DOC: Loading sphinxcontrib.spelling to sphinx only if it's available (#21396)#21397
DOC: Loading sphinxcontrib.spelling to sphinx only if it's available (#21396)#21397jorisvandenbossche merged 1 commit intopandas-dev:masterfrom datapythonista:spelling_deps
Conversation
Good call: I agree that it is an optional dependency. You don't need it to install |
ci/requirements-optional-conda.txt
Outdated
| s3fs | ||
| scipy | ||
| seaborn | ||
| sphinxcontrib-spelling |
There was a problem hiding this comment.
these should go in requirements_dev.txt instead (only)
ci/requirements_dev.txt
Outdated
| pytz | ||
| setuptools>=24.2.0 | ||
| sphinx | ||
| sphinxcontrib-spelling |
There was a problem hiding this comment.
this file indicates it is autogenerated
|
As said on the issue, given the installation troubles, I would make this an optional one. |
jorisvandenbossche
left a comment
There was a problem hiding this comment.
Apart from the comment where to put the dep, looks good to me!
| - pytz | ||
| - setuptools>=24.2.0 | ||
| - sphinx | ||
| - sphinxcontrib-spelling |
There was a problem hiding this comment.
I think we should rather put it in the requirements-pip/conda-optional.txt
There was a problem hiding this comment.
@jreback, you said it should be in requirements_dev.txt, may be you want to explain why? I also thought the optional files made more sense, but you two surely know better than me.
There was a problem hiding this comment.
requirements_dev are for pure dev things
but i think things have gotten confused
There was a problem hiding this comment.
OK, let's discuss that in another one then
|
Just FYI, I just followed the instructions to create a development environment with Conda. On the Is that desired behavior or should the |
git diff upstream/master -u -- "*.py" | flake8 --diffI think the optional dependencies go directly into the
.txtfiles for pip and conda, and not into the yaml file. Please correct me if I'm wrong.Seems like besides
sphinxcontrib-spelling, a binary programenchantis required, and it's not available in conda for what I've seen. Not sure what's the right approach for it.