unstructured-ingest 0.3.3__tar.gz → 0.3.5__tar.gz

{unstructured_ingest-0.3.3 → unstructured_ingest-0.3.5}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: unstructured-ingest
-Version: 0.3.3
+Version: 0.3.5
 Summary: A library that prepares raw documents for downstream ML tasks.
 Home-page: https://github.com/Unstructured-IO/unstructured-ingest
 Author: Unstructured Technologies
@@ -22,13 +22,13 @@ 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: pandas
 Requires-Dist: click
-Requires-Dist: dataclasses_json
 Requires-Dist: opentelemetry-sdk
-Requires-Dist: pydantic>=2.7
+Requires-Dist: python-dateutil
 Requires-Dist: tqdm
-Requires-Dist: pandas
+Requires-Dist: dataclasses_json
+Requires-Dist: pydantic>=2.7
 Provides-Extra: remote
 Requires-Dist: unstructured-client>=0.26.1; extra == "remote"
 Provides-Extra: csv
@@ -66,16 +66,16 @@ Requires-Dist: pyairtable; extra == "airtable"
 Provides-Extra: astradb
 Requires-Dist: astrapy; extra == "astradb"
 Provides-Extra: azure
-Requires-Dist: adlfs; extra == "azure"
 Requires-Dist: fsspec; extra == "azure"
+Requires-Dist: adlfs; extra == "azure"
 Provides-Extra: azure-ai-search
 Requires-Dist: azure-search-documents; extra == "azure-ai-search"
 Provides-Extra: biomed
-Requires-Dist: requests; extra == "biomed"
 Requires-Dist: bs4; extra == "biomed"
+Requires-Dist: requests; 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: chromadb; extra == "chroma"
 Provides-Extra: clarifai
@@ -91,8 +91,8 @@ Requires-Dist: deltalake; extra == "delta-table"
 Provides-Extra: discord
 Requires-Dist: discord-py; extra == "discord"
 Provides-Extra: dropbox
-Requires-Dist: dropboxdrivefs; extra == "dropbox"
 Requires-Dist: fsspec; extra == "dropbox"
+Requires-Dist: dropboxdrivefs; extra == "dropbox"
 Provides-Extra: elasticsearch
 Requires-Dist: elasticsearch[async]; extra == "elasticsearch"
 Provides-Extra: gcs
@@ -107,8 +107,8 @@ 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
@@ -122,13 +122,13 @@ Requires-Dist: pymilvus; extra == "milvus"
 Provides-Extra: mongodb
 Requires-Dist: pymongo; extra == "mongodb"
 Provides-Extra: notion
+Requires-Dist: httpx; extra == "notion"
 Requires-Dist: htmlBuilder; extra == "notion"
 Requires-Dist: backoff; extra == "notion"
-Requires-Dist: httpx; extra == "notion"
 Requires-Dist: notion-client; extra == "notion"
 Provides-Extra: onedrive
-Requires-Dist: Office365-REST-Python-Client; extra == "onedrive"
 Requires-Dist: bs4; extra == "onedrive"
+Requires-Dist: Office365-REST-Python-Client; extra == "onedrive"
 Requires-Dist: msal; extra == "onedrive"
 Provides-Extra: opensearch
 Requires-Dist: opensearch-py; extra == "opensearch"
@@ -152,13 +152,13 @@ Requires-Dist: msal; extra == "sharepoint"
 Provides-Extra: salesforce
 Requires-Dist: simple-salesforce; extra == "salesforce"
 Provides-Extra: sftp
-Requires-Dist: paramiko; extra == "sftp"
 Requires-Dist: fsspec; extra == "sftp"
+Requires-Dist: paramiko; extra == "sftp"
 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: wikipedia
 Requires-Dist: wikipedia; extra == "wikipedia"
 Provides-Extra: weaviate
@@ -172,8 +172,8 @@ Requires-Dist: requests; extra == "vectara"
 Provides-Extra: embed-huggingface
 Requires-Dist: sentence-transformers; extra == "embed-huggingface"
 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
@@ -181,8 +181,8 @@ Requires-Dist: voyageai; extra == "embed-voyageai"
 Provides-Extra: embed-mixedbreadai
 Requires-Dist: mixedbread-ai; extra == "embed-mixedbreadai"
 Provides-Extra: openai
-Requires-Dist: tiktoken; extra == "openai"
 Requires-Dist: openai; extra == "openai"
+Requires-Dist: tiktoken; extra == "openai"
 Provides-Extra: bedrock
 Requires-Dist: boto3; extra == "bedrock"
 Provides-Extra: togetherai

