'
 Provides-Extra: salesforce
 Requires-Dist: simple-salesforce ; extra == 'salesforce'
 Provides-Extra: sftp
-Requires-Dist: fsspec ; 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

{unstructured_ingest-0.0.0.dist-info → unstructured_ingest-0.0.1.dist-info}/RECORD

@@ -1,9 +1,9 @@
 unstructured_ingest/__init__.py,sha256=U4S_2y3zgLZVfMenHRaJFBW8yqh2mUBuI291LGQVOJ8,35
-unstructured_ingest/__version__.py,sha256=Q49HKCptFgT2OeWLy_cQ73sq2sMFreeYIt3GaSMpXf8,42
+unstructured_ingest/__version__.py,sha256=SI019rW6paHw93e6fOWFzF9TruLom8o9HrgZsjGZvaE,42
 unstructured_ingest/error.py,sha256=qDncnJgbf5ils956RcO2CGlAKYDT5OaEM9Clv1JVTNc,1448
 unstructured_ingest/evaluate.py,sha256=R-mKLFXbVX1xQ1tjGsLHjdP-TbSSV-925IHzggW_bIg,9793
 unstructured_ingest/interfaces.py,sha256=uS8L5mS0mXD8I4XTfVlKZxAwqnpJ4yrRqn4vxWVRhQI,31107
-unstructured_ingest/logger.py,sha256=P5KVgFSRN4uSSNmf5S00zr_TdlL7uAhjxn_26tcNWxI,4480
+unstructured_ingest/logger.py,sha256=TrhyH7VbCWO5VVuhvL0yUyXxuem3b4pzbqj2uQHUwZk,4480
 unstructured_ingest/main.py,sha256=82G_7eG4PNhc_xIqj4Y_sFbDV9VI-nwSfsfJQMzovMk,169
 unstructured_ingest/processor.py,sha256=XKKrvbxsb--5cDzz4hB3-GfWZYyIjJ2ah8FpzQKF_DM,2760
 unstructured_ingest/cli/__init__.py,sha256=9kNcBOHuXON5lB1MJU9QewEhwPmId56vXqB29-kqEAA,302
@@ -104,7 +104,7 @@ unstructured_ingest/connector/fsspec/sftp.py,sha256=x2w8JGM81S_HXww7Aa-bTY1LjZSi
 unstructured_ingest/connector/notion/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 unstructured_ingest/connector/notion/client.py,sha256=vU1GE64ktEAM4b-jo8UnMAwz60KSiQ6iRI3De3ixNdI,8689
 unstructured_ingest/connector/notion/connector.py,sha256=E-t7q5XAiYP9xk-1aqIqcGwdJOH8UNgiE0HcH9Oc4i4,17475
-unstructured_ingest/connector/notion/helpers.py,sha256=5SbQbNxIenMHyxEAMfrsVsXpNcAKPHo3gwWQVi1NUOc,20702
+unstructured_ingest/connector/notion/helpers.py,sha256=jqg5-cPcrjm5G7dBF4jOjHxdJN1CI7yBeTefiX4hBoM,20702
 unstructured_ingest/connector/notion/interfaces.py,sha256=SrTT-9c0nvk0fMqVgudYF647r04AdMKi6wkIkMy7Szw,563
 unstructured_ingest/connector/notion/types/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 unstructured_ingest/connector/notion/types/block.py,sha256=AKOY-o6CTFC-caWlkLfKskMuFemH4-Vdrhv7HnRkS8w,3009
@@ -253,7 +253,7 @@ unstructured_ingest/utils/string_and_date_utils.py,sha256=hnGglD8Z626vLhH_UV4Qyb
 unstructured_ingest/utils/table.py,sha256=aWjcowDVSClNpEAdR6PY3H7khKu4T6T3QqQE6GjmQ_M,3469
