refactored; added output directory in colors.py

Author: Shuo Han

colors_script/calc_colormap.py

@@ -1,40 +1,69 @@
 #!/usr/bin/env python
 # -*- coding: utf-8 -*-
 
+
 import sys
 sys.path.append('../mindboggle/mio')
 import colors
 import numpy as np
 import os
-import pickle
+import argparse
+import shutil
+
+
+def parse_inputs():
+
+    des = 'calculate colormap for labeled image, store the colors in the file'\
+        + ' colors.npy'
+    parser = argparse.ArgumentParser(description=des)
+    parser.add_argument('label_filename', help='the path of the label image file')
+    help = 'path of folder to store intermediate files and final results'
+    parser.add_argument('output_dir', help=help)
+    args = parser.parse_args()
+    return args
+
 
-filename = 'label.nii.gz'
-matrix_filename = 'matrix.pickle'
-colormap_filename = 'colormap.npy'
+def main(args):
+
+    if not os.path.isdir(args.output_dir):
+        os.makedirs(args.output_dir)
+        
+    matrix_filename = os.path.join(args.output_dir, 'matrix.npy')
+    colormap_filename = os.path.join(args.output_dir, 'colormap.npy')
+    IDs_filename = os.path.join(args.output_dir, 'IDs.npy')
+    colors_filename = os.path.join(args.output_dir, 'colors.npy')
 
-if not os.path.isfile(matrix_filename):
     print('finding adjacency maps')
-    IDs, matrix, output = colors.label_adjacency_matrix(filename, ignore_values=[])
-    pickle.dump(matrix, open(matrix_filename, 'wb'))
-    np.save('IDs', IDs)
-else:
-    IDs = np.load('IDs.npy')
-    matrix = pickle.load(open(matrix_filename, 'rb'))
-
-matrix = matrix.as_matrix()
-matrix = matrix[:, 1:]
-
-print('finding colormap')
-if not os.path.isfile(colormap_filename):
-    colormap = colors.distinguishable_colors(ncolors=len(IDs),
-                                            backgrounds=[[0,0,0],[1,1,1]],
-                                            save_csv=False)
-    np.save(colormap_filename, colormap)
-else:
-    colormap = np.load(colormap_filename)
-
-print(colormap.shape)
-
-colors = colors.group_colors(colormap, 'cerebellum', IDs=IDs,
-                             adjacency_matrix=matrix)
-np.save('colors', colors)
+    if not os.path.isfile(matrix_filename) and not os.path.isfile(IDs_filename):
+        IDs, matrix, output = colors.label_adjacency_matrix(args.label_filename,
+                                                            out_dir=args.output_dir)
+        matrix = matrix.as_matrix()[:, 1:]
+        np.save(matrix_filename, matrix)
+        np.save(IDs_filename, IDs)
+    else:
+        IDs = np.load(IDs_filename)
+        matrix = np.load(matrix_filename)
+
+    print('finding colormap')
+    if not os.path.isfile(colormap_filename):
+        ncolors = len(IDs)
+        colormap = colors.distinguishable_colors(ncolors=ncolors,
+                                                 plot_colormap=False,
+                                                 save_csv=False,
+                                                 out_dir=args.output_dir)
+        np.save(colormap_filename, colormap)
+    else:
+        colormap = np.load(colormap_filename)
+
+    print('finding label colors')
+    label_colors = colors.group_colors(colormap, args.label_filename, IDs=IDs,
+                                       adjacency_matrix=matrix, 
+                                       out_dir=args.output_dir,
+                                       plot_colors=True,
+                                       plot_graphs=False)
+    np.save(colors_filename, label_colors)
+
+
+if __name__ == "__main__":
+    args = parse_inputs()
+    main(args)

