Review python dependencies

rabies/analysis_pkg/diagnosis_pkg/analysis_QC.py

@@ -1,8 +1,7 @@
 import SimpleITK as sitk
 import numpy as np
 import matplotlib.pyplot as plt
-import nilearn
-from rabies.visualization import otsu_scaling, plot_3d
+from rabies.visualization import otsu_scaling, plot_3d, cold_hot
 from rabies.analysis_pkg.analysis_math import elementwise_spearman, elementwise_corrcoef, dice_coefficient
 from rabies.utils import recover_3D
 from rabies.confound_correction_pkg.utils import smooth_image
@@ -180,7 +179,7 @@ def masked_plot(fig,axes, img, scaled, mask_img, vmax=None):
     plot_3d(axes,scaled,fig,vmin=0,vmax=1,cmap='gray', alpha=1, cbar=False, num_slices=6, planes=planes)
     # resample to match template
     sitk_img = sitk.Resample(masked, scaled)
-    cbar_list = plot_3d(axes,sitk_img,fig,vmin=-vmax,vmax=vmax,cmap='cold_hot', alpha=1, cbar=True, threshold=vmax*0.001, num_slices=6, planes=planes)
+    cbar_list = plot_3d(axes,sitk_img,fig,vmin=-vmax,vmax=vmax,cmap=cold_hot, alpha=1, cbar=True, threshold=vmax*0.001, num_slices=6, planes=planes)
     return cbar_list
 
 

rabies/analysis_pkg/diagnosis_pkg/diagnosis_functions.py

@@ -5,7 +5,6 @@
 from rabies.utils import copyInfo_3DImage, recover_3D
 from rabies.analysis_pkg import analysis_functions
 import SimpleITK as sitk
-import nilearn.plotting
 from .analysis_QC import masked_plot, threshold_top_percent
 
 def compute_spatiotemporal_features(CR_data_dict, sub_maps_data_dict, common_maps_data_dict, analysis_dict, 
@@ -465,7 +464,7 @@ def scan_diagnosis(CR_data_dict, maps_data_dict, temporal_info, spatial_info, pl
     fig2, axes2 = plt.subplots(nrows=nrows, ncols=3, figsize=(12*3, 2*nrows))
     plt.tight_layout()
 
-    from rabies.visualization import otsu_scaling, plot_3d
+    from rabies.visualization import otsu_scaling, plot_3d, cold_hot
 
     axes = axes2[0, :]
     scaled = otsu_scaling(template_file)
@@ -534,7 +533,7 @@ def scan_diagnosis(CR_data_dict, maps_data_dict, temporal_info, spatial_info, pl
     vector = spatial_info['GS_cov'].flatten()
     vector.sort()
     vmax = vector[int(len(vector)*0.95)]
-    cbar_list = plot_3d(axes, sitk_img, fig2, vmin=-vmax, vmax=vmax, cmap='cold_hot',
+    cbar_list = plot_3d(axes, sitk_img, fig2, vmin=-vmax, vmax=vmax, cmap=cold_hot,
             alpha=1, cbar=True, num_slices=6)
     for cbar in cbar_list:
         cbar.ax.get_yaxis().labelpad = 20

rabies/confound_correction_pkg/mod_ICA_AROMA/classification_plots.py

@@ -53,13 +53,13 @@ def classification_plot(myinput, outDir):
         df3 = pd.DataFrame.from_records([["True", 1., 1., 0., 0.],
                                         ["True", 1., 1., 0., 0.],
                                         ["True", 1., 1., 0., 0.]])
-        df = df.append(df3, ignore_index=True)
+        df = pd.concat([df, df3], ignore_index=True)
     tmp = df.loc[df[0] == "False"]
     if len(tmp) < 3:
         df3 = pd.DataFrame.from_records([["False", 0., 0., 0., 0.],
                                         ["False", 0., 0., 0., 0.],
                                         ["False", 0., 0., 0., 0.]])
-        df = df.append(df3, ignore_index=True)
+        df = pd.concat([df, df3], ignore_index=True)
 
