Rewrite EDGE replicant in numpy

[mishtaiwi]

Ben, this is now in a much better state than previously reported on progress. Essentially this is done. there're few items that still needed, and highlighted in a "TODO" comments but these are to do with the reporting, checking and exporting. i.e. the core of the program is now done and QA'ed.

Happy to have a chat if you think is needed.

[BenTaylor-TfN]

Noting down some errors I've spotted:

• Output .csv have headings of "stn_zone_id" instead of "model_zone_id"
• When converting to CUBE .MAT files, only the first row of attractions is correctly converting, with all other rows being 0
  • This is potentially to do with the outputs being out of order. csvs need to be sorted by [col1, col2]. Perhaps an update to csv_to_mat() needed?

[BenTaylor-TfN]

@mishtaiwi , please can you re-open a PR and request pulling this into master? I've merged the EDGE Replicant branch into master now, so we have the working version (with a version number) there as a reference for NPR sensitivities work