 unstructured_ingest/v2/__init__.py,sha256=U4S_2y3zgLZVfMenHRaJFBW8yqh2mUBuI291LGQVOJ8,35
 unstructured_ingest/v2/example.py,sha256=qkwmpMxUlaJXdDNKQ4LlUt3XGxgTUU3CXGGO57eW5Gs,1644
-unstructured_ingest/v2/logger.py,sha256=tI_PtlibnJmv1MMajHT5GSZhQ77dv30UADEWaXWgynA,4324
+unstructured_ingest/v2/logger.py,sha256=akcghdHwpKM3CfoeFzir0zmc7R9Hk7zjquU-X-gwUIw,4324
 unstructured_ingest/v2/main.py,sha256=WFdLEqEXRy6E9_G-dF20MK2AtgX51Aan1sp_N67U2B8,172
 unstructured_ingest/v2/cli/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 unstructured_ingest/v2/cli/cli.py,sha256=qHXIs-PcvMgDZhP1AR9iDMxh8FXBMJCEDksPBfiMULE,648
@@ -264,7 +264,7 @@ unstructured_ingest/v2/cli/base/cmd.py,sha256=qVHmquVsVDoYyPByKdUTVCwAFfILMYBw5w
 unstructured_ingest/v2/cli/base/dest.py,sha256=YMbVIHmYDqvOtxZeEY93stmF2p2ImjuJts7-u-NznYw,2887
 unstructured_ingest/v2/cli/base/importer.py,sha256=nRt0QQ3qpi264-n_mR0l55C2ddM8nowTNzT1jsWaam8,1128
 unstructured_ingest/v2/cli/base/src.py,sha256=7LnZh9FgUX9rerBH6cizVtTWmM6R2sRkxatnGsxYHG0,2410
-unstructured_ingest/v2/cli/cmds/__init__.py,sha256=aOcJb2FLQaUOU-vdu4xHr5_BJQme6ADlPaRjCSHL1Io,2590
+unstructured_ingest/v2/cli/cmds/__init__.py,sha256=DWPMD6Wqus22sSoIEyTSiOJAm97aNjvdpdrXgsL4uQ0,2647
 unstructured_ingest/v2/cli/cmds/astra.py,sha256=L-GR2KSP_cFQkQm0aVcdiXmgYMJZCVKIAH794y8qT1M,2590
 unstructured_ingest/v2/cli/cmds/azure_cognitive_search.py,sha256=VTCSUYeIYKnP60lC7DeBYqoqAJnWuBZrwevCXbeIEzw,2248
 unstructured_ingest/v2/cli/cmds/chroma.py,sha256=RinNOPripk2zRYx1Rt-u-jywXbwh7JsidVia4F0-wyU,3359
@@ -272,6 +272,7 @@ unstructured_ingest/v2/cli/cmds/databricks_volumes.py,sha256=53d9A7UunJLYZFwwwHE
 unstructured_ingest/v2/cli/cmds/elasticsearch.py,sha256=joUfnV992fAwEDCtFVJaABwgpyQiWeDl1ZCBEudRtnk,5258
 unstructured_ingest/v2/cli/cmds/google_drive.py,sha256=mXozabpi8kjRFb0S7kw-xMGtEuFoVUxnvefwL5ZIPHc,2334
 unstructured_ingest/v2/cli/cmds/local.py,sha256=lGBFOVDRlrcCtPFjyk0IAYHLRWg95Kunu1Kli7t0ZK4,1899
+unstructured_ingest/v2/cli/cmds/milvus.py,sha256=PB1ib1rFGGH_-KDi1bSIO3BIiVcqSJEHCBFFrzQrnmI,1998
 unstructured_ingest/v2/cli/cmds/mongodb.py,sha256=oyV6tacuuxm3dN-AXQgbxvYJiDYo2OOWQKRSBCUGj0E,1823
 unstructured_ingest/v2/cli/cmds/onedrive.py,sha256=DKqhQyyF-swZxs3C9G5W8ECleq8sWpDbpTuiAHXukXQ,2781
 unstructured_ingest/v2/cli/cmds/opensearch.py,sha256=7zl8dUXzxs24MDRRASKfNc14IDM798qOXRl2FZdXG1I,3064
