如何在python中设置ply的yacc优先级？

Question

I need to make a AST from a regular expression using ply. For example, if the RE is (a|b*)abc, I want to make a pared tuple as (':', (':', (':', ('|', 'a', ('*', 'b')), 'a'), 'b'), 'c') <-- ':' means just split a string in two parts.

Here is my code.

tokens = (
    "SYMBOL",
    "LBRACKET",
    "RBRACKET",
    "STAR",
    "UNION"
)

t_ignore = ' \t\n'
t_SYMBOL = r'[a-zA-Z0-9]'
t_LBRACKET = r'\('
t_RBRACKET = r'\)'
t_STAR = r'\*'
t_UNION = r'\|'

def t_error(t):
    raise TypeError("Unknown text '%s'" % t.value)

def p_multiple(p) :
    '''string : SYMBOL SYMBOL
              | string SYMBOL
              | string string'''
    p[0] = (':', p[1], p[2])    

def p_union(p) :
    '''string : string UNION string'''
    p[0] = ('|', p[1], p[3])

def p_star(p) :
    '''string : string STAR'''
    p[0] = ('*', p[1])

def p_brackets(p) :
    '''string : LBRACKET string RBRACKET'''
    p[0] = p[2]


def p_symbol(p) :
    '''string : SYMBOL'''
    p[0] = p[1]

def p_error(p):
    print ("Syntax error at '%s'" % p.value)

lex.lex()
yacc.yacc()

lex.input("(a|b*)abc")
parsed = yacc.parse(RE)

Answer 1

To resurrect this question, ply allows you to set a precedence variable that contains the precedence for the symbols. For example:

precedence = (
    ('left', 'PLUS', 'MINUS'),
    ('left', 'TIMES', 'DIVIDE'),
)

You can find a full tutorial and documentation here: http://www.dabeaz.com/ply/ply.html#ply_nn27 (I had to use find functionality for finding the right spot)

Answer 2

I can't answer your question directly. However, you could perhaps consider using greenery .

>>> from greenery.lego import parse
>>> thingy = parse('(a|b*)abc')
>>> thingy.to_fsm()
fsm(alphabet = {'b', 'c', anything_else, 'a'}, states = {0, 1, 2, 3, 4, 5}, initial = 0, finals = {5}, map = {0: {'b': 2, 'a': 1}, 1: {'b': 4, 'a': 3}, 2: {'b': 2, 'a': 3}, 3: {'b': 4}, 4: {'c': 5}, 5: {}})

I think you could recover what you want by stepping backwards recursively from the each of the finals states in map .

If you down-vote please indicate why.