bihealth / sodar-server

SODAR: System for Omics Data Access and Retrieval
https://github.com/bihealth/sodar-server
MIT License
14 stars 3 forks source link

Simplify internal iRODS path handling for landing zones #1873

Closed mikkonie closed 8 months ago

mikkonie commented 8 months ago

While looking into #1869, I noticed inconsistencies in how we handle iRODS path internally in landing zone creation taskflow setup. Example:

This behaviour is possibly a remnant from the days of the separate sodar_taskflow microservice in an attempt to recude redundant transferred data. Nowadays it doesn't make any sense.

For the aforementioned case the solution is simple: just use full paths throughout the process and skip the conversions. However, I should look into other instances and see if 1) similar refactoring should be done and 2) what is the consistent pattern to follow.

mikkonie commented 8 months ago

Looking into this in more detail, the conversions are actually needed as we store the full assay path, which of course will note be the same as the zone path.

Probably the optimal solution would be to store subpaths in the cache as well as expect those to be provided by the assay plugins. But that's quite a lot of refactoring and also updating existing cache data. Probably not worth it at this point. Closing the issue, at least for now.