gistfile1.py

def tokenize(string):

    # Regular expression to extract blocks
    _recipient = re.compile(r'^(.+?)\s(in|to|about)\b',
                            re.IGNORECASE)
    _time = re.compile(r'\bin\s(\d+\s\b.+?\b)(?:\sto\b|\sabout\b|$)',
                            re.IGNORECASE)
    _message = re.compile(r'\b(?:to|about)\s(.+?)(?:\sin\b|$)',
                            re.IGNORECASE)

    recipientmatch = _recipient.search(string)
    timematch = _time.search(string)
    messagematch = _message.search(string)

    if recipientmatch is None:
        raise ReminderSyntaxError("Remind who?")

    if timematch is None:
        raise ReminderSyntaxError("Remind when?")

    if messagematch is None:
        raise ReminderSyntaxError("Remind what..?")

    # Fetch groups as tokens, in case of multiples
    # groups required
    recipientblock = recipientmatch.group(1)
    timeblock = timematch.group(1)
    messageblock = messagematch.group(1)

    return {"recipient": recipientblock,
        "time": timeblock, "message": messageblock }