Add ability to open a plain text GFF

[cmdcolin]

I think this will become more important once we have for example JBrowse desktop. Users should be able to open a plain text gff without having to open e.g. tabix, genome tools to sort it, etc (those are i think difficult barriers to entry)

We could either pre-process the file to a more efficient format and write to disk (maybe a desktop only thing, but may also be useful for users of CLI...they would also be relieved of burden of pre-processing their GFFs) or, we do the simple thing and load the whole thing into memory

[cmdcolin]

Note that the human NCBI refseq GFF is pretty large coming in uncompressed at 537MB

https://s3.amazonaws.com/jbrowse.org/genomes/hg19/ncbi_refseq/GRCh37_latest_genomic.sort.gff.gz

It is quite detailed though and many GFFs might not share the same large size but it does suggest that loading it into memory could have a substantial cost

[cmdcolin]

Note also that this has historically I think been a stumbling block for jbrowse 1 in some cases (people try to open a plain text gff, browser freezes, no fun, and then they post a thread like this where they look for a different tool https://www.biostars.org/p/490779/#491309)

[rbuels]

could parse everything and just keep in memory, or could just hold the file text in memory and scan it and parse chunks on the fly to satisfy region queries

[rbuels]

or could get really fancy and just make some kind of tabix-like index when opened and hold that index in memory