CI testing approaches V2

[RoySegall]

In order to develop better we need to have a second try of the tests. Unit tests are pain in the a$$ and they really test the work with the DB. For that the new approach will be:

• for each service we will have tests for the endpoints/graphql resolvers
• each service will fired up with docker compose
• Adding unit tests for the front and backoffice components
• Once the tests OK, run another github task to fire the services and check the health of services
• Once that passes, set up the front and the backoffice and check the styles is OK, the backoffice allow login, the front allow search with empty array
• after the service will be merged to master we will fire up the services in staging env and run a basic flow of e2e: login, upload all the files, verify the data is processes. Another test will seed all the information and verify the graphs works
• if that pass - cool, if that fail an we will trigger a warning and see what's the problem.

This will need to have:

• creating users outside the container (exposing the user data via bash args)
• using testim