krcrouse commented on PR #13126: URL: https://github.com/apache/arrow/pull/13126#issuecomment-1236365597
@jorisvandenbossche , Here's the run down of the most recent push to the branch, which includes updates from upstream as of the end of last week: Quick Summary Points (many that condense prior one-off comments): * Based on your comment, I moved `python/pyarrow/_rstutils.py` into `python/scripts/lib/arrowdoc.py` as it is not needed. I don't love a sub-lib of `scripts`, but it's at least straightforward to then have the `generate_sources.py` script reference it. * As I had replied earlier, there should eventually have a step added to the build or CI process that verifies the committed version of `python/pyarrow/_compute_generated.py` is the same thing that is output from the `generate_sources.py` script. * The `archery lint --rat` test is failing because the rst files in `python/docs/additions/compute/` don't have license text and I do not know how do that in a sensible way and don't see it in docs or examples. As for the points you brought up about `docutils` - I wonder if you are misunderstanding its role and how I use it in the PR. It's probably more straightforward with the reorganization in this round. Now, the `docutils` modules is only used in the `scripts/lib/arrodoc.py` file, and there it is only used to parse the reST files in `python/docs/additions/compute/` into the document tree that is then merged/replaced/appended to the docs pulled from the C++ libraries. I did add it to the `requirements-build` and `requirements-wheel-build` files for the PR, but neither of those are really accurate. It's required solely to run the script, and so I felt that it would need to go into some sort of "requirements," but if you're not regenerating the compute functions for python, it's not required for anything. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
