slike

'slike, perfect...maybe

this is somewhere between proof of concept and in progress, i'll put some more words here later

TODO

• weather API retry/backoff before failing
• reuse the same token if its still good
• remainder of the hour ttl cache for given coords for weather request
• use the actual logger print is not a logger cmon
• dockerize it
  • intermediate image for scipy/numpy because they take forever to install
• local build 'n run script
• run on fly
• deploy script
• migrate from yaml to fly secrets, populate in deploy script and plain env vars in buildnrun script
• local docker compose orchestration with redis
• fly/upstash redis for prod
• read scoring.json into redis at startup, use it to read from when scoring
• admin endpoint for scoring ingest to populate redis instead
  • split the API call caching and scoring caching, bust scoring when profile is updated
• replace async-cache with redis for forecast
  • endpoint to force evict all caches
• out of memory crashes
• figure out the redis reconnection timeouts
• multiple scoring profiles
  • UI select from available, persist selection in localstorage, include in API calls
  • parameterize ingest, read from profiles, everything
• choose location instead of gps
  • use openweathermap api for geocoding
• clean old caches values from redis? upstash retention settings?
• remove field scores from forecast response, only tag most impactful
• admin routes auth
• make a website that don't suck

WISHLIST

• add aqi fields
  • nobody has hourly data on this, let alone forecasts, so comparisons are mostly pointless

lat long decimal place accuracy

in case you were wondering

decimal places	degrees	distance
0	1	111 km
1	0.1	11.1 km
2	0.01	1.11 km
3	0.001	111 m
4	0.0001	11.1 m
5	0.00001	1.11 m
6	0.000001	11.1 cm
7	0.0000001	1.11 cm
8	0.00000001	1.11 mm