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