{unstructured_ingest-0.3.3 → unstructured_ingest-0.3.5}/test/integration/connectors/test_lancedb.py

@@ -12,6 +12,7 @@ from lancedb import AsyncConnection
 from upath import UPath
 
 from test.integration.connectors.utils.constants import DESTINATION_TAG
+from unstructured_ingest.v2.constants import RECORD_ID_LABEL
 from unstructured_ingest.v2.interfaces.file_data import FileData, SourceIdentifiers
 from unstructured_ingest.v2.processes.connectors.lancedb.aws import (
     LanceDBAwsAccessConfig,
@@ -43,7 +44,6 @@ DATABASE_NAME = "database"
 TABLE_NAME = "elements"
 DIMENSION = 384
 NUMBER_EXPECTED_ROWS = 22
-NUMBER_EXPECTED_COLUMNS = 10
 S3_BUCKET = "s3://utic-ingest-test-fixtures/"
 GS_BUCKET = "gs://utic-test-ingest-fixtures-output/"
 AZURE_BUCKET = "az://utic-ingest-test-fixtures-output/"
@@ -54,9 +54,9 @@ REQUIRED_ENV_VARS = {
     "local": (),
 }
 
-
 SCHEMA = pa.schema(
     [
+        pa.field(RECORD_ID_LABEL, pa.string()),
         pa.field("vector", pa.list_(pa.float16(), DIMENSION)),
         pa.field("text", pa.string(), nullable=True),
         pa.field("type", pa.string(), nullable=True),
@@ -69,6 +69,7 @@ SCHEMA = pa.schema(
         pa.field("metadata-page_number", pa.int32(), nullable=True),
     ]
 )
+NUMBER_EXPECTED_COLUMNS = len(SCHEMA.names)
 
 
 @pytest_asyncio.fixture
@@ -116,7 +117,7 @@ async def test_lancedb_destination(
     file_data = FileData(
         source_identifiers=SourceIdentifiers(fullpath=upload_file.name, filename=upload_file.name),
         connector_type=CONNECTOR_TYPE,
-        identifier="mock file data",
+        identifier="mock-file-data",
     )
     stager = LanceDBUploadStager()
     uploader = _get_uploader(uri)
@@ -129,17 +130,52 @@ async def test_lancedb_destination(
 
     await uploader.run_async(path=staged_file_path, file_data=file_data)
 
-    table = await connection.open_table(TABLE_NAME)
-    table_df: pd.DataFrame = await table.to_pandas()
+    # Test upload to empty table
+    with await connection.open_table(TABLE_NAME) as table:
+        table_df: pd.DataFrame = await table.to_pandas()
 
     assert len(table_df) == NUMBER_EXPECTED_ROWS
     assert len(table_df.columns) == NUMBER_EXPECTED_COLUMNS
 
+    assert table_df[RECORD_ID_LABEL][0] == file_data.identifier
     assert table_df["element_id"][0] == "2470d8dc42215b3d68413b55bf00fed2"
     assert table_df["type"][0] == "CompositeElement"
     assert table_df["metadata-filename"][0] == "DA-1p-with-duplicate-pages.pdf.json"
     assert table_df["metadata-text_as_html"][0] is None
 
+    # Test upload of the second file, rows should be appended
+    file_data.identifier = "mock-file-data-2"
+    staged_second_file_path = stager.run(
+        elements_filepath=upload_file,
+        file_data=file_data,
+        output_dir=tmp_path,
+        output_filename=f"{upload_file.stem}-2{upload_file.suffix}",
+    )
+    await uploader.run_async(path=staged_second_file_path, file_data=file_data)
+    with await connection.open_table(TABLE_NAME) as table:
+        appended_table_df: pd.DataFrame = await table.to_pandas()
+    assert len(appended_table_df) == 2 * NUMBER_EXPECTED_ROWS
+
+    # Test re-upload of the first file, rows should be overwritten, not appended
+    await uploader.run_async(path=staged_file_path, file_data=file_data)
+    with await connection.open_table(TABLE_NAME) as table:
+        overwritten_table_df: pd.DataFrame = await table.to_pandas()
+    assert len(overwritten_table_df) == 2 * NUMBER_EXPECTED_ROWS
+
+
+class TestPrecheck:
+    @pytest.mark.tags(CONNECTOR_TYPE, DESTINATION_TAG)
+    @pytest.mark.parametrize("connection_with_uri", ["local", "s3", "gcs", "az"], indirect=True)
+    def test_succeeds(
+        self,
+        upload_file: Path,
+        connection_with_uri: tuple[AsyncConnection, str],
+        tmp_path: Path,
+    ) -> None:
+        _, uri = connection_with_uri
+        uploader = _get_uploader(uri)
+        uploader.precheck()
+
 
