Garnett - command not found error

[hexhowells]

Running the Garnett - classify cells tool in galaxy produces the following error:

Fatal error: Exit code 127 ()

/data/jwd05e/main/061/429/61429169/tool_script.sh: line 26: garnett_classify_cells.R: command not found

Seems some of the backend scripts are missing and are unable to run.

[pcm32]

if we try with the container that would be used by galaxy on a container based setup:

docker run -it quay.io/biocontainers/garnett-cli:0.0.5--hdfd78af_1 bash

root@8aa407ff1d47:/# which garnett_classify_cells.R
/usr/local/bin/garnett_classify_cells.R

so I would suggest to check whether the conda env was created correctly, or even better to avoid any issues with all conda packages, switch to docker or singularity for dependency resolution in Galaxy.

[hexhowells]

I'm not too sure I understand what you mean by checking the conda env, I am trying to run the tool in a Galaxy workflow, here's a link to a Galaxy history that replicates the issue:

https://humancellatlas.usegalaxy.eu/u/hexhowells/h/garnett-error

[pcm32]

Apologies, I though that you were running your own (private) Galaxy instance and hence controlled how dependencies are resolved.

Hi @bgruening, I suspect that this is an issue with the conda environment being used on usegalaxy.eu. Could someone take a look please? I was assuming that usegalaxy.eu was resolving against containers rather than conda environments. At least the biocontainer seems to have the missing command, so I'm surprised. Thanks! Let me know if I should be opening this as an issue somewhere in the galaxy.eu github repos.

[bgruening]

@hexhowells can you please try again?
@pcm32 we are not yet running all tools in containers. It's a slow migration of 3000 tools :)

[hexhowells]

I have run it again and the error looks to be resolved thank you, however the tool still doesn't seem to be running correctly, I have opened another issue #301 detailing the new error.