cache for pseudobulk same as for single cell?

[stemangiola]

@myushen I noticed that if I query pseudobulk

my_metadata |>
     get_pseudobulk(assays = "counts")

the data is stored in the wrong cache?

Downloading https://object-store.rc.nectar.org.au/v1/AUTH_06d6e008e3e642da99d806ba3ea629c5/pseudobulk-0.1.0/original/GSE122960/se.rds to /Users/mangiola.s/Library/Caches/org.R-project.R/R/CuratedAtlasQueryR/0.2.1/original/GSE122960/se.rds

I think we should use pseudobulk-0.1.0 for the cache , not 0.2.1

[myushen]

good point. this is associated with the cache directory structure strategy implementation. we might need to think about if we want to adapt the same structure strategy as the future cloud.
cellxgene

• date (yy-mm-dd)
• • original
• • cpm
• another date (yy-mm-dd)
• • original
• • cpm

fibrosis

• date (yy-mm-dd)
• • original
• • cpm

pseudobulk

• date (yy-mm-dd)
• • original
• • quantile_normalised

[stemangiola]

How about this hierarchy

• cellxgene

  • date (yy-mm-dd)

    • single_cell
      • original
      • cpm
    • pseudobulk
      • original
      • cpm
    • metacells_2
      • original
      • cpm
    • metacells_4
      • ...

  • ANOTHER date (yy-mm-dd)

    • single_cell
      • original
      • cpm
    • pseudobulk
      • original
      • cpm
    • metacells_2
      • original
      • cpm
    • metacells_4
      • ...

• Fibrosis

  • date (yy-mm-dd)
    • single_cell
      • original
      • cpm
    • pseudobulk
      • original
      • cpm
    • metacells_2
      • original
      • cpm
    • metacells_4
      • ...

...