Document Localization Studio (Free Stack)

A structure-aware document localization project for the GitHub Copilot CLI Challenge, built only with free libraries.

Supported inputs

• .txt
• .docx
• text-based .pdf
• screenshots/images: .png, .jpg, .jpeg (OCR pipeline)

Supported locales (requested set)

• de_de, es_es, fr_fr, it_it, ja_jp, ko_kr, pt_br, zh_cn, zh_tw

Feature coverage

• Language/term localization (rule-based)
• Currency + locale default USD FX (editable)
• Date/time + timezone conversion
• Measurement/unit conversion (mi -> km, lb -> kg, F -> C)
• Address/phone/postal adaptation by locale
• Tax/VAT/GST label adaptation + compliance labels
• Legal clause lock/protect zones ([[LOCK]]...[[/LOCK]])
• Terminology memory (term_memory.json)
• Style/tone presets (formal, legal, technical, marketing)
• Table overflow risk hints (DOCX)
• Cross-reference/TOC/page-reference QA warnings
• Font/script fallback QA checks for CJK
• Approval workflow states (Draft, Legal Review, Final)

UI highlights

• Attractive dashboard UI (hero header, styled cards)
• Animated Before/After scorecards
• Visual side-by-side diff
• Layout risk heatmap

Stack

• streamlit
• python-docx
• pypdf
• reportlab
• pymupdf (layout-preserving PDF localization)
• pillow
• pytesseract

Setup

cd "/Users/swatigoyal/Documents/New project/document_localizer_challenge"
python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt

Run UI

streamlit run app.py

Run CLI

python -m localizer.cli input.docx output.docx --locale de_de
python -m localizer.cli input.pdf output.pdf --locale ja_jp --source-timezone America/Los_Angeles --tone legal
python -m localizer.cli screenshot.png localized.txt --locale fr_fr --workflow "Legal Review"

OCR note

Screenshot OCR requires a local Tesseract binary in addition to pytesseract.

• macOS (example): brew install tesseract

Demo script for challenge

1. Upload DOCX/PDF/screenshot.
2. Change locale and watch default FX auto-update.
3. Adjust tone + workflow state.
4. Run localization and show scorecards, diff, heatmap, QA.
5. Download localized output + QA report.