import py
import datetime

SSH = 'wyvern'
REMOTE_PATH = '/var/www/localhost/htdocs/irc-logs/pypy'
R_BASENAME = '.*(2\d\d\d)(\d\d)(\d\d)$'
R_TOPIC_CHANGE = '\[(\d+):(\d+)\] Topic changed on #pypy by [^:]+: [^(]+\(irc-logs: http://[^)]+\)(.*)'
MIN_ALIVE_TIME = 30*60   # half an hour

channel = py.execnet.SshGateway(SSH).remote_exec("""
    import os, re
    path = channel.receive()
    r_basename = re.compile(channel.receive())
    r_topic_change = re.compile(channel.receive())
    files = os.listdir(path)
    files.sort()
    for basename in files:
        match1 = r_basename.match(basename)
        if match1:
            f = open(os.path.join(path, basename))
            lines = f.readlines()
            f.close()
            for line in lines:
                match = r_topic_change.match(line)
                if match:
                    channel.send(match1.groups() + match.groups())
""")

channel.send(REMOTE_PATH)
channel.send(R_BASENAME)
channel.send(R_TOPIC_CHANGE)

starttime = []
texts = []
for year, month, day, hour, minute, text in channel:
    text = text.strip().lstrip('|').strip()
    if text:
        starttime.append((datetime.datetime(int(year),
                                            int(month),
                                            int(day),
                                            int(hour),
                                            int(minute))))
        texts.append(text)

endtime = starttime[1:] + [datetime.datetime.now()]

seen = {}
for start, end, text in zip(starttime, endtime, texts):
    if end - start >= datetime.timedelta(seconds=MIN_ALIVE_TIME):
        atom = object()
        if seen.setdefault(text.upper(), atom) is atom:
            print text