Implementation of the conn_fcd_corr + test + doc

Author: EtienneCmb

docs/source/api/api_connectivity.rst

@@ -22,12 +22,30 @@ Connectivity metrics
 Utility functions
 +++++++++++++++++
 
+Reshaping connectivity outputs
+==============================
+
 .. autosummary::
    :toctree: generated/
 
    conn_reshape_undirected
    conn_reshape_directed
    conn_ravel_directed
    conn_get_pairs
+
+Metrics to apply on FC
+======================
+
+.. autosummary::
+   :toctree: generated/
+
+   conn_fcd_corr
+
+Define sliding windows
+======================
+
+.. autosummary::
+   :toctree: generated/
+
    define_windows
    plot_windows

frites/conn/conn_fcd_corr.py

@@ -55,8 +55,7 @@ def conn_fcd_corr(conn, roi='roi', times='times', tskip=1, estimator=None,
     conn = conn.loc[:, ::tskip, :]
 
     # get coordinates
-    supp_c, roi_c = conn[supp_dim].data, conn[roi].data
-    times_c = conn[times].data
+    supp_c, times_c = conn[supp_dim].data, conn[times].data
 
     # _______________________________ ESTIMATOR _______________________________
     if estimator is None:

frites/conn/tests/test_fcd_corr.py

@@ -0,0 +1,54 @@
+"""Test conn_fcd_corr"""
+import numpy as np
+import xarray as xr
+
+from frites.estimator import DcorrEstimator
+from frites.conn import conn_dfc, conn_fcd_corr, define_windows
+
+# sample data
+x = np.random.rand(10, 3, 1000)
+trials = np.arange(10)
+roi = ['roi_1', 'roi_0', 'roi_0']
+times = (np.arange(1000) - 10) / 64.
+x = xr.DataArray(x, dims=('trials', 'roi', 'times'),
+                 coords=(trials, roi, times))
+win, _ = define_windows(times, slwin_len=.5, slwin_step=.1)
+dfc = conn_dfc(x, times='times', roi='roi', win_sample=win, verbose=False)
+
+
+class TestFCDCorr(object):
+
+    def test_smoke(self):
+        """Test that it's working (overall)."""
+        # test on full network
+        corr = conn_fcd_corr(dfc, roi='roi', times='times', verbose=False)
+        assert corr.shape[0] == len(trials)
+        assert corr.shape[1] == corr.shape[2] == len(win)
+
+        # test on single time-point
+        corr = conn_fcd_corr(dfc.isel(times=[0]), roi='roi', times='times',
+                             verbose=False)
+        assert corr.shape[0] == len(trials)
+        assert corr.shape[1] == corr.shape[2] == 1
+
+        # test internal reshaping
+        corr = conn_fcd_corr(dfc.transpose('times', 'trials', 'roi'),
+                             roi='roi', times='times', verbose=False)
+        assert corr.shape[0] == len(trials)
+        assert corr.shape[1] == corr.shape[2] == len(win)
+
+    def test_kwargs(self):
+        """Test with a custom estimator."""
+        # testing estimator and tskip
+        est = DcorrEstimator()
+        corr = conn_fcd_corr(dfc, roi='roi', times='times', verbose=False,
+                             estimator=est, tskip=10)
+        assert np.nanmin(corr.data) > 0
+        assert corr.shape[0] == len(trials)
+        assert corr.shape[1] == corr.shape[2] == len(win[::10])
+
+        # testing diagonal filling
+        corr = conn_fcd_corr(dfc, roi='roi', times='times', fill_diagonal=-1,
+                             verbose=False)
+        cm = corr.mean('trials')
+        np.testing.assert_array_equal(np.diag(cm), np.full((len(win),), -1))