Merge pull request #982 from EmmaRenauld/easy_tests_reconst

Easy tests reconst

scilpy/reconst/fodf.py

@@ -15,18 +15,23 @@
 cvx, have_cvxpy, _ = optional_package("cvxpy")
 
 
-def get_ventricles_max_fodf(data, fa, md, zoom, sh_basis, small_dims,
-                            fa_threshold, md_threshold, is_legacy=True):
+def get_ventricles_max_fodf(data, fa, md, zoom, sh_basis,
+                            fa_threshold, md_threshold,
+                            small_dims=False, is_legacy=True):
     """
     Compute mean maximal fodf value in ventricules. Given heuristics thresholds
     on FA and MD values, finds the voxels of the ventricules or CSF and
     computes a mean fODF value. This is described in
     Dell'Acqua et al. HBM 2013.
 
+    Ventricles are searched in a window in the middle of the data to increase
+    speed. No need to scan the whole image.
+
     Parameters
     ----------
     data: ndarray (x, y, z, ncoeffs)
-         Input fODF file in spherical harmonics coefficients.
+         Input fODF file in spherical harmonics coefficients. Uses sphere
+         'repulsion100' to convert to SF values.
     fa: ndarray (x, y, z)
          FA (Fractional Anisotropy) volume from DTI
     md: ndarray (x, y, z)
@@ -38,15 +43,16 @@ def get_ventricles_max_fodf(data, fa, md, zoom, sh_basis, small_dims,
         Either 'tournier07' or 'descoteaux07'
     small_dims: bool
         If set, takes the full range of data to search the max fodf amplitude
-        in ventricles. Useful when the data has small dimensions.
+        in ventricles, rather than a window center in the data. Useful when the
+        data has small dimensions.
     fa_threshold: float
         Maximal threshold of FA (voxels under that threshold are considered
-        for evaluation).
+        for evaluation). Suggested value: 0.1.
     md_threshold: float
         Minimal threshold of MD in mm2/s (voxels above that threshold are
-        considered for evaluation).
+        considered for evaluation). Suggested value: 0.003.
     is_legacy : bool, optional
-        Whether or not the SH basis is in its legacy form.
+        Whether the SH basis is in its legacy form.
 
     Returns
     -------
@@ -57,26 +63,17 @@ def get_ventricles_max_fodf(data, fa, md, zoom, sh_basis, small_dims,
     order = find_order_from_nb_coeff(data)
     sphere = get_sphere('repulsion100')
     b_matrix, _ = sh_to_sf_matrix(sphere, order, sh_basis, legacy=is_legacy)
-    sum_of_max = 0
-    count = 0
-
     mask = np.zeros(data.shape[:-1])
 
-    if np.min(data.shape[:-1]) > 40:
-        step = 20
-    else:
-        if np.min(data.shape[:-1]) > 20:
-            step = 10
-        else:
-            step = 5
-
     # 1000 works well at 2x2x2 = 8 mm3
     # Hence, we multiply by the volume of a voxel
     vol = (zoom[0] * zoom[1] * zoom[2])
     if vol != 0:
         max_number_of_voxels = 1000 * 8 // vol
     else:
         max_number_of_voxels = 1000
+    logging.debug("Searching for ventricle voxels, up to a maximum of {} "
+                  "voxels.".format(max_number_of_voxels))
 
     # In the case of 2D-like data (3D data with one dimension size of 1), or
     # a small 3D dataset, the full range of data is scanned.
@@ -85,14 +82,27 @@ def get_ventricles_max_fodf(data, fa, md, zoom, sh_basis, small_dims,
         all_j = list(range(0, data.shape[1]))
         all_k = list(range(0, data.shape[2]))
     # In the case of a normal 3D dataset, a window is created in the middle of
-    # the image to capture the ventricules. No need to scan the whole image.
+    # the image to capture the ventricles. No need to scan the whole image.
+    # (Automatic definition of window's radius based on the shape of the data.)
     else:
-        all_i = list(range(int(data.shape[0]/2) - step,
-                           int(data.shape[0]/2) + step))
-        all_j = list(range(int(data.shape[1]/2) - step,
-                           int(data.shape[1]/2) + step))
-        all_k = list(range(int(data.shape[2]/2) - step,
-                           int(data.shape[2]/2) + step))
+        if np.min(data.shape[:-1]) > 40:
+            radius = 20
+        else:
+            if np.min(data.shape[:-1]) > 20:
+                radius = 10
+            else:
+                radius = 5
+
+        all_i = list(range(int(data.shape[0]/2) - radius,
+                           int(data.shape[0]/2) + radius))
+        all_j = list(range(int(data.shape[1]/2) - radius,
+                           int(data.shape[1]/2) + radius))
+        all_k = list(range(int(data.shape[2]/2) - radius,
+                           int(data.shape[2]/2) + radius))
+
+    # Ok. Now find ventricle voxels.
+    sum_of_max = 0
+    count = 0
     for i in all_i:
         for j in all_j:
             for k in all_k:

