C[omp]ute

From: "andrew cooke" <andrew@...>

Date: Mon, 12 Jan 2009 00:54:55 -0300 (CLST)

Not ready for release yet, but I've just got a new parser, written in
Python, to the point where it's useful.

It includes full backtracing (parse forests etc), the (untested) ability
to 'automatically' control resource use (think 'maximum backtrace stack')
and enough syntactic sugar to rot your teeth :o)

This test:


from logging import basicConfig, DEBUG
from unittest import TestCase

from lepl.match import *
from lepl.node import Node


class NodeTest(TestCase):


  def test_node(self):
    basicConfig(level=DEBUG)

    class Term(Node): pass
    class Factor(Node): pass
    class Expression(Node): pass

    expression  = Delayed()
    number      = Digit()[1:,...]                   > 'number'
    term        = (number | '(' / expression / ')') > Term
    muldiv      = Any('*/')                         > 'operator'
    factor      = (term / (muldiv / term)[0:])      > Factor
    addsub      = Any('+-')                         > 'operator'
    expression += (factor / (addsub / factor)[0:])  > Expression

    (ast, _) = next(expression.match_string('1 + 2 * (3 + 4 - 5)'))
    print(ast[0])
    (ast, _) = next(expression('1 + 2 * (3 + 4 - 5)'))
    print(ast[0])


Prints the following (twice):

Expression
 +- Factor
 |   +- Term
 |   |   `- number=1
 |   `- ' '
 +- operator=+
 +- ' '
 `- Factor
     +- Term
     |   `- number=2
     +- ' '
     +- operator=*
     +- ' '
     `- Term
         +- '('
         +- Expression
         |   +- Factor
         |   |   +- Term
         |   |   |   `- number=3
         |   |   `- ' '
         |   +- operator=+
         |   +- ' '
         |   +- Factor
         |   |   +- Term
         |   |   |   `- number=4
         |   |   `- ' '
         |   +- operator=-
         |   +- ' '
         |   `- Factor
         |       `- Term
         |           `- number=5
         `- ')'

Andrew