Indexing

[kescobo]

This reworks indexing of CommunityProfiles. Previously, I had features and samples set as == if their names were the same. I thought this was a clever way to get indexing into the underlying AxisArray with sample/feature names for free, but might be causing problems with actually constructing them.

This PR removes the equality definitions, so things now have to have their underlying components ==, and then adds methods for indexing with Strings and Regexs. For stings, indexing is based on exact matches on the name, and for regexes, it returns all contains(samplename, regex).

examples:

julia> comm
CommunityProfile{Float64, Taxon, MicrobiomeSample} with 10 features in 5 samples

Feature names:
taxon1, taxon2, taxon3...taxon9, taxon10

Sample names:
sample1, sample2, sample3, sample4, sample5

julia> comm["taxon1", "sample1"]
CommunityProfile{Float64, Taxon, MicrobiomeSample} with 1 features in 1 samples

Feature names:
taxon1

Sample names:
sample1



julia> comm[r"taxon1", "sample1"]
CommunityProfile{Float64, Taxon, MicrobiomeSample} with 2 features in 1 samples

Feature names:
taxon1, taxon10

Sample names:
sample1



julia> comm[r"taxon1", r"sample[15]"]
CommunityProfile{Float64, Taxon, MicrobiomeSample} with 2 features in 2 samples

Feature names:
taxon1, taxon10

Sample names:
sample1, sample5