unstructured-ingest 0.5.18__tar.gz → 0.5.20__tar.gz

{unstructured_ingest-0.5.18/unstructured_ingest.egg-info → unstructured_ingest-0.5.20}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: unstructured-ingest
-Version: 0.5.18
+Version: 0.5.20
 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,<3.14
 Description-Content-Type: text/markdown
 License-File: LICENSE.md
-Requires-Dist: pydantic>=2.7
 Requires-Dist: python-dateutil
+Requires-Dist: pandas
+Requires-Dist: opentelemetry-sdk
 Requires-Dist: click
-Requires-Dist: dataclasses_json
+Requires-Dist: pydantic>=2.7
 Requires-Dist: tqdm
-Requires-Dist: opentelemetry-sdk
-Requires-Dist: pandas
+Requires-Dist: dataclasses_json
 Provides-Extra: remote
 Requires-Dist: unstructured-client>=0.30.0; extra == "remote"
 Provides-Extra: csv
@@ -71,11 +71,11 @@ 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: fsspec; extra == "box"
 Requires-Dist: boxfs; extra == "box"
+Requires-Dist: fsspec; extra == "box"
 Provides-Extra: chroma
 Requires-Dist: chromadb; extra == "chroma"
 Provides-Extra: clarifai
@@ -91,26 +91,31 @@ Requires-Dist: boto3; 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: duckdb
 Requires-Dist: duckdb; extra == "duckdb"
 Provides-Extra: elasticsearch
 Requires-Dist: elasticsearch[async]; extra == "elasticsearch"
 Provides-Extra: gcs
-Requires-Dist: gcsfs; extra == "gcs"
 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
 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: ibm-watsonx-s3
+Requires-Dist: pyarrow; extra == "ibm-watsonx-s3"
+Requires-Dist: tenacity; extra == "ibm-watsonx-s3"
+Requires-Dist: pyiceberg; extra == "ibm-watsonx-s3"
+Requires-Dist: httpx; extra == "ibm-watsonx-s3"
 Provides-Extra: jira
 Requires-Dist: atlassian-python-api; extra == "jira"
 Provides-Extra: kafka
@@ -124,23 +129,23 @@ Requires-Dist: pymilvus; extra == "milvus"
 Provides-Extra: mongodb
 Requires-Dist: pymongo; extra == "mongodb"
 Provides-Extra: neo4j
-Requires-Dist: networkx; extra == "neo4j"
-Requires-Dist: cymple; extra == "neo4j"
 Requires-Dist: neo4j-rust-ext; extra == "neo4j"
+Requires-Dist: cymple; extra == "neo4j"
+Requires-Dist: networkx; extra == "neo4j"
 Provides-Extra: notion
-Requires-Dist: notion-client; extra == "notion"
-Requires-Dist: backoff; 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: msal; 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"
 Provides-Extra: outlook
-Requires-Dist: Office365-REST-Python-Client; extra == "outlook"
 Requires-Dist: msal; extra == "outlook"
+Requires-Dist: Office365-REST-Python-Client; extra == "outlook"
 Provides-Extra: pinecone
 Requires-Dist: pinecone-client>=3.7.1; extra == "pinecone"
 Provides-Extra: postgres
@@ -155,18 +160,18 @@ Provides-Extra: s3
 Requires-Dist: s3fs; extra == "s3"
 Requires-Dist: fsspec; extra == "s3"
 Provides-Extra: sharepoint
-Requires-Dist: Office365-REST-Python-Client; extra == "sharepoint"
 Requires-Dist: msal; extra == "sharepoint"
+Requires-Dist: Office365-REST-Python-Client; 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: snowflake-connector-python; extra == "snowflake"
 Requires-Dist: psycopg2-binary; extra == "snowflake"
+Requires-Dist: snowflake-connector-python; extra == "snowflake"
 Provides-Extra: wikipedia
 Requires-Dist: wikipedia; extra == "wikipedia"
 Provides-Extra: weaviate
@@ -178,12 +183,12 @@ Requires-Dist: databricks-sql-connector; extra == "databricks-delta-tables"
 Provides-Extra: singlestore
 Requires-Dist: singlestoredb; extra == "singlestore"
 Provides-Extra: vectara
-Requires-Dist: requests; extra == "vectara"
 Requires-Dist: aiofiles; extra == "vectara"
+Requires-Dist: requests; extra == "vectara"
 Requires-Dist: httpx; extra == "vectara"
 Provides-Extra: vastdb
-Requires-Dist: vastdb; extra == "vastdb"
 Requires-Dist: pyarrow; extra == "vastdb"
+Requires-Dist: vastdb; extra == "vastdb"
 Requires-Dist: ibis; extra == "vastdb"
 Provides-Extra: zendesk
 Requires-Dist: bs4; extra == "zendesk"
@@ -204,8 +209,8 @@ Provides-Extra: openai
 Requires-Dist: tiktoken; extra == "openai"
 Requires-Dist: openai; extra == "openai"
 Provides-Extra: bedrock
-Requires-Dist: boto3; extra == "bedrock"
 Requires-Dist: aioboto3; extra == "bedrock"
+Requires-Dist: boto3; extra == "bedrock"
 Provides-Extra: togetherai
 Requires-Dist: together; extra == "togetherai"
 Dynamic: author

