Add support for writing minimal Anndata conversion + tools for querying

[dosumis]

Minimal representation:

• [x] cellannotation/cas-tools#6

  • input CAS JSON schema file; AnnData file (ideally AnnData file location will be specified by resolvable path in CAS file)
  • Checks before writing:
    • Do all barcodes (cell IDs) in CAS file exists in AnnData file - if not, warn if not with option to terminate.
    • Do all barcodes (cell IDs) in AnnData file exist in CAS file - if not, warn if not with option to terminate.
    • Do any obs keys in the AnnData file match labelset names in CAS - if they do warn - with options to overwrite or terminate.
  • Store JSON in uns (dropping barcodes)
  • Write labelset:cell_label pairs in obs - this replaces barcode lookup.

• [ ] Write Python tools for extracting and displaying additional annotation data on request.

  • function to write a dataframe of annotations
  • function to return details of any single annotation
  • function to return general metadata about annotations
  • function to return general metadata about labesets

Note - For better modularisation, we may move code to here https://github.com/cellannotation/castool (TBD)

[hkir-dev]

I have a project for CAS operations. We can improve it together if it will help. https://github.com/hkir-dev/cell-type-annotation-tools/tree/tdt_sprint2 https://pypi.org/project/ctat/

I am using this project to:

• convert user data to CAS compliant format using a config
• exporting CAS formatted data in json format (to be used by the taxonomy development tools primarily)
• exporting CAS formatted data into tables (to be used by the taxonomy development tools primarily)

[dosumis]

@hkir-dev - strongly agree we should discuss merging efforts.