jbrowse2: fix for pif; deprecate blastxml #1529
Conversation
…m/fubar2/nftoolmaker It actually does work. Sort of. Hope this is useful to someone.
…ply to all future generated tools :)
Allows single ended fastq or fasta to be mapped, or a pair of forward and reverse - both must be fastq
…dio replacement...
Pity the built in fails in biocontainers.
revert to paf until sorted
…y is broken tabix does not do this.
…om the old JB1 converter so deprecating everything Fixes to pif finally working Turns out that the SHELL env variable must be set when calling jbrowse command line functions to prevent node.js from asking for some ID which does not end well in the bare bones biocontainer.
|
@bgruening It was passing tests because everything works except display - the inputs convert, but when loaded as a GFF3 JBrowse2 track, show validation errors with some (not all!) blastxml samples I tried. There have been a number of changes to the GFF parsing noted in recent release notes and it looks like it is now more strict. A GFF3 track loaded into Jbrowse2 from converting a blastxml using that existing tool shows the same errors I saw when loading the same blastxml before I removed it:
Point being, the blastxml input option has been redundant since I started because of the existing coverter. That tool code is from Users can use the existing converter and then load the GFF3 in JBrowse2. The converter can be fixed separately if anyone needs it, but there's no point in having identical complicated code that needs subtle fixing in 2 places - one is more than enough and it is way beyond my skills to fix. |
|
For the record, |
|
Ok this makes sense. Thanks for writing this down! |
@bgruening:
Finally fixed the pif problem
blastxml is now non-functional and has been deprecated because the old JB1 converter emits gff that 2.15.4 refuses to display with an improved parser. All removed including tests...