parse_record.py

#!/usr/bin/env python
"""
http://ru.stackoverflow.com/q/427088/23044
"""
import grako # $ pip install grako

grammar_ebnf = """
    record = type name teacher auditory $ ;
    type = "pr." | "lab." | "c." ;
    name = { word }+ ;
    teacher = abbr word ;
    auditory = /[0-9]{3}/ ;
    (* add negative lookahead for the dot to avoid matching start of the abbr *)
    word = { CHAR }+ !/\./;
    abbr = CHAR [ CHAR ] /\./ ;
    CHAR = /[a-zA-Z]/ ;
"""

class Semantics:
    def name(self, ast):
        return ' '.join(ast)
    def teacher(self, ast):
        return ' '.join(ast)
    def auditory(self, ast):
        return int(ast)
    def word(self, ast):
        return ''.join(ast)
    def abbr(self, ast):
        return ''.join(ast)

model = grako.genmodel("Record", grammar_ebnf)
ast = model.parse("lab. Subject name Au. Diachin 511", "record",
                  semantics=Semantics())
from pprint import pprint
pprint(dict(zip("type name teacher auditory".split(), ast)))