Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Proposal for versioned documentation #13

Open
wants to merge 7 commits into
base: master
Choose a base branch
from
Open

Proposal for versioned documentation #13

wants to merge 7 commits into from

Conversation

riedgar-ms
Copy link
Member

Simple proposal for getting started on versioned documentation


1. Update links on the custom-designed landing page to go to this new `main` directory.

1. Create a `v0.4.6` directory in `fairlearn.github.io`, and do a documentation build corresponding to the `v0.4.6` tag in this directory. This may well have to be done manually.
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Everything is built from the fairlearn repo, so we wouldn't manually create these directories, but rather instruct our pipeline to put the results of the builds there. We should keep treating the website repo as something we can delete at any point and it gets fully recreated with the next push.

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Problem then is getting the v0.4.6 documentation (although your point about automation is well taken). Right now, a huge pain point is that people install from PyPI, and then try reading the docs here.

If we're willing to leave that problem to one side (in the hopes of v0.4.7 coming out soon after whatever work this requires), then full automation is certainly an option.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I'm not sure I understand. This problem is unrelated to any particular version since we just need to start at some point, right?

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It's just that v0.4.6 is a particular problem right now. Having said that, the packages you listed below do appear to support pattern matching on tags and the like, so that might be able to do the necessary.

@@ -0,0 +1,23 @@
# Versioned Documentation
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Please link this somewhere as well fairlearn/fairlearn#482

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This mention links it :-)

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Once checking in it won't

1. Create a `v0.4.6` directory in `fairlearn.github.io`, and do a documentation build corresponding to the `v0.4.6` tag in this directory. This may well have to be done manually.

The last stage will at least give us a set of documentation corresponding to the version of Fairlearn on PyPI, which we can share with people who run into difficulties.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Will do some more reading. When I was writing this, I was primarily thinking about the v0.4.6 issue (not to mention the even older versions for which all the sphinx docs will vanish).

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I don't think we need to care. Scikit learn starts somewhere around 0.17 or so and that's fine.

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I didn't mean to go back to the beginning, but I would like to be able to generate docs for v0.4.6.

I've been reading the docs for both sphinxcontrib-versioning and sphinx-multiversion. Both look like they can go back to older versions just fine.

Comparing:
https://github.com/sphinx-contrib/sphinxcontrib-versioning
and
https://github.com/Holzhaus/sphinx-multiversion
I'm a little concerned that the former hasn't had any activity since last December, and it was mostly 'done' back in 2016. I'll start by seeing that sphinx-multiversion can do.

Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Been trying to persuade sphinx-multiversion to work... have wound up filing an issue on it :-/

Signed-off-by: Richard Edgar <[email protected]>
Signed-off-by: Richard Edgar <[email protected]>
Signed-off-by: Richard Edgar <[email protected]>
@MiroDudik
Copy link
Member

Is our expectation that all of our examples will be continuously updated to work with the most recent version? I wonder whether it wouldn't be better to allow installation instructions at the beginning of examples to use, e.g.,
pip install -I fairlearn==v0.4.5.

Especially, if we allow examples to rely on arbitrary visualization packages, it might be better to have a frozen version of the installation instructions for each example notebook?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants