v0.2.1 - Cross-Platform Support

🎉 Cross-Platform Release

Nowe funkcje:

✅ Wsparcie dla macOS i Linux

• Skrypt instalacyjny dla systemów Unix (install.sh)
• Automatyczna instalacja Homebrew na macOS
• Integracja z Finder przez Automator Quick Actions
• Konfiguracja PATH w shell RC files (.zshrc, .bashrc)

✅ Instrukcja dla laika (README_LAIK.md)

• Instalacja jedną komendą dla Windows i macOS
• Proste wyjaśnienia z emoji
• FAQ i troubleshooting
• Trzy metody użycia (GUI, menu kontekstowe, CLI)

Instalacja:

Windows:

iwr https://raw.githubusercontent.com/gacabartosz/presidio-local-anonymizer/main/scripts/install.ps1 | iex

macOS:

curl -fsSL https://raw.githubusercontent.com/gacabartosz/presidio-local-anonymizer/main/scripts/install.sh | bash

Linux:

curl -fsSL https://raw.githubusercontent.com/gacabartosz/presidio-local-anonymizer/main/scripts/install.sh | bash

Platforma:

• ✅ Windows 10/11
• ✅ macOS 10.15+ (Catalina lub nowszy)
• ✅ Linux (Ubuntu 20.04+, Debian 10+, Fedora 30+)

Wszystkie funkcje:

• 📄 DOCX, ODT, PDF (text layer)
• 🖼️ OCR dla skanów (PDF, PNG, JPG, TIFF)
• 🎨 GUI (tkinter)
• 🖱️ Menu kontekstowe (prawy przycisk)
• ⌨️ CLI

🤖 Generated with Claude Code

v0.2.0 - PDF + OCR + GUI 🎨📄

Presidio Local Anonymizer v0.2.0

🎉 Major Release - New Features!

📄 PDF Support

• ✅ Text Layer PDFs - extract and anonymize embedded text
• ✅ Scanned PDFs (OCR) - automatic detection and OCR processing
• ✅ Auto-detection - system automatically chooses the right processor

🔍 OCR for Images

• ✅ Tesseract OCR integration for scanned documents
• ✅ Image formats - PNG, JPG, JPEG, TIFF
• ✅ Smart redaction - black boxes over detected PII on images
• ✅ Polish language support - pol.traineddata included

🎨 GUI (Graphical User Interface)

• ✅ Simple tkinter interface - no more command line!
• ✅ Batch processing - select multiple files or entire folders
• ✅ Progress bar - real-time progress tracking
• ✅ Live logs - see what's happening
• ✅ Drag & Drop ready - install tkinterdnd2 for full functionality

📦 Installation

One-liner PowerShell:

powershell -NoProfile -ExecutionPolicy Bypass -Command "& {Invoke-WebRequest -Uri 'https://raw.githubusercontent.com/gacabartosz/presidio-local-anonymizer/main/scripts/install.ps1' -OutFile '$env:TEMP\presidio_install.ps1'; & '$env:TEMP\presidio_install.ps1'; Remove-Item '$env:TEMP\presidio_install.ps1' -Force}"

What's installed:

• Python 3.11
• Git
• Tesseract OCR (new! for scans)
• All Python dependencies
• SpaCy Polish model
• Tesseract Polish model (new!)

Installation time: ~10-15 minutes

🚀 Usage

GUI (new!)

anonymize-gui.cmd

Command Line

# DOCX/ODT (as before)
anonymize.cmd document.docx

# PDF (new\!)
anonymize.cmd report.pdf

# Images (new\!)
anonymize.cmd scan.png

# Batch process folder
anonymize.cmd C:\Documents\

Context Menu

Right-click on any supported file → Anonimizuj (Presidio)

📊 Supported Formats

Format	Extension	OCR Required	Status
Word	.docx	No	✅
OpenDocument	.odt	No	✅
PDF (text)	.pdf	No	✅ NEW
PDF (scan)	.pdf	Yes	✅ NEW
PNG	.png	Yes	✅ NEW
JPEG	.jpg, .jpeg	Yes	✅ NEW
TIFF	.tiff, .tif	Yes	✅ NEW