 def _get_uri(target: Literal["local", "s3", "gcs", "az"], local_base_path: Path) -> str:
     if target == "local":
@@ -158,11 +194,12 @@ def _get_uploader(
     uri: str,
 ) -> Union[LanceDBAzureUploader, LanceDBAzureUploader, LanceDBAwsUploader, LanceDBGSPUploader]:
     target = uri.split("://", maxsplit=1)[0] if uri.startswith(("s3", "az", "gs")) else "local"
+    upload_config = LanceDBUploaderConfig(table_name=TABLE_NAME)
     if target == "az":
         azure_connection_string = os.getenv("AZURE_DEST_CONNECTION_STR")
         access_config_kwargs = _parse_azure_connection_string(azure_connection_string)
         return LanceDBAzureUploader(
-            upload_config=LanceDBUploaderConfig(table_name=TABLE_NAME),
+            upload_config=upload_config,
             connection_config=LanceDBAzureConnectionConfig(
                 access_config=LanceDBAzureAccessConfig(**access_config_kwargs),
                 uri=uri,
@@ -171,7 +208,7 @@ def _get_uploader(
 
     elif target == "s3":
         return LanceDBAwsUploader(
-            upload_config=LanceDBUploaderConfig(table_name=TABLE_NAME),
+            upload_config=upload_config,
             connection_config=LanceDBAwsConnectionConfig(
                 access_config=LanceDBAwsAccessConfig(
                     aws_access_key_id=os.getenv("S3_INGEST_TEST_ACCESS_KEY"),
@@ -182,7 +219,7 @@ def _get_uploader(
         )
     elif target == "gs":
         return LanceDBGSPUploader(
-            upload_config=LanceDBUploaderConfig(table_name=TABLE_NAME),
+            upload_config=upload_config,
             connection_config=LanceDBGCSConnectionConfig(
                 access_config=LanceDBGCSAccessConfig(
                     google_service_account_key=os.getenv("GCP_INGEST_SERVICE_KEY")
@@ -192,7 +229,7 @@ def _get_uploader(
         )
     else:
         return LanceDBLocalUploader(
-            upload_config=LanceDBUploaderConfig(table_name=TABLE_NAME),
+            upload_config=upload_config,
             connection_config=LanceDBLocalConnectionConfig(
                 access_config=LanceDBLocalAccessConfig(),
                 uri=uri,

{unstructured_ingest-0.3.3 → unstructured_ingest-0.3.5}/test/integration/connectors/test_pinecone.py

@@ -1,4 +1,5 @@
 import json
+import math
 import os
 import re
 import time
@@ -19,6 +20,7 @@ from unstructured_ingest.v2.interfaces import FileData, SourceIdentifiers
 from unstructured_ingest.v2.logger import logger
 from unstructured_ingest.v2.processes.connectors.pinecone import (
     CONNECTOR_TYPE,
+    MAX_QUERY_RESULTS,
     PineconeAccessConfig,
     PineconeConnectionConfig,
     PineconeUploader,
@@ -118,7 +120,10 @@ def validate_pinecone_index(
             f"retry attempt {i}: expected {expected_num_of_vectors} != vector count {vector_count}"
         )
         time.sleep(interval)
-    assert vector_count == expected_num_of_vectors
+    assert vector_count == expected_num_of_vectors, (
+        f"vector count from index ({vector_count}) doesn't "
+        f"match expected number: {expected_num_of_vectors}"
+    )
 
 
 @requires_env(API_KEY)
@@ -147,10 +152,7 @@ async def test_pinecone_destination(pinecone_index: str, upload_file: Path, temp
     uploader = PineconeUploader(connection_config=connection_config, upload_config=upload_config)
     uploader.precheck()
 
