Hi,
I think this is fixed on the master branch. With master, I get:
>>> pc.utf8_trim("aba", characters="a")
>>> pc.ascii_trim("aba", characters="a")
while with released pyarrow 4.0.1, the second one indeed raises an error
(not directly sure when it was fixed).
For the documentation, contributions are very welcome! The API
documentation does get generated from the docstrings (for python, they are
listed here: https://arrow.apache.org/docs/python/api/compute.html). But
currently those are only hosted online for the released version.
Best,
Joris
On Thu, 1 Jul 2021 at 07:11, Ying Zhou wrote:
> Hi,
>
> It seems that pyarrow.compute.ascii_trim can not be used without a
> TrimOption. However a TrimOption can not be given as a keyword only
> argument either. This looks like a bug since utf8_trim does not have this
> problem. Is my understanding correct?
>
> Also it seems that there is a lot of Python documentation we need to write
> especially in compute, ORC and CSV (writer). I would like to do my part.
> May I ask whether C++ and Python API documentation on the site are
> regularly generated from the docstrings and comment blocks in released code
> for Doxygen? Thanks!
