Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
    Marius
    @mvdbeek:matrix.org
    [m]
    ๐Ÿ˜† it's the tool shed, how much more can I possibly break it ?
    Lucille Delisle
    @lldelisle
    ;-)
    Nate Coraor
    @natefoo:matrix.org
    [m]
    a lot!
    Marius
    @mvdbeek:matrix.org
    [m]
    haha, ok, this isn't urgent, but I won't get around to doing it next week
    dannon
    @dannon:matrix.org
    [m]
    Hold onto that thought! I'm working on ripping all the client code out of the main build and expect to break it again soon.
    (the toolshed)
    wm75 (Wolfgang Maier)
    @wm75:matrix.org
    [m]
    I am currently facing the following issue: I have a tool (https://usegalaxy.eu/root?tool_id=toolshed.g2.bx.psu.edu/repos/iuc/vapor/vapor/1.0.2+galaxy2) that expects a multi-sequence FASTA file and, in my case, a list of PE data (representing different samples. The tool will then find the fasta sequence that best matches the reads in each input and will produce a results collection equal in number of elements to the number of PE input samples.
    So far, so good.
    Now, however, I would like to extend this to a collection of multisequence fastas, and I want a run of each sample against each such fasta resulting in a nested list with e.g. samples as the elements of the outer list and assigned best match for that sample for each of the multi-sequence Fastas.
    After many attempts I'm about to conclude that there's no elegant way to achieve this, but maybe I'm wrong?
    The only rather stupid-looking solution I have is to extract, as part of a workflow, each multi-sequence fasta (it's always eight of them) with __EXTRACT_DATASET__, then run vapor on each of those datasets and the list:paired collection.
    wm75 (Wolfgang Maier)
    @wm75:matrix.org
    [m]
    I can live with that if it's the only way, but don't want to build something dumb.
    Marius
    @mvdbeek:matrix.org
    [m]
    So instead of lining up collections to be processed match by match you want to do all possible combinations?
    wm75 (Wolfgang Maier)
    @wm75:matrix.org
    [m]
    yes
    Marius
    @mvdbeek:matrix.org
    [m]
    IIRC we do support that in the backend, but haven't enabled this for lack of a user interface. Just to be sure, can you share a history with your workaround ? I wonder if this could also be done by nesting your collection deeper
    wm75 (Wolfgang Maier)
    @wm75:matrix.org
    [m]
    I tried that actually, but can share, yes
    https://usegalaxy.eu/u/wolfgang-maier/h/vapor-nested-attempts
    Collections 11 and 15 are the relevant inputs.
    Lucille Delisle
    @lldelisle
    Looks too good to be true:
    image.png
    pvanheusden
    @pvanheusden:matrix.org
    [m]
    hey workflow gurus! staramr specifies species using a drop down (this is necessary to run PointFinder) - while the user could set this by changing the "default tool parameters" at pipeline invocation time, is there any way to ask for a user input parameter for the pipeline (so it shows up right at the top like other pipeline inputs) and connect this to the dropdown?
    Marius
    @mvdbeek:matrix.org
    [m]
    Yes, you can click on the add connection to module button, then connect a text parameter to that input
    It'll be shown as a select list
    Note that only works for selects, not conditionals, since conditionals can add arbitrary inputs and outputs, breaking workflows
    pvanheusden
    @pvanheusden:matrix.org
    [m]
    "add connection to module" button is the <-> ?
    Marius
    @mvdbeek:matrix.org
    [m]
    yes
    pvanheusden
    @pvanheusden:matrix.org
    [m]
    ah damn, the staramr input is a conditional - instead of making it "use pointfinder: yes / no" and then "which database", they (PHAC-NML) combined it into one
    Marius
    @mvdbeek:matrix.org
    [m]
    I think this particular conditional could be flattened by moving the unvalidated one into the main select, then checking whether or not the custom flag is needed in cheetah
    Marius
    @mvdbeek:matrix.org
    [m]
    The next workflows call is in 60 minutes:
    Workflow WG call
    Thursday, 17 November ยท 16:00 โ€“ 17:00
    Google Meet joining info
    Video call link: https://meet.google.com/myu-vozy-bsj
    Lucille Delisle
    @lldelisle
    I won't be able to join because I have another seminar at the same time but please keep me updated if you find an agreement on the multiple workflow directory.
    Marius
    @mvdbeek:matrix.org
    [m]
    Sounds like there's some movement on the workflowhub side, but will do!
    Marius
    @mvdbeek:matrix.org
    [m]
    wm75 (Wolfgang Maier)
    @wm75:matrix.org
    [m]
    Sorry, trapped in a grant meeting the whole afternoon.
    Marius
    @mvdbeek:matrix.org
    [m]
    I think I got most of the backend modalities worked out for conditional steps, if you want to have a look at the test case and assumptions, they're in lib/galaxy_test/api/test_workflows.py here: https://github.com/galaxyproject/galaxy/pull/14846/files#diff-62218e13fbfad227cca222eb36d1801285ceb01eb2f9eb0c14aa99175bddd762R1896 ... no UI yet, that's the next thing I'll work on
    Lucille Delisle
    @lldelisle
    This looks super great!
    Thanks
    Just for information, this could be part of 23.0? Is that the idea?
    Marius
    @mvdbeek:matrix.org
    [m]
    Yes, I hope so!
    Guess we got ~ 2-3 weeks to get the user interface in, it seems doable if there's no too many unexpected things popping up
    Lucille Delisle
    @lldelisle
    Some iuc tools are not up to date in usegalaxy.org, this makes some upgrade PR failing, would it be possible to update them (toolshed.g2.bx.psu.edu/repos/iuc/rgrnastar/rna_star/2.7.8a+galaxy1 and toolshed.g2.bx.psu.edu/repos/devteam/bowtie2/bowtie2/2.5.0+galaxy0).
    Thanks
    Marius
    @mvdbeek:matrix.org
    [m]
    I tried this morning, but we have to wait for the bioconductor rebuild to finish
    otherwise we may not have all dependencies on usegalaxy.org
    Lucille Delisle
    @lldelisle
    ok no problem, I did not know you were aware
    Delphine Lariviere
    @delphine-l:matrix.org
    [m]
    what is the rule for subworkflows in the iwc repo? should i have a separate folder with tests or is it enought that they are tested in the main workflows ?
    Marius
    @mvdbeek:matrix.org
    [m]
    If they're useful on their own it certainly is easier for you to have separate tests, yeah
    Delphine Lariviere
    @delphine-l:matrix.org
    [m]
    ok
    M Bernt
    @bernt-matthias:matrix.org
    [m]
    What can/should be done about tool versions used in a workflow that are missing on usegalaxy.org? If someone tells me I would add it here: https://github.com/galaxyproject/iwc/blob/main/workflows/README.md#optional-updating-tools :)
    (its the newest version of the NCBI dataset download tool)
    Marius
    @mvdbeek:matrix.org
    [m]
    that's here: https://github.com/galaxyproject/usegalaxy-tools#requesting-a-new-tool-or-updating-a-tool ... but don't add it, we'll automate this, that's just extra cognitive load
    (also we should be able to install tools as well during tests)
    Lucille Delisle
    @lldelisle
    (When we set up the readme we did not dare putting 'ask Marius if a tool is missing') ;-)
    Marius
    @mvdbeek:matrix.org
    [m]