Market Arbitrage Python Script

I’m looking for a Python script that polls Polymarket and Kalshi every 30–60 seconds, pulls every open market, and flags genuine cross-platform arbitrage opportunities. A “good match” is the exact same bet even if the wording is different, so the matching logic must go beyond simple string equality—tokenisation, Jaccard similarity or any technique you prefer is fine as long as it reliably pairs bets that are 100 % the same event.

For each confirmed overlap the script has to compile and push a JSON payload to my existing API containing:
• full market questions / titles,
• current Yes/No prices on both sites,
• volume and liquidity metrics,
• the calculated arbitrage margin in %,
• category, close date and any other valuable metadata.

I also need lightweight logging plus a CSV or JSON backup of every signal. The code should be clean, well-commented and make sensible use of requests or httpx. Handle rate limits and other API quirks gracefully; we’ll discuss any constraints you uncover before you start coding so nothing breaks in production.

Deliverables I’ll review against:
1. Python 3 script (single file or small package) with clear docstrings and inline comments.
2. Config section or .env example for API keys, polling interval, and endpoints.
3. Sample output files (one pushed JSON signal, one local backup).
4. Brief README explaining setup, matching approach, and how you deal with rate limiting and errors.

If that’s clear, tell me how you’d tackle the matching logic and any rate-limit strategies you’d put in place, then we can get started.