Points DRF API

1. create .env file based on .env.dist
2. make run-app to start the app at http://localhost:8000
3. make run-migrate to apply migrations
4. make run-import-users-data to import users from users.csv file (or use python manage.py import_users_data <path_to_file> in docker container)
5. make run-tests to run tests

• List of users with their points balance: http://localhost:8000/api/points/
• Subpage with form to add/remove points is available from list view.
• Swagger documentation is available at http://localhost:8000/swagger

Assumptions:

• first name, last name, email are required fields,
• user's email is unique and is used to identify user,
• if users.csv contains invalid data (e.g. invalid balance) it will be skipped and logged,
• if users.csv contains invalid referral user (e.g. user does not exist in system) it will be skipped and logged,
• ids (row numbers) from users.csv are not imported (database ids are used)
• points balance can be negative as described in the task, no maximum/minimum value is set,
• points sources are kept in separate table to allow adding new sources in the future (via admin panel), initial sources are created during migration,
• changing balance requires existing points source,
• points history keeps points changes made with service after account creation
• permissions are not implemented (all endpoints are public),

Possible improvements:

• when balance is changed receiver is using account's post_save signal when handler will raise exception it will rollback transaction - should be clarified with product owner if this is the desired behaviour, sending emails could be moved to queue and executed asynchronously,
• add points maximum/minimum value and validate it during points addition/subtraction,
• separate sources by deposit/withdrawal and validate if user has enough points to withdraw,
• using custom user model based on Django more suited to the project scope (authentication, registration, account management),
• using context manager for signal disconnect/connect,
• add csv schema validation,
• adding context to source (e.g. referred user id, tournament id etc.)
• worth consideration is disconnecting balance change signals in tests environment,
• testing receivers should mock method called inside handler (e.g. send_email),

Known issues:

• users.csv contains row with invalid balance ('59O' - should be '590'),
• users.csv contains row with invalid referral user ('[email protected]' does not exist in system),
• api is using html templates so requests from swagger are not accepted (fixable by using json renderer),