Parser for Dutch natural language time strings (Python 2.5, pyPEG 1.5)

parse_time_pyPEG1.5.py

import re, fileinput
import pyPEG
from pyPEG import parse, parseLine
from pyPEG import keyword, _and, _not, ignore

import datetime

#pyPEG.print_trace = True

numbers = [
	"nul",
	"een",
	"twee",
	"drie",
	"vier",
	"vijf",
	"zes",
	"zeven",
	"acht",
	"negen",
	"tien",
	"elf",
	"twaalf",
	"dertien",
	"veertien",
	"vijftien",
	"zestien",
	"zeventien",
	"achttien",
	"negentien",
	"twintig",
	"eenentwintig",
	"tweeentwintig",
	"drieentwintig",
	"vierentwintig",
	"vijfentwintig",
	"zesentwintig",
	"zevenentwintig",
	"achtentwintig",
	"negenentwintig",
	"dertig",
	"eenendertig",
	"tweeendertig",
	"drieendertig",
	"vierendertig",
	"vijfendertig",
	"zesendertig",
	"zevenendertig",
	"achtendertig",
	"negenendertig",
	"veertig",
	"eenenveertig",
	"tweeenveertig",
	"drieenveertig",
	"vierenveertig",
	"vijfenveertig",
	"zesenveertig",
	"zevenenveertig",
	"achtenveertig",
	"negenenveertig",
	"vijftig",
	"eenenvijftig",
	"tweeenvijftig",
	"drieenvijftig",
	"vierenvijftig",
	"vijfenvijftig",
	"zesenvijftig",
	"zevenenvijftig",
	"achtenvijftig",
	"negenenvijftig",	
	"zestig"	
]

times = [
	"12 56",
	"12:56",
	"twaalf dertig",
	"negen vijftien",
	"kwart voor twee",
	"tien voor twee",
	"drie over half tien",
	"13 uur 52",
	"half drie",
	"zevenenveertig over tien",
	"zeven voor half vier",
	"achttien vijftien",
	"zeven",
	"negen uur",
	"8 uur",
	"achttien uur vijftien"
]

def number():	return re.compile(r"\w+")
def half():		return re.compile(r"half")
def hours():	return -1, half, number, -1, keyword("uur")
def sign():		return [re.compile(r"voor"), re.compile(r"over")]
def minutes():	return number
def time():		return [		
					(minutes, sign, hours),
					(minutes, sign, hours),
					(hours, -1, ":", minutes),
					hours
				]

def string_to_int(str):
	if str == "kwart":
		return 15		
	for i in range(0, 60):
		if str == numbers[i]:
			return i
		
	return int(str)

def to_time(ast):
	minutes_str = ""
	hours_str = ""
	half = False
	sign = 1
	
	# ast is tuple (ast, ''). skip weird '' part:
	ast = ast[0]

	for symbol in ast:
		name = symbol[0]
		value = symbol[1]
		if name == "hours":
			if len(value) == 2:
				# Has 'half'
				half = True
				hours_str = value[1][1]
			else:
				hours_str = value[0][1]
		elif name == "minutes":
			minutes_str = value[0][1]
		elif name == "sign":
			if value[0] == "voor":
				sign = -1

	minutes = 0

	if len(hours_str) > 0:
		minutes = string_to_int(hours_str) * 60

	if half:
		minutes -= 30
		
	if len(minutes_str) > 0:
		minutes += sign * string_to_int(minutes_str)

	hours = minutes // 60
	minutes = minutes - (60 * hours)
	
	today = datetime.date.today() + datetime.timedelta(days=1)
	return datetime.datetime.combine(today, datetime.time(hours, minutes))

for time_str in times:	
	ast = parseLine(textline=time_str, pattern=time(), resultSoFar=[])	
	print time_str, " => ", to_time(ast)