[ARROW-17819] [Doc] Remove dependency on PyArrow from rst build process

• Type: Improvement
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Component/s: Documentation
• External issue URL: https://github.com/apache/arrow/issues/33041
• Assignee: Unassigned
• Reporter: Larry White
• Created: 22/Sep/22 17:35
• Updated: 11/Jan/23 11:55

Description

The process for building the tutorial documentation (see: apache/arrow/docs/source/developers/documentation.rst) identifies PyArrow as a requirement

Ideally, it would not be necessary to install PyArrow to modify text documentation for Java or any other language. 

FWIW, On my machine, the PyArrow install has been running for over 20 minutes with repeated failure messages:

conda install -c conda-forge pyarrow

Collecting package metadata (current_repodata.json): done
Solving environment: failed with initial frozen solve. Retrying with flexible solve.
Solving environment: failed with repodata from current_repodata.json, will retry with next repodata source.
Collecting package metadata (repodata.json): done
Solving environment: failed with initial frozen solve. Retrying with flexible solve.
Solving environment: |

At this point I cancelled this and used pip to install, which finished successfully almost immediately. However, the next step (run python setup.py install) also failed. And failed again when I tried to run it in a dedicated Conda environment.  

Rather than continue with that, I tried to run the java documentation script: (run make java_tutorial in the arrow docs folder) and it was able to execute without issue, so I could see my changes as HTML.  

Issue Links

• relates to: ARROW-17076 [Python][Docs] Enable building documentation with pyarrow nightly builds