     # rename columns
     df = df.rename(index=str, columns={0: 'Motion',

rabies/confound_correction_pkg/utils.py

@@ -642,8 +642,6 @@ def phase_randomized_regressors(regressors_array, frame_mask, TR):
 
 
 def smooth_image(img, fwhm, mask_img):
-    # apply nilearn's Gaussian smoothing on a SITK image
-    from nilearn.image.image import _smooth_array
     from rabies.utils import copyInfo_4DImage, copyInfo_3DImage
 
     dim = img.GetDimension()
@@ -706,4 +704,94 @@ def sitk_affine_lps(img):
     affine = np.eye(dim + 1)
     affine[:dim, :dim] = direction @ spacing
     affine[:dim, dim] = origin
-    return affine
+    return affine
+
+
+"""
+Reproduction of Nilearn's nilearn.image.smooth_img smoothing function (version 0.7.1)
+
+Original reference:
+Nilearn developers. Nilearn: Statistical learning for neuroimaging in Python.
+URL: https://nilearn.github.io/
+"""
+def _smooth_array(arr, affine, fwhm=None, ensure_finite=True, copy=True):
+    """Smooth images by applying a Gaussian filter.
+
+    Apply a Gaussian filter along the three first dimensions of `arr`.
+
+    Parameters
+    ----------
+    arr : :class:`numpy.ndarray`
+        4D array, with image number as last dimension. 3D arrays are also
+        accepted.
+
+    affine : :class:`numpy.ndarray`
+        (4, 4) matrix, giving affine transformation for image. (3, 3) matrices
+        are also accepted (only these coefficients are used).
+        If `fwhm='fast'`, the affine is not used and can be None.
+
+    fwhm : scalar, :class:`numpy.ndarray`/:obj:`tuple`/:obj:`list`, 'fast' or None, optional
+        Smoothing strength, as a full-width at half maximum, in millimeters.
+        If a nonzero scalar is given, width is identical in all 3 directions.
+        A :class:`numpy.ndarray`, :obj:`tuple`, or :obj:`list` must have 3 elements,
+        giving the FWHM along each axis.
+        If any of the elements is zero or None, smoothing is not performed
+        along that axis.
+        If  `fwhm='fast'`, a fast smoothing will be performed with a filter
+        [0.2, 1, 0.2] in each direction and a normalisation
+        to preserve the local average value.
+        If fwhm is None, no filtering is performed (useful when just removal
+        of non-finite values is needed).
+
+    ensure_finite : :obj:`bool`, optional
+        If True, replace every non-finite values (like NaNs) by zero before
+        filtering. Default=True.
+
+    copy : :obj:`bool`, optional
+        If True, input array is not modified. True by default: the filtering
+        is not performed in-place. Default=True.
+
+    Returns
+    -------
+    :class:`numpy.ndarray`
+        Filtered `arr`.
+
+    Notes
+    -----
+    This function is most efficient with arr in C order.
+
+    """
+    # Here, we have to investigate use cases of fwhm. Particularly, if fwhm=0.
+    # See issue #1537
+    if isinstance(fwhm, (int, float)) and (fwhm == 0.0):
+        import warnings
+        warnings.warn("The parameter 'fwhm' for smoothing is specified "
+                      "as {0}. Setting it to None "
+                      "(no smoothing will be performed)"
+                      .format(fwhm))
+        fwhm = None
+    if arr.dtype.kind == 'i':
+        if arr.dtype == np.int64:
+            arr = arr.astype(np.float64)
+        else:
+            arr = arr.astype(np.float32)  # We don't need crazy precision.
+    if copy:
+        arr = arr.copy()
+    if ensure_finite:
+        # SPM tends to put NaNs in the data outside the brain
+        arr[np.logical_not(np.isfinite(arr))] = 0
+    if isinstance(fwhm, str) and (fwhm == 'fast'):
+        raise NotImplementedError("The 'fast' option for fwhm is not implemented in rabies confound correction. Please specify a numeric value for fwhm or set to None for no smoothing.")
+        arr = _fast_smooth_array(arr)
+    elif fwhm is not None:
+        from scipy.ndimage import gaussian_filter1d
+        fwhm = np.asarray([fwhm]).ravel()
+        fwhm = np.asarray([0. if elem is None else elem for elem in fwhm])
+        affine = affine[:3, :3]  # Keep only the scale part.
+        fwhm_over_sigma_ratio = np.sqrt(8 * np.log(2))  # FWHM to sigma.
+        vox_size = np.sqrt(np.sum(affine ** 2, axis=0))
+        sigma = fwhm / (fwhm_over_sigma_ratio * vox_size)
+        for n, s in enumerate(sigma):
+            if s > 0.0:
+                gaussian_filter1d(arr, s, output=arr, axis=n)
+    return arr

rabies/preprocess_pkg/preprocess_visual_QC.py

@@ -65,7 +65,6 @@ def template_info(anat_template, opts, out_dir,figure_format):
     import SimpleITK as sitk
     # set default threader to platform to avoid freezing with MultiProc https://github.com/SimpleITK/SimpleITK/issues/1239
     sitk.ProcessObject_SetGlobalDefaultThreader('Platform')
-    from nilearn import plotting
     import matplotlib.pyplot as plt
     from rabies.visualization import plot_3d, otsu_scaling
     os.makedirs(out_dir, exist_ok=True)
@@ -99,7 +98,6 @@ def template_masking(template, mask, out_dir, figure_format):
     import SimpleITK as sitk
     # set default threader to platform to avoid freezing with MultiProc https://github.com/SimpleITK/SimpleITK/issues/1239
     sitk.ProcessObject_SetGlobalDefaultThreader('Platform')
-    from nilearn import plotting
     import matplotlib.pyplot as plt
     from rabies.visualization import plot_3d, otsu_scaling
 

rabies/visualization.py

@@ -18,6 +18,46 @@
     "savefig.facecolor": "black",
     "savefig.edgecolor": "black"})
 
