unstructured-ingest 0.6.0__tar.gz → 0.6.1__tar.gz

{unstructured_ingest-0.6.0/unstructured_ingest.egg-info → unstructured_ingest-0.6.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: unstructured-ingest
-Version: 0.6.0
+Version: 0.6.1
 Summary: A library that prepares raw documents for downstream ML tasks.
 Home-page: https://github.com/Unstructured-IO/unstructured-ingest
 Author: Unstructured Technologies
@@ -22,12 +22,12 @@ Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
 Requires-Python: >=3.9.0,<3.14
 Description-Content-Type: text/markdown
 License-File: LICENSE.md
-Requires-Dist: dataclasses_json
-Requires-Dist: python-dateutil
-Requires-Dist: pydantic>=2.7
 Requires-Dist: tqdm
-Requires-Dist: click
 Requires-Dist: opentelemetry-sdk
+Requires-Dist: python-dateutil
+Requires-Dist: click
+Requires-Dist: dataclasses_json
+Requires-Dist: pydantic>=2.7
 Requires-Dist: pandas
 Requires-Dist: numpy
 Provides-Extra: remote
@@ -103,8 +103,8 @@ Requires-Dist: astrapy; extra == "astradb"
 Requires-Dist: pandas; extra == "astradb"
 Requires-Dist: numpy; extra == "astradb"
 Provides-Extra: azure
-Requires-Dist: adlfs; extra == "azure"
 Requires-Dist: fsspec; extra == "azure"
+Requires-Dist: adlfs; extra == "azure"
 Requires-Dist: pandas; extra == "azure"
 Requires-Dist: numpy; extra == "azure"
 Provides-Extra: azure-ai-search
@@ -112,8 +112,8 @@ Requires-Dist: azure-search-documents; extra == "azure-ai-search"
 Requires-Dist: pandas; extra == "azure-ai-search"
 Requires-Dist: numpy; extra == "azure-ai-search"
 Provides-Extra: biomed
-Requires-Dist: bs4; extra == "biomed"
 Requires-Dist: requests; extra == "biomed"
+Requires-Dist: bs4; extra == "biomed"
 Requires-Dist: pandas; extra == "biomed"
 Requires-Dist: numpy; extra == "biomed"
 Provides-Extra: box
@@ -130,8 +130,8 @@ Requires-Dist: clarifai; extra == "clarifai"
 Requires-Dist: pandas; extra == "clarifai"
 Requires-Dist: numpy; extra == "clarifai"
 Provides-Extra: confluence
-Requires-Dist: requests; extra == "confluence"
 Requires-Dist: atlassian-python-api; extra == "confluence"
+Requires-Dist: requests; extra == "confluence"
 Requires-Dist: pandas; extra == "confluence"
 Requires-Dist: numpy; extra == "confluence"
 Provides-Extra: couchbase
@@ -139,8 +139,8 @@ Requires-Dist: couchbase; extra == "couchbase"
 Requires-Dist: pandas; extra == "couchbase"
 Requires-Dist: numpy; extra == "couchbase"
 Provides-Extra: delta-table
-Requires-Dist: deltalake; extra == "delta-table"
 Requires-Dist: boto3; extra == "delta-table"
+Requires-Dist: deltalake; extra == "delta-table"
 Requires-Dist: pandas; extra == "delta-table"
 Requires-Dist: numpy; extra == "delta-table"
 Provides-Extra: discord
@@ -148,8 +148,8 @@ Requires-Dist: discord.py; extra == "discord"
 Requires-Dist: pandas; extra == "discord"
 Requires-Dist: numpy; extra == "discord"
 Provides-Extra: dropbox
-Requires-Dist: dropboxdrivefs; extra == "dropbox"
 Requires-Dist: fsspec; extra == "dropbox"
+Requires-Dist: dropboxdrivefs; extra == "dropbox"
 Requires-Dist: pandas; extra == "dropbox"
 Requires-Dist: numpy; extra == "dropbox"
 Provides-Extra: duckdb
@@ -161,9 +161,9 @@ Requires-Dist: elasticsearch[async]; extra == "elasticsearch"
 Requires-Dist: pandas; extra == "elasticsearch"
 Requires-Dist: numpy; extra == "elasticsearch"
 Provides-Extra: gcs
-Requires-Dist: bs4; extra == "gcs"
 Requires-Dist: fsspec; extra == "gcs"
 Requires-Dist: gcsfs; extra == "gcs"
+Requires-Dist: bs4; extra == "gcs"
 Requires-Dist: pandas; extra == "gcs"
 Requires-Dist: numpy; extra == "gcs"
 Provides-Extra: github
@@ -180,15 +180,15 @@ Requires-Dist: google-api-python-client; extra == "google-drive"
 Requires-Dist: pandas; extra == "google-drive"
 Requires-Dist: numpy; extra == "google-drive"
 Provides-Extra: hubspot
-Requires-Dist: urllib3; extra == "hubspot"
 Requires-Dist: hubspot-api-client; extra == "hubspot"
+Requires-Dist: urllib3; extra == "hubspot"
 Requires-Dist: pandas; extra == "hubspot"
 Requires-Dist: numpy; extra == "hubspot"
 Provides-Extra: ibm-watsonx-s3
 Requires-Dist: tenacity; extra == "ibm-watsonx-s3"
-Requires-Dist: httpx; extra == "ibm-watsonx-s3"
-Requires-Dist: pyarrow; extra == "ibm-watsonx-s3"
 Requires-Dist: pyiceberg; extra == "ibm-watsonx-s3"
+Requires-Dist: pyarrow; extra == "ibm-watsonx-s3"
+Requires-Dist: httpx; extra == "ibm-watsonx-s3"
 Requires-Dist: pandas; extra == "ibm-watsonx-s3"
 Requires-Dist: numpy; extra == "ibm-watsonx-s3"
 Provides-Extra: jira
@@ -217,20 +217,20 @@ Requires-Dist: pandas; extra == "mongodb"
 Requires-Dist: numpy; 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: pandas; extra == "neo4j"
 Requires-Dist: numpy; extra == "neo4j"
 Provides-Extra: notion
-Requires-Dist: notion-client; extra == "notion"
 Requires-Dist: httpx; extra == "notion"
-Requires-Dist: backoff; extra == "notion"
 Requires-Dist: htmlBuilder; extra == "notion"
+Requires-Dist: backoff; extra == "notion"
+Requires-Dist: notion-client; extra == "notion"
 Requires-Dist: pandas; extra == "notion"
 Requires-Dist: numpy; extra == "notion"
 Provides-Extra: onedrive
-Requires-Dist: bs4; extra == "onedrive"
 Requires-Dist: msal; extra == "onedrive"
+Requires-Dist: bs4; extra == "onedrive"
 Requires-Dist: Office365-REST-Python-Client; extra == "onedrive"
 Requires-Dist: pandas; extra == "onedrive"
 Requires-Dist: numpy; extra == "onedrive"
@@ -318,15 +318,15 @@ Requires-Dist: aiofiles; extra == "vectara"
 Requires-Dist: pandas; extra == "vectara"
 Requires-Dist: numpy; extra == "vectara"
 Provides-Extra: vastdb
-Requires-Dist: vastdb; extra == "vastdb"
 Requires-Dist: ibis; extra == "vastdb"
+Requires-Dist: vastdb; extra == "vastdb"
 Requires-Dist: pyarrow; extra == "vastdb"
 Requires-Dist: pandas; extra == "vastdb"
 Requires-Dist: numpy; extra == "vastdb"
 Provides-Extra: zendesk
+Requires-Dist: aiofiles; extra == "zendesk"
 Requires-Dist: bs4; extra == "zendesk"
 Requires-Dist: httpx; extra == "zendesk"
-Requires-Dist: aiofiles; extra == "zendesk"
 Requires-Dist: pandas; extra == "zendesk"
 Requires-Dist: numpy; extra == "zendesk"
 Provides-Extra: embed-huggingface
@@ -356,8 +356,8 @@ Requires-Dist: tiktoken; extra == "openai"
 Requires-Dist: pandas; extra == "openai"
 Requires-Dist: numpy; extra == "openai"
 Provides-Extra: bedrock
-Requires-Dist: aioboto3; extra == "bedrock"
 Requires-Dist: boto3; extra == "bedrock"
+Requires-Dist: aioboto3; extra == "bedrock"
 Requires-Dist: pandas; extra == "bedrock"
 Requires-Dist: numpy; extra == "bedrock"
 Provides-Extra: togetherai

