Helper function to get the dimensions and storage format

Distinguishes sparse (CSR/CSC) from dense storage. For dense X the orientation is inferred by comparing the matrix dims against the obs and var lengths (DENSE_ROW = cells x genes, DENSE_COL = genes x cells). Ties (no_obs == no_var) fall back to the AnnData convention (DENSE_ROW).

Usage

get_h5ad_dimensions(f_path)

Arguments

f_path

File path to the .h5ad file.

Value

A list with dims (named integer c(obs, var)) and type (one of "CSR", "CSC", "DENSE_ROW", "DENSE_COL").

Examples

# dimensions and storage layout without reading the counts
data <- generate_single_cell_test_data(
  syn_data_params = params_sc_synthetic_data(n_cells = 200L, n_genes = 40L)
)
f_path <- tempfile(fileext = ".h5ad")
write_h5ad_sc(f_path, data$counts, data$obs, data$var, .verbose = FALSE)
get_h5ad_dimensions(f_path)
#> $dims
#> obs var 
#> 200  40 
#> 
#> $type
#> [1] "CSR"
#> 

unlink(f_path)