but it's possible that this wont work with range requests (referencing https://github.com/GMOD/jbrowse-components/discussions/3172)
i'm not sure i'm getting the same errors as @dprat is reporting though
this is what i'm seeing just for reference https://github.com/cmdcolin/django-jbrowse2-nonworking-example
might be that using a apache server deployed in parallel with the django server or similar might be best for the situation https://docs.djangoproject.com/en/4.1/howto/static-files/deployment/#serving-static-files-in-production
this is what I get
you are able to get past the screen i used because you use fasta index, which only needs a non-range-request to boot up (just requests the full fasta index .fai file with a http 200) and then you hit the range request issue which turning on the bedtabix file
actually, it is slightly different but related
it gives an internal error called (failed)net::ERR_CONTENT_DECODING_FAILED
and that is because it is applying Content-Encoding: gzip which jbrowse does not want to happen
if you renamed the .gz extension to .somethingelse
then it would get past that error but then give you 206 errors
i would consider hosting static files on a different server
that would mean using plaintext bed, plaintext gff
and we would need to create a plaintext fasta adapter (does not exist right now)
but it would not allow any indexed data files like BAM,CRAM,BigWig,Tabix (tabix can be used for VCF,GFF,BED,etc), etc. as these are large and use range requests
Hi everyone, I haven't been here for a while but I have a question regarding the most efficient way to configure dynamically a large number of tracks, and periodically update the track list using jbrowse. I'm not sure it would be appropriate for an issue in GitHub so I thought I'd try here first.
are there any specific details about your setup e.g. the dynamic part?
most of the time, a "static file" with all the tracks is configured (e.g. trackList.json) in jb1 (config.json in jb2)
but you can certainly also write a REST API that generates that tracklist too
that would remove the need to write any configuration files to disk
and all of the changes would be instant
Hi again, so when I was working on a diagnostic lab, I set up jb1 to visualize the outputs of our WES experiments. Each time a new batch of samples were processed, I would automatically generate a new trackList.json incorporating the new data. That way lab members could search directly in jbrowse any sample, and I could craft automatic url links to the data as well.
Now the problem is a bit different, because I'm starting again and could use jb2 directly, the data sits in S3 instead of our local compute disks, and the number of samples is in the (high) tens of thousands, instead of few thousands. I would like a jb2(/1?) instance from which I could directly visualize the CRAM files sitting in S3.
express.js can be a good way to make small REST API code
sure thing :) after the REST API is created, you can use e.g. with jb1 ?data=http://yourhost/path/to/your/restapi (similar idea with jb2)
Hello I am working on a project where I need to be able to upload data files into Jb1 with just a link. Everytime I try to do this nothing happens. No error message but also nothing ever uploads. Wondering if anyone has had the same issue and been able to work around it somehow
"The "name" field (column 4) of a BED file can contain GFF3 style key-value attribute tags by specifying "gffTags=on" on the track line. These attributes will be displayed in the mouse hover popup text." - I often use this capability (https://software.broadinstitute.org/software/igv/feature_track_options) happily in desktop IGV, and less happily in igv.js (https://github.com/igvteam/igv.js/issues/1550).
Is it reasonable to hope it might be (someday?) be supported in jbrowse?
just a random note, we don't try to interpret "track" lines at all...wish bed format was better standardized in general
hi @cmdcolin - no particular reason - that's just what I have at the end of a particular pipeline.
That said, does jbrowse display GFF column 9 attributes on hover? Further, is there any possibility of controlling the label displayed on the glyph in the browser. I presume that by default it displays the value of the GFF name attribute, but is there any override of this default? FWIW - IGV provides for choosing alternate attribute to display as label.
@malcook you can create a javascript callback that will display any information you want. For example, for several WormBase variant tracks (that should come from VCF but actually come from GFF). I pull a bunch of data from the GFF’s column 9. For example, see https://wormbase.org/tools/genome/jbrowse-simple/?data=data%2Fc_elegans_PRJNA13758&loc=III%3A1855046..1860848&tracks=Change-of-function%20alleles%2CChange-of-function%20polymorphisms%2CHigh-throughput%20alleles%2CMillion%20Mutation%20Project%2CPolymorphisms&highlight= and mouseover any features. The function that generates these hovers is variationsBalloon at https://github.com/WormBase/website-genome-browsers/blob/af7ed9a60061e8b0d241fe1cad480f1c8f7ec291/jbrowse/jbrowse/data/functions.conf#L368 and the config bit that makes it work looks like this onClick json: https://github.com/WormBase/website-genome-browsers/blob/af7ed9a60061e8b0d241fe1cad480f1c8f7ec291/jbrowse/jbrowse/data/c_elegans/includes/variations_million_mutation_project.json#L25-L30