-    if uploader.is_async():
-        await uploader.run_async(path=new_upload_file, file_data=file_data)
-    else:
-        uploader.run(path=new_upload_file, file_data=file_data)
+    uploader.run(path=new_upload_file, file_data=file_data)
     with new_upload_file.open() as f:
         staged_content = json.load(f)
     expected_num_of_vectors = len(staged_content)
@@ -160,10 +162,59 @@ async def test_pinecone_destination(pinecone_index: str, upload_file: Path, temp
     )
 
     # Rerun uploader and make sure no duplicates exist
-    if uploader.is_async():
-        await uploader.run_async(path=new_upload_file, file_data=file_data)
-    else:
-        uploader.run(path=new_upload_file, file_data=file_data)
+    uploader.run(path=new_upload_file, file_data=file_data)
+    logger.info("validating second upload")
+    validate_pinecone_index(
+        index_name=pinecone_index, expected_num_of_vectors=expected_num_of_vectors
+    )
+
+
+@requires_env(API_KEY)
+@pytest.mark.asyncio
+@pytest.mark.tags(CONNECTOR_TYPE, DESTINATION_TAG)
+@pytest.mark.skip(reason="TODO: get this to work")
+async def test_pinecone_destination_large_index(
+    pinecone_index: str, upload_file: Path, temp_dir: Path
+):
+    new_file = temp_dir / "large_file.json"
+    with upload_file.open() as f:
+        upload_content = json.load(f)
+
+    min_entries = math.ceil((MAX_QUERY_RESULTS * 2) / len(upload_content))
+    new_content = (upload_content * min_entries)[: (2 * MAX_QUERY_RESULTS)]
+    print(f"Creating large index content with {len(new_content)} records")
+    with new_file.open("w") as f:
+        json.dump(new_content, f)
+
+    expected_num_of_vectors = len(new_content)
+    file_data = FileData(
+        source_identifiers=SourceIdentifiers(fullpath=new_file.name, filename=new_file.name),
+        connector_type=CONNECTOR_TYPE,
+        identifier="pinecone_mock_id",
+    )
+    connection_config = PineconeConnectionConfig(
+        index_name=pinecone_index,
+        access_config=PineconeAccessConfig(api_key=get_api_key()),
+    )
+    stager_config = PineconeUploadStagerConfig()
+    stager = PineconeUploadStager(upload_stager_config=stager_config)
+    new_upload_file = stager.run(
+        elements_filepath=new_file,
+        output_dir=temp_dir,
+        output_filename=new_file.name,
+        file_data=file_data,
+    )
+
+    upload_config = PineconeUploaderConfig()
+    uploader = PineconeUploader(connection_config=connection_config, upload_config=upload_config)
+    uploader.precheck()
+
+    uploader.run(path=new_upload_file, file_data=file_data)
+    validate_pinecone_index(
+        index_name=pinecone_index, expected_num_of_vectors=expected_num_of_vectors
+    )
+    # Rerun uploader and make sure no duplicates exist
+    uploader.run(path=new_upload_file, file_data=file_data)
     logger.info("validating second upload")
     validate_pinecone_index(
         index_name=pinecone_index, expected_num_of_vectors=expected_num_of_vectors

