powered by
In the python implementation, they preprocess all 1's into () and then parse it into the equivalent of a GroupAstNode. Here, we won't do that
parse_einops_ast(tokens)
EinopsAst object
EinopsTokenSequence object from the lexer