fix: builtins/publish: keep xs+xsi NS in cleanup

[vladimir-mencl-eresearch]

Fixes #333

The lxml.etree.cleanup_namespaces function only considers namespaces used in XML Element and Attribute names, but not in attribute values. The xs namespace is used only in values of xsi:type attributes and is thus not considered as used by LXML, and would get removed.

Keep it explicitly to avoid creating invalid metadata.

All Submissions:

• Have you checked to ensure there aren't other open Pull Requests for the same update/change?
• Have you added an explanation of what problem you are trying to solve with this PR?
• Have you added information on what your changes do and why you chose this as your solution?
• [NotYet] Have you written new tests for your changes?
• [N/A] Does your submission pass tests?
• This project follows PEP8 style guide. Have you run your code against the 'flake8' linter?

[vladimir-mencl-eresearch]

FYI @theseal @btmattsson

[vladimir-mencl-eresearch]

Hi,

I have looked into this a bit deeper - I was surprised #333 broke only conventional metadata but not MDQ.

Turns out the namespace cleanup was skipping over MDQ (cleanup_namespaces does nothing when invoked on an element that is not the root element).

So I have moved the namespace cleanup to where the document is created - that solves it also for MDQ files (and I have kept keep_ns_prefixes that fix for #333).

Please let me know what you think. If you'd rather just fix #333 only with no other changes, the first commit will do it. The second commit makes NS cleanup apply also to MDQ files.