Apppublic

Maverick9876/document-extraction-workbench-cloud

sourceHugging Faceupdated 5mo agoView on Hugging Face

0likes

App README

Document Extraction Workbench Cloud

Hosted cloud version of the original three-tool app:

—SharePoint OCR
—Certificate Extraction
—Name Matching

Key behavior

—This is not the HF OCR test app.
—This app expects each user to provide their own OpenAI or Gemini API key for SharePoint OCR and certificate AI fallback.
—Name Matching works without an API key.
—The app runs one job at a time on free CPU infrastructure for reliability.
—The app is tuned for smaller batches on free hosting.

Hosting target

This folder is intended for a Hugging Face Docker Space.

Files

—app.py: Gradio UI for the hosted app
—Dockerfile: build/runtime definition for the Space
—main_sharepoint_ocr.py: copied original SharePoint OCR logic
—credential_extractor.py: copied original certificate extraction logic
—name_matcherv2.py: copied original matcher logic
—ai_provider.py: shared AI provider helper
—workbench_env.py: environment loader (safe if .env is absent)

Notes

—No live .env file is included here.
—If you want owner-side defaults later, add them as Space secrets/variables, not as committed files.
—Free Hugging Face Spaces sleep when idle, so cold starts are expected.
—Large OCR batches may time out on free CPU hardware; smaller CSVs are recommended.