🔧 Technical Details

New Dependencies

• PyPDF2==3.0.1 - PDF text extraction
• reportlab==4.0.7 - PDF generation
• pytesseract==0.3.10 - OCR wrapper
• pdf2image==1.16.3 - PDF to image conversion
• Pillow==10.1.0 - Image manipulation
• img2pdf==0.5.1 - Image to PDF conversion
• tkinterdnd2==0.3.0 - GUI drag & drop

Architecture

PDF file
  ↓
Check if has text layer
  ↓
YES → pdf_processor.py (fast)
NO  → ocr_processor.py (Tesseract)
  ↓
Presidio Analyzer → Detect PII
  ↓
Anonymize:
  - Text PDFs: replace text
  - Scanned: black boxes on image

⚠️ Known Limitations

PDF Formatting:

• Text PDFs: Complex layouts simplified
• Scanned PDFs: Output is image-based

OCR Accuracy:

• Depends on scan quality
• Works best with 300 DPI or higher
• Polish language model included

GUI:

• Requires tkinter (included with Python)
• Drag & Drop requires tkinterdnd2 (optional)

🐛 Bug Fixes

• Improved error handling for missing libraries
• Better logging for debugging
• Graceful degradation if OCR unavailable

📝 Changelog

Added

• PDF text layer support (pdf_processor.py)
• OCR for scanned PDFs and images (ocr_processor.py)
• Tesseract OCR auto-installation in install.ps1
• Simple GUI with batch processing (app/gui.py)
• Support for PNG, JPG, TIFF formats
• Auto-detection: text vs scanned PDFs

Changed

• Updated main.py to route files to appropriate processors
• Extended requirements.txt with OCR libraries
• Updated README with new features
• Installation time increased to 10-15 min (Tesseract download)

Roadmap Progress

• PDF support
• OCR support
• Image anonymization
• Basic GUI
• DOCX formatting preservation (v0.3.0)
• Parallel processing (v0.3.0)

🙏 Credits

• Microsoft Presidio - PII detection engine
• Tesseract OCR - OCR engine
• SpaCy - NLP for Polish language
• UB-Mannheim - Tesseract Windows builds

📚 Documentation

• README
• Technical Docs
• Configuration Guide

---

Full Changelog: v0.1.0...v0.2.0

🤖 Generated with Claude Code

MVP Release v0.1.0

Presidio Local Anonymizer v0.1.0 (MVP)

Główne funkcje

✅ Offline - wszystkie operacje wykonywane lokalnie
✅ Obsługa polskich danych - PESEL, NIP, imiona, nazwiska
✅ Formaty - DOCX i ODT
✅ Integracja z Windows - menu kontekstowe
✅ Przetwarzanie wsadowe - pliki i foldery
✅ Raporty JSON - szczegółowe statystyki

Wykrywane dane osobowe

• PERSON (imiona i nazwiska)
• EMAIL_ADDRESS
• PHONE_NUMBER
• PL_PESEL
• PL_NIP
• LOCATION
• DATE_TIME
• URL
• IP_ADDRESS

Instalacja

One-liner PowerShell:

powershell -NoProfile -ExecutionPolicy Bypass -Command "& {Invoke-WebRequest -Uri 'https://raw.githubusercontent.com/gacabartosz/presidio-local-anonymizer/main/scripts/install.ps1' -OutFile '$env:TEMP\presidio_install.ps1'; & '$env:TEMP\presidio_install.ps1'; Remove-Item '$env:TEMP\presidio_install.ps1' -Force}"

Znane ograniczenia MVP

⚠️ Formatowanie dokumentów:

• DOCX: Style inline mogą się uprościć
• ODT: Struktura jest linearyzowana

⚠️ Wydajność:

• Przetwarzanie sekwencyjne (nie równoległe)

⚠️ Wykrywanie:

• Możliwe false positives/negatives

Co dalej?

Zobacz roadmap w README.md