Automate testing of em-public-dashboard plots

[iantei]

• Initial setup of Github Action
  • Introduce test_with_docker.yml file
  • Calls when we have git push or pull request in main branch
  • Calls up docker-compose.tests.yml which will subsequently call start_tests.sh, that will trigger pytest for test_plots.py
• Initial unit test cases for plots.py using pytest and coverage
  • Since pytest will be directly run through python file, rather than notebook. Updated plots.py to allow this.
  • Execute unit test using pytest for initial function of calculate_pct() and merge_small_entries()

Current coverage progress:

notebook-server-1  | ======================== 6 passed, 12 warnings in 1.94s ========================
notebook-server-1  | Name                                    Stmts   Miss  Cover
notebook-server-1  | -----------------------------------------------------------
notebook-server-1  | /usr/src/app/saved-notebooks/plots.py     303    248    18%
notebook-server-1  | test_plots.py                              48      0   100%
notebook-server-1  | -----------------------------------------------------------
notebook-server-1  | TOTAL                                     351    248    29%

[JGreenlee]

LGTM, great start for coverage!

Testing the notebooks themselves will be a separate challenge but I think this PR is a strong first step because it introduces pytest and a GH workflow

Is there an issue / PR somewhere to discuss further automated testing? I think pytest-notebook looks promising, or even just having assert statements embedded in the notebooks

[iantei]

Is there an issue / PR somewhere to discuss further automated testing? I think pytest-notebook looks promising, or even just having assert statements embedded in the notebooks

#171

We have the above Issue where we have discussed about the high level overview of automated testing for the public dashboard