unstructured_ingest-0.3.5/test/integration/embedders/test_azure_openai.py

@@ -0,0 +1,59 @@
+import json
+import os
+from dataclasses import dataclass
+from pathlib import Path
+
+from test.integration.embedders.utils import validate_embedding_output, validate_raw_embedder
+from test.integration.utils import requires_env
+from unstructured_ingest.embed.azure_openai import (
+    AzureOpenAIEmbeddingConfig,
+    AzureOpenAIEmbeddingEncoder,
+)
+from unstructured_ingest.v2.processes.embedder import Embedder, EmbedderConfig
+
+API_KEY = "AZURE_OPENAI_API_KEY"
+ENDPOINT = "AZURE_OPENAI_ENDPOINT"
+
+
+@dataclass(frozen=True)
+class AzureData:
+    api_key: str
+    endpoint: str
+
+
+def get_azure_data() -> AzureData:
+    api_key = os.getenv(API_KEY, None)
+    assert api_key
+    endpoint = os.getenv(ENDPOINT, None)
+    assert endpoint
+    return AzureData(api_key, endpoint)
+
+
+@requires_env(API_KEY, ENDPOINT)
+def test_azure_openai_embedder(embedder_file: Path):
+    azure_data = get_azure_data()
+    embedder_config = EmbedderConfig(
+        embedding_provider="azure-openai",
+        embedding_api_key=azure_data.api_key,
+        embedding_azure_endpoint=azure_data.endpoint,
+    )
+    embedder = Embedder(config=embedder_config)
+    results = embedder.run(elements_filepath=embedder_file)
+    assert results
+    with embedder_file.open("r") as f:
+        original_elements = json.load(f)
+    validate_embedding_output(original_elements=original_elements, output_elements=results)
+
+
+@requires_env(API_KEY, ENDPOINT)
+def test_raw_azure_openai_embedder(embedder_file: Path):
+    azure_data = get_azure_data()
+    embedder = AzureOpenAIEmbeddingEncoder(
+        config=AzureOpenAIEmbeddingConfig(
+            api_key=azure_data.api_key,
+            azure_endpoint=azure_data.endpoint,
+        )
+    )
+    validate_raw_embedder(
+        embedder=embedder, embedder_file=embedder_file, expected_dimensions=(1536,)
+    )

unstructured_ingest-0.3.5/unstructured_ingest/__version__.py

@@ -0,0 +1 @@
+__version__ = "0.3.5"  # pragma: no cover

unstructured_ingest-0.3.5/unstructured_ingest/embed/azure_openai.py

@@ -0,0 +1,31 @@
+from dataclasses import dataclass
+from typing import TYPE_CHECKING
+
+from pydantic import Field
+
+from unstructured_ingest.embed.openai import OpenAIEmbeddingConfig, OpenAIEmbeddingEncoder
+from unstructured_ingest.utils.dep_check import requires_dependencies
+
+if TYPE_CHECKING:
+    from openai import AzureOpenAI
+
+
+class AzureOpenAIEmbeddingConfig(OpenAIEmbeddingConfig):
+    api_version: str = Field(description="Azure API version", default="2024-06-01")
+    azure_endpoint: str
+    embedder_model_name: str = Field(default="text-embedding-ada-002", alias="model_name")
+
+    @requires_dependencies(["openai"], extras="openai")
+    def get_client(self) -> "AzureOpenAI":
+        from openai import AzureOpenAI
+
+        return AzureOpenAI(
+            api_key=self.api_key.get_secret_value(),
+            api_version=self.api_version,
+            azure_endpoint=self.azure_endpoint,
+        )
+
+
+@dataclass
+class AzureOpenAIEmbeddingEncoder(OpenAIEmbeddingEncoder):
+    config: AzureOpenAIEmbeddingConfig

{unstructured_ingest-0.3.3 → unstructured_ingest-0.3.5}/unstructured_ingest/v2/processes/connectors/couchbase.py

@@ -219,6 +219,9 @@ class CouchbaseIndexer(Indexer):
 
 
 class CouchbaseDownloaderConfig(DownloaderConfig):
+    collection_id: str = Field(
+        default="id", description="The unique key of the id field in the collection"
+    )
     fields: list[str] = field(default_factory=list)
 
 
@@ -250,7 +253,7 @@ class CouchbaseDownloader(Downloader):
     def generate_download_response(
         self, result: dict, bucket: str, file_data: FileData
     ) -> DownloadResponse:
-        record_id = result["id"]
+        record_id = result[self.download_config.collection_id]
         filename_id = self.get_identifier(bucket=bucket, record_id=record_id)
         filename = f"{filename_id}.txt"
         download_path = self.download_dir / Path(filename)

{unstructured_ingest-0.3.3 → unstructured_ingest-0.3.5}/unstructured_ingest/v2/processes/connectors/elasticsearch/elasticsearch.py

@@ -142,8 +142,6 @@ class ElasticsearchIndexer(Indexer):
     def precheck(self) -> None:
         try:
             with self.connection_config.get_client() as client:
-                if not client.ping():
-                    raise SourceConnectionError("cluster not detected")
                 indices = client.indices.get_alias(index="*")
                 if self.index_config.index_name not in indices:
                     raise SourceConnectionError(
@@ -393,11 +391,9 @@ class ElasticsearchUploader(Uploader):
     def precheck(self) -> None:
         try:
             with self.connection_config.get_client() as client:
-                if not client.ping():
-                    raise DestinationConnectionError("cluster not detected")
                 indices = client.indices.get_alias(index="*")
                 if self.upload_config.index_name not in indices:
-                    raise SourceConnectionError(
+                    raise DestinationConnectionError(
                         "index {} not found: {}".format(
                             self.upload_config.index_name, ", ".join(indices.keys())
                         )

