Improve user messaging, logging and error/exception handling

[jsheunis]

The implementation of logging, exceptions, and datalad results rendering in this extension is inconsistent. This stems from my unfamiliarity with these concepts as I started developing. The code base could benefit a lot from an experienced datalad developer going through the code to identify places where the current implementation needs changes.

The modules where this would be most applicable are:

• catalog.py
• meta_item.py
• webcatalog.py
• utils.py
• extractors/datacite_gin.py

[adswa]

I'm noting a few general TODOs and antipatterns that need to be worked on.

• f-strings: Logging and result messages should not be be done with f-strings - instead, we can hand down the template and the arguments as two separate arguments. This increases speed as the string formatting is decided upon by the logger/result renderer and only made when necessary. Instead of

  msg = f"here is something of importance for {var}"
  yield dict(
  ... ,
  message = mgs,
  )

  do

  yield dict(
  ... ,
  message = (
       "Here is something of importance for %s",
       var
  )
  )

• action names: Many of the currently existing result records are populated with action names by hand, such as "catalog validate". We should unify action names and make them comply to the design docs (no whitespace, underscores between words) http://docs.datalad.org/en/latest/design/result_records.html
• The path argument of result records often points to something like abspath(curdir), which might not in all cases refer to the entity that the command acted on. We should change it into something more meaningful, e.g. for catalog_validate not the the path the command is run in, but the metadata file that is validated