The module documentation for robotparser and its Python 3 counterpart, urllib.robotparser, mention that they use the original specification. This specification does not have an Allow directive; that is a non-standard extension. Some major crawlers support it, but you (obviously) don’t have to support it to claim compliance.