[CWM] Design a functional back-end

[rogup]

Description

We have a basic NodeJS back-end server running, which returns mock API responses.

The next step is to create a functional back-end server and implement the API methods properly, so that they return real responses for any given dataset.

Acceptance Criteria

• [ ] We have a document that proposes a design (possibly with alternative options) for the following:
  • The initiatives data file format (e.g. CSV or other format, which fields, any pre-generated indexes?)
  • How the back-end reads and stores the data at runtime
  • Overall app architecture (e.g. classes, state)
  • Data flows between the BE/FE when a client map is initially launched, and then when it performs subsequent searches.
  • How the back-end may be extended in the future to also handle other Mykomaps
• [ ] The above document is shared and reviewed by dev team, and all design decisions are made
• [ ] (maybe after implementation) the design is summarised in the codebase docs.

[rogup]

Design doc WIP https://docs.google.com/document/d/1Y2MZ-Q92bhFCit-SoivpriOsxQZIbJop4ewiCqkYRNU/edit?usp=sharing

[rogup]

@wu-lee Please could you review this back-end design? In particular, I had some questions (in yellow) about where the data should actually live. I was wondering whether we actually need a separate api.maps.coop and data.maps.coop, since we'll just be duplicating data anyway. It makes intuitive sense to me that the data factory pipeline is the thing that generates the data in the correct format to be consumed by the back-end.