Yes, exactly. Since that is our setup, it would be...
# configs
m
Yes, exactly. Since that is our setup, it would be preferable to have a test that is as close to our setup as possible. But any test helps to narrow it down. Since I can see some profiles (those like
singularity
,
test
,
wave
which are declared right in the main
nextflow.config
), I know that the problem must be the
includeConfig
of the institutional profiles that is no longer happening. I have already checked that
$NXF_OFFLINE
is unset and
params.custom_config_base
is set. The latter is, however, entered in the Pipeline Parameters box and may therefore only be transmitted once I actually launch the pipeline from the GUI and not yet, when I the agent clones the pipeline to cache to evaluate the config and retrieves available profiles. Alternatively, it could be related to the HPC agent setup (that modern pipelines work on say AWS) or the local installation (that pipelines pulled from GitHub directly work, whereas downloaded ones do not).