{unstructured_ingest-0.3.3 → unstructured_ingest-0.3.5}/unstructured_ingest/v2/processes/connectors/lancedb/lancedb.py

@@ -15,6 +15,7 @@ from unstructured_ingest.error import DestinationConnectionError
 from unstructured_ingest.logger import logger
 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
 from unstructured_ingest.v2.interfaces.connector import ConnectionConfig
 from unstructured_ingest.v2.interfaces.file_data import FileData
 from unstructured_ingest.v2.interfaces.upload_stager import UploadStager, UploadStagerConfig
@@ -84,7 +85,7 @@ class LanceDBUploadStager(UploadStager):
 
         df = pd.DataFrame(
             [
-                self._conform_element_contents(element_contents)
+                self._conform_element_contents(element_contents, file_data)
                 for element_contents in elements_contents
             ]
         )
@@ -94,9 +95,10 @@ class LanceDBUploadStager(UploadStager):
 
         return output_path
 
-    def _conform_element_contents(self, element: dict) -> dict:
+    def _conform_element_contents(self, element: dict, file_data: FileData) -> dict:
         return {
             "vector": element.pop("embeddings", None),
+            RECORD_ID_LABEL: file_data.identifier,
             **flatten_dict(element, separator="-"),
         }
 
@@ -134,6 +136,14 @@ class LanceDBUploader(Uploader):
         async with self.get_table() as table:
             schema = await table.schema()
             df = self._fit_to_schema(df, schema)
+            if RECORD_ID_LABEL not in schema.names:
+                logger.warning(
+                    f"Designated table doesn't contain {RECORD_ID_LABEL} column of type"
+                    " string which is required to support overwriting updates on subsequent"
+                    " uploads of the same record. New rows will be appended instead."
+                )
+            else:
+                await table.delete(f'{RECORD_ID_LABEL} = "{file_data.identifier}"')
             await table.add(data=df)
 
     def _fit_to_schema(self, df: pd.DataFrame, schema) -> pd.DataFrame:

{unstructured_ingest-0.3.3 → unstructured_ingest-0.3.5}/unstructured_ingest/v2/processes/connectors/pinecone.py

@@ -31,6 +31,7 @@ CONNECTOR_TYPE = "pinecone"
 MAX_PAYLOAD_SIZE = 2 * 1024 * 1024  # 2MB
 MAX_POOL_THREADS = 100
 MAX_METADATA_BYTES = 40960  # 40KB https://docs.pinecone.io/reference/quotas-and-limits#hard-limits
+MAX_QUERY_RESULTS = 10000
 
 
 class PineconeAccessConfig(AccessConfig):
@@ -84,7 +85,7 @@ ALLOWED_FIELDS = (
 
 class PineconeUploadStagerConfig(UploadStagerConfig):
     metadata_fields: list[str] = Field(
-        default=str(ALLOWED_FIELDS),
+        default=list(ALLOWED_FIELDS),
         description=(
             "which metadata from the source element to map to the payload metadata being sent to "
             "Pinecone."
@@ -137,7 +138,6 @@ class PineconeUploadStager(UploadStager):
             flatten_lists=True,
             remove_none=True,
         )
-        metadata[RECORD_ID_LABEL] = file_data.identifier
         metadata_size_bytes = len(json.dumps(metadata).encode())
         if metadata_size_bytes > MAX_METADATA_BYTES:
             logger.info(
@@ -146,6 +146,8 @@ class PineconeUploadStager(UploadStager):
             )
             metadata = {}
 
+        metadata[RECORD_ID_LABEL] = file_data.identifier
+
         return {
             "id": str(uuid.uuid4()),
             "values": embeddings,
@@ -213,6 +215,18 @@ class PineconeUploader(Uploader):
             f"from pinecone index: {resp}"
         )
 
