Priceflow/task.md

Task Board: Catalog Image Fix

🚀 Current Focus

• Investigate Comparator Price Extraction (User reported false prices)
• Fix Catalog Images (Generic/Missing icons)

📋 Master Plan

• Analyze verify_extraction_logic.py, ai_service.py, ai_schema.py, tracking_scraper_service.py.
• Proposal Phase: Vision Priority accepted.
• Implementation Phase:
  • Update ai_schema.py.
  • Fix scheduler_service.py.
• Verification Phase: Verified with verify_vision_priority.py.
• Catalog Retrieval Fix:
  • Reproduce failure (Confirmed Browserless issue).
  • Fix cataloguemate_scraper.py with HTTP fallback.
  • Cleanup debug files.
• Catalog Image Fix:
  • Analyze page HTML for correct image selectors (Found Thumbor URL).
  • Check PDF availability (Browser Task).
  • Refine Scraper Logic:
    • Update cataloguemate_scraper.py to grab src and prioritize Thumbor.
    • Update routers/catalogues.py to clean up bad images.
  • Cleanup Bad Data:
    • Run cleanup endpoint.
  • Verify fix with new scrape.
• [/] Comparator Price Fix:
  • Reproduce "False Price" issue with script.
  • Audit improved_search_service.py extraction logic (Lowest Price vs AI).
  • Implement robust price validation/filtering.
  • Verify fix.

📝 Progress Log

• 2025-12-22: Vision Priority implemented and verified.
• 2025-12-22: Fixed Catalog Retrieval using HTTP fallback.
• 2025-12-22: Investigating generic icon issue. Found images are served via Thumbor.
• 2025-12-22: FIXED: Scraper now targets Thumbor images. Wiped bad catalogs.
• 2025-12-22: User reported incorrect prices in Comparator. Investigating.