Fixes (#6)

datumaro/cli/contexts/project/__init__.py

@@ -447,7 +447,8 @@ def filter_command(args):
     if not args.filter:
         raise CliException("Expected a filter expression ('-e' argument)")
 
-    dataset.filter_project(save_dir=dst_dir, expr=args.filter, **filter_args)
+    dataset.filter_project(save_dir=dst_dir,
+        filter_expr=args.filter, **filter_args)
 
     log.info("Subproject has been extracted to '%s'" % dst_dir)
 
@@ -565,6 +566,8 @@ def diff_command(args):
 
     return 0
 
+_ediff_default_if = ['id', 'group'] # avoid https://bugs.python.org/issue16399
+
 def build_ediff_parser(parser_ctor=argparse.ArgumentParser):
     parser = parser_ctor(help="Compare projects for equality",
         description="""
@@ -583,9 +586,9 @@ def build_ediff_parser(parser_ctor=argparse.ArgumentParser):
         help="Ignore item attribute (repeatable)")
     parser.add_argument('-ia', '--ignore-attr', action='append',
         help="Ignore annotation attribute (repeatable)")
-    parser.add_argument('-if', '--ignore-field',
-        action='append', default=['id', 'group'],
-        help="Ignore annotation field (repeatable, default: %(default)s)")
+    parser.add_argument('-if', '--ignore-field', action='append',
+        help="Ignore annotation field (repeatable, default: %s)" % \
+            _ediff_default_if)
     parser.add_argument('--match-images', action='store_true',
         help='Match dataset items by images instead of ids')
     parser.add_argument('--all', action='store_true',
@@ -600,6 +603,8 @@ def ediff_command(args):
     first_project = load_project(args.project_dir)
     second_project = load_project(args.other_project_dir)
 
+    if args.ignore_field:
+        args.ignore_field = _ediff_default_if
     comparator = ExactComparator(
         match_images=args.match_images,
         ignored_fields=args.ignore_field,

datumaro/plugins/image_dir.py

@@ -42,8 +42,8 @@ def __init__(self, url):
         for dirpath, _, filenames in os.walk(url):
             for name in filenames:
                 path = osp.join(dirpath, name)
+                image = Image(path=path)
                 try:
-                    image = Image(path)
                     # force loading
                     image.data # pylint: disable=pointless-statement
                 except Exception:

datumaro/plugins/transforms.py

@@ -308,6 +308,9 @@ class RandomSplit(Transform, CliPlugin):
     |s|s%(prog)s --subset train:.67 --subset test:.33
     """
 
+    # avoid https://bugs.python.org/issue16399
+    _default_split = [('train', 0.67), ('test', 0.33)]
+
     @staticmethod
     def _split_arg(s):
         parts = s.split(':')
@@ -321,14 +324,17 @@ def build_cmdline_parser(cls, **kwargs):
         parser = super().build_cmdline_parser(**kwargs)
         parser.add_argument('-s', '--subset', action='append',
             type=cls._split_arg, dest='splits',
-            default=[('train', 0.67), ('test', 0.33)],
-            help="Subsets in the form of: '<subset>:<ratio>' (repeatable)")
+            help="Subsets in the form: '<subset>:<ratio>' "
+                "(repeatable, default: %s)" % dict(cls._default_split))
         parser.add_argument('--seed', type=int, help="Random seed")
         return parser
 
     def __init__(self, extractor, splits, seed=None):
         super().__init__(extractor)
 
+        if splits is None:
+            splits = self._default_split
+
         assert 0 < len(splits), "Expected at least one split"
         assert all(0.0 <= r and r <= 1.0 for _, r in splits), \
             "Ratios are expected to be in the range [0; 1], but got %s" % splits

datumaro/util/image.py

@@ -41,6 +41,8 @@ def load_image(path):
     else:
         raise NotImplementedError()
 
+    if image is None:
+        raise ValueError("Can't open image '%s'" % path)
     assert len(image.shape) in {2, 3}
     if len(image.shape) == 3:
         assert image.shape[2] in {3, 4}
@@ -206,6 +208,8 @@ def __init__(self, data=None, path=None, loader=None, cache=None,
         self._path = path
 
         assert data is not None or path or loader, "Image can not be empty"
+        if data is not None:
+            assert callable(data) or isinstance(data, np.ndarray), type(data)
         if data is None and (path or loader):
             if osp.isfile(path) or loader:
                 data = lazy_image(path, loader=loader, cache=cache)