unstructured_ingest-0.6.1/unstructured_ingest/__version__.py

@@ -0,0 +1 @@
+__version__ = "0.6.1"  # pragma: no cover

{unstructured_ingest-0.6.0 → unstructured_ingest-0.6.1}/unstructured_ingest/v2/processes/chunker.py

@@ -6,6 +6,7 @@ from typing import Any, Optional
 from pydantic import BaseModel, Field, SecretStr
 
 from unstructured_ingest.utils.chunking import assign_and_map_hash_ids
+from unstructured_ingest.utils.data_prep import get_json_data
 from unstructured_ingest.utils.dep_check import requires_dependencies
 from unstructured_ingest.v2.interfaces.process import BaseProcess
 from unstructured_ingest.v2.logger import logger
@@ -92,9 +93,11 @@ class Chunker(BaseProcess, ABC):
     @requires_dependencies(dependencies=["unstructured"])
     def run(self, elements_filepath: Path, **kwargs: Any) -> list[dict]:
         from unstructured.chunking import dispatch
-        from unstructured.staging.base import elements_from_json
+        from unstructured.staging.base import elements_from_dicts
 
-        elements = elements_from_json(filename=str(elements_filepath))
+        element_dicts = get_json_data(elements_filepath)
+
+        elements = elements_from_dicts(element_dicts=element_dicts)
         if not elements:
             return [e.to_dict() for e in elements]
         local_chunking_strategies = ("basic", "by_title")

