bump unstructured-inference (#1140)

Pulls in fix from unstructured-inference==0.5.12:

When a pdf page doesn't have much data, it may get buffered in the write to a tempfile. If this happens, we'll hit an error reading the file back. Open to suggestions for a way to unit test this - I was creating some test files with pypdf but I couldn't trigger the error.

CHANGELOG.md

@@ -1,6 +1,8 @@
-## 0.10.1-dev2
+## 0.10.1-dev3
 
 ### Enhancements
+* Bump unstructured-inference==0.5.12:
+  - fix to avoid trace for certain PDF's
 * Bump unstructured-inference==0.5.11:
   - better defaults for DPI for hi_res and  Chipper
 * Bump unstructured-inference==0.5.10:

requirements/constraints.in

@@ -26,4 +26,4 @@ Pillow<10.0.0
 # AttributeError: 'ResourcePath' object has no attribute 'collection'
 Office365-REST-Python-Client<2.4.3
 # NOTE(christine) Pinned to set the `unstructured-inference` version
-unstructured-inference==0.5.11
+unstructured-inference==0.5.12

requirements/extra-pdf-image.txt

@@ -205,7 +205,7 @@ typing-extensions==4.7.1
     #   torch
 tzdata==2023.3
     # via pandas
-unstructured-inference==0.5.11
+unstructured-inference==0.5.12
     # via
     #   -c requirements/constraints.in
     #   -r requirements/extra-pdf-image.in

unstructured/__version__.py

@@ -1 +1 @@
-__version__ = "0.10.1-dev2"  # pragma: no cover
+__version__ = "0.10.1-dev3"  # pragma: no cover