refactor

nilearn/glm/tests/_utils.py

@@ -53,6 +53,17 @@ def _modulated_block_paradigm():
     return events
 
 
+def _spm_paradigm(block_duration):
+    frame_times = np.linspace(0, 99, 100)
+    conditions = ["c0", "c0", "c0", "c1", "c1", "c1", "c2", "c2", "c2"]
+    onsets = [30, 50, 70, 10, 30, 80, 30, 40, 60]
+    durations = block_duration * np.ones(len(onsets))
+    events = pd.DataFrame(
+        {"trial_type": conditions, "onset": onsets, "duration": durations}
+    )
+    return events, frame_times
+
+
 def _design_with_null_duration():
     durations = _durations()
     durations[2] = 0

nilearn/glm/tests/test_dmtx.py

@@ -28,6 +28,7 @@
     _block_paradigm,
     _modulated_block_paradigm,
     _modulated_event_paradigm,
+    _spm_paradigm,
 )
 
 # load the spm file to test cosine basis
@@ -145,18 +146,10 @@ def test_design_matrix_regressors_provided_manually_errors(
         )
 
 
-def test_convolve_regressors():
+def test_convolve_regressors(frame_times):
     # tests for convolve_regressors helper function
-    conditions = ["c0", "c1"]
-    onsets = [20, 40]
-    duration = [1, 1]
-    events = pd.DataFrame(
-        {"trial_type": conditions, "onset": onsets, "duration": duration}
-    )
-    # names not passed -> default names
-    frame_times = np.arange(100)
-    f, names = _convolve_regressors(events, "glover", frame_times)
-    assert names == ["c0", "c1"]
+    _, names = _convolve_regressors(basic_paradigm(), "glover", frame_times)
+    assert names == ["c0", "c1", "c2"]
 
 
 def test_design_matrix_basic_paradigm_glover_hrf(frame_times):
@@ -287,7 +280,7 @@ def test_design_matrix_FIR_column_1_3_and_11(frame_times):
     assert_array_almost_equal(X[onset + 4, 11], np.ones(3))
 
 
-def test_design_matrix_FIR_time_shift(frame_times, n_frames):
+def test_design_matrix_FIR_time_shift(frame_times):
     # Check that the first column of FIR design matrix is OK after a 1/2
     # time shift
     tr = 1.0
@@ -453,37 +446,18 @@ def test_csv_io(tmp_path, frame_times):
     assert names == names_
 
 
-def test_spm_1():
+@pytest.mark.parametrize(
+    "block_duration, array", [(1, "arr_0"), (10, "arr_1")]
+)
+def test_compare_design_matrix_to_spm(block_duration, array):
     # Check that the nistats design matrix is close enough to the SPM one
     # (it cannot be identical, because the hrf shape is different)
-    frame_times = np.linspace(0, 99, 100)
-    conditions = ["c0", "c0", "c0", "c1", "c1", "c1", "c2", "c2", "c2"]
-    onsets = [30, 50, 70, 10, 30, 80, 30, 40, 60]
-    durations = 1 * np.ones(9)
-    events = pd.DataFrame(
-        {"trial_type": conditions, "onset": onsets, "duration": durations}
-    )
+    events, frame_times = _spm_paradigm(block_duration=block_duration)
     X1 = make_first_level_design_matrix(frame_times, events, drift_model=None)
     _, matrix, _ = check_design_matrix(X1)
-    spm_design_matrix = DESIGN_MATRIX["arr_0"]
-    assert ((spm_design_matrix - matrix) ** 2).sum() / (
-        spm_design_matrix**2
-    ).sum() < 0.1
 
+    spm_design_matrix = DESIGN_MATRIX[array]
 
-def test_spm_2():
-    # Check that the nistats design matrix is close enough to the SPM one
-    # (it cannot be identical, because the hrf shape is different)
-    frame_times = np.linspace(0, 99, 100)
-    conditions = ["c0", "c0", "c0", "c1", "c1", "c1", "c2", "c2", "c2"]
-    onsets = [30, 50, 70, 10, 30, 80, 30, 40, 60]
-    durations = 10 * np.ones(9)
-    events = pd.DataFrame(
-        {"trial_type": conditions, "onset": onsets, "duration": durations}
-    )
-    X1 = make_first_level_design_matrix(frame_times, events, drift_model=None)
-    spm_design_matrix = DESIGN_MATRIX["arr_1"]
-    _, matrix, _ = check_design_matrix(X1)
     assert ((spm_design_matrix - matrix) ** 2).sum() / (
         spm_design_matrix**2
     ).sum() < 0.1