+"""
+Standalone reproduction of Nilearn's 'cold_hot' colormap.
+
+Original reference:
+Nilearn developers. Nilearn: Statistical learning for neuroimaging in Python.
+URL: https://nilearn.github.io/
+"""
+from matplotlib.colors import LinearSegmentedColormap
+# Segment data extracted from the actual Nilearn 'cold_hot'
+_cold_hot_segments = {
+    'red': [
+        (0.0, 1.0, 1.0),
+        (0.126984, 0.0, 0.0),
+        (0.5, 0.0, 0.0),
+        (0.5, 0.0416, 0.0416),
+        (0.682540, 1.0, 1.0),
+        (1.0, 1.0, 1.0),
+    ],
+    'green': [
+        (0.0, 1.0, 1.0),
+        (0.126984, 1.0, 1.0),
+        (0.317461, 0.0, 0.0),
+        (0.5, 0.0, 0.0),
+        (0.5, 0.0, 0.0),
+        (0.682540, 0.0, 0.0),
+        (0.873016, 1.0, 1.0),
+        (1.0, 1.0, 1.0),
+    ],
+    'blue': [
+        (0.0, 1.0, 1.0),
+        (0.317461, 1.0, 1.0),
+        (0.5, 0.0416, 0.0416),
+        (0.5, 0.0, 0.0),
+        (0.873016, 0.0, 0.0),
+        (1.0, 1.0, 1.0),
+    ]
+}
+# Create the LinearSegmentedColormap
+cold_hot = LinearSegmentedColormap('cold_hot', _cold_hot_segments, N=256)
+
 
 def otsu_scaling(input_image):
     from skimage.filters import threshold_multiotsu

rabies_environment.dev.yml

@@ -9,7 +9,6 @@ dependencies:
   - future
   - matplotlib
   - nibabel
-  - nilearn
   - nipype
   - numpy
   - pandas

rabies_environment.yml

@@ -9,7 +9,6 @@ dependencies:
   - future
   - matplotlib=3.3.4
   - nibabel=3.2.1
-  - nilearn=0.7.1
   - nipype=1.10.0
   - numpy=1.26.4
   - pandas=1.2.4

setup.py

@@ -17,26 +17,24 @@
 URL = 'https://github.com/CoBrALab/RABIES'
 EMAIL = '[email protected]'
 AUTHOR = 'CoBrALab'
-REQUIRES_PYTHON = '>=3.9.0'
+REQUIRES_PYTHON = '>=3.9.0, <3.13' # the package was tested up to 3.12
 
-# What packages are required for this module to be executed?
 REQUIRED = [
     # 'requests', 'maya', 'records',
     'pip>=23.0',
-    'matplotlib>=3.3.4',
-    'nibabel>=3.2.1',
-    'nilearn>=0.7.1',
-    'nipype>=1.6.1',
-    'numpy>=1.20.1',
-    'pandas>=1.2.4',
-    'pathos>=0.2.7',
-    'pybids==0.16.3',
-    'scikit-learn>=0.24.1',
-    'scikit-image>=0.18.2',
-    'scipy>=1.8.1',
-    'seaborn>=0.11.1',
+    'matplotlib>=3.3.4', # tested up to 3.10.8
+    'nibabel>=3.2.1', # tested up to 5.4.0
+    'nipype>=1.6.1', # tested up to 1.11.0
+    'numpy>=1.20.1', # tested up to 2.4.3
+    'pandas>=1.2.4', # tested up to 3.0.1
+    'pathos>=0.2.7', # tested up to 0.3.5
+    'pybids==0.16.3', # only tested with 0.16.3, newer versions have some issues with the BIDS layout of our data
+    'scikit-learn>=0.24.1', # tested up to 1.8.0
+    'scikit-image>=0.18.2', # tested up to 0.24.0
+    'scipy>=1.8.1', # tested up to 1.17.1
+    'seaborn>=0.11.1', # tested up to 0.13.2
     'simpleitk>=2.5.0',
-    'qbatch==2.3',
+    'qbatch==2.3', # tested only with 2.3
     'networkx<3',
     'traits<7.0',
     'tqdm',