Get text data from arimaa.com WC pages

#!/usr/bin/python
import requests
import re
import sys
from collections import defaultdict

# maximum number of losses before elimination
MAX_LOSSES = 3 - 1

if len(sys.argv)>1:
    year = int(sys.argv[1])
    url = "http://arimaa.com/arimaa/wc/%i/showGames.cgi" % year
    page = requests.get(url).text
else:
    page = sys.stdin.read()

# The regex matches eg.
#    <nobr>* <a href="javascript:playerPage('novacat')">novacat</a><br><font size=-1>2002.3</font></nobr>
#    <nobr> <a href="javascript:playerPage('RmznA')">RmznA</a><br><font size=-1>1819.7</font></nobr>
#     <nobr><a href="javascript:playerPage('rabbits')">rabbits</a> *<br><font size=-1>2259.6</font></nobr>
regex = r"""(?P<line>.+?playerPage\('(?P<player>[\w_]+)'\)">(?P=player)</a>.+?<font size=-1>)"""
# "-BYE-" is sometimes used as a player, but doesn't match the player group
regex+=r"|(?P<bye>(?:bye|>-BYE-<))"
#break the games into rounds ("2nd place", "3rd place" etc. won't be matched)
regex+=r"|(?:Round (?P<round>\d+))"

regex  = re.compile(regex)
matches = re.finditer(regex, page)

nb_loss_for = defaultdict(lambda:0)
rounds = dict()
round_id = -1

while True:
    try:
        match_next = matches.next()
        # Rounds can come left-to-right or right-to-left.
        # The titles are always above the games.
        if match_next.group("round"):
            if round_id > 0:
                rounds[round_id] = new_round
            round_id =  int(match_next.group("round"))
            new_round =  []
            continue
        match_gold = match_next
        match_silver = matches.next()
        if match_silver.group("bye") or match_gold.group("bye"):
            continue

        gold = match_gold.group("player")
        silver = match_silver.group("player")

        if "*" in match_gold.group("line"):
            result = "g"
        elif "*" in match_silver.group("line"):
            result = "s"
        else:
            result = "?"
        new_round.append( (gold,silver,result) )

    except StopIteration:
        rounds[round_id] = new_round
        break

for round_id in sorted(rounds):
    print "\n### Round %i ###"  % round_id
    for gold, silver, result in rounds[round_id]:
        if (nb_loss_for[gold] > MAX_LOSSES):
            swiss = "!"
        else:
            swiss = ""
        if result=="g":
            nb_loss_for[silver]+=1
        elif result=="s":
            nb_loss_for[gold]+=1
        print "%-24s%-24s%s" % (swiss + gold, silver, result)