Fix the -e flag of moabb.run splitting its value into characters

docs/source/whats_new.rst

@@ -35,7 +35,7 @@ Requirements
 
 Bugs
 ~~~~
-- None yet.
+- Fix the ``-e``/``--evaluations`` flag of ``python -m moabb.run``, which used ``type=list`` and so split its value into single characters: ``-e WithinSession`` reached :func:`moabb.benchmark` as ``['W', 'i', 't', ...]`` and raised ``KeyError: 'W'``. It now takes one or more evaluation names, space separated (by `Iain`_)
 
 Code health
 ~~~~~~~~~~~

moabb/run.py

@@ -34,10 +34,11 @@ def parser_init():
         "-e",
         "--evaluations",
         dest="evaluations",
-        type=list,
+        type=str,
+        nargs="+",
         default=None,
-        help="Evaluation types to be run. Must be given as a list. "
-        'Options - ["WithinSession","CrossSession","CrossSubject"]'
+        help="Evaluation types to be run, space separated. "
+        'Options - "WithinSession", "CrossSession", "CrossSubject". '
         "By default, all 3 types of evaluations will be done",
     )
     parser.add_argument(

moabb/tests/test_run.py

@@ -0,0 +1,15 @@
+import pytest
+
+from moabb.run import parser_init
+
+
+@pytest.mark.parametrize(
+    "argv,expected",
+    [
+        (["-e", "WithinSession"], ["WithinSession"]),
+        (["-e", "WithinSession", "CrossSession"], ["WithinSession", "CrossSession"]),
+        ([], None),
+    ],
+)
+def test_parser_evaluations(argv, expected):
+    assert parser_init().parse_args(argv).evaluations == expected