Bring catalog-core extractor into main and add same for catalog-runprov

[jsheunis]

The metalad_core extractor was previously ported into datalad-catalog as a standalone script, specifically in the abcdj branch..

See:

• 35b1394
• 7f37391

The idea is to "have a self-contained script that could in future just be ripped and replaced with whatever new functionality supersedes" it, and the goal of the script would be to extract runrecord metadata from a datalad dataset and output it as a catalog-ready entry, without having to depend on metalad or jq bindings used by datalad-catalog's translation component.