Milestone 2 Feedback

[tonyliang19]

Milestone 2 Feedback Group 11

Criterias for Milestone 2:

Functions are well documented

• No function mentions input types and the parameter meanings -4 #30

Helper data for tests is tractable and reproducible

• Messy and repeated helper data that could be separated into a different file and import it -2 #31

Tests

Tests are written for each function, and work expected

• test_show_correct() in test_grapher failed -1 #32

Tests are well documented

• Quite generic test documentation across all tests (you tend to have similar messages and some are not as described) -2 #35

Tests suite is of high quality

• All tests of functions are too simple, either one or two assert statements (isnt quite robust), not checking type of data after transformation, or validating input types -8 #36

Analysis can be run reproducibly using doc in README

• HTML could not be found after running your command in the README.md, again missed option for different OS systems. -8 #33
• Limit option to build image locally (This took EXTREMELY long, and could take even longer once you have more dependencies), you could've provide either of build or pull the image directly from your DockerHub. -3 #34

Side note: I highly do not recommend having python binary as your image, this makes it hard to run the codes or modular files since it lacks interactivity

[sgskung]

Hi @tonyliang19, thank you for the feedback for Milestone 2 release, I have a few questions about the reproducibility section if it's not too much trouble:

1. What other options exist for a python project besides the python binary?
2. To account for other OS systems, would we provide instructions in the README.md or account for it within the build itself (Dockerfile, etc.).

Thank you so much for your time and sorry for the inconvenience!

[tonyliang19]

1. Use a Jupyter based image given you are using the *ipynb file as analysis, unless you want to transfer your analysis into pure *.py file (that is one python script that runs the whole analysis there) then keep with your python binary.
2. You could provide instructions in the README.md, since the latter one is harder (this assumes you have good understanding of docker container and differences on OS systems)