Andrew Jiang
|
5319d4d868
|
Initial commit: Twitter-Telegram Profile Matching System
This module provides comprehensive Twitter-to-Telegram profile matching
and verification using 10 different matching methods and LLM verification.
Features:
- 10 matching methods (phash, usernames, bio handles, URL resolution, fuzzy names)
- URL resolution integration for t.co → t.me links
- Async LLM verification with GPT-5-mini
- Interactive menu system with real-time stats
- Threaded candidate finding (~1.5 contacts/sec)
- Comprehensive documentation and guides
Key Components:
- find_twitter_candidates.py: Core matching logic (10 methods)
- find_twitter_candidates_threaded.py: Threaded implementation
- verify_twitter_matches_v2.py: LLM verification (V5 prompt)
- review_match_quality.py: Analysis and quality review
- main.py: Interactive menu system
- Complete documentation (README, CHANGELOG, QUICKSTART)
Performance:
- Candidate finding: ~16-18 hours for 43K contacts
- LLM verification: ~23 hours for 43K users
- Cost: ~$130 for full verification
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
|
2025-11-04 22:56:25 -08:00 |
|