+    def delete_by_query(self, index: "PineconeIndex", query_params: dict) -> None:
+        while True:
+            query_results = index.query(**query_params)
+            matches = query_results.get("matches", [])
+            if not matches:
+                break
+            ids = [match["id"] for match in matches]
+            delete_params = {"ids": ids}
+            if namespace := self.upload_config.namespace:
+                delete_params["namespace"] = namespace
+            index.delete(**delete_params)
+
     def serverless_delete_by_record_id(self, file_data: FileData) -> None:
         logger.debug(
             f"deleting any content with metadata "
@@ -221,29 +235,25 @@ class PineconeUploader(Uploader):
         )
         index = self.connection_config.get_index(pool_threads=MAX_POOL_THREADS)
         index_stats = index.describe_index_stats()
+        dimension = index_stats["dimension"]
         total_vectors = index_stats["total_vector_count"]
         if total_vectors == 0:
             return
-        dimension = index_stats["dimension"]
-        query_params = {
-            "filter": {self.upload_config.record_id_key: {"$eq": file_data.identifier}},
-            "vector": [0] * dimension,
-            "top_k": total_vectors,
-        }
-        if namespace := self.upload_config.namespace:
-            query_params["namespace"] = namespace
-        while True:
-            query_results = index.query(**query_params)
-            matches = query_results.get("matches", [])
-            if not matches:
-                break
-            ids = [match["id"] for match in matches]
-            delete_params = {"ids": ids}
+        while total_vectors > 0:
+            top_k = min(total_vectors, MAX_QUERY_RESULTS)
+            query_params = {
+                "filter": {self.upload_config.record_id_key: {"$eq": file_data.identifier}},
+                "vector": [0] * dimension,
+                "top_k": top_k,
+            }
             if namespace := self.upload_config.namespace:
-                delete_params["namespace"] = namespace
-            index.delete(**delete_params)
-        logger.debug(
-            f"deleted any content with metadata "
+                query_params["namespace"] = namespace
+            self.delete_by_query(index=index, query_params=query_params)
+            index_stats = index.describe_index_stats()
+            total_vectors = index_stats["total_vector_count"]
+
+        logger.info(
+            f"deleted {total_vectors} records with metadata "
             f"{self.upload_config.record_id_key}={file_data.identifier} "
             f"from pinecone index"
         )

{unstructured_ingest-0.3.3 → unstructured_ingest-0.3.5}/unstructured_ingest/v2/processes/connectors/weaviate/__init__.py

@@ -10,8 +10,6 @@ from .embedded import CONNECTOR_TYPE as EMBEDDED_WEAVIATE_CONNECTOR_TYPE
 from .embedded import weaviate_embedded_destination_entry
 from .local import CONNECTOR_TYPE as LOCAL_WEAVIATE_CONNECTOR_TYPE
 from .local import weaviate_local_destination_entry
-from .weaviate import CONNECTOR_TYPE as WEAVIATE_CONNECTOR_TYPE
-from .weaviate import weaviate_destination_entry
 
 add_destination_entry(
     destination_type=LOCAL_WEAVIATE_CONNECTOR_TYPE, entry=weaviate_local_destination_entry
@@ -22,4 +20,3 @@ add_destination_entry(
 add_destination_entry(
     destination_type=EMBEDDED_WEAVIATE_CONNECTOR_TYPE, entry=weaviate_embedded_destination_entry
 )
-add_destination_entry(destination_type=WEAVIATE_CONNECTOR_TYPE, entry=weaviate_destination_entry)

{unstructured_ingest-0.3.3 → unstructured_ingest-0.3.5}/unstructured_ingest/v2/processes/connectors/weaviate/weaviate.py

@@ -22,7 +22,6 @@ from unstructured_ingest.v2.interfaces import (
     UploadStagerConfig,
 )
 from unstructured_ingest.v2.logger import logger
-from unstructured_ingest.v2.processes.connector_registry import DestinationRegistryEntry
 
 if TYPE_CHECKING:
     from weaviate.classes.init import Timeout
@@ -288,12 +287,3 @@ class WeaviateUploader(Uploader, ABC):
                         vector=vector,
                     )
             self.check_for_errors(client=weaviate_client)
-
-
-weaviate_destination_entry = DestinationRegistryEntry(
-    connection_config=WeaviateConnectionConfig,
-    uploader=WeaviateUploader,
-    uploader_config=WeaviateUploaderConfig,
-    upload_stager=WeaviateUploadStager,
-    upload_stager_config=WeaviateUploadStagerConfig,
-)