plomlombot.py

   1 #!/usr/bin/python3
   2
   3 import argparse
   4 import socket
   5 import datetime
   6 import select
   7 import time
   8 import re
   9 import requests
  10 import bs4
  11 import random
  12 import hashlib
  13 import os
  14 import plomsearch
  15
  16 # Defaults, may be overwritten by command line arguments.
  17 SERVER = "irc.freenode.net"
  18 PORT = 6667
  19 TIMEOUT = 240
  20 USERNAME = "plomlombot"
  21 NICKNAME = USERNAME
  22 TWTFILE = ""
  23 DBDIR = os.path.expanduser("~/plomlombot_db")
  24
  25
  26 class ExceptionForRestart(Exception):
  27     pass
  28
  29
  30 class IO:
  31
  32     def __init__(self, server, port, timeout):
  33         self.timeout = timeout
  34         self.socket = socket.socket()
  35         self.socket.connect((server, port))
  36         self.socket.setblocking(0)
  37         self.line_buffer = []
  38         self.rune_buffer = ""
  39         self.last_pong = time.time()
  40         self.servername = self.recv_line(send_ping=False).split(" ")[0][1:]
  41
  42     def _pingtest(self, send_ping=True):
  43         if self.last_pong + self.timeout < time.time():
  44             print("SERVER NOT ANSWERING")
  45             raise ExceptionForRestart
  46         if send_ping:
  47             self.send_line("PING " + self.servername)
  48
  49     def send_line(self, msg):
  50         msg = msg.replace("\r", " ")
  51         msg = msg.replace("\n", " ")
  52         if len(msg.encode("utf-8")) > 510:
  53             print("NOT SENT LINE TO SERVER (too long): " + msg)
  54         print("LINE TO SERVER: "
  55               + str(datetime.datetime.now()) + ": " + msg)
  56         msg = msg + "\r\n"
  57         msg_len = len(msg)
  58         total_sent_len = 0
  59         while total_sent_len < msg_len:
  60             sent_len = self.socket.send(bytes(msg[total_sent_len:], "UTF-8"))
  61             if sent_len == 0:
  62                 print("SOCKET CONNECTION BROKEN")
  63                 raise ExceptionForRestart
  64             total_sent_len += sent_len
  65
  66     def _recv_line_wrapped(self, send_ping=True):
  67         if len(self.line_buffer) > 0:
  68             return self.line_buffer.pop(0)
  69         while True:
  70             ready = select.select([self.socket], [], [], int(self.timeout / 2))
  71             if not ready[0]:
  72                 self._pingtest(send_ping)
  73                 return None
  74             self.last_pong = time.time()
  75             received_bytes = self.socket.recv(1024)
  76             try:
  77                 received_runes = received_bytes.decode("UTF-8")
  78             except UnicodeDecodeError:
  79                 received_runes = received_bytes.decode("latin1")
  80             if len(received_runes) == 0:
  81                 print("SOCKET CONNECTION BROKEN")
  82                 raise ExceptionForRestart
  83             self.rune_buffer += received_runes
  84             lines_split = str.split(self.rune_buffer, "\r\n")
  85             self.line_buffer += lines_split[:-1]
  86             self.rune_buffer = lines_split[-1]
  87             if len(self.line_buffer) > 0:
  88                 return self.line_buffer.pop(0)
  89
  90     def recv_line(self, send_ping=True):
  91         line = self._recv_line_wrapped(send_ping)
  92         if line:
  93             print("LINE FROM SERVER " + str(datetime.datetime.now()) + ": " +
  94                   line)
  95         return line
  96
  97
  98 def handle_command(command, argument, notice, target, session):
  99     hash_string = hashlib.md5(target.encode("utf-8")).hexdigest()
 100     quotesfile_name = session.dbdir + "/quotes_" + hash_string
 101
 102     def addquote():
 103         if not os.access(quotesfile_name, os.F_OK):
 104             quotesfile = open(quotesfile_name, "w")
 105             quotesfile.write("QUOTES FOR " + target + ":\n")
 106             quotesfile.close()
 107         quotesfile = open(quotesfile_name, "a")
 108         quotesfile.write(argument + "\n")
 109         quotesfile.close()
 110         quotesfile = open(quotesfile_name, "r")
 111         lines = quotesfile.readlines()
 112         quotesfile.close()
 113         notice("ADDED QUOTE #" + str(len(lines) - 1))
 114
 115     def quote():
 116
 117         def help():
 118             notice("SYNTAX: !quote [int] OR !quote search QUERY")
 119             notice("QUERY may be a boolean grouping of quoted or unquoted " +
 120                    "search terms, examples:")
 121             notice("!quote search foo")
 122             notice("!quote search foo AND (bar OR NOT baz)")