{unstructured_ingest-0.6.0 → unstructured_ingest-0.6.1/unstructured_ingest.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: unstructured-ingest
-Version: 0.6.0
+Version: 0.6.1
 Summary: A library that prepares raw documents for downstream ML tasks.
 Home-page: https://github.com/Unstructured-IO/unstructured-ingest
 Author: Unstructured Technologies
@@ -22,12 +22,12 @@ Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
 Requires-Python: >=3.9.0,<3.14
 Description-Content-Type: text/markdown
 License-File: LICENSE.md
-Requires-Dist: dataclasses_json
-Requires-Dist: python-dateutil
-Requires-Dist: pydantic>=2.7
 Requires-Dist: tqdm
-Requires-Dist: click
 Requires-Dist: opentelemetry-sdk
+Requires-Dist: python-dateutil
+Requires-Dist: click
+Requires-Dist: dataclasses_json
+Requires-Dist: pydantic>=2.7
 Requires-Dist: pandas
 Requires-Dist: numpy
 Provides-Extra: remote
@@ -103,8 +103,8 @@ Requires-Dist: astrapy; extra == "astradb"
 Requires-Dist: pandas; extra == "astradb"
 Requires-Dist: numpy; extra == "astradb"
 Provides-Extra: azure
-Requires-Dist: adlfs; extra == "azure"
 Requires-Dist: fsspec; extra == "azure"
+Requires-Dist: adlfs; extra == "azure"
 Requires-Dist: pandas; extra == "azure"
 Requires-Dist: numpy; extra == "azure"
 Provides-Extra: azure-ai-search
@@ -112,8 +112,8 @@ Requires-Dist: azure-search-documents; extra == "azure-ai-search"
 Requires-Dist: pandas; extra == "azure-ai-search"
 Requires-Dist: numpy; extra == "azure-ai-search"
 Provides-Extra: biomed
-Requires-Dist: bs4; extra == "biomed"
 Requires-Dist: requests; extra == "biomed"
+Requires-Dist: bs4; extra == "biomed"
 Requires-Dist: pandas; extra == "biomed"
 Requires-Dist: numpy; extra == "biomed"
 Provides-Extra: box
@@ -130,8 +130,8 @@ Requires-Dist: clarifai; extra == "clarifai"
 Requires-Dist: pandas; extra == "clarifai"
 Requires-Dist: numpy; extra == "clarifai"
 Provides-Extra: confluence
-Requires-Dist: requests; extra == "confluence"
 Requires-Dist: atlassian-python-api; extra == "confluence"
+Requires-Dist: requests; extra == "confluence"
 Requires-Dist: pandas; extra == "confluence"
 Requires-Dist: numpy; extra == "confluence"
 Provides-Extra: couchbase
@@ -139,8 +139,8 @@ Requires-Dist: couchbase; extra == "couchbase"
 Requires-Dist: pandas; extra == "couchbase"
 Requires-Dist: numpy; extra == "couchbase"
 Provides-Extra: delta-table
-Requires-Dist: deltalake; extra == "delta-table"
 Requires-Dist: boto3; extra == "delta-table"
+Requires-Dist: deltalake; extra == "delta-table"
 Requires-Dist: pandas; extra == "delta-table"
 Requires-Dist: numpy; extra == "delta-table"
 Provides-Extra: discord
@@ -148,8 +148,8 @@ Requires-Dist: discord.py; extra == "discord"
 Requires-Dist: pandas; extra == "discord"
 Requires-Dist: numpy; extra == "discord"
 Provides-Extra: dropbox
-Requires-Dist: dropboxdrivefs; extra == "dropbox"
 Requires-Dist: fsspec; extra == "dropbox"
+Requires-Dist: dropboxdrivefs; extra == "dropbox"
 Requires-Dist: pandas; extra == "dropbox"
 Requires-Dist: numpy; extra == "dropbox"
 Provides-Extra: duckdb
@@ -161,9 +161,9 @@ Requires-Dist: elasticsearch[async]; extra == "elasticsearch"
 Requires-Dist: pandas; extra == "elasticsearch"
 Requires-Dist: numpy; extra == "elasticsearch"
 Provides-Extra: gcs
-Requires-Dist: bs4; extra == "gcs"