scilpy/reconst/frf.py

@@ -19,10 +19,11 @@
 def compute_ssst_frf(data, bvals, bvecs, b0_threshold=DEFAULT_B0_THRESHOLD,
                      mask=None, mask_wm=None, fa_thresh=0.7, min_fa_thresh=0.5,
                      min_nvox=300, roi_radii=10, roi_center=None):
-    """Compute a single-shell (under b=1500), single-tissue single Fiber
-    Response Function from a DWI volume.
-    A DTI fit is made, and voxels containing a single fiber population are
-    found using a threshold on the FA.
+    """
+    Computes a single-shell (under b=1500), single-tissue single Fiber
+    Response Function from a DWI volume. A DTI fit is made, and voxels
+    containing a single fiber population are found using either a threshold on
+    the FA, inside a white matter mask.
 
     Parameters
     ----------
@@ -43,7 +44,7 @@ def compute_ssst_frf(data, bvals, bvecs, b0_threshold=DEFAULT_B0_THRESHOLD,
         3D mask with shape (X,Y,Z)
         Binary white matter mask. Only the data inside this mask and above the
         threshold defined by fa_thresh will be used to estimate the fiber
-        response function.
+        response function. If not given, all voxels inside `mask` will be used.
     fa_thresh : float, optional
         Use this threshold as the initial threshold to select single fiber
         voxels. Defaults to 0.7
@@ -63,7 +64,7 @@ def compute_ssst_frf(data, bvals, bvecs, b0_threshold=DEFAULT_B0_THRESHOLD,
 
     Returns
     -------
-    full_reponse : ndarray
+    full_response : ndarray
         Fiber Response Function, with shape (4,)
 
     Raises
@@ -139,10 +140,11 @@ def compute_msmt_frf(data, bvals, bvecs, btens=None, data_dti=None,
                      fa_thr_wm=0.7, fa_thr_gm=0.2, fa_thr_csf=0.1,
                      md_thr_gm=0.0007, md_thr_csf=0.003, min_nvox=300,
                      roi_radii=10, roi_center=None, tol=20):
-    """Compute a multi-shell, multi-tissue single Fiber
-    Response Function from a DWI volume.
-    A DTI fit is made, and voxels containing a single fiber population are
-    found using a threshold on the FA and MD.
+    """
+    Computes a multi-shell, multi-tissue single Fiber Response Function from a
+    DWI volume. A DTI fit is made, and voxels containing a single fiber
+    population are found using a threshold on the FA and MD, inside a mask of
+    each tissue type.
 
     Parameters
     ----------
@@ -304,33 +306,47 @@ def compute_msmt_frf(data, bvals, bvecs, btens=None, data_dti=None,
     return responses, frf_masks
 
 
-def replace_frf(old_frf, new_frf, no_factor):
+def replace_frf(old_frf, new_frf, no_factor=False):
     """
-    Replace old_frf with new_frf
+    Replaces the 3 first values of old_frf with new_frf. Formats the new_frf
+    from a string value and verifies that the number of shells corresponds.
 
     Parameters
     ----------
     old_frf: np.ndarray
-        A loaded frf file, of shape (n, 4).
-    new_frf: tuple
-        The new frf, to be interpreted with a 10**-4 factor. Ex: (15,4,4)
+        A loaded frf file, of shape (N, 4), where N is the number of shells.
+    new_frf: str
+        The new frf, to be interpreted with a 10**-4 factor. Ex: 15,4,4. With
+        multishell: all values, concatenated into one string.
+        Ex: 15,4,4,13,5,5,12,5,5.
     no_factor: bool
         If true, the fiber response function is evaluated without the
         10**-4 factor.
+
+    Returns
+    -------
+    response: np.ndarray
+        Formatted new frf, of shape (n, 4)
     """
-    old_frf = old_frf.T
-    new_frf = np.array(literal_eval(new_frf), dtype=np.float64)
+    if len(old_frf.shape) == 1:   # When loading from one shell, we get (4, )
+        old_frf = old_frf[None, :]
+    old_nb_shells = old_frf.shape[0]
+    b0_mean = old_frf[:, 3]
 
+    new_frf = np.array(literal_eval(new_frf), dtype=np.float64)
     if not no_factor:
         new_frf *= 10 ** -4
-    b0_mean = old_frf[3]
 
     if new_frf.shape[0] % 3 != 0:
         raise ValueError('Inputed new frf is not valid. There should be '
                          'three values per shell, and thus the total number '
                          'of values should be a multiple of three.')
 
     nb_shells = int(new_frf.shape[0] / 3)
+    if nb_shells != old_nb_shells:
+        raise ValueError("The old frf contained {} shell(s). Cannot replace "
+                         "with {} shell(s).".format(old_nb_shells, nb_shells))
+
     new_frf = new_frf.reshape((nb_shells, 3))
 
     response = np.empty((nb_shells, 4))

scilpy/reconst/tests/test_fodf.py

@@ -1,9 +1,41 @@
 # -*- coding: utf-8 -*-
+import numpy as np
+from dipy.data import get_sphere
+from dipy.reconst.shm import sh_to_sf_matrix
+
+from scilpy.reconst.fodf import get_ventricles_max_fodf
+from scilpy.reconst.utils import find_order_from_nb_coeff
+from scilpy.tests.arrays import fodf_3x3_order8_descoteaux07
 
 
 def test_get_ventricles_max_fodf():
-    # toDO
-    pass
+    fake_fa = np.ones((3, 3, 1))  # High FA
+    fake_fa[1:3, 0:2, 0] = 0      # Low in ventricles
+    fake_md = np.zeros((3, 3, 1))  # Low MD
+    fake_md[0:2, 0:2, 0] = 1       # High in ventricles
+    zoom = [1, 1, 1]
+    fa_threshold = 0.5
+    md_threshold = 0.5
+    sh_basis = 'descoteaux07'
+
+    # Should find that the only 2 ventricle voxels are at [1, 0:2, 0]
+    mean, mask = get_ventricles_max_fodf(
+        fodf_3x3_order8_descoteaux07, fake_fa, fake_md, zoom, sh_basis,
+        fa_threshold, md_threshold, small_dims=True)
+
+    expected_mask = np.logical_and(~fake_fa.astype(bool), fake_md)
+    assert np.count_nonzero(mask) == 2
+    assert np.array_equal(mask.astype(bool), expected_mask)
+
+    # Reconstruct SF values same as in method.
+    order = find_order_from_nb_coeff(fodf_3x3_order8_descoteaux07)
+    sphere = get_sphere('repulsion100')
+    b_matrix, _ = sh_to_sf_matrix(sphere, order, sh_basis, legacy=True)
+
+    sf1 = np.dot(fodf_3x3_order8_descoteaux07[1, 0, 0], b_matrix)
+    sf2 = np.dot(fodf_3x3_order8_descoteaux07[1, 1, 0], b_matrix)
+
+    assert mean == np.mean([np.max(sf1), np.max(sf2)])
 
 
 def test_fit_from_model():

scilpy/reconst/tests/test_frf.py

@@ -1,16 +1,79 @@
 # -*- coding: utf-8 -*-
+import os
+import tempfile
+
+import nibabel as nib
+import numpy as np
+from dipy.io import read_bvals_bvecs
+
+from scilpy import SCILPY_HOME
+from scilpy.io.fetcher import fetch_data, get_testing_files_dict
+from scilpy.reconst.frf import compute_ssst_frf, compute_msmt_frf, replace_frf
+
+# If they already exist, this only takes 5 seconds (check md5sum)
+fetch_data(get_testing_files_dict(), keys=['processing.zip'])
+tmp_dir = tempfile.TemporaryDirectory()
+in_dwi = os.path.join(SCILPY_HOME, 'processing', 'dwi_crop.nii.gz')
+in_bval = os.path.join(SCILPY_HOME, 'processing', 'dwi.bval')
+in_bvec = os.path.join(SCILPY_HOME, 'processing', 'dwi.bvec')
 
 
 def test_compute_ssst_frf():
-    # toDO
-    pass
+    # Uses data from our test data.
+    # To use a smaller subset, we need to ensure that it has at least one
+    # voxel with FA higher than 0.7. Quite fast as is, so, ok.
+    dwi = nib.load(in_dwi).get_fdata()  # Shape: 57, 67, 56, 64
+    bvals, bvecs = read_bvals_bvecs(in_bval, in_bvec)
+
+    result = compute_ssst_frf(dwi, bvals, bvecs)
+
+    # Value with current data at the date of test creation:
+    expected_result = [1.03068237e-03, 2.44994949e-04,
+                       2.44994949e-04, 3.26903486e+03]
+    assert np.allclose(result, expected_result)
 
 
 def test_compute_msmt_frf():
-    # toDO
-    pass
+    # Uses data from our test data.
+    # To use a smaller subset, we need to ensure that it has at least one
+    # voxel with each tissue type.
+    dwi = nib.load(in_dwi).get_fdata()  # Shape: 57, 67, 56, 64
+    bvals, bvecs = read_bvals_bvecs(in_bval, in_bvec)
+
+    responses, masks = compute_msmt_frf(dwi, bvals, bvecs)
+
+    # Value with current data at the date of test creation:
+    expected_result_wm = [[1.56925332e-03, 4.68706503e-04,
+                           4.68706503e-04, 3.26903486e+03],
+                          [1.15181122e-03, 3.75303294e-04,
+                           3.75303294e-04, 3.26903486e+03],
+                          [8.61299793e-04, 3.14541494e-04,
+                           3.14541494e-04, 3.26903486e+03]]
+    expected_result_gm = [[9.74471606e-04, 8.34628732e-04,
+                           8.34628732e-04, 3.42007686e+03],
+                          [7.76991313e-04, 6.89550835e-04,
+                           6.89550835e-04, 3.42007686e+03],
+                          [6.26617550e-04, 5.73389066e-04,
+                           5.73389066e-04, 3.42007686e+03]]
+    expected_result_csf = [[9.33140592e-04, 8.31445917e-04,
+                            8.31445917e-04, 3.62805637e+03],
+                           [7.69894406e-04, 7.07255607e-04,
+                            7.07255607e-04, 3.62805637e+03],
+                           [6.34735398e-04, 5.96451860e-04,
+                            5.96451860e-04, 3.62805637e+03]]
+    assert np.allclose(responses[0], expected_result_wm)
+    assert np.allclose(responses[1], expected_result_gm)
+    assert np.allclose(responses[2], expected_result_csf)
+
+    assert np.count_nonzero(masks[0]) == 845   # wm
+    assert np.count_nonzero(masks[1]) == 1779  # gm
+    assert np.count_nonzero(masks[2]) == 449   # csf
 
 
 def test_replace_frf():
-    # toDo
-    pass
+    old_frf = np.random.rand(4)
+    new_frf = "15,4,4"
+    result = replace_frf(old_frf, new_frf, no_factor=True)
+
+    # Rounds to float64
+    assert np.allclose(result, [15, 4, 4, old_frf[-1]])

scripts/scil_fodf_max_in_ventricles.py

@@ -89,8 +89,8 @@ def main():
     sh_basis, is_legacy = parse_sh_basis_arg(args)
 
     value, mask = get_ventricles_max_fodf(fodf, fa, md, zoom, sh_basis,
-                                          args.small_dims, args.fa_threshold,
-                                          args.md_threshold,
+                                          args.fa_threshold, args.md_threshold,
+                                          small_dims=args.small_dims,
                                           is_legacy=is_legacy)
 
     if args.mask_output:

scripts/tests/test_frf_set_diffusivities.py

@@ -20,26 +20,23 @@ def test_help_option(script_runner):
 
 def test_execution_processing_ssst(script_runner, monkeypatch):
     monkeypatch.chdir(os.path.expanduser(tmp_dir.name))
-    in_frf = os.path.join(SCILPY_HOME, 'processing',
-                          'frf.txt')
+    in_frf = os.path.join(SCILPY_HOME, 'processing', 'frf.txt')
     ret = script_runner.run('scil_frf_set_diffusivities.py', in_frf,
                             '15,4,4', 'new_frf.txt', '-f')
     assert ret.success
 
 
 def test_execution_processing_msmt(script_runner, monkeypatch):
     monkeypatch.chdir(os.path.expanduser(tmp_dir.name))
-    in_frf = os.path.join(SCILPY_HOME, 'commit_amico',
-                          'wm_frf.txt')
+    in_frf = os.path.join(SCILPY_HOME, 'commit_amico', 'wm_frf.txt')
     ret = script_runner.run('scil_frf_set_diffusivities.py', in_frf,
                             '15,4,4,13,4,4,12,5,5', 'new_frf.txt', '-f')
     assert ret.success
 
 
 def test_execution_processing__wrong_input(script_runner, monkeypatch):
     monkeypatch.chdir(os.path.expanduser(tmp_dir.name))
-    in_frf = os.path.join(SCILPY_HOME, 'commit_amico',
-                          'wm_frf.txt')
+    in_frf = os.path.join(SCILPY_HOME, 'commit_amico', 'wm_frf.txt')
     ret = script_runner.run('scil_frf_set_diffusivities.py', in_frf,
                             '15,4,4,13,4,4', 'new_frf.txt', '-f')
     assert not ret.success