 123             notice("!quote search \"foo\\\"bar\" AND ('NOT\"' AND \"'foo'\"" +
 124                    " OR 'bar\\'baz')")
 125
 126         if "" == argument:
 127             tokens = []
 128         else:
 129             tokens = argument.split(" ")
 130         if (len(tokens) > 1 and tokens[0] != "search") or \
 131             (len(tokens) == 1 and
 132                 (tokens[0] == "search" or not tokens[0].isdigit())):
 133             help()
 134             return
 135         if not os.access(quotesfile_name, os.F_OK):
 136             notice("NO QUOTES AVAILABLE")
 137             return
 138         quotesfile = open(quotesfile_name, "r")
 139         lines = quotesfile.readlines()
 140         quotesfile.close()
 141         lines = lines[1:]
 142         if len(tokens) == 1:
 143             i = int(tokens[0])
 144             if i == 0 or i > len(lines):
 145                 notice("THERE'S NO QUOTE OF THAT INDEX")
 146                 return
 147             i = i - 1
 148         elif len(tokens) > 1:
 149             query = str.join(" ", tokens[1:])
 150             try:
 151                 results = plomsearch.search(query, lines)
 152             except plomsearch.LogicParserError as err:
 153                 notice("FAILED QUERY PARSING: " + str(err))
 154                 return
 155             if len(results) == 0:
 156                 notice("NO QUOTES MATCHING QUERY")
 157             else:
 158                 for result in results:
 159                     notice("QUOTE #" + str(result[0] + 1) + " : " + result[1])
 160             return
 161         else:
 162             i = random.randrange(len(lines))
 163         notice("QUOTE #" + str(i + 1) + ": " + lines[i])
 164
 165     def markov():
 166         from random import choice, shuffle
 167         select_length = 2
 168         selections = []
 169
 170         def markov(snippet):
 171             usable_selections = []
 172             for i in range(select_length, 0, -1):
 173                 for selection in selections:
 174                     add = True
 175                     for j in range(i):
 176                         j += 1
 177                         if snippet[-j] != selection[-(j+1)]:
 178                             add = False
 179                             break
 180                     if add:
 181                         usable_selections += [selection]
 182                 if [] != usable_selections:
 183                     break
 184             if [] == usable_selections:
 185                 usable_selections = selections
 186             selection = choice(usable_selections)
 187             return selection[select_length]
 188
 189         hash_string = hashlib.md5(target.encode("utf-8")).hexdigest()
 190         markovfeed_name = session.dbdir + "/markovfeed_" + hash_string
 191         if not os.access(markovfeed_name, os.F_OK):
 192             notice("NOT ENOUGH TEXT TO MARKOV.")
 193             return
 194
 195         # Lowercase incoming lines, ensure they end in a sentence end mark.
 196         file = open(markovfeed_name, "r")
 197         lines = file.readlines()
 198         file.close()
 199         tokens = []
 200         sentence_end_markers = ".!?)("
 201         for line in lines:
 202             line = line.lower().replace("\n", "")
 203             if line[-1] not in sentence_end_markers:
 204                 line += "."
 205             tokens += line.split()
 206         if len(tokens) <= select_length:
 207             notice("NOT ENOUGH TEXT TO MARKOV.")
 208             return
 209
 210         # Replace URLs with escape string for now, so that the Markov selector
 211         # won't see them as different strings. Stash replaced URLs in urls.
 212         urls = []
 213         url_escape = "\nURL"
 214         url_starts = ["http://", "https://", "<http://", "<https://"]
 215         for i in range(len(tokens)):
 216             for url_start in url_starts:
 217                 if tokens[i][:len(url_start)] == url_start:
 218                     length = len(tokens[i])
 219                     if url_start[0] == "<":
 220                         try:
 221                             length = tokens[i].index(">") + 1
 222                         except ValueError:
 223                             pass
 224                     urls += [tokens[i][:length]]
 225                     tokens[i] = url_escape + tokens[i][length:]
 226                     break
 227
 228         # For each snippet of select_length, use markov() to find continuation
 229         # token from selections. Replace present users' names with malkovich.
 230         # Start snippets with the beginning of a sentence, if possible.
 231         for i in range(len(tokens) - select_length):
 232             token_list = []
 233             for j in range(select_length + 1):
 234                 token_list += [tokens[i + j]]
 235             selections += [token_list]
 236         snippet = []
 237         for i in range(select_length):
 238             snippet += [""]
 239         shuffle(selections)
 240         for i in range(len(selections)):
 241             if selections[i][0][-1] in sentence_end_markers:
 242                 for i in range(select_length):
 243                     snippet[i] = selections[i][i + 1]
 244                 break
 245         msg = ""
 246         malkovich = "malkovich"
 247         while 1:
 248             new_end = markov(snippet)
 249             for name in session.users_in_chan:
 250                 if new_end[:len(name)] == name.lower():
 251                     new_end = malkovich + new_end[len(name):]
 252                     break
 253             if len(msg) + len(new_end) > 200:
 254                 break
 255             msg += new_end + " "
 256             for i in range(select_length - 1):
 257                 snippet[i] = snippet[i + 1]
 258             snippet[select_length - 1] = new_end
 259
 260         # Replace occurences of url escape string with random choice from urls.
 261         while True:
 262             index = msg.find(url_escape)
 263             if index < 0:
 264                 break
 265             msg = msg.replace(url_escape, choice(urls), 1)
 266
 267         # More meaningful ways to randomly end sentences.
 268         notice(msg + malkovich + ".")
 269
 270     def twt():
 271         def try_open(mode):
 272             try:
 273                 twtfile = open(session.twtfile, mode)
 274             except (PermissionError, FileNotFoundError) as err:
 275                 notice("CAN'T ACCESS OR CREATE TWT FILE: " + str(err))
 276                 return None
 277             return twtfile
 278
 279         from datetime import datetime
 280         if not os.access(session.twtfile, os.F_OK):
 281             twtfile = try_open("w")
 282             if None == twtfile:
 283                 return
 284             twtfile.close()
 285         twtfile = try_open("a")
 286         if None == twtfile:
 287             return
 288         twtfile.write(datetime.utcnow().isoformat() + "\t" + argument + "\n")
 289         twtfile.close()
 290         notice("WROTE TWT.")
 291
 292     if "addquote" == command:
 293         addquote()
 294     elif "quote" == command:
 295         quote()
 296     elif "markov" == command:
 297         markov()
 298     elif "twt" == command:
 299         twt()
 300
 301
 302 def handle_url(url, notice, show_url=False):
 303
 304     def mobile_twitter_hack(url):
 305         re1 = 'https?://(mobile.twitter.com/)[^/]+(/status/)'
 306         re2 = 'https?://mobile.twitter.com/([^/]+)/status/([^\?/]+)'
 307         m = re.search(re1, url)
 308         if m and m.group(1) == 'mobile.twitter.com/' \
 309                 and m.group(2) == '/status/':
 310             m = re.search(re2, url)
 311             url = 'https://twitter.com/' + m.group(1) + '/status/' + m.group(2)
 312             handle_url(url, notice, True)
 313             return True
 314
 315     try:
 316         r = requests.get(url, timeout=15)
 317     except (requests.exceptions.TooManyRedirects,
 318             requests.exceptions.ConnectionError,
 319             requests.exceptions.InvalidURL,
 320             UnicodeError,
 321             requests.exceptions.InvalidSchema) as error:
 322         notice("TROUBLE FOLLOWING URL: " + str(error))
 323         return
 324     if mobile_twitter_hack(url):
 325         return
 326     title = bs4.BeautifulSoup(r.text, "html5lib").title
 327     if title and title.string:
 328         prefix = "PAGE TITLE: "
 329         if show_url:
 330             prefix = "PAGE TITLE FOR <" + url + ">: "
 331         notice(prefix + title.string.strip())
 332     else:
 333         notice("PAGE HAS NO TITLE TAG")
 334
 335
 336 class Session:
 337
 338     def __init__(self, io, username, nickname, channel, twtfile, dbdir):
 339         self.io = io
 340         self.nickname = nickname
 341         self.channel = channel
 342         self.users_in_chan = []
 343         self.twtfile = twtfile
 344         self.dbdir = dbdir
 345         self.io.send_line("NICK " + self.nickname)
 346         self.io.send_line("USER " + username + " 0 * : ")
 347         self.io.send_line("JOIN " + self.channel)
 348         hash_string = hashlib.md5(self.channel.encode("utf-8")).hexdigest()
 349         self.logdir = self.dbdir + "/irclogs_" + hash_string + "/"
 350         if not os.path.exists(self.logdir):
 351             os.makedirs(self.logdir)
 352
 353     def loop(self):
 354
 355         def log(line):
 356             now = datetime.datetime.utcnow()
 357             logfile = open(self.logdir + now.strftime("%Y-%m-%d") + ".txt", "a")
 358             form = "%Y-%m-%d %H:%M:%S UTC\t"
 359             logfile.write(now.strftime(form) + " " + line + "\n")
 360             logfile.close()
 361
 362         def handle_privmsg(tokens):
 363
 364             def handle_input(msg, target):
 365
 366                 def notice(msg):
 367                     self.io.send_line("NOTICE " + target + " :" + msg)
 368
 369                 matches = re.findall("(https?://[^\s>]+)", msg)
 370                 for i in range(len(matches)):
 371                     handle_url(matches[i], notice)
 372                 if "!" == msg[0]:
 373                     tokens = msg[1:].split()
 374                     argument = str.join(" ", tokens[1:])
 375                     handle_command(tokens[0], argument, notice, target, self)
 376                     return
 377                 hash_string = hashlib.md5(target.encode("utf-8")).hexdigest()
 378                 markovfeed_name = self.dbdir + "/markovfeed_" + hash_string
 379                 file = open(markovfeed_name, "a")
 380                 file.write(msg + "\n")
 381                 file.close()
 382
 383             sender = ""
 384             for rune in tokens[0]:
 385                 if rune == "!":
 386                     break
 387                 if rune != ":":
 388                     sender += rune
 389             receiver = ""
 390             for rune in tokens[2]:
 391                 if rune == "!":
 392                     break
 393                 if rune != ":":
 394                     receiver += rune
 395             target = sender
 396             if receiver != self.nickname:
 397                 target = receiver
 398             msg = str.join(" ", tokens[3:])[1:]
 399             if target == self.channel:
 400                 log("<" + sender + "> " + msg)
 401             handle_input(msg, target)
 402
 403         def name_from_join_or_part(tokens):
 404             token = tokens[0][1:]
 405             index_cut = token.find("@")
 406             index_ex = token.find("!")
 407             if index_ex > 0 and index_ex < index_cut:
 408                 index_cut = index_ex
 409             return token[:index_cut]
 410
 411         while True:
 412             line = self.io.recv_line()
 413             if not line:
 414                 continue
 415             tokens = line.split(" ")
 416             if len(tokens) > 1:
 417                 if tokens[0] == "PING":
 418                     self.io.send_line("PONG " + tokens[1])
 419                 elif tokens[1] == "PRIVMSG":
 420                     handle_privmsg(tokens)
 421                 elif tokens[1] == "353":
 422                     names = tokens[5:]
 423                     names[0] = names[0][1:]
 424                     log("PRESENT: " + str.join(", ", names))
 425                     for i in range(len(names)):
 426                         names[i] = names[i].replace("@", "").replace("+", "")
 427                     self.users_in_chan += names
 428                 elif tokens[1] == "JOIN":
 429                     name = name_from_join_or_part(tokens)
 430                     if name != self.nickname:
 431                         self.users_in_chan += [name]
 432                     log(line)
 433                 elif tokens[1] == "PART":
 434                     name = name_from_join_or_part(tokens)
 435                     del(self.users_in_chan[self.users_in_chan.index(name)])
 436                     log(line)
 437                 else:
 438                     log(line)
 439
 440
 441 def parse_command_line_arguments():
 442     parser = argparse.ArgumentParser()
 443     parser.add_argument("-s, --server", action="store", dest="server",
 444                         default=SERVER,
 445                         help="server or server net to connect to (default: "
 446                         + SERVER + ")")
 447     parser.add_argument("-p, --port", action="store", dest="port", type=int,
 448                         default=PORT, help="port to connect to (default : "
 449                         + str(PORT) + ")")
 450     parser.add_argument("-w, --wait", action="store", dest="timeout",
 451                         type=int, default=TIMEOUT,
 452                         help="timeout in seconds after which to attempt " +
 453                         "reconnect (default: " + str(TIMEOUT) + ")")
 454     parser.add_argument("-u, --username", action="store", dest="username",
 455                         default=USERNAME, help="username to use (default: "
 456                         + USERNAME + ")")
 457     parser.add_argument("-n, --nickname", action="store", dest="nickname",
 458                         default=NICKNAME, help="nickname to use (default: "
 459                         + NICKNAME + ")")
 460     parser.add_argument("-t, --twtxtfile", action="store", dest="twtfile",
 461                         default=TWTFILE, help="twtxt file to use (default: "
 462                         + TWTFILE + ")")
 463     parser.add_argument("-d, --dbdir", action="store", dest="dbdir",
 464                         default=DBDIR, help="directory to store DB files in")
 465     parser.add_argument("CHANNEL", action="store", help="channel to join")
 466     opts, unknown = parser.parse_known_args()
 467     return opts
 468
 469
 470 opts = parse_command_line_arguments()
 471 while True:
 472     try:
 473         io = IO(opts.server, opts.port, opts.timeout)
 474         session = Session(io, opts.username, opts.nickname, opts.CHANNEL,
 475             opts.twtfile, opts.dbdir)
 476         session.loop()
 477     except ExceptionForRestart:
 478         io.socket.close()
 479         continue