Convert text dumps to json format

[emdupre]

Comes from #24 (comment)

Right now, csstepdata is generated by evaluating variables and saving out the result to a text file. Something similar happens in seldict, although there the result is returned as a dictionary.

It would be great to instead zip together a list of the variables with descriptive identifiers. Then, if the dictionary is being saved out, we could use json.dumps() to write out a structured instead of plain text file.

[rmarkello]

This was partially resolved with #22!

[emdupre]

Yes, I realized in doing it that the issue with seldict is actually slightly different. I might be ok to close this for now and address seldict later as part of #16. WDYT ?

[rmarkello]

That seems fine with me! I would, however, potentially consider raising a separate issue for reorganizing the comp_table.txt file generated via tedana.interfaces.tedana.writect() to something a bit more JSON-like, too?

[emdupre]

Yes, I think that's fair. That might even be a more accurate description of this issue... seldict is coming up in #35 if I want to keep linting

[tsalo]

This was originally about the component table format, right? Given that we're now using a csv format for those tables, should we close this?

[emdupre]

Agreed ! Yes, this was still open if we wanted to use JSON instead, but that's overkill at this point I think. Closing :)