RSE keys don't look very stable

The data that is exported to RSE "looks" a bit scary. See for example: https://github.com/research-software-ecosystem/content/blob/master/data/abricate/abricate.galaxy.json

Those data are meant to be parsed by scripts to do fancy stuff. However, if the keys change frequently over time, this can become cumbersome.

I guess the keys in this file do match 1:1 to column names? Would it be better to map column names to some more stable entities, so that you can change column names but don't change the export?

If not, can we maybe change the column names to no include `()` and such. E.g. convert `"No._of_tool_users_(all_time)_(usegalaxy.org.au)":` to `"no_of_tool_users_all_time_usegalaxy.org.au":`?
All lower case would also help :)

Begin with data/abricate/abricate.galaxy.json and compare its RSE keys with the source column names. Trace how the export is produced, decide whether stable mapping or normalized keys is the intended scope, and define completion as a documented, consistent key policy demonstrated in the example.

Written by the indexing model from the issue text.