Implement --doctest-plus-generate-diff to fix existing docs #227
Conversation
|
This tries to address gh-107. EDIT: Sorry for failing to run the normal tests on this... Hopefully fixed now. |
Can you use or add to https://github.com/scientific-python/pytest-doctestplus/tree/main/tests ? |
seberg
force-pushed
the
generate-diff-option
branch
from
06e4bed to
cb17e00
Compare
|
Added one test, checking the very basic modes. I also tried to make sure that we imply So unlike the very first push, I think this now should work and ready for a closer look, which doesn't mean we may not want to reorganize the approach! |
|
@bsipocz , are you able to review this? I feel like this is going to be very useful when numpy 2.0 is released and people want to update their docs to match. 🙏 |
There was a problem hiding this comment.
Thank you so much @seberg!
Minimal wordsmithing and bike-shedding are attached.
The bottom line, I tried this, and it works wonderfully, as expected. I feel the suggestions below make some sense, but feel free to take or leave any of them.
The one more thing that would be nice is a line in the changelog file (CHANGES.rst), but, we certainly can do that at release time.
| "folder and use `git diff -p` to generate a diff."), | ||
| choices=["diff", "overwrite"], | ||
| action="store", nargs="?", default=False, const="diff") | ||
|
|
||
| parser.addini("text_file_format", |
There was a problem hiding this comment.
I wonder whether a parser.addini is needed for this new option, too?
There was a problem hiding this comment.
Not sure, that would mean you could specify things as a config file? To some degree, I am not sure I even want that (because I don't really think it is a good habit to run this in CI). But I don't mind adding it either.
pytest_doctestplus/plugin.py
Outdated
| "causes editing of the original files.\n" | ||
| "NOTE: Unless an in-pace build is picked up, python " | ||
| "file paths may point to unexpected places. " | ||
| "If `"overwrite"` is not used, will create a temporary " |
There was a problem hiding this comment.
| "If `"overwrite"` is not used, will create a temporary " | |
| "If 'overwrite' is not used, will create a temporary " |
Just going to use single ' and no backtick.
Thanks for the other suggestions and fixes, they looked great!
Had tested the diff creation, but missed testing the normal operation...
Co-authored-by: Brigitta Sipőcz <b.sipocz@gmail.com>
seberg
force-pushed
the
generate-diff-option
branch
from
7e814f3 to
269873e
Compare
|
Thanks @seberg! |
This implements an approach to allow fixing existing docs (both
.rstfiles and inline docs).pytest-doctestis not always good at guessing where code came from, which makes using this for NumPy unfortunately a bit hard:__module__which confuses things (the best way to fix this is really to just make the decorator(s) that does this a no-op).As documented, I think this type of tool should always be single-shot use and there should not be any guarantees for it working without any changes.
(In other words, I hope that it is clear that the hooks as implemented here could change!)
It was a while ago that I first wrote this code. @pllim or @bsipocz do you think there is a good way to add some very minimal tests, or do you just want to try it out and see how it goes?
Fix #107