Support multi-column filtering for DF2TFilter

[martinholters]

Fixes #372. I mostly wanted to see whether #372 could be tackled in a non-breaking way. The answer is yes, so this could wait until after release of v0.8. OTOH, it seems simple enough that it's unlikely to stall v0.8 even more, so here it is.

Note that while this is API-wise orthogonal to #599, there will be some clean-up possible after both of them have been merged. (Or rather, after the first one merges, the second one can be extended to include the clean-up.)

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 97.91%. Comparing base (be65e83) to head (307428a).
Report is 1 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #602      +/-   ##
==========================================
+ Coverage   97.90%   97.91%   +0.01%     
==========================================
  Files          19       19              
  Lines        3246     3262      +16     
==========================================
+ Hits         3178     3194      +16     
  Misses         68       68              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[wheeheee]

LGTM, apart from documentation and some formatting.

[martinholters]

How do you like the rewritten docstring?

[wheeheee]

The explanation about what coldims and sitype do is much clearer after the rewrite.
Actually, I think I was just unfamiliar with what the square brackets imply about the number / configurations of optional arguments that can be provided. So we could just merge now, or follow what JuliaLang/julia#4902 proposes, since maybe not everyone is that familiar with the bracket syntax.

[martinholters]

Yes, worth considering. (For docstrings being edited anyway; I wouldn't want to go through the whole code base and change this.)

This could become

    DF2TFilter(coef::FilterCoefficients{:z}, coldims::Tuple = ())
    DF2TFilter(coef::FilterCoefficients{:z}, sitype::Type, coldims::Tuple = ())
    DF2TFilter(coef::FilterCoefficients{:z}, si)

or without using default args

    DF2TFilter(coef::FilterCoefficients{:z})
    DF2TFilter(coef::FilterCoefficients{:z}, coldims::Tuple)
    DF2TFilter(coef::FilterCoefficients{:z}, sitype::Type)
    DF2TFilter(coef::FilterCoefficients{:z}, sitype::Type, coldims::Tuple)
    DF2TFilter(coef::FilterCoefficients{:z}, si)

The latter seems excessive, although I like that the one-argument case is more explicit, as it will probably by the one most often used. So maybe

    DF2TFilter(coef::FilterCoefficients{:z})
    DF2TFilter(coef::FilterCoefficients{:z}, coldims::Tuple)
    DF2TFilter(coef::FilterCoefficients{:z}, sitype::Type, coldims::Tuple = ())
    DF2TFilter(coef::FilterCoefficients{:z}, si)

?
I think I prefer the last option. I'll push a change with that, but am open to switching to one of the others.

[wheeheee]

Looks good!

[martinholters]

Ok, will merge once CI is green.