TODOs

[mahmoudmoravej]

• Fixing the GraphQL server error to include data
• Fixing GraphQL request for OPTIONS to avoid returning data (it causes to get data twice)
• The authentication solution is a scrappy one. Probably using ID token (JWT token) for authentication is not correct. We need an access token to use for Berear token. All in all, we need a better, best practiced approach for this kind of SSO approach.
• we need to remove token at logout and also store token at login.
• Running yarn graphql returns error when the server is Ruby.
• Mocking GraphQL for vitest: https://vitest.dev/guide/mocking.html#requests
• Both client and server need a centeral auth management. So, if the user is not authenticated, they are being redirected to login page. Current basic approach in the Managers route is not acceptable and should be replaced
• It is better to load the GraphQL schema file and pass it to graphql.config.ts. Otherwise it wants to load from the main endpoint each time which could affect our develelopment speed.
• Use https://github.com/Shopify/quilt/tree/main/packages/react-form or something similar for form data management.
• Currently we have used "network-only" poliy for GraphQL, if we want to rely on cache, we should fix and find a good solution for it. For now, even updating the "cache" manually, only affect one query not all the object instances (due to unknown behaviour that we haven't explored it yet)
• Maybe better to move to "https://mui.com/material-ui/" (not fully open-source but based on MaterialUI: https://mui.com/pricing/) or "https://mantine.dev/" ? One example: there is no AutoComplete component in the current lib.
• Consider multi-organization model. Each person has its own (hidden) organization by default. But it is important to know if we want to consider "null" as personal org, or we need a real org_id for it. Real org_id stops us to merge both personal and org level expectations. and also sharding based on org_id could be challenging.
• We also need to make a clear diff between org level and personal level settings. If I am CTO of an org while I am using it as my personal coach, how could I differ things clearly?
• We need a refactor in Visions structure. some examples are:
  • Having tons of visions but each individual could have some on his trajectory. So, we probably need some kind of assignment to fix that. In other words, we can have "parent_vision_id" where individuals can take a vision from their organization.
  • Also, a manager wants to assign a vision for an individual. How could do they that? They are still "organizational" vision, but is defined by other people.
• Run Sidekiq service to fetch and process data (activities, visions) on the background.