Get pseudobulk api

[myushen]

solve issue #136
get_pseudobulk API takes pre-calculated summarized experiment assays for selected sample IDs, combines them and returns a merged summarized experiment. The pre-calculated summarized experiment assays will be stored in a cloud container in the future.

[multimeric]

Note also that SingleCellExperiment is a subclass of SummarisedExperiment, so if it's possible to write code that works on a SummarisedExperiment, then it will automatically work on both. This can help you simplify code and also documentation.

[stemangiola]

I did not export fibrosis pseudobulk yet, @myushen did you test with fibrosis?

[myushen]

I did not export fibrosis pseudobulk yet, @myushen did you test with fibrosis?

No, it is only based on cellxgene.

[multimeric]

I'm still concerned that group_to_sme is a near-exact copy of group_to_sce and the same for get_pseudobulk and get_single_cell_experiment. I see little reason why you can't consolidate these pairs of functions into one function, considering that you are working with a SummarizedExperiment in both cases.

[myushen]

I'm still concerned that group_to_sme is a near-exact copy of group_to_sce and the same for get_pseudobulk and get_single_cell_experiment. I see little reason why you can't consolidate these pairs of functions into one function, considering that you are working with a SummarizedExperiment in both cases.

I'll wrap up into one function, but do we want to separate function names for get_sce and get_pseudobulk? @stemangiola

[stemangiola]

I'm still concerned that group_to_sme is a near-exact copy of group_to_sce and the same for get_pseudobulk and get_single_cell_experiment. I see little reason why you can't consolidate these pairs of functions into one function, considering that you are working with a SummarizedExperiment in both cases.

I'll wrap up into one function, but do we want to separate function names for get_sce and get_pseudobulk? @stemangiola

Are we asking here if

get_single_cell_experiment and get_pseudobulk should be the same function.

If this is the question I think that is beneficial keeping them separate, from the user perspective. What happens in the backend is another story. In my code I have a lot of utility backend function, that are called in different flavours in the front end, and the user will never know that two different methods, come from a generic function underneath.

[multimeric]

Needs a test for the get_pseudobulk workflow.

Also this isn't a blocker, but I'd like to see some of the assay-specific logic (if (type == "sce")) moved into the calling function like get_single_cell_experiment, so that get_summarized_experiment remains decoupled from the subclass we're dealing with.

[stemangiola]

Needs a test for the get_pseudobulk workflow.

Also this isn't a blocker, but I'd like to see some of the assay-specific logic (if (type == "sce")) moved into the calling function like get_single_cell_experiment, so that get_summarized_experiment remains decoupled from the subclass we're dealing with.

Thanks. Maybe @myushen do the tests, and we tackle the second point from @multimeric in the future (we can convert it to github issue for now). We need to proceed, with other tasks.

[stemangiola]

Great team work fellas!