 Requires-Dist: fsspec; extra == "gcs"
 Requires-Dist: gcsfs; extra == "gcs"
+Requires-Dist: bs4; extra == "gcs"
 Requires-Dist: pandas; extra == "gcs"
 Requires-Dist: numpy; extra == "gcs"
 Provides-Extra: github
@@ -180,15 +180,15 @@ Requires-Dist: google-api-python-client; extra == "google-drive"
 Requires-Dist: pandas; extra == "google-drive"
 Requires-Dist: numpy; extra == "google-drive"
 Provides-Extra: hubspot
-Requires-Dist: urllib3; extra == "hubspot"
 Requires-Dist: hubspot-api-client; extra == "hubspot"
+Requires-Dist: urllib3; extra == "hubspot"
 Requires-Dist: pandas; extra == "hubspot"
 Requires-Dist: numpy; extra == "hubspot"
 Provides-Extra: ibm-watsonx-s3
 Requires-Dist: tenacity; extra == "ibm-watsonx-s3"
-Requires-Dist: httpx; extra == "ibm-watsonx-s3"
-Requires-Dist: pyarrow; extra == "ibm-watsonx-s3"
 Requires-Dist: pyiceberg; extra == "ibm-watsonx-s3"
+Requires-Dist: pyarrow; extra == "ibm-watsonx-s3"
+Requires-Dist: httpx; extra == "ibm-watsonx-s3"
 Requires-Dist: pandas; extra == "ibm-watsonx-s3"
 Requires-Dist: numpy; extra == "ibm-watsonx-s3"
 Provides-Extra: jira
@@ -217,20 +217,20 @@ Requires-Dist: pandas; extra == "mongodb"
 Requires-Dist: numpy; 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: pandas; extra == "neo4j"
 Requires-Dist: numpy; extra == "neo4j"
 Provides-Extra: notion
-Requires-Dist: notion-client; extra == "notion"
 Requires-Dist: httpx; extra == "notion"
-Requires-Dist: backoff; extra == "notion"
 Requires-Dist: htmlBuilder; extra == "notion"
+Requires-Dist: backoff; extra == "notion"
+Requires-Dist: notion-client; extra == "notion"
 Requires-Dist: pandas; extra == "notion"
 Requires-Dist: numpy; extra == "notion"
 Provides-Extra: onedrive
-Requires-Dist: bs4; extra == "onedrive"
 Requires-Dist: msal; extra == "onedrive"
+Requires-Dist: bs4; extra == "onedrive"
 Requires-Dist: Office365-REST-Python-Client; extra == "onedrive"
 Requires-Dist: pandas; extra == "onedrive"
 Requires-Dist: numpy; extra == "onedrive"
@@ -318,15 +318,15 @@ Requires-Dist: aiofiles; extra == "vectara"
 Requires-Dist: pandas; extra == "vectara"
 Requires-Dist: numpy; extra == "vectara"
 Provides-Extra: vastdb
-Requires-Dist: vastdb; extra == "vastdb"
 Requires-Dist: ibis; extra == "vastdb"
+Requires-Dist: vastdb; extra == "vastdb"
 Requires-Dist: pyarrow; extra == "vastdb"
 Requires-Dist: pandas; extra == "vastdb"
 Requires-Dist: numpy; extra == "vastdb"
 Provides-Extra: zendesk
+Requires-Dist: aiofiles; extra == "zendesk"
 Requires-Dist: bs4; extra == "zendesk"
 Requires-Dist: httpx; extra == "zendesk"
-Requires-Dist: aiofiles; extra == "zendesk"
 Requires-Dist: pandas; extra == "zendesk"
 Requires-Dist: numpy; extra == "zendesk"
 Provides-Extra: embed-huggingface
@@ -356,8 +356,8 @@ Requires-Dist: tiktoken; extra == "openai"
 Requires-Dist: pandas; extra == "openai"
 Requires-Dist: numpy; extra == "openai"
 Provides-Extra: bedrock
-Requires-Dist: aioboto3; extra == "bedrock"
 Requires-Dist: boto3; extra == "bedrock"
+Requires-Dist: aioboto3; extra == "bedrock"
 Requires-Dist: pandas; extra == "bedrock"
 Requires-Dist: numpy; extra == "bedrock"
 Provides-Extra: togetherai

unstructured_ingest-0.6.0/unstructured_ingest/__version__.py

@@ -1 +0,0 @@
-__version__ = "0.6.0"  # pragma: no cover