Explore conversion of data model to LinkML

[aclayton555]

Good opportunity to consider aligning our MC2 data model with parallel efforts, including those in HTAN, to adapt our data model to LinkML.

Current: csv -> jsonld
Potential: LinkML (yaml -> can be ceonverted to several formats, including jsonld)

In addition to alignment with HTAN, LinkML would offer a lot more flexibility over the current "flat" format (e.g. hierarchy, etc).

More to come to assess lift required here. FYI @Bankso and @aditigopalan in case you want to brain dump into this ticket for now, or link any existing resources from ANV.

[aclayton555]

24-11/12 Cross-pollinate efforts with HTAN: ncihtan/data-models#372

[aditigopalan]

Update 24-11/12 , in parallel with efforts in HTAN:

• Working on a pure LinkML subset model for the CDS Imaging Template here: scratch branch

• The current focus is on developing a "subset model" for the CDS Imaging Template.

• The goal is to create modular CDS components for template integration and HTAN modules to support a more flexible data model.

• Starting with the CDS model lays the foundation for a "common cancer data model," which can later incorporate MC2 and HTAN terms.

• Since Schematic is not LinkML-compatible, the NF model is a makeshift version tailored to work with Schematic (requesting Tom's help to make Curator fully LinkML-compatible and eliminate this workaround).

• The subset model under development is pure LinkML.

• ANV is running some scripts to test out the current subset model (meeting with her in 2 weeks to discuss)

[aclayton555]

25-1 Moving forward, no blockers. Aditi meeting with ANV this week. Starting to identify schematic development needs, but may need to check assumptions about whether MC2, HTAN, and NF can actually be written in the same way