RhodiumGroup / docker_images

Docker images for Rhodium's jupyterlab deployments
2 stars 2 forks source link

pep8 nudeges #105

Open delgadom opened 5 years ago

delgadom commented 5 years ago

anything we can do on the image side to push users in a good coding style direction?

anything else?

@bolliger32

bolliger32 commented 5 years ago

Include a best-practices markdown file in the home directory, including telling people how to use flake8?

brews commented 5 years ago

Ian pointed me to this issue. This might be a solution for yall https://github.com/dfm/black_nbconvert?

bolliger32 commented 5 years ago

this is not pep8 related, but could be useful for us...

https://medium.com/@kevinkreuzer/the-way-to-fully-automated-releases-in-open-source-projects-44c015f38fd6

brews commented 5 years ago

I actually think we're pretty close to fully automated releases on python packages, unless you're looking for something specific. The only parts that aren't automated, 1) developer still needs to tag, push commits, 2) conda packages arn't 100% automated in a fool proof way.

As for pep8 stuff. I got to play with black and black_nbconvert last week. I think we could implement either as an optional part of your jupyterhub environments for users, or as a more automated check/autoformat for actual packaged code. I wouldn't recommend it unless you have decent automated unit tests, as I was able to trick black into breaking code. It's very good but it's not bomb proof (and I don't think we should think of any auto-styler as bomb-proof).