Improves error handling when "shear" but no metadata passed #525
Conversation
|
Fixed the linter error I introduced, but left the other two alone to keep the diff minimal. I can fix these other two here if you like. |
There was a problem hiding this comment.
Thank you @ChrisKeefe! I have one small comment -- I think we could catch this error a bit earlier, but in any case this is really useful to have fixed.
I am not sure about those linting errors -- from what I can tell, your code passes all of the linting stuff on my system (running make stylecheck in the root of the repository), and it seems now to be passing in our CI. Maybe our flake8 versions are slightly different? (Looks like flake8 v3.5.0 and v3.9.2 both accept this PR's code without a problem now.) Not a big deal, though!
EDIT: Actually omg I think I figured out why you got linting errors and I didn't -- so what's happening is the difference is our shell, and/or the make version we have installed. On my system (Ubuntu, using the fish shell), flake8 empress/*.py only expands to catch the python files one level down -- so it doesn't catch the python files in the empress/scripts/ directory. However, I think your system is being smart, and actually is expanding that glob to include all of the stuff in empress/scripts/, which has apparently never been linted until now :O
... So, if you wouldn't mind adding empress/scripts/*.py to the stylecheck Makefile directive, and fixing those two linting errors (it's just adding a few blank lines), that would solve that problem. Thank you for catching that!!! 🌳 🌳 🌳
|
Thanks @ChrisKeefe for catching/fixing the Since you are a first time contributor, I will go ahead an approve the workflows so that our CI can run and after that I this should be good to merge into master. |
When
tree_plotis run withshear-to-feature-metadata, but no metadata is passed, the following error is produced.This PR explicitly handles that case, improving the error message:
