Merge branch 'master' of https://github.com/scilus/scilpy into new_co…

…nnectivity_pairwise_metric

.github/workflows/build.yml

@@ -0,0 +1,28 @@
+name: Build Docker with selected version
+
+on:
+  workflow_dispatch:
+    inputs:
+      scilpy_commit:
+        description: Scilpy commit id
+        required: true
+
+jobs:
+  Build_Docker:
+
+    runs-on: ubuntu-latest
+
+    steps:
+      - uses: actions/checkout@v2
+        name: Check out repository
+      - name: Change scilpy version
+        run: sed -i '/ENV SCILPY_VERSION=/c\ENV SCILPY_VERSION=${{ github.event.inputs.scilpy_commit }}' containers/Dockerfile
+      - uses: mr-smithers-excellent/[email protected]
+        name: Docker Build & Push
+        with:
+          image: scilus/scilpy
+          tag: dev
+          dockerfile: containers/Dockerfile
+          registry: docker.io
+          username: ${{ secrets.DOCKER_USERNAME }}
+          password: ${{ secrets.DOCKER_PASSWORD }}

Jenkinsfile

@@ -4,21 +4,11 @@ pipeline {
     stages {
         stage('Build') {
             stages {
-                stage('Python3.6') {
-                    steps {
-                        withPythonEnv('CPython-3.6') {
-                            sh '''
-                                pip3 install numpy==1.18.* wheel
-                                pip3 install -e .
-                            '''
-                        }
-                    }
-                }
                 stage('Python3.7') {
                     steps {
                         withPythonEnv('CPython-3.7') {
                             sh '''
-                                pip3 install numpy==1.18.* wheel
+                                pip3 install numpy==1.20.* wheel
                                 pip3 install -e .
                             '''
                         }
@@ -31,7 +21,7 @@ pipeline {
             steps {
                 withPythonEnv('CPython-3.7') {
                     sh '''
-                        pip3 install numpy==1.18.* wheel
+                        pip3 install numpy==1.20.* wheel
                         pip3 install -e .
                         export MPLBACKEND="agg"
                         export OPENBLAS_NUM_THREADS=1
@@ -55,7 +45,12 @@ pipeline {
             cleanWs()
             script {
                 if (env.CHANGE_ID) {
-                    pullRequest.createReviewRequests(['arnaudbore'])
+                    if (pullRequest.createdBy != "arnaudbore"){
+                        pullRequest.createReviewRequests(['arnaudbore'])
+                    }
+                    else{
+                        pullRequest.createReviewRequests(['GuillaumeTh'])
+                    }
                 }
             }
         }

docs/requirements.txt

@@ -6,7 +6,7 @@ matplotlib==2.2.*
 nibabel==3.0.*
 nilearn==0.6.*
 numpy==1.18.*
-Pillow==7.1.*
+Pillow==8.2.*
 pybids==0.10.*
 pyparsing==2.2.*
 python-dateutil==2.7.*

requirements.txt

@@ -3,15 +3,15 @@ coloredlogs==10.0.*
 cycler==0.10.*
 Cython==0.29.*
 dipy==1.3.*
-fury==0.6.*
+fury==0.7.*
 future==0.17.*
 h5py==2.10.*
 kiwisolver==1.0.*
 matplotlib==2.2.*
 nibabel==3.0.*
 nilearn==0.6.*
-numpy==1.18.*
-Pillow==7.1.*
+numpy==1.20.*
+Pillow==8.2.*
 bids-validator==1.6.0
 pybids==0.10.*
 pyparsing==2.2.*

scilpy/image/utils.py

@@ -34,7 +34,7 @@ def count_non_zero_voxels(image):
     return nb_voxels
 
 
-def volume_iterator(img, blocksize=1):
+def volume_iterator(img, blocksize=1, start=0, end=0):
     """Generator that iterates on volumes of data.
 
     Parameters
@@ -43,24 +43,34 @@ def volume_iterator(img, blocksize=1):
         Image of a 4D volume with shape X,Y,Z,N
     blocksize : int, optional
         Number of volumes to return in a single batch
+    start : int, optional
+        Starting iteration index in the 4D volume
+    end : int, optional
+        Stopping iteration index in the 4D volume
+        (the volume at this index is excluded)
 
     Yields
     -------
     tuple of (list of int, ndarray)
         The ids of the selected volumes, and the selected data as a 4D array
     """
+    assert end <= img.shape[-1], "End limit provided is greater than the " \
+                                 "total number of volumes in image"
+
     nb_volumes = img.shape[-1]
+    end = end if end else img.shape[-1]
 
     if blocksize == nb_volumes:
-        yield list(range(nb_volumes)), img.get_fdata(dtype=np.float32)
+        yield list(range(start, end)), \
+              img.get_fdata(dtype=np.float32)[..., start:end]
     else:
-        start, end = 0, 0
-        for i in range(0, nb_volumes - blocksize, blocksize):
-            start, end = i, i + blocksize
-            logging.info("Loading volumes {} to {}.".format(start, end - 1))
-            yield list(range(start, end)), img.dataobj[..., start:end]
+        stop = start
+        for i in range(start, end - blocksize, blocksize):
+            start, stop = i, i + blocksize
+            logging.info("Loading volumes {} to {}.".format(start, stop - 1))
+            yield list(range(start, stop)), img.dataobj[..., start:stop]
 
