Unit testing

[leoglonz]

Summary

By opening dMG and hydro_data_dev (potentially hydroDL2) to active development by other users, changes need to be validated so that they do not break or bias existing code. Unit tests will focus on data handling, configuration, and core model components, to catch errors early and support maintainability

Actions

• Define the core testing scope; identify priority modules and functions for testing, particularly those related to configuration parsing, data loading, Trainers, and any other modules (inputs/outputs)
• Write modular, isolated tests for individual functions and classes.
  • Check edge cases
  • Ideally these would be general enough as to be applicable to any dMG modules.
• Integrate with testing framework like pytest so that test execution can be run automatically when changes are pushed to Github. A separate script could also exist in dMG to be run locally by end-users.

Considerations

• Details of unit testing should eventually be included in documentation provided here.
• Test frameworks like pyscripter add complexity, so ideally this and other unit testing infrastructure should be hidden away.

Assignments: @JiangtaoLiud, @taddyb, @leoglonz