build: 0.0.77 release; bump to unstructured==0.15.5 (#450)

### Summary

Bumps to `unstructured==0.15.5`. Also pulls in the latest version of the
`wolfi` base image.

Author: MthwRobinson

CHANGELOG.md

@@ -1,3 +1,7 @@
+## 0.0.77
+
+* Bump to `unstructured` 0.15.5
+
 ## 0.0.76
 * Use the library's `detect_filetype` in API to determine mimetype
 * Add content_type api parameter

Dockerfile

@@ -1,5 +1,5 @@
 # syntax=docker/dockerfile:experimental
-FROM quay.io/unstructured-io/base-images:wolfi-base-e48da6b@sha256:8ad3479e5dc87a86e4794350cca6385c01c6d110902c5b292d1a62e231be711b as base
+FROM quay.io/unstructured-io/base-images:wolfi-base-latest as base
 
 # NOTE(crag): NB_USER ARG for mybinder.org compat:
 #             https://mybinder.readthedocs.io/en/latest/tutorials/dockerfile.html

prepline_general/api/app.py

@@ -13,7 +13,7 @@
 app = FastAPI(
     title="Unstructured Pipeline API",
     summary="Partition documents with the Unstructured library",
-    version="0.0.76",
+    version="0.0.77",
     docs_url="/general/docs",
     openapi_url="/general/openapi.json",
     servers=[

prepline_general/api/general.py

@@ -619,7 +619,7 @@ async def stream_response(self, send: Send) -> None:
         )
         async for chunk in self.body_iterator:
             if not isinstance(chunk, bytes):
-                chunk = chunk.encode(self.charset)
+                chunk = chunk.encode(self.charset)  # type: ignore
                 chunk = b64encode(chunk)
             await send(
                 {"type": "http.response.body", "body": self.build_part(chunk), "more_body": True}
@@ -649,7 +649,7 @@ def return_content_type(filename: str):
 
 
 @router.get("/general/v0/general", include_in_schema=False)
-@router.get("/general/v0.0.76/general", include_in_schema=False)
+@router.get("/general/v0.0.77/general", include_in_schema=False)
 async def handle_invalid_get_request():
     raise HTTPException(
         status_code=status.HTTP_405_METHOD_NOT_ALLOWED, detail="Only POST requests are supported."
@@ -664,7 +664,7 @@ async def handle_invalid_get_request():
     description="Description",
     operation_id="partition_parameters",
 )
-@router.post("/general/v0.0.76/general", include_in_schema=False)
+@router.post("/general/v0.0.77/general", include_in_schema=False)
 def general_partition(
     request: Request,
     # cannot use annotated type here because of a bug described here:

preprocessing-pipeline-family.yaml

@@ -1,2 +1,2 @@
 name: general
-version: 0.0.76
+version: 0.0.77

requirements/base.txt

@@ -26,7 +26,7 @@ certifi==2024.7.4
     #   httpx
     #   requests
     #   unstructured-client
-cffi==1.16.0
+cffi==1.17.0
     # via cryptography
 chardet==5.2.0
     # via unstructured
@@ -65,7 +65,7 @@ et-xmlfile==1.1.0
     # via openpyxl
 exceptiongroup==1.2.2
     # via anyio
-fastapi==0.112.0
+fastapi==0.112.1
     # via -r requirements/base.in
 filelock==3.15.4
     # via
@@ -84,7 +84,7 @@ fsspec==2024.6.1
     #   torch
 google-api-core[grpc]==2.19.1
     # via google-cloud-vision
-google-auth==2.32.0
+google-auth==2.33.0
     # via
     #   google-api-core
     #   google-cloud-vision
@@ -136,7 +136,7 @@ langdetect==1.0.9
     # via unstructured
 layoutparser==0.3.4
     # via unstructured-inference
-lxml==5.2.2
+lxml==5.3.0
     # via
     #   pikepdf
     #   python-docx
@@ -150,7 +150,7 @@ marshmallow==3.21.3
     # via
     #   dataclasses-json
     #   unstructured-client
-matplotlib==3.9.0
+matplotlib==3.9.2
     # via
     #   pycocotools
     #   unstructured-inference
@@ -191,7 +191,7 @@ onnx==1.16.2
     # via
     #   unstructured
     #   unstructured-inference
-onnxruntime==1.18.1
+onnxruntime==1.19.0
     # via unstructured-inference
 opencv-python==4.10.0.84
     # via
@@ -208,7 +208,6 @@ packaging==24.1
     #   matplotlib
     #   onnxruntime
     #   pikepdf
-    #   pytesseract
     #   transformers
     #   unstructured-client
     #   unstructured-pytesseract
@@ -224,9 +223,9 @@ pdfminer-six==20231228
     # via
     #   pdfplumber
     #   unstructured
-pdfplumber==0.11.2
+pdfplumber==0.11.3
     # via layoutparser
-pikepdf==9.1.0
+pikepdf==9.1.1
     # via unstructured
 pillow==10.4.0
     # via
@@ -236,7 +235,6 @@ pillow==10.4.0
     #   pdfplumber
     #   pikepdf
     #   pillow-heif
-    #   pytesseract
     #   python-pptx
     #   torchvision
     #   unstructured-pytesseract
@@ -288,8 +286,6 @@ pypdf==4.3.1
     #   unstructured-client
 pypdfium2==4.30.0
     # via pdfplumber
-pytesseract==0.3.10
-    # via unstructured
 python-dateutil==2.9.0.post0
     # via
     #   matplotlib
@@ -305,18 +301,18 @@ python-multipart==0.0.9
     # via unstructured-inference
 python-oxmsg==0.0.1
     # via unstructured
-python-pptx==0.6.23
+python-pptx==1.0.2
     # via unstructured
 pytz==2024.1
     # via pandas
-pyyaml==6.0.1
+pyyaml==6.0.2
     # via
     #   huggingface-hub
     #   layoutparser
     #   omegaconf
     #   timm
     #   transformers
-rapidfuzz==3.9.5
+rapidfuzz==3.9.6
     # via
     #   unstructured
     #   unstructured-inference
@@ -354,11 +350,11 @@ sniffio==1.3.1
     # via
     #   anyio
     #   httpx
-soupsieve==2.5
+soupsieve==2.6
     # via beautifulsoup4
-starlette==0.37.2
+starlette==0.38.2
     # via fastapi
-sympy==1.13.1
+sympy==1.13.2
     # via
     #   onnxruntime
     #   torch
@@ -387,7 +383,7 @@ tqdm==4.66.5
     #   nltk
     #   transformers
     #   unstructured
-transformers==4.43.4
+transformers==4.44.0
     # via unstructured-inference
 typing-extensions==4.12.2
     # via
@@ -401,6 +397,7 @@ typing-extensions==4.12.2
     #   pypdf
     #   python-docx
     #   python-oxmsg
+    #   python-pptx
     #   torch
     #   typing-inspect
     #   unstructured
@@ -412,19 +409,19 @@ typing-inspect==0.9.0
     #   unstructured-client
 tzdata==2024.1
     # via pandas
-unstructured[all-docs]==0.15.1
+unstructured[all-docs]==0.15.5
     # via -r requirements/base.in
-unstructured-client==0.25.2
+unstructured-client==0.25.5
     # via unstructured
 unstructured-inference==0.7.36
     # via unstructured
-unstructured-pytesseract==0.3.12
+unstructured-pytesseract==0.3.13
     # via unstructured
 urllib3==2.2.2
     # via
     #   requests
     #   unstructured-client
-uvicorn==0.30.5
+uvicorn==0.30.6
     # via -r requirements/base.in
 wrapt==1.16.0
     # via