Open bernt-matthias opened 8 months ago
I'm going to look into this for the next release. I think name
is just something that shouldn't be there. I think this may have been a workaround for regex-based paths, where there is some variable prefix (like in reference indices/databases).
Would it be possible to set the
name
argument, e.g. over here [https://github.com/qiime2/galaxy-tools/blob/65e4952f33eb335528e8553150e9097e5ea8f556/tools/suite_qiime2_core__tools/qiime2_core__tools__import.xml#L245C78-L245C78] automatically based on the correspondingdata
parameter?In Galaxy you could use
data.element_identifier
. This would be equal todata.name
for normal datasets, but different for collection elements (whereelement_identfier
is the name of the collection element).Edit: Maybe use the element identifier in case if
name
is left empty (which leads currently to an error).Edit: So, the actual problem is that we see the following error if the
name
is something different than the name of the dataset. When running the tool manually that is "no problem" (but still a bit inconvenient and redundant), but when running in a workflow we do not know the name of the dataset...