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

Get the transform docs into the API pages. #3690

Merged
merged 12 commits into from
Nov 26, 2019

Conversation

rpgoldman
Copy link
Contributor

I'm marking this as WIP, because the entries for the transform instances in the docs are not formatted as well as I would like.

Someone should also vet my additional exports. I found that the transforms were not being exported from distributions into pymc3 -- some of them were dropped and only exported from distributions.

If we want to have a hierarchical namespace like this, we should probably spend a little time deciding how it will be laid out.

The doc strings for the transforms were not linked into the documentation page for distributions.
from .transforms import logodds
from .transforms import log
from .transforms import sum_to_1
from .transforms import *
Copy link
Member

Choose a reason for hiding this comment

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

I like this better than the current method. Two things:

  1. This is technically a breaking change, since pm.distributions.transform will no longer work. I bet this messes up 0 people, but we should add a deprecation warning before removing.
  2. As a suggestion, I think we should surface transforms as pm.transforms, so that you could use pm.transforms.ordered() instead of pm.distributions.ordered() (In fact, I think right now pm.log might be a transform, and not a logarithm?). I think we would do this by not importing anything from .transforms here, but in the top level init, import .distributions.transforms as transforms. How does that sound?

Copy link
Contributor Author

Choose a reason for hiding this comment

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

WRT your (1), I think we could just not export transform, since it's just an alias for Transform. That's also theoretically a breaking change, but seems even less likely to cause troubles. Especially since one should not use transform as the name of a class in python.

(2) Sounds good to me.

LMK what you think about my responses and I will make the changes in the PR, or I believe you could do them yourself, if that's what you prefer.

Copy link
Member

Choose a reason for hiding this comment

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

Both of those sound good to me. Feel free to ping after you mess with those imports I'll kick the tires locally (and add tests if I find anything :) ) and then merge. Ooh, I'll check the doc thing you just mentioned, too.

Copy link
Contributor Author

Choose a reason for hiding this comment

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

Wait -- why is pm.distributions.transform no longer going to work? It's in the __all__ list in transforms.py, so I think it's still ok, isn't it?
But I like the idea of deprecating it, since for a class Transform is preferable. Also transform, not being an initialized entity, doesn't serve the same purpose as the other lower-case names.

@rpgoldman
Copy link
Contributor Author

When I look at the formatted docs for the transforms, the parameters are displayed like this

**foo**type

instead of

***foo** : type

as they should. Please have a look and see if this works for you -- it could be just a local fail, in which case we can ignore it.

I really don't know what to do about what sphinx is doing with the ::autodata::.

@rpgoldman
Copy link
Contributor Author

Actually, my autosummary also seems broken.

The transforms code has lower-case alternatives to the upper-cased
class names.  Sometimes these are instances of the classes (e.g.,
`log` is an instance of the `Log` transform), but sometimes they are
aliases (e.g., `lowerbound` is an alias for `LowerBound`).

I couldn't find a good way to document these in the source file, so
they are hand-documented in transforms.rst.
@codecov
Copy link

codecov bot commented Nov 24, 2019

Codecov Report

Merging #3690 into master will increase coverage by <.01%.
The diff coverage is 100%.

Impacted file tree graph

@@            Coverage Diff             @@
##           master    #3690      +/-   ##
==========================================
+ Coverage    89.9%   89.91%   +<.01%     
==========================================
  Files         134      134              
  Lines       20184    20189       +5     
==========================================
+ Hits        18147    18152       +5     
  Misses       2037     2037
Impacted Files Coverage Δ
pymc3/distributions/continuous.py 80.85% <ø> (ø) ⬆️
pymc3/distributions/__init__.py 100% <ø> (ø) ⬆️
pymc3/distributions/transforms.py 87.65% <100%> (ø) ⬆️
pymc3/sampling.py 83.09% <0%> (ø) ⬆️
pymc3/tests/test_sampling.py 99.54% <0%> (+0.01%) ⬆️

Thomas caught a typo of mine, and when fixing it, I picked off some grammar errors and improved some typesetting.
@twiecki
Copy link
Member

twiecki commented Nov 25, 2019

Did you compile things locally to ensure everything looks right?

@twiecki
Copy link
Member

twiecki commented Nov 25, 2019

And needs a line in release-notes.

@rpgoldman
Copy link
Contributor Author

Did you compile things locally to ensure everything looks right?

I did, but .... they don't. See my discussion with @ColCarroll above:

  1. The autosummary doesn't work for me. This is probably at least partly due to my hand-documenting the transform aliases that are exported. I do not know how to fix this. If anyone else does, that would be great. Otherwise, I can hand-hack a toctree here.
  2. Parameters are not properly typeset for me. I have seen this before, but don't know what causes it. I don't know whether this is a problem for others, or something wrong with my sphinx set up. For me it seems to be all over the PyMC3 docs, not just here. Here's an example from the distributions api pages:

image

@rpgoldman
Copy link
Contributor Author

And needs a line in release-notes.

Done!

Fix typesetting of parameters in the docstrings.
I still don't know how to get the title of *this* page *not* to appear in its table of contents.
Unfortunately, `autosummary` does not work at all.
While I was at it, fixed the punctuation of parameters in the docstrings so that Sphinx will format them correctly.
@rpgoldman
Copy link
Contributor Author

I've taken this as far as I can go with it. .. autosummary:: doesn't work because of something to do with the aliases that I had to document by hand (couldn't figure out how to make Sphinx do that properly).

I'm going to take off the "WIP" -- @ColCarroll will you please have one last look and then squash merge?

@rpgoldman rpgoldman changed the title WIP: Get the transform docs into the API pages. Get the transform docs into the API pages. Nov 26, 2019
@ColCarroll
Copy link
Member

Built docs here -- I found the transform docs, but would love if you confirm they're what you expected!

https://pymc3.colindcarroll.com/

@rpgoldman
Copy link
Contributor Author

That looks as I had expected, thanks, @ColCarroll

@ColCarroll ColCarroll merged commit 9c4b740 into pymc-devs:master Nov 26, 2019
@ColCarroll
Copy link
Member

thanks @rpgoldman -- careful and valuable work as always.

@rpgoldman
Copy link
Contributor Author

Thanks for the kind words -- couldn't have done it with the help of a bunch of the devs' kind help.

@rpgoldman rpgoldman deleted the transform-docs branch November 26, 2019 19:03
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