@@ -304,7 +305,7 @@ unstructured_ingest/v2/interfaces/processor.py,sha256=uHVHeKo5Gt_zFkaEXw7xgaCBDT
 unstructured_ingest/v2/interfaces/upload_stager.py,sha256=SylhDl9pK6qa7hvfrhpabCkjwE03yIlI6oM-mQnqtho,1220
 unstructured_ingest/v2/interfaces/uploader.py,sha256=bzfx3Ei4poXKu-hsgjAB4sj4jKij9CoaRSadUM5LtGk,1083
 unstructured_ingest/v2/pipeline/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-unstructured_ingest/v2/pipeline/interfaces.py,sha256=3zRcu6sc-2rnm3gALOzXA9mI39m2RUPrREZGTd9x77c,6352
+unstructured_ingest/v2/pipeline/interfaces.py,sha256=Zz76fLHNKw6BDsBSYQXiRa6CvyW91ulvZU0yw5vVQ5M,6544
 unstructured_ingest/v2/pipeline/pipeline.py,sha256=r8jRMZI2RF8GQIuTcjIFBDeFtMnqpOJmKhEriy6Vo5Y,11616
 unstructured_ingest/v2/pipeline/utils.py,sha256=oPAitfdnITqh2O8Z0uf6VOHg9BTJhitRzNmKXqTwPxg,422
 unstructured_ingest/v2/pipeline/steps/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -322,7 +323,7 @@ unstructured_ingest/v2/processes/connector_registry.py,sha256=KOrvJNNRdpBPyqFwmT
 unstructured_ingest/v2/processes/embedder.py,sha256=QjAsiXAjWtZzh6lJ4D5LsTMBD81zuMBkegXNWq-FZt0,3308
 unstructured_ingest/v2/processes/partitioner.py,sha256=f6UQoQHVKjl8rmM5J9EcuP30RTFLSLrArGdC6qh-ffE,7645
 unstructured_ingest/v2/processes/uncompress.py,sha256=x-JZYNs1zJOtRS7xNgiMyrYoAbzKM0p18O8NAl7avCA,1631
-unstructured_ingest/v2/processes/connectors/__init__.py,sha256=cuPXXs7__tztof9Z8jE4b7UPQFc6p616hH1BECzBK1M,3762
+unstructured_ingest/v2/processes/connectors/__init__.py,sha256=7QMKd8gtEJTIuK352Ho6XyoFvLLhrWIzgdu0dXwXWOE,3960
 unstructured_ingest/v2/processes/connectors/astra.py,sha256=TSI_3GHnEh3gYAC30RTG4b2eEB07agroEFmJ38GnQY4,4903
 unstructured_ingest/v2/processes/connectors/azure_cognitive_search.py,sha256=PT02ZKiJuHMrmBClxqBsyDS0aXUQYLVg02Ns2qh1hD4,7935
 unstructured_ingest/v2/processes/connectors/chroma.py,sha256=nYzNz-8oq-DN0Z4r7lHQFmlved76IaYeRvm7-EmbGUE,6998
@@ -330,10 +331,11 @@ unstructured_ingest/v2/processes/connectors/databricks_volumes.py,sha256=MTLK7Sv
 unstructured_ingest/v2/processes/connectors/elasticsearch.py,sha256=6QBvVzPk3mWj9ZqJZN7NvhcJaOO6nSLqLwU6zggP59A,14864
 unstructured_ingest/v2/processes/connectors/google_drive.py,sha256=IkLVafUu280OOoqYmdfdfMB6zlpiWjs2Z5J31ZzJOj4,12681
 unstructured_ingest/v2/processes/connectors/local.py,sha256=maAXVKpRRXj_jseC6EPLTosMgw6ll-0lnGsDdAFLWAE,6646
+unstructured_ingest/v2/processes/connectors/milvus.py,sha256=FWH4FH-zns7gh8sITg9pLYE9uKm_3GeOXJ4wjY6PMno,6776
 unstructured_ingest/v2/processes/connectors/mongodb.py,sha256=ErZWAD-su3OCRGv1h84X1PpAWleUPVZcFDEIYjtyP4E,4310
 unstructured_ingest/v2/processes/connectors/onedrive.py,sha256=WDDoFEfd8M_QBTpkGNI2zZGZZ_CR1rQiCsBWYOO2JoA,8311
 unstructured_ingest/v2/processes/connectors/opensearch.py,sha256=HNRZVQsWnjLLm0yAGiIyHRbhAsBnGSXBO_VkUfIdwdE,5463
-unstructured_ingest/v2/processes/connectors/pinecone.py,sha256=VkPYmGmFKbgsmmrWV09roxztAv5LlTBVHizPeyPoFVc,5746
+unstructured_ingest/v2/processes/connectors/pinecone.py,sha256=PtAodxemYgiBZESx-g9a8fcL6XagJd9DIDQjrhE8aPk,5746
 unstructured_ingest/v2/processes/connectors/salesforce.py,sha256=Cz4qEtnbsD9-m1DXANxnVRZTHX2ZaUUBPVFPu5wnFRk,10832
 unstructured_ingest/v2/processes/connectors/sharepoint.py,sha256=SNovgGUE5tHdfX_lF5zwM_QRZK7mahHzLZKhnqfk6Tc,17696
 unstructured_ingest/v2/processes/connectors/singlestore.py,sha256=upF2O4hJ2uiBhDRrpQ8CSJUvzmqu2j5H1b_QbReHJpw,5168
@@ -349,8 +351,8 @@ unstructured_ingest/v2/processes/connectors/fsspec/gcs.py,sha256=RYZq_8hKF7bRxuB
 unstructured_ingest/v2/processes/connectors/fsspec/s3.py,sha256=7lOm5hjb0LBkbe-OWXnV3wDC-3mM_GWwwmdKW0xzh8c,5333
 unstructured_ingest/v2/processes/connectors/fsspec/sftp.py,sha256=J7Ej-j7dtXAluHunwynUfHlNsYwymb-LsrGUFcljcsA,5700
 unstructured_ingest/v2/processes/connectors/fsspec/utils.py,sha256=jec_Qfe2hbfahBuY-u8FnvHuv933AI5HwPFjOL3kEEY,456
-unstructured_ingest-0.0.0.dist-info/METADATA,sha256=XkFXINFQaqCmpnfNfymilYFVadQCXgkbXUm236ko_so,21501
-unstructured_ingest-0.0.0.dist-info/WHEEL,sha256=GJ7t_kWBFywbagK5eo9IoUwLW6oyOeTKmQ-9iHFVNxQ,92
-unstructured_ingest-0.0.0.dist-info/entry_points.txt,sha256=gUAAFnjFPnBgThJSEbw0N5ZjxtaKlT1s9e05_arQrNw,70
-unstructured_ingest-0.0.0.dist-info/top_level.txt,sha256=QaTxTcjfM5Hr9sZJ6weOJvSe5ESQc0F8AWkhHInTCf8,20
-unstructured_ingest-0.0.0.dist-info/RECORD,,
+unstructured_ingest-0.0.1.dist-info/METADATA,sha256=Qru27Cxrf0C-vFe7MqfaKOfavazrWYTTRif6loKf71o,21568
+unstructured_ingest-0.0.1.dist-info/WHEEL,sha256=GJ7t_kWBFywbagK5eo9IoUwLW6oyOeTKmQ-9iHFVNxQ,92
+unstructured_ingest-0.0.1.dist-info/entry_points.txt,sha256=gUAAFnjFPnBgThJSEbw0N5ZjxtaKlT1s9e05_arQrNw,70
+unstructured_ingest-0.0.1.dist-info/top_level.txt,sha256=QaTxTcjfM5Hr9sZJ6weOJvSe5ESQc0F8AWkhHInTCf8,20
+unstructured_ingest-0.0.1.dist-info/RECORD,,