Add data_frame type to polyphemus metis loaders

[dtm2451]

Adds ability to fill in attributes for multiple records from csv/tsv files + a map of column name <-> attribute pairs.

ToDo:

• [x] Build UI
  • [x] SelectAttributes input (in polyphemus)
  • [x] etna MUI Autocomplete input with property 'multiple' set to true
• [x] json schema for this loader type
• [x] python executor
  • [x] tests for this
• [x] Also work for Table Models
  • [x] UI additions
  • replace/append switch
  • identifier pointers --> parent
  • [x] python additions
  • alternative paths / error catches

[graft]

Testing by trying to put data into ipi demographics model, which is a table (somewhat of a strange case), I get: Job failed with error: Exception('MetisLoaderError("demographics-1.csv is missing column(s) targetted by demographic data_frame loader \'column_map\': id.")')

I'm curious to see how the successful result arrives; I suspect it will be append-only (i.e., it will add items to the demographics table, not overwrite them). Might need an option to allow replacement-of or append-to existing table rows, although I suspect there will remain some further edge cases about this.

[dtm2451]

Oh big oops... As you may have already have figured out or intuited from that error message, this isn't designed with table models in mind at the moment. I had only 'standard' models in mind. I've not worked with table models enough, so definitely worth a pair to ensure I understand the workings and edge-cases to update for!