Legal Document to MediaWiki Converter

Converts PDF and DOCX legal documents into MediaWiki markup.

Installation

It's recommended to use a virtual environment to avoid system Python restrictions.

python3 -m venv .venv
. .venv/bin/activate
pip install pdfplumber python-docx tqdm

Usage

python main.py <input_dir> <output_dir>

Example:

python main.py attachments laws_output

Importing .wiki pages into MediaWiki

This repository includes a simple dockerized MediaWiki setup and a helper script to bulk import the generated .wiki files as pages.

Steps:

1. Start MediaWiki with the database

docker-compose up -d

2. Complete the web installer once (first time only)

• Open http://localhost:8080
• Database settings to use during install:
  • DB host: db
  • DB name: mediawiki
  • DB user: wiki
  • DB password: change_me
• When the installer offers LocalSettings.php, place it into the wiki root (the container volume /var/www/html persists it). If prompted to download, upload it back via the container or bind mount; with this compose, it should be saved automatically in the volume.

3. Import all .wiki files

• The folder laws_output/ is mounted read-only inside the MediaWiki container at /import.
• Run the helper script to import all files. It uses each filename (without .wiki) as the page title, converting underscores to spaces.

chmod +x import_wiki.sh
./import_wiki.sh

Optionally, import into a specific namespace (e.g., Laws):

./import_wiki.sh Laws

Notes:

• The script uses maintenance/importTextFiles.php with --overwrite, so re-running will update existing pages with the latest content.
• If you see an error about LocalSettings.php missing, finish the installer in step 2 first.
• Access the wiki at http://localhost:8080

Output Format

• Filename: sanitized document title (max 200 chars)
• Extension: .wiki
• Content: MediaWiki markup with proper hierarchy

Notes on formatting

• Title consistency: the infobox now uses the same clean title shown in the page banner. If no title is detected, it falls back to "Untitled Bill" (not {{PAGENAME}}).
• Introduced by: stray lines like Introduced By: Name are captured into the infobox and removed from the article body.
• Section symbols (§): lines with multiple sections like §1 ... §2 ... are split so each section begins on its own line, improving readability.