Add old script for creating dataset TOML from HDF5

scripts/create_dataset_toml.py

@@ -0,0 +1,77 @@
+#!/usr/bin/env python
+# -*- coding: utf-8 -*-
+
+
+from __future__ import print_function
+
+import glob
+import os
+import re
+
+import argparse
+import h5py
+import numpy as np
+import pytoml as toml
+
+from diluvian.util import get_nonzero_aabb
+
+
+def create_dataset_conf_from_files(path, file_pattern, name_regex, name_format, mask_bounds=True):
+    pathspec = path + file_pattern
+    name_regex = re.compile(name_regex)
+
+    datasets = []
+
+    for pathname in glob.iglob(pathspec):
+        filename = os.path.basename(pathname)
+        name = name_format.format(*name_regex.match(filename).groups())
+        ds = {
+                'name': name,
+                'hdf5_file': pathname,
+                'image_dataset': 'volumes/raw',
+                'label_dataset': 'volumes/labels/neuron_ids',
+                'mask_dataset': 'volumes/labels/mask',
+                'resolution': [40, 4, 4],
+        }
+
+        if mask_bounds:
+            print('Finding mask bounds for {}'.format(filename))
+            f = h5py.File(pathname, 'r')
+            d = f[ds['mask_dataset']]
+            mask_data = d[:]
+            mask_min, mask_max = get_nonzero_aabb(mask_data)
+
+            ds['mask_bounds'] = [mask_min, mask_max]
+            f.close()
+
+        datasets.append(ds)
+
+    return {'dataset': datasets}
+
+
+if __name__ == '__main__':
+    parser = argparse.ArgumentParser(description='Create a dataset TOML from a directory of HDF5 files.')
+
+    parser.add_argument(
+        '--file-pattern', dest='file_pattern', default='sample_[ABC]*hdf',
+        help='Glob for HDF5 volume filenames.')
+    parser.add_argument(
+        '--name-regex', dest='name_regex', default=r'sample_([ABC])(.*).hdf',
+        help='Regex for extracting volume name from filenames.')
+    parser.add_argument(
+        '--name-format', dest='name_format', default='Sample {} ({})',
+        help='Format string for creating volume names from name regex matches.')
+    parser.add_argument(
+        'path', default=None,
+        help='Path to the HDF5 volume files.')
+    parser.add_argument(
+        'dataset_file', default=None,
+        help='Name for the TOML dataset file that will be created.')
+
+    args = parser.parse_args()
+
+    conf = create_dataset_conf_from_files(args.path, args.file_pattern, args.name_regex, args.name_format)
+    print('Found {} datasets.'.format(len(conf['dataset'])))
+
+    with open(args.dataset_file, 'wb') as tomlfile:
+        tomlfile.write(toml.dumps(conf))