ABSTRACTOR

Automated PDF parsing and assembly with a Streamlit UI. Upload client PDFs, auto-extract fields, review/edit, and generate a complete Bradley Abstract with a filled cover page plus merged documents.

Quick start

1) Python env and packages

Use the provided script:

bash setup_and_run.sh

This creates/activates a virtualenv (prefers .venv) and installs dependencies from requirements.txt.

System OCR deps (for scanned PDFs) are listed in packages.txt and should be installed already in this devcontainer:

• tesseract-ocr, tesseract-ocr-eng
• poppler-utils

If running elsewhere, install them via your OS package manager.

2) Run Streamlit

• VS Code: Command Palette → “Run Task” → “Run Streamlit app”
• Or manually:

./.venv/bin/python -m streamlit run streamlit_app.py --server.headless true --server.port 8501

Then open the URL shown in the terminal (typically http://localhost:8501).

3) Use the app

• Upload one or more PDF files in the sidebar.
• Click “Extract Data from PDFs” to parse and prefill fields.
• Edit fields as needed.
• Click “Generate Cover Page” to create the cover and assemble the final abstract PDF.
• Download from the button that appears, or find output files in output/.

Internals

• src/parser.py: PDF text extraction (PyPDF2). Falls back to OCR (pytesseract + pdf2image) when text quality is low.
• src/field_extractor.py: Regex-based field extraction.
• src/cover_page_generator.py: Generates the Bradley Abstract cover using PyMuPDF (fitz) and ReportLab.
  • Adapter BradleyAbstractCoverPage.generate_cover_page(data, output_path) is used by the UI.
• src/pdf_assembler.py: Merges cover + (optional) form + original docs into a single PDF via PyPDF2.
• Template file: templates/bradley_abstract_cover.pdf.

Git MCP server (optional)

A local MCP server exposing Git tools is included for MCP-capable clients.

• Location: tools/git-mcp-server/
• Start via VS Code task: “Start Git MCP server” (or “Start Git MCP server (dev)”)
• Exposed tools: git_status, git_log, git_diff, git_commit, git_push

Troubleshooting

• OCR doesn’t work: ensure tesseract-ocr and poppler-utils are installed on the system.
• Import errors for PyMuPDF/ReportLab: re-run bash setup_and_run.sh.
• Template missing: confirm templates/bradley_abstract_cover.pdf exists.
• Streamlit CORS/XSRF warning: informational by default; app still runs. If you need cross-origin embedding, disable server.enableXsrfProtection in Streamlit config (not recommended unless you know the implications).

License

This repository’s code is provided as-is by the owner. See repository settings for license details.