Feat/view fix

[anngvu]

For really relevant context, review PLFM-4862 and PLFM-5869. The discussions there explain:

• Why view schemas break
• Why you don't get notified and just come upon broken ones occasionally
• Challenges and considerations for adjusting view schemas automatically
• Other technical details good to be aware of overall (budgeting of the schema)

The more recent issue that brought this up again is https://github.com/nf-osi/dcc-site/issues/669, so figured we can just take stab at this. Useful for:

• If working in programmatic context, can fix without having to go to UI
• Automatically fix things if they are broken so they can be snapshotted (versioned)

To see what this looks like in action:

library(nfportalutils)
syn_login(Sys.getenv("SYNAPSE_AUTH_TOKEN"))

# Broken view with three adjustments needed
test_view <- "syn28499567"
adjust_view(test_view)

# After fixing, reset back to the broken schema model
schema <- .syn$get(test_view)
schema <- swap_col(schema, old = "82416", new = "110715") # back to assay with too small size 15
schema <- swap_col(schema, old = "185618", new = "200178") # back to contentType with with too small size 2
schema <- swap_col(schema, old = "200109", new = "200114") # back to tumorType with too small list len 2

[allaway]

Hi @anngvu , neat idea! I'll take a look at the code now. Just as an FYI we recently reopened one of the several old Jira issues about this: https://sagebionetworks.jira.com/browse/PLFM-5346 (a different one than what you linked), so it's possible we might get some Synapse side fix or tools that we could use in the future.

[allaway]

Code and test LGTM, thanks!