Viewing empty chromosome for CRAM produces large seqFetch

[cmdcolin]

If a particular chromosome for a CRAM file is empty, the stats estimation finds it is "small enough" to display the data, but since it decides to display it, it triggers a seqFetch of the entire chromosome corresponding with this bug on cram-js GMOD/cram-js#79

[cmdcolin]

we could (a) make a special case for CRAM (b) make it so alignments tracks never display such a high zoom level by default (back to bpPerPx limits, with force load option) or (c) fix the seqfetch option (maybe can lazily fetch sequence)

[cmdcolin]

example way to create this scenario

create small snippet of a larger file

samtools view -T 'https://jbrowse.org/genomes/hg19/fasta/hg19.fa.gz' 'https://s3.amazonaws.com/jbrowse.org/genomes/hg19/ultra-long-ont_hs37d5_phased.cram' 22:1-100 -o out.cram

load track

jbrowse add-track out.cram --out test_data/config_demo.json -a hg19 --load copy

then visit chr1 in the browser