MacTrack

After spending too much time scrolling through mac.bid, I decided it would be a more effective use of my time to write a tool to more effectively use mac.bid.

This webapp aims to achieve 2 goals:

• notify you when searches matching your criteria are met
• provide historical data to help inform current purchase decisions

Roadmap

This is subject to change, I don't even have a working E2E prototype

Scraper

• ALL datetime in UTC (unverified whether this is actually done)
• sanity checks: how many lots were open before/after scrape
• CLI integration + async rewrite
• queuing system (SAQ)?
  • see here

Filters

• manual running
• "term preview"
  • chips on the cards, + and -, single line only
• queuing system (SAQ)?
• node render process for notification emails
  • see: https://github.com/markfinger/python-react
• quick view - view results of your filter
• join auctionlot_idx with auctionlot for faster querying
  • the former is fast for fts, slow for everything else (nothing can be indexed besides text)

Auth

• make it so that auth actually works
  • set current user if page reload
  • direct to sign in page if logged out (when redirecting user, make sure we don't return them to /current-user)
• load current user on page refresh regardless of auth required or not

General

• deploy backend on GCP
  • get SSL certs (caddy)
  • see: https://github.com/dadatuputi/bitwarden_gcloud/tree/master/caddy
• deploy frontend to netlify (or others)
• logging infrastructure
• PAGINATION

Architecture

Backend

• SQLAlchemy for ORM
• SQLite + FTS5 for DB

Frontend

• React + MUI