feat(predictions): add Kalshi as prediction market data source

[koala73]

Summary

Add Kalshi as a second prediction market source alongside Polymarket. Kalshi is a CFTC-regulated exchange with a comprehensive API for event contracts covering politics, economics, climate, tech, and more.

API Reference

• Docs: https://docs.kalshi.com/welcome
• OpenAPI spec: https://docs.kalshi.com/openapi.yaml
• AsyncAPI spec (WebSocket): https://docs.kalshi.com/asyncapi.yaml
• LLM-friendly docs: https://docs.kalshi.com/llms.txt

Key Endpoints

Endpoint	Purpose
GET /events	List events (categories, titles)
GET /markets	List markets with status filter (open/closed/settled)
GET /markets/{ticker}	Market detail (prices, volume, open interest)
GET /market/{ticker}/orderbook	Current bid/ask
GET /market/{ticker}/candlesticks	OHLCV data

Auth & Limits

• Auth: RSA key-pair based API keys
• Rate limits: Tiered — check via "Get Account API Limits" endpoint
• WebSocket: Real-time market ticker, orderbook deltas, public trades (ping/pong every 10s)
• Demo environment: Available for testing

Data Model

• Binary yes/no contracts (same as Polymarket)
• Prices in cents (0-100 range) — maps directly to probability %
• Fields: ticker, title, yes_price, volume, open_interest, close_time, settlement status

Implementation Plan

1. Proto Updates

• Add source field to PredictionMarket message (e.g., POLYMARKET, KALSHI)
• Add Kalshi-specific fields if needed (e.g., ticker, open_interest)

2. Backend

• New seed script scripts/seed-kalshi-markets.mjs or extend existing seed
• Fetch from GET /events + GET /markets (filter open, non-sports)
• Apply same filtering logic: volume threshold, probability discrepancy >5%
• Write to Redis under prediction:markets-bootstrap:v1 (merge with Polymarket data)
• Consider: separate Redis key prediction:kalshi:v1 vs combined key

3. Frontend

• Show source badge/icon on each prediction item (Polymarket vs Kalshi)
• Kalshi URLs point to https://kalshi.com/markets/<ticker>
• Same probability bar visualization works (both are 0-100%)

4. Considerations

• Auth: Kalshi requires API key authentication (RSA key-pair) — need to store credentials securely
• Rate limits: Tiered, need to respect limits and add appropriate delays
• Dedup: Some markets may overlap with Polymarket (e.g., election outcomes) — consider dedup by similarity
• Categories: Kalshi categories may differ from Polymarket tags — normalize to shared categories

References

• Current Polymarket implementation: server/worldmonitor/prediction/v1/list-prediction-markets.ts
• Prediction service: src/services/prediction/index.ts
• Seed script: scripts/seed-prediction-markets.mjs

[Rau1CS]

hey! I'm taking this one if that's ok (I'll push the PR in a bit)