Seems that one can set the number of CPUs in medaka with an env variable: https://github.com/nanoporetech/medaka#usage
Seems to default to
NPROC=$(nproc) (which is all processors)
OK. Only applies to
medaka_consensus
So the issue with medaka was that I had a redundant scipy dependency declared in the
variant wrapper. Beats me why it was the issue exactly but it seems to have been because removing that dependency removes both the issue with discrepancy in output between conda and biocontainer resolution (which makes sense) and the issue with segfaulting (which makes less sense to me)
something that is linking against a library that is not supported and not pinned? Who knows. Glad you figured that out!
A data manager can both create entries in multiple data tables (checkout the bowtie2 data manager for that). If you want to add multiple entries the produced data_manager.json should contain the additional entries.
{
"data_tables": {
"bowtie2_indexes": [
{
"dbkey": "phiX174",
"name": "phiX174",
"path": "phiX174",
"value": "phiX174"
}
],
"tophat2_indexes": [
{
"dbkey": "phiX174",
"name": "phiX174",
"path": "phiX174",
"value": "phiX174"
}
]
}
}
for example
{
"data_tables": {
"bowtie2_indexes": [
{
"dbkey": "phiX174",
"name": "phiX174",
"path": "phiX174",
"value": "phiX174"
},
{
"dbkey": "phiX175",
"name": "phiX175",
"path": "phiX175",
"value": "phiX175"
},
],
}
}
ah...
--skip_client_build
are the filters that can be used with
<options from_data_table...> documented somewhere?
ah thanks. I see
sort_by does not have an order attribute... pity... because I'd really like to specify descending order...
Hi all -- Did I guess the answer right? Is a bit sparse and not exact, maybe someone can help add more details? It seems that they are not pulling the correct version of python into their environment. https://help.galaxyproject.org/t/conda-detected-but-not-work/5844 Thanks!
I was wondering if there were plans to wrap the unimap tool? This is now recommended for some types of alignments over minimap2 (as mentioned in the latest minimap2 release notes: https://github.com/lh3/minimap2/releases/tag/v2.18). Thanks if possible!
I replied to @jennaj on https://help.galaxyproject.org/t/conda-detected-but-not-work/5844
I wanted know how you launch a job ? Do you use popen ?
Best regards
I wanted know how you launch a job ? Do you use popen ?
Best regards
@studyfranco Does your conda command (
/opt/conda/bin/conda create -y --quiet --override-channels --channel conda-forge --channel bioconda --channel iuc --channel defaults --channel R --name mulled-v1-5b73db27209dd7e83d26bd4dedeadbcf356c9297c08644dec23df66a734a236c python=2.7 scipy matplotlib Pillow reportlab tk ConfigParser) succeed on the command line ?
You may have combined incompatible versions, and then Conda can’t create the environment
Also please don’t set
conda_exec
You also don’t want to set
conda_auto_init
The conda_prefix must match your conda setup, so I guess that should be
/opt/conda
I'm writing a wrapper for
nextclade which has a bunch of optional file inputs (e.g. --input-qc-config). If I make these visible as <param type="data" optional="true"> it seems they are assigned an input even if the chooser shows "Nothing selected". Is there a way to have optional dataset inputs without guarding each one in a <conditional> to activate / deactive it?
If I make these visible as <param type="data" optional="true"> it seems they are assigned an input even if the chooser shows "Nothing selected".
This should be totally fine.
what do you mean with "are assigned an input"?
Galaxy picks up the most recent dataset of the type of the input and assigns that...
so for example
input-qc-config is type JSON input... it picks up a JSON from the history and adds that to the command line
let me make a minimal example to share.
@mvdbeek Yep, the command succeed when I do it manually. Galaxy detect this environment.
In fact, the command: . '/home/galaxy/.conda/bin/activate' '/home/galaxy/.conda/envs/mulled-v1-5b73db27209dd7e83d26bd4dedeadbcf356c9297c08644dec23df66a734a236c' > conda_activate.log 2>&1 work if I use it.
And when I launch the script command, it work fine too. This is why I am lost.
In fact, the command: . '/home/galaxy/.conda/bin/activate' '/home/galaxy/.conda/envs/mulled-v1-5b73db27209dd7e83d26bd4dedeadbcf356c9297c08644dec23df66a734a236c' > conda_activate.log 2>&1 work if I use it.
And when I launch the script command, it work fine too. This is why I am lost.
conda_prefix: /home/galaxy/.conda
conda_exec: /opt/conda/bin/conda (I have a symbolic link: ln -s /opt/conda/bin /home/galaxy/.conda/)conda_prefix to /opt/conda/
Remove
conda_exec too please