Editoast, chartos and api merge proposal

[multun]

Description of the new architecture

The only change to the previous architecture is the merge of chartos, editoast and the api into a single rust service.

Upsides

• there would be single web service queried by the front-end, unifying chartos, editoast and the django api service.
• the postgres database wont be shared by multiple services anymore
• no more invalidating the chartos cache from editoast
• easier refactorings across codebases
• less tech stacks: we would switch from django + fastapi + java + rust to just java + rust
• speedup all kind of processing currently done by the API, such as post-processing simulation results
• consistent policies / authentication are easier to implement

Drawbacks

• migrating costs time
• no more django admin
• rust ORMs aren't yet as fancy as the django ORM
• no more django rest framework
• rust has a steeper learning curve than python

Migration plan

• until the plan starts, avoid adding new code into the django api
• rename editoast to FIXME
• port chartos to rust, merge it into FIXME
• port api to rust, merge it into FIXME
  • migrate all non-trivial views, such as pathfinding and simulation to FIXME
  • move the migration and database ownership to FIXME

[flomonster]

This has been done