Introduce a step to update study metadata

[forus]

Add support of update any of the study attributes described in the Meta File Formats Documentation without reloading the entire study.

[migbro]

Not sure if this sounds silly, but perhaps allowing for update of any meta file, especially description as perhaps if a clarifying detail needed to be added about the nature of any of the clinical data files or the genomic files, one could easily patch that rather than having to reload the entire study.

[forus]

@migbro I’ve realized that including the update of study attributes as part of incremental upload doesn’t make much sense. Incremental upload operates within the patient-sample hierarchy, where entities and their child entities are removed and reloaded based on the supplied files. For example, when incrementally uploading patients, all their associated samples are removed and reloaded.

Following this logic, applying incremental upload to studies would require removing all patients and their samples when a study is incrementally uploaded. This essentially mirrors the behavior of a full reload, which defeats the purpose of incremental uploads. This is why incremental upload for studies was not supported in the first place.

That said, we could implement study attribute updates as a separate update or patch command.

By the way, is there a description field in clinical meta files currently? I couldn’t find one.

All platform-based molecular data types share the following fields in their meta files: show_profile_in_analysis_tab, profile_name, and profile_description. Let me know if you’d like to update these fields, and I can create a separate ticket for it.

[forus]

@migbro When updating a study's metadata, how should the absence of an attribute be interpreted? Should the missing attribute be removed, or should it remain unchanged?

For example, consider study A, which initially included a citation field when it was first uploaded. If the updated metadata file does not include the citation field, does this imply that the citation should be reset (emptied) or left as it is?

[forus]

It won't be possible to update cancer_study_identifier attribute with this command as we rely on this field to lookup previous record from the database.

[forus]

Not sure how to deal with add_global_case_list attribute updates and what it'd mean yet.

[migbro]

Honestly, I was more concerned with, when a meta file has a description filed, or as you pointed out a citation field, being able to update that. Even just say, replacing a complete meta file is fine, it's the hope of avoiding reloading the whole study when all you wanted to do was update text that provides context or add al link. Other fields, especially in meta_study like reference_genome I agree are a bit heavy-hitting or have greater implications.

[forus]

Here is the implementation #68