get_pseudobulk now handles any sparse format

saezlab · Jul 30, 2023 · 7752273 · 7752273
1 parent 4190f6f
commit 7752273
Show file tree

Hide file tree

Showing 2 changed files with 11 additions and 3 deletions.
diff --git a/decoupler/tests/test_utilsanndata.py b/decoupler/tests/test_utilsanndata.py
@@ -1,7 +1,7 @@
 import pytest
 import numpy as np
 import pandas as pd
-from scipy.sparse import csr_matrix
+from scipy.sparse import csr_matrix, csc_matrix
 from anndata import AnnData
 from ..utils_anndata import get_acts, swap_layer, extract_psbulk_inputs, check_X
 from ..utils_anndata import format_psbulk_inputs, psbulk_profile, compute_psbulk, get_unq_dict, get_pseudobulk
@@ -68,6 +68,9 @@ def test_extract_psbulk_inputs():
         extract_psbulk_inputs(adata, obs=None, layer=None, use_raw=True)
     with pytest.raises(ValueError):
         extract_psbulk_inputs(df, obs=None, layer=None, use_raw=False)
+    csc_adata = adata.copy()
+    csc_adata.X = csc_matrix(csc_adata.X).copy()
+    extract_psbulk_inputs(csc_adata, obs=None, layer=None, use_raw=False)
 
 
 def test_check_X():

diff --git a/decoupler/utils_anndata.py b/decoupler/utils_anndata.py
@@ -4,7 +4,7 @@
 """
 
 import numpy as np
-from scipy.sparse import csr_matrix
+from scipy.sparse import csr_matrix, issparse
 import pandas as pd
 import sys
 
@@ -115,8 +115,13 @@ def extract_psbulk_inputs(adata, obs, layer, use_raw):
 
     # Sort genes
     msk = np.argsort(var.index)
+    X = X[:, msk]
+    var = var.iloc[msk]
 
-    return X[:, msk], obs, var.iloc[msk]
+    if issparse(X) and not isinstance(X, csr_matrix):
+        X = csr_matrix(X)
+
+    return X, obs, var
 
 
 def check_X(X, mode='sum', skip_checks=False):