[P2] Redfin scraper — data normalizer

Parse and normalize Redfin Apify output:
- Map Redfin fields → properties table
- Extract price history → property_history (TimescaleDB)
- Extract image URLs → queue for MinIO download
- Store raw JSON blob — never lose data
- Handle duplicates (upsert by address + source)

No files or tests are named. Start by locating the TypeScript entry point that receives Redfin Apify output and the code handling properties, property_history, MinIO downloads, and queueing. Done means all listed fields are normalized, price history and images are handled, raw JSON is retained, and duplicates upsert by address and source.

Written by the indexing model from the issue text.