{unstructured_ingest-0.5.18 → unstructured_ingest-0.5.20}/setup.py

@@ -102,6 +102,7 @@ connectors_reqs = {
     "gitlab": load_requirements("requirements/connectors/gitlab.in"),
     "google-drive": load_requirements("requirements/connectors/google-drive.in"),
     "hubspot": load_requirements("requirements/connectors/hubspot.in"),
+    "ibm-watsonx-s3": load_requirements("requirements/connectors/ibm-watsonx-s3.in"),
     "jira": load_requirements("requirements/connectors/jira.in"),
     "kafka": load_requirements("requirements/connectors/kafka.in"),
     "kdbai": load_requirements("requirements/connectors/kdbai.in"),

{unstructured_ingest-0.5.18 → unstructured_ingest-0.5.20}/test/integration/connectors/test_redis.py

@@ -23,20 +23,22 @@ from unstructured_ingest.v2.processes.connectors.redisdb import (
 )
 
 
-async def delete_record(client: Redis, element_id: str) -> None:
-    await client.delete(element_id)
+async def delete_record(client: Redis, element_id: str, key_prefix: str) -> None:
+    key_with_prefix = f"{key_prefix}{element_id}"
+    await client.delete(key_with_prefix)
 
 
-async def validate_upload(client: Redis, first_element: dict):
+async def validate_upload(client: Redis, first_element: dict, key_prefix: str) -> None:
     element_id = first_element["element_id"]
+    key_with_prefix = f"{key_prefix}{element_id}"
     expected_text = first_element["text"]
     expected_embeddings = first_element["embeddings"]
     async with client.pipeline(transaction=True) as pipe:
         try:
-            response = await pipe.json().get(element_id, "$").execute()
+            response = await pipe.json().get(key_with_prefix, "$").execute()
             response = response[0][0]
         except redis_exceptions.ResponseError:
-            response = await pipe.get(element_id).execute()
+            response = await pipe.get(key_with_prefix).execute()
             response = json.loads(response[0])
 
     embedding_similarity = np.linalg.norm(
@@ -53,6 +55,7 @@ async def redis_destination_test(
     upload_file: Path,
     tmp_path: Path,
     connection_kwargs: dict,
+    uploader_config: dict,
     uri: Optional[str] = None,
     password: Optional[str] = None,
 ):
@@ -60,8 +63,9 @@ async def redis_destination_test(
         connection_config=RedisConnectionConfig(
             **connection_kwargs, access_config=RedisAccessConfig(uri=uri, password=password)
         ),
-        upload_config=RedisUploaderConfig(batch_size=10),
+        upload_config=RedisUploaderConfig(batch_size=10, **uploader_config),
     )
+    key_prefix = uploader.upload_config.key_prefix
 
     file_data = FileData(
         source_identifiers=SourceIdentifiers(fullpath=upload_file.name, filename=upload_file.name),
@@ -78,20 +82,32 @@ async def redis_destination_test(
 
         if uri:
             async with from_url(uri) as client:
-                await validate_upload(client=client, first_element=first_element)
+                await validate_upload(
+                    client=client,
+                    first_element=first_element,
+                    key_prefix=key_prefix,
+                )
         else:
             async with Redis(**connection_kwargs, password=password) as client:
-                await validate_upload(client=client, first_element=first_element)
+                await validate_upload(
+                    client=client,
+                    first_element=first_element,
+                    key_prefix=key_prefix,
+                )
     except Exception as e:
         raise e
     finally:
         if uri:
             async with from_url(uri) as client:
-                tasks = [delete_record(client, element["element_id"]) for element in elements]
+                tasks = [
+                    delete_record(client, element["element_id"], key_prefix) for element in elements
+                ]
                 await asyncio.gather(*tasks)
         else:
             async with Redis(**connection_kwargs, password=password) as client:
-                tasks = [delete_record(client, element["element_id"]) for element in elements]
+                tasks = [
+                    delete_record(client, element["element_id"], key_prefix) for element in elements
+                ]
                 await asyncio.gather(*tasks)
 
 
@@ -105,8 +121,13 @@ async def test_redis_destination_azure_with_password(upload_file: Path, tmp_path
         "db": 0,
         "ssl": True,
     }
+    uploader_config = {
+        "key_prefix": "test_ingest:",
+    }
     redis_pw = os.environ["AZURE_REDIS_INGEST_TEST_PASSWORD"]
-    await redis_destination_test(upload_file, tmp_path, connection_kwargs, password=redis_pw)
+    await redis_destination_test(
+        upload_file, tmp_path, connection_kwargs, uploader_config, password=redis_pw
+    )
 
 
 @pytest.mark.asyncio
@@ -114,6 +135,9 @@ async def test_redis_destination_azure_with_password(upload_file: Path, tmp_path
 @requires_env("AZURE_REDIS_INGEST_TEST_PASSWORD")
 async def test_redis_destination_azure_with_uri(upload_file: Path, tmp_path: Path):
     connection_kwargs = {}
+    uploader_config = {
+        "key_prefix": "test_ingest:",
+    }
     redis_pw = os.environ["AZURE_REDIS_INGEST_TEST_PASSWORD"]
     uri = f"rediss://:{redis_pw}@utic-dashboard-dev.redis.cache.windows.net:6380/0"
-    await redis_destination_test(upload_file, tmp_path, connection_kwargs, uri=uri)
+    await redis_destination_test(upload_file, tmp_path, connection_kwargs, uploader_config, uri=uri)