Enable prohibiting landing zone uploads of specific file types

[mmilek]

Problem

for some projects managed by the external sodar instance, the sharing of primary human sequence data (ie identifying data) is not allowed. We would need to prevent the upload of certain file types such as fastq to sodar

Solution

perhaps by file suffix or by content validation?

[mikkonie]

This seems to be a reasonable feature to add. There are some assumptions and questions that come to mind:

• This should obviously be at least project specfic
  • Is more granularity needed? E.g. assay specific?
    • If so, the controls for this would be ISA-Tab / assay plugin specific
    • Best choice would be to configure this with SODAR specific ISA-Tab comments
• Are file suffix checks sufficient?
  • If not, what else would be needed?
• How and where to communicate these limitations to the user?
  • In the end, this will depend on how and where exactly we choose to configure this

See also #2065.

[mmilek]

i think project specific is enough, assay specific maybe not needed

file suffix checks could be the initial solution but i think content validation is the way to go. for fastq the format is well-defined.

the data steward is responsible for creating a project with the necessary limitations and communicating with the user eg via project description.

[mikkonie]

Thanks for the feedback. In that case, I'd add a simple project setting similar to BAM/VCF file omission and check against file suffixes in landing zone validation.

We can of course make this fancier later on, but with this initial implementation this should be trivial to get in v1.1. Tagging the issue in the milestone.