DOC: Clarify allowed values for on_bad_lines in read_csv

[wjandrea]

Move the callable options out of the version added/changed tags and improve the flow.

• [ ] closes #xxxx (Replace xxxx with the GitHub issue number)
• [ ] Tests added and passed if fixing a bug or adding a new feature
• All code checks passed.
• [ ] Added type annotations to new arguments/methods/functions.
• [ ] Added an entry in the latest doc/source/whatsnew/vX.X.X.rst file if fixing a bug or adding a new feature.

Check my work to make sure I got the details right, please. One thing is, I'm not sure if "process" is the right term for the PyArrow invalid_row_handler since all it does is say "skip" or "error".

BTW, invalid_row_handler isn't mentioned specifically, so at first I thought "signature as described in pyarrow documentation" was referring to the signature of the ParseOptions constructor.

Also, the description should probably state explicitly what happens when the function returns a list of strings with the right number of elements (with engine='python'). I haven't used it myself, but I'm inferring it gets forwarded to the parser that turns a table of strings into a dataframe.

[mroeschke]

Thanks @wjandrea