Updated tests for previous additions

decoupler/tests/test_decouple.py

@@ -1,5 +1,6 @@
 import pytest
 import numpy as np
+import numpy.testing as npt
 import pandas as pd
 from anndata import AnnData
 from ..decouple import get_wrappers, run_methods, parse_methods, decouple, run_consensus
@@ -10,13 +11,27 @@ def test_get_wrappers():
 
 
 def test_run_methods():
-    m = np.array([[7., 1., 1., 1.], [4., 2., 1., 2.], [1., 2., 5., 1.], [1., 1., 6., 2.]])
+    m = np.array([[7., 1., 1., 1., 0.1], [4., 2., 1., 2., 0.1], [1., 2., 5., 1., 0.1], [1., 1., 6., 2., 0.1]])
     r = np.array(['S1', 'S2', 'S3', 'S4'])
-    c = np.array(['G1', 'G2', 'G3', 'G4'])
+    c = np.array(['G1', 'G2', 'G3', 'G4', 'G5'])
     df = pd.DataFrame(m, index=r, columns=c)
     net = pd.DataFrame([['T1', 'G1', 1], ['T1', 'G2', 2], ['T2', 'G3', -3], ['T2', 'G4', 4]],
                        columns=['source', 'target', 'weight'])
-    run_methods(df, net, 'source', 'target', 'weight', ['mlm', 'ulm'], {}, 0, True, False)
+    run_methods(AnnData(df), net, 'source', 'target', 'weight', ['gsva'], {}, 0, False, False, False)
+    run_methods(df, net, 'source', 'target', 'weight', ['gsva'], {}, 0, False, False, False)
+    wo_args = run_methods(df, net, 'source', 'target', 'weight', ['ulm', 'aucell'], {}, 0, False, False, False)
+    wt_args = run_methods(df, net, 'source', 'target', 'weight', ['ulm', 'aucell'],
+                          {'aucell': {'n_up': 3}}, 0, False, False, False)
+    assert np.all(wo_args['aucell_estimate'].values != wt_args['aucell_estimate'].values)
+    # Run dense
+    df.loc['S1', ['G3', 'G4']] = 0
+    dens = run_methods(df, net, 'source', 'target', 'weight', ['ulm', 'aucell'], {}, 0, False, False, True)
+    assert wt_args['ulm_estimate'].loc['S1', 'T2'] != dens['ulm_estimate'].loc['S1', 'T2']
+    npt.assert_allclose(wt_args['ulm_estimate'].values[1:, :], dens['ulm_estimate'].values[1:, :], rtol=1e-6)
+    # Check that estimate_loc works
+    res = run_methods(df, net, 'source', 'target', 'weight', ['ulm', 'wsum'],
+                      {'wsum': {'estimate_loc': 3}}, 0, False, False, True)
+    npt.assert_allclose(res['wsum_estimate'], res['wsum_pvals'])
 
 
 def test_parse_methods():

decoupler/tests/test_pre.py

@@ -22,6 +22,8 @@ def test_check_mat():
     assert nr.size == 2
     assert nc.size == 2
     assert type(nm) is not csr_matrix and isinstance(nm, np.ndarray)
+    with pytest.raises(ValueError):
+        check_mat(m, r, np.array(['G1', 'G2', 'G1']))
     m = csr_matrix(np.array([[1, 0, 2], [np.nan, 0, 3], [0, 0, 0]]))
     with pytest.raises(ValueError):
         check_mat(m, r, c)

decoupler/tests/test_utils.py

@@ -1,12 +1,15 @@
 import pytest
 import numpy as np
+from scipy.sparse import csr_matrix
+import numpy.testing as npt
 import pandas as pd
 import os
 from anndata import AnnData
 from ..utils import m_rename, melt, show_methods, check_corr, get_toy_data, summarize_acts
 from ..utils import assign_groups, p_adjust_fdr, dense_run, shuffle_net, read_gmt
 from ..method_mlm import run_mlm
 from ..method_ora import run_ora
+from ..method_gsva import run_gsva
 
 
 def test_m_rename():
@@ -32,6 +35,7 @@ def test_melt():
     melt(pvals)
     melt(res_dict)
     melt(res_list)
+    melt({estimate.name: estimate})
     with pytest.raises(ValueError):
         melt({0, 1, 2, 3})
 
@@ -85,20 +89,26 @@ def test_p_adjust_fdr():
 
 
 def test_denserun():
-    mat = pd.DataFrame([[0, 2, 3, 4, 5, 6], [1, 0, 0, 0, 0, 0]], columns=['G01', 'G02', 'G03', 'G06', 'G07', 'G08'])
+    mat = pd.DataFrame([[0, 2, 3, 4, 5, 6], [1, 0, 0, 0, 0, 0]], index=['S1', 'S2'],
+                       columns=['G01', 'G02', 'G03', 'G06', 'G07', 'G08'])
     net = pd.DataFrame([['T1', 'G01', 1], ['T1', 'G02', 1], ['T2', 'G06', 1], ['T2', 'G07', 0.5],
                         ['T3', 'G06', -0.5], ['T3', 'G07', -3]], columns=['source', 'target', 'weight'])
     acts, _ = dense_run(run_mlm, mat, net, min_n=2, verbose=True)
     assert acts.shape[1] == 2
     assert not np.all(np.isnan(acts.values[0]))
     assert np.all(np.isnan(acts.values[1]))
+    spr = AnnData(mat.astype(np.float32))
+    spr.X = csr_matrix(spr.X).copy()
+    dense_run(run_mlm, spr, net, min_n=2, verbose=True, use_raw=False)
+    npt.assert_allclose(spr.obsm['mlm_estimate'], acts)
 
     mat = AnnData(mat.astype(np.float32))
     dense_run(run_ora, mat, net, min_n=2, verbose=True, use_raw=False)
     acts = mat.obsm['ora_estimate']
     assert acts.shape[1] == 2
     assert not np.all(np.isnan(acts.values[0]))
     assert np.all(np.isnan(acts.values[1]))
+    dense_run(run_gsva, mat, net, min_n=2, verbose=True, use_raw=False)
 
 
 def test_shuffle_net():