-        if end < nb_volumes:
+        if stop < end:
             logging.info(
-                "Loading volumes {} to {}.".format(end, nb_volumes - 1))
-            yield list(range(end, nb_volumes)), img.dataobj[..., end:]
+                "Loading volumes {} to {}.".format(stop, end - 1))
+            yield list(range(stop, end)), img.dataobj[..., stop:end]

scilpy/reconst/utils.py

@@ -15,6 +15,22 @@ def find_order_from_nb_coeff(data):
     return int((-3 + np.sqrt(1 + 8 * shape[-1])) / 2)
 
 
+def get_sh_order_and_fullness(ncoeffs):
+    """
+    Get the order of the SH basis from the number of SH coefficients
+    as well as a boolean indicating if the basis is full.
+    """
+    # the two curves (sym and full) intersect at ncoeffs = 1, in what
+    # case both bases correspond to order 1.
+    sym_order = (-3.0 + np.sqrt(1.0 + 8.0 * ncoeffs)) / 2.0
+    if sym_order.is_integer():
+        return sym_order, False
+    full_order = np.sqrt(ncoeffs) - 1.0
+    if full_order.is_integer():
+        return full_order, True
+    raise ValueError('Invalid number of coefficients for SH basis.')
+
+
 def _honor_authorsnames_sh_basis(sh_basis_type):
     sh_basis = sh_basis_type
     if sh_basis_type == 'fibernav':

scilpy/segment/voting_scheme.py

@@ -288,13 +288,13 @@ def __call__(self, input_tractogram_path, nbr_processes=1, seeds=None):
                                                 slr_transform_type, seed])
 
         tmp_dir, tmp_memmap_filenames = streamlines_to_memmap(wb_streamlines)
+        del wb_streamlines
         comb_param_cluster = product(self.tractogram_clustering_thr, seeds)
 
         # Clustring is now parallelize
         pool = multiprocessing.Pool(nbr_processes)
         all_rbx_dict = pool.map(single_clusterize_and_rbx_init,
-                                zip(repeat(wb_streamlines),
-                                    repeat(tmp_memmap_filenames),
+                                zip(repeat(tmp_memmap_filenames),
                                     comb_param_cluster,
                                     repeat(self.nb_points)))
         pool.close()
@@ -363,8 +363,6 @@ def single_clusterize_and_rbx_init(args):
 
     Parameters
     ----------
-    wb_streamlines : list or ArraySequence
-        All streamlines of the tractogram to segment.
     tmp_memmap_filename: tuple (3)
         Temporary filename for the data, offsets and lengths.
 
@@ -381,11 +379,11 @@ def single_clusterize_and_rbx_init(args):
     rbx : dict
         Initialisation of the recobundles class using specific parameters.
     """
-    wb_streamlines = args[0]
-    tmp_memmap_filename = args[1]
-    clustering_thr = args[2][0]
-    seed = args[2][1]
-    nb_points = args[3]
+    tmp_memmap_filename = args[0]
+    wb_streamlines = reconstruct_streamlines_from_memmap(tmp_memmap_filename)
+    clustering_thr = args[1][0]
+    seed = args[1][1]
+    nb_points = args[2]
 
     rbx = {}
     base_thresholds = [45, 35, 25]

scilpy/tractanalysis/features.py

@@ -7,9 +7,52 @@
 from dipy.segment.metric import ResampleFeature
 from dipy.segment.metric import AveragePointwiseEuclideanMetric
 from dipy.tracking import metrics as tm
+from scilpy.tracking.tools import resample_streamlines_num_points
 import numpy as np
 
 
+def detect_ushape(sft, minU, maxU):
+    """
+    Extract streamlines depending of their "u-shapeness".
+    Parameters
+    ----------
+    sft: Statefull tractogram
+        Tractogram used to extract streamlines depending on their ushapeness.
+    minU: Float
+        Minimum ufactor of a streamline.
+    maxU: Float
+        Maximum ufactor of a streamline.
+
+    Returns
+    -------
+    list: the ids of clean streamlines
+        Only the ids are returned so proper filtering can be done afterwards.
+    """
+    ids = []
+    new_sft = resample_streamlines_num_points(sft, 4)
+    for i, s in enumerate(new_sft.streamlines):
+        if len(s) == 4:
+            first_point = s[0]
+            last_point = s[-1]
+            second_point = s[1]
+            third_point = s[2]
+
+            v1 = first_point - second_point
+            v2 = second_point - third_point
+            v3 = third_point - last_point
+
+            v1 = v1 / np.linalg.norm(v1)
+            v2 = v2 / np.linalg.norm(v2)
+            v3 = v3 / np.linalg.norm(v3)
+
+            val = np.dot(np.cross(v1, v2), np.cross(v2, v3))
+
+            if minU <= val <= maxU:
+                ids.append(i)
+
+    return ids
+
+
 def remove_loops_and_sharp_turns(streamlines,
                                  max_angle,
                                  use_qb=False,

scilpy/tractanalysis/todi.py

@@ -200,7 +200,8 @@ def smooth_todi_spatial(self, sigma=0.5):
                 new_todi = deepcopy(tmp_todi)
             else:
                 new_todi = np.hstack((new_todi, tmp_todi))
-            self.todi = np.delete(self.todi, range(0, chunk_size), axis=1)
+            self.todi = np.delete(self.todi, range(
+                0, min(self.todi.shape[1], chunk_size)), axis=1)
             chunk_count -= 1
 
         self.mask = new_mask