Closed maximlt closed 5 months ago
All modified and coverable lines are covered by tests :white_check_mark:
Comparison is base (
8fb873c
) 85.65% compared to head (28065e5
) 85.65%.:exclamation: Current head 28065e5 differs from pull request most recent head fbc5ff8. Consider uploading reports for the commit fbc5ff8 to get more accurate results
Ugh! Thanks for dealing with that.
The doc build was broken with:
It turns out Datashader doesn't use the
sphinx.ext.napoleon
extension to process its docstrings but numpydoc, which does more validation (than I would have expected at least) and wasn't happy with some recent changes made in xarray (see https://github.com/pydata/xarray/issues/8596). I tried:sphinx.ext.napoleon
but the API page didn't look the same and I didn't want to find what changed exactly and if it was better or worse.So I ended up figuring out how to disable all the validations numpydoc performs. We probably need to revert that some day.
EDIT: Meh I must have done something wrong in my testing, this doesn't work yet.
EDIT2:
Datashader uses
sphinx.ext.autodoc
(e.g. theautomodule
directive) for its API reference and automatically includes a module that contains theImage
class.Image
inherits fromxr.DataArray
. Datashader uses numpydoc to parse the docstrings. It turns out xarray affected numpydoc in some way (see https://github.com/pydata/xarray/issues/8596). I found a (bad) hacky way to work around that.