colors_script/convert_to_mipav_lut.py

@@ -2,28 +2,40 @@
 # -*- coding: utf-8 -*-
 
 import numpy as np
+import argparse
 
-colormap_filename = 'colors.npy'
-labels_filename = 'IDs.npy'
-lut_filename = 'colors.lut'
-
-labels = np.load(labels_filename)
-colors = np.load(colormap_filename)
-colors = colors * 255
-
-contents = list()
-contents.append('<LUT>')
-contents.append('256\t# Size of LUT Arrays')
-
-for i in range(256):
-    if i in labels:
-        idx = np.where(labels == i)
-        c = colors[idx].tolist()[0]
-    else:
-        c = [0.0, 0.0, 0.0]
-    c_str = [str(cc) for cc in c]
-    line = '\t'.join([str(i), '1.0', *c_str])
-    contents.append(line)
-
-with open(lut_filename, 'w') as file:
-    file.write('\n'.join(contents))
+def parse_inputs():
+    des = 'convert colormap to mipav lut file.'
+    parser = argparse.ArgumentParser(description=des)
+    parser.add_argument('colormap_filename')
+    parser.add_argument('ids_filename')
+    parser.add_argument('output_filename')
+    args = parser.parse_args()
+    return args
+
+def main(args):
+
+    labels = np.load(args.ids_filename)
+    colors = np.load(args.colormap_filename) * 255
+
+    contents = list()
+    contents.append('<LUT>')
+    contents.append('256\t# Size of LUT Arrays')
+
+    for i in range(256):
+        if i in labels:
+            idx = np.where(labels == i)[0][0]
+            c = colors[idx].tolist()
+        else:
+            c = [0.0, 0.0, 0.0]
+        c_str = [str(cc) for cc in c]
+        line = '\t'.join([str(i), '1.0', *c_str])
+        contents.append(line)
+
+    with open(args.output_filename, 'w') as file:
+        file.write('\n'.join(contents))
+
+
+if __name__ == "__main__":
+    args = parse_inputs()
+    main(args)

mindboggle/mio/colors.py

@@ -9,11 +9,13 @@
 
 """
 import sys
-sys.path.append('/Users/shuo/Code/git_repo/mindboggle')
+sys.path.append('../../mindboggle')
+import os
 
 
 def distinguishable_colors(ncolors, backgrounds=[[0,0,0],[1,1,1]],
-                           save_csv=True, plot_colormap=True, verbose=True):
+                           save_csv=True, plot_colormap=True, verbose=True,
+                           out_dir='.'):
     """
     Create a colormap of perceptually distinguishable colors.
 
@@ -76,6 +78,7 @@ def distinguishable_colors(ncolors, backgrounds=[[0,0,0],[1,1,1]],
     from colormath.color_diff import delta_e_cie2000
 
     filename = "colormap_of_{0}_distinguishable_colors".format(ncolors)
+    filename = os.path.join(out_dir, filename)
 
     # ------------------------------------------------------------------------
     # Generate a sizable number of RGB triples. This represents our space of
@@ -160,7 +163,7 @@ def distinguishable_colors(ncolors, backgrounds=[[0,0,0],[1,1,1]],
             if verbose:
                 print(rgb)
             plt.barh(0, 50, 1, 0, color=rgb)
-            plt.savefig(filename + ".png")
+        plt.savefig(filename + ".png")
         if verbose:
             print("Colormap image saved to {0}".format(filename + ".png"))
 
@@ -178,7 +181,7 @@ def distinguishable_colors(ncolors, backgrounds=[[0,0,0],[1,1,1]],
 
 def label_adjacency_matrix(label_file, ignore_values=[-1, 999], add_value=0,
                            save_table=True, output_format='csv',
-                           verbose=True):
+                           verbose=True, out_dir='.'):
     """
     Extract surface or volume label boundaries, find unique label pairs,
     and write adjacency matrix (useful for constructing a colormap).
@@ -282,6 +285,8 @@ def label_adjacency_matrix(label_file, ignore_values=[-1, 999], add_value=0,
     else:
         raise IOError("Use appropriate input file type.")
 
+    output_table = os.path.join(out_dir, output_table)
+
     # Find unique pairs (or first two of each list):
     pairs = []
     for pair in label_pairs: