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 signal
  15 import plomsearch
  16 import irclog
  17
  18 # Defaults, may be overwritten by command line arguments.
  19 SERVER = "irc.freenode.net"
  20 PORT = 6667
  21 TIMEOUT = 240
  22 USERNAME = "plomlombot"
  23 NICKNAME = USERNAME
  24 TWTFILE = ""
  25 DBDIR = os.path.expanduser("~/plomlombot_db")
  26
  27
  28 def write_to_file(path, mode, text):
  29     f = open(path, mode)
  30     f.write(text)
  31     f.close()
  32
  33
  34 class ExceptionForRestart(Exception):
  35     pass
  36
  37
  38 class Line:
  39
  40     def __init__(self, line):
  41         self.line = line
  42         self.tokens = line.split(" ")
  43         self.sender = ""
  44         if self.tokens[0][0] == ":":
  45             for rune in self.tokens[0][1:]:
  46                 if rune in {"!", "@"}:
  47                     break
  48                 self.sender += rune
  49         self.receiver = ""
  50         if len(self.tokens) > 2:
  51             for rune in self.tokens[2]:
  52                 if rune in {"!", "@"}:
  53                     break
  54                 if rune != ":":
  55                     self.receiver += rune
  56
  57
  58 class Log:
  59
  60     def __init__(self, chandir, nickname, username, channel, rmlogs):
  61         self.nickname = nickname
  62         self.username = username
  63         self.channel = channel
  64         self.chandir = chandir
  65         self.rmlogcycle = rmlogs
  66         self.rawlogdir = chandir + "raw_logs/"
  67         self.logdir = chandir + "logs/"
  68         if not os.path.exists(self.logdir):
  69             os.makedirs(self.logdir)
  70         if not os.path.exists(self.rawlogdir):
  71             os.makedirs(self.rawlogdir)
  72
  73     def log(self, line, sent=False):
  74         identity = ""
  75         separator = " > "
  76         if sent:
  77             separator = " "
  78             line = Line("< " + line)
  79             line.sender = self.nickname
  80             identity = self.username + "@localhost"
  81         else:
  82             if type(line) == str:
  83                 line = Line(line)
  84         now = datetime.datetime.utcnow()
  85         form = "%Y-%m-%d %H:%M:%S UTC"
  86         write_to_file(self.rawlogdir + now.strftime("%Y-%m-%d") + ".txt", "a",
  87                       now.strftime(form) + separator + line.line + "\n")
  88         to_log = irclog.format_logline(line, self.channel, identity)
  89         if to_log != None:
  90             write_to_file(self.logdir + now.strftime("%Y-%m-%d") + ".txt", "a",
  91                           now.strftime(form) + " " + to_log + "\n")
  92
  93     def rmlogs(self):
  94         if self.rmlogcycle > 0:
  95             for f in os.listdir(self.logdir):
  96                 f = os.path.join(self.logdir, f)
  97                 if os.path.isfile(f) and \
  98                         os.stat(f).st_mtime < time.time() - self.rmlogcycle:
  99                     os.remove(f)
 100
 101     def separator_line(self):
 102         now = datetime.datetime.utcnow()
 103         write_to_file(self.logdir + now.strftime("%Y-%m-%d") + ".txt", "a",
 104                       "-----------------------\n")
 105
 106
 107 class IO:
 108
 109     def __init__(self, server, port, timeout):
 110         self.log = None
 111         self.timeout = timeout
 112         self.socket = socket.socket()
 113         try:
 114             self.socket.connect((server, port))
 115         except TimeoutError:
 116             raise ExceptionForRestart
 117         self.socket.setblocking(0)
 118         self.line_buffer = []
 119         self.rune_buffer = ""
 120         self.last_pong = time.time()
 121         line = self.recv_line(send_ping=False)
 122         if not line or len(line) < 1:
 123             raise ExceptionForRestart
 124         self.servername = line.split(" ")[0][1:]
 125
 126     def _pingtest(self, send_ping=True):
 127         if self.last_pong + self.timeout < time.time():
 128             print("SERVER NOT ANSWERING")
 129             raise ExceptionForRestart
 130         if send_ping:
 131             self.send_line("PING " + self.servername)
 132
 133     def send_line(self, msg):
 134         msg = msg.replace("\r", " ")
 135         msg = msg.replace("\n", " ")
 136         if len(msg.encode("utf-8")) > 510:
 137             print("NOT SENT LINE TO SERVER (too long): " + msg)
 138         print("LINE TO SERVER: "
 139               + str(datetime.datetime.now()) + ": " + msg)
 140         if self.log != None:
 141             self.log.log(msg, True)
 142         msg = msg + "\r\n"
 143         msg_len = len(msg)
 144         total_sent_len = 0
 145         while total_sent_len < msg_len:
 146             sent_len = self.socket.send(bytes(msg[total_sent_len:], "UTF-8"))
 147             if sent_len == 0:
 148                 print("SOCKET CONNECTION BROKEN")
 149                 raise ExceptionForRestart
 150             total_sent_len += sent_len
 151
 152     def _recv_line_wrapped(self, send_ping=True):
 153         if len(self.line_buffer) > 0:
 154             return self.line_buffer.pop(0)
 155         while True:
 156             ready = select.select([self.socket], [], [], int(self.timeout / 2))
 157             if not ready[0]:
 158                 self._pingtest(send_ping)
 159                 return None
 160             self.last_pong = time.time()
 161             received_bytes = self.socket.recv(1024)
 162             try:
 163                 received_runes = received_bytes.decode("UTF-8")
 164             except UnicodeDecodeError:
 165                 received_runes = received_bytes.decode("latin1")
 166             if len(received_runes) == 0:
 167                 print("SOCKET CONNECTION BROKEN")
 168                 raise ExceptionForRestart
 169             self.rune_buffer += received_runes
 170             lines_split = str.split(self.rune_buffer, "\r\n")
 171             self.line_buffer += lines_split[:-1]
 172             self.rune_buffer = lines_split[-1]
 173             if len(self.line_buffer) > 0:
 174                 return self.line_buffer.pop(0)
 175
 176     def recv_line(self, send_ping=True):
 177         line = self._recv_line_wrapped(send_ping)
 178         if line:
 179             if self.log != None:
 180                 self.log.log(line)
 181             print("LINE FROM SERVER " + str(datetime.datetime.now()) + ": " +
 182                   line)
 183         return line
 184
 185
 186 def handle_command(command, argument, notice, target, session):
 187
 188     def addquote():
 189         if not os.access(session.quotesfile, os.F_OK):
 190             write_to_file(session.quotesfile, "w",
 191                           "QUOTES FOR " + target + ":\n")
 192         write_to_file(session.quotesfile, "a", argument + "\n")
 193         quotesfile = open(session.quotesfile, "r")
 194         lines = quotesfile.readlines()
 195         quotesfile.close()
 196         notice("added quote #" + str(len(lines) - 1))
 197
 198     def quote():
 199
 200         def help():
 201             notice("syntax: !quote [int] OR !quote search QUERY "
 202                    "OR !quote offset-search [int] QUERY")
 203             notice("QUERY may be a boolean grouping of quoted or unquoted " +
 204                    "search terms, examples:")
 205             notice("!quote search foo")
 206             notice("!quote search foo AND (bar OR NOT baz)")
 207             notice("!quote search \"foo\\\"bar\" AND ('NOT\"' AND \"'foo'\"" +
 208                    " OR 'bar\\'baz')")
 209             notice("The offset-search int argument defines how many matches "
 210                    "to skip (useful if results are above maximum number to "
 211                    "display).")
 212
 213         if "" == argument:
 214             tokens = []
 215         else:
 216             tokens = argument.split(" ")
 217         if (len(tokens) == 1 and not tokens[0].isdigit()) or \
 218            (tokens > 1 and
 219             tokens[0] not in {"search", "offset-search"} or
 220             (tokens[0] == "offset-search" and
 221              ((not len(tokens) > 2) or (not tokens[1].isdigit())))):
 222             help()
 223            return
 224         if not os.access(session.quotesfile, os.F_OK):
 225             notice("no quotes available")
 226             return
 227         quotesfile = open(session.quotesfile, "r")
 228         lines = quotesfile.readlines()
 229         quotesfile.close()
 230         lines = lines[1:]
 231         if len(tokens) == 1:
 232             i = int(tokens[0])
 233             if i == 0 or i > len(lines):
 234                 notice("there's no quote of that index")
 235                 return
 236             i = i - 1
 237         elif len(tokens) > 1:
 238             to_skip = 0
 239             if tokens[0] == "search":
 240                 query = str.join(" ", tokens[1:])
 241             elif tokens[0] == "offset-search":
 242                 to_skip = int(tokens[1])
 243                 query = str.join(" ", tokens[2:])
 244             try:
 245                 results = plomsearch.search(query, lines)
 246             except plomsearch.LogicParserError as err:
 247                 notice("failed query parsing: " + str(err))
 248                 return
 249             if len(results) == 0:
 250                 notice("no quotes matching query")
 251             else:
 252                 if to_skip >= len(results):
 253                     notice("skipped all quotes matching query")
 254                 else:
 255                     notice("found %s matches, showing max. 3, skipping %s"
 256                            % (len(results), to_skip))
 257                 for i in range(len(results)):
 258                     if i >= to_skip and i < to_skip + 3:
 259                         result = results[i]
 260                         notice("quote #" + str(result[0] + 1) + ": "
 261                                + result[1][:-1])
 262             return
 263         else:
 264             i = random.randrange(len(lines))
 265         notice("quote #" + str(i + 1) + ": " + lines[i][:-1])
 266
 267     def markov():
 268
 269         def help():
 270             notice("syntax: !markov [integer from 1 to infinite]")
 271
 272         def markov(snippet):
 273             usable_selections = []
 274             for i in range(select_length, 0, -1):
 275                 for selection in selections:
 276                     add = True
 277                     for j in range(i):
 278                         j += 1
 279                         if snippet[-j] != selection[-(j+1)]:
 280                             add = False
 281                             break
 282                     if add:
 283                         usable_selections += [selection]
 284                 if [] != usable_selections:
 285                     break
 286             if [] == usable_selections:
 287                 usable_selections = selections
 288             selection = choice(usable_selections)
 289             return selection[select_length]
 290
 291         if "" == argument:
 292             tokens = []
 293         else:
 294             tokens = argument.split(" ")
 295         if (len(tokens) > 1 or (len(tokens) == 1 and not tokens[0].isdigit())):
 296             help()
 297             return
 298
 299         from random import choice, shuffle
 300         select_length = 2
 301         if len(tokens) == 1:
 302             n = int(tokens[0])
 303             if n > 0:
 304                 select_length = n
 305             else:
 306                 notice("bad value, using default: " + str(select_length))
 307         selections = []
 308
 309         if not os.access(session.markovfile, os.F_OK):
 310             notice("not enough text to markov for selection length")
 311             return
 312
 313         # Lowercase incoming lines, ensure they end in a sentence end mark.
 314         file = open(session.markovfile, "r")
 315         lines = file.readlines()
 316         file.close()
 317         tokens = []
 318         sentence_end_markers = ".!?)("
 319         for line in lines:
 320             line = line.lower().replace("\n", "")
 321             if line[-1] not in sentence_end_markers:
 322                 line += "."
 323             tokens += line.split()
 324         if len(tokens) - 1 <= select_length:
 325             notice("not enough text to markov")
 326             return
 327
 328         # Replace URLs with escape string for now, so that the Markov selector
 329         # won't see them as different strings. Stash replaced URLs in urls.
 330         urls = []
 331         url_escape = "\nURL"
 332         url_starts = ["http://", "https://", "<http://", "<https://"]
 333         for i in range(len(tokens)):
 334             for url_start in url_starts:
 335                 if tokens[i][:len(url_start)] == url_start:
 336                     length = len(tokens[i])
 337                     if url_start[0] == "<":
 338                         try:
 339                             length = tokens[i].index(">") + 1
 340                         except ValueError:
 341                             pass
 342                     urls += [tokens[i][:length]]
 343                     tokens[i] = url_escape + tokens[i][length:]
 344                     break
 345
 346         # For each snippet of select_length, use markov() to find continuation
 347         # token from selections. Replace present users' names with malkovich.
 348         # Start snippets with the beginning of a sentence, if possible.
 349         for i in range(len(tokens) - select_length):
 350             token_list = []
 351             for j in range(select_length + 1):
 352                 token_list += [tokens[i + j]]
 353             selections += [token_list]
 354         snippet = []
 355         for i in range(select_length):
 356             snippet += [""]
 357         shuffle(selections)
 358         for i in range(len(selections)):
 359             if selections[i][0][-1] in sentence_end_markers:
 360                 for j in range(select_length):
 361                     snippet[j] = selections[j][j + 1]
 362                 break
 363         msg = ""
 364         malkovich = "malkovich"
 365         while 1:
 366             new_end = markov(snippet)
 367             for name in session.users_in_chan:
 368                 if new_end[:len(name)] == name.lower():
 369                     new_end = malkovich + new_end[len(name):]
 370                     break
 371             if len(msg) + len(new_end) > 200:
 372                 break
 373             msg += new_end + " "
 374             for i in range(select_length - 1):
 375                 snippet[i] = snippet[i + 1]
 376             snippet[select_length - 1] = new_end
 377
 378         # Replace occurences of url escape string with random choice from urls.
 379         while True:
 380             index = msg.find(url_escape)
 381             if index < 0:
 382                 break
 383             msg = msg.replace(url_escape, choice(urls), 1)
 384
 385         # More meaningful ways to randomly end sentences.
 386         notice(msg + malkovich + ".")
 387
 388     def twt():
 389         def try_open(mode):
 390             try:
 391                 twtfile = open(session.twtfile, mode)
 392             except (PermissionError, FileNotFoundError) as err:
 393                 notice("can't access or create twt file: " + str(err))
 394                 return None
 395             return twtfile
 396
 397         from datetime import datetime
 398         if not os.access(session.twtfile, os.F_OK):
 399             twtfile = try_open("w")
 400             if None == twtfile:
 401                 return
 402             twtfile.close()
 403         twtfile = try_open("a")
 404         if None == twtfile:
 405             return
 406         twtfile.write(datetime.utcnow().isoformat() + "\t" + argument + "\n")
 407         twtfile.close()
 408         notice("wrote twt.")
 409
 410     if "addquote" == command:
 411         addquote()
 412     elif "quote" == command:
 413         quote()
 414     elif "markov" == command:
 415         markov()
 416     elif "twt" == command:
 417         twt()
 418
 419
 420 def handle_url(url, notice, show_url=False):
 421
 422     def mobile_twitter_hack(url):
 423         re1 = 'https?://(mobile.twitter.com/)[^/]+(/status/)'
 424         re2 = 'https?://mobile.twitter.com/([^/]+)/status/([^\?/]+)'
 425         m = re.search(re1, url)
 426         if m and m.group(1) == 'mobile.twitter.com/' \
 427                 and m.group(2) == '/status/':
 428             m = re.search(re2, url)
 429             url = 'https://twitter.com/' + m.group(1) + '/status/' + m.group(2)
 430             handle_url(url, notice, True)
 431             return True
 432
 433     class TimeOut(Exception):
 434         pass
 435
 436     def timeout_handler(ignore1, ignore2):
 437         raise TimeOut("timeout")
 438
 439     signal.signal(signal.SIGALRM, timeout_handler)
 440     signal.alarm(15)
 441     try:
 442         r = requests.get(url, headers = {'User-Agent': 'plomlombot'}, stream=True)
 443         r.raw.decode_content = True
 444         text = r.raw.read(10000000+1)
 445         if len(text) > 10000000:
 446             raise ValueError('Too large a response')
 447     except (requests.exceptions.TooManyRedirects,
 448             requests.exceptions.ConnectionError,
 449             requests.exceptions.InvalidURL,
 450             TimeOut,
 451             UnicodeError,
 452             ValueError,
 453             requests.exceptions.InvalidSchema) as error:
 454         signal.alarm(0)
 455         notice("trouble following url: " + str(error))
 456         return False
 457     signal.alarm(0)
 458     if mobile_twitter_hack(url):
 459         return True
 460     title = bs4.BeautifulSoup(text, "html5lib").title
 461     if title and title.string:
 462         prefix = "page title: "
 463         if show_url:
 464             prefix = "page title for <" + url + ">: "
 465         notice(prefix + title.string.strip())
 466     else:
 467         notice("page has no title tag")
 468     return True
 469
 470
 471 class Session:
 472
 473     def __init__(self, io, username, nickname, channel, twtfile, dbdir, rmlogs,
 474                  markov_input):
 475         self.io = io
 476         self.nickname = nickname
 477         self.users_in_chan = []
 478         self.twtfile = twtfile
 479         hash_channel = hashlib.md5(channel.encode("utf-8")).hexdigest()
 480         chandir = dbdir + "/" + hash_channel + "/"
 481         self.markov_input = markov_input
 482         self.markovfile = chandir + "markovfeed"
 483         self.quotesfile = chandir + "quotes"
 484         self.log = Log(chandir, self.nickname, username, channel, rmlogs)
 485         self.io.send_line("NICK " + self.nickname)
 486         self.io.send_line("USER " + username + " 0 * : ")
 487         self.io.send_line("JOIN " + channel)
 488         self.io.log = self.log
 489         self.log.separator_line()
 490
 491     def loop(self):
 492
 493         def handle_privmsg(line):
 494
 495             def notice(msg):
 496                 line = "NOTICE " + target + " :" + msg
 497                 self.io.send_line(line)
 498
 499             target = line.sender
 500             if line.receiver != self.nickname:
 501                 target = line.receiver
 502             msg = str.join(" ", line.tokens[3:])[1:]
 503             matches = re.findall("(https?://[^\s>]+)", msg)
 504             url_count = 0
 505             for i in range(len(matches)):
 506                 if handle_url(matches[i], notice):
 507                     url_count += 1
 508                     if url_count == 3:
 509                         notice("maximum number of urls to parse per message "
 510                                "reached")
 511                         break
 512             if "!" == msg[0] and len(msg) > 1:
 513                 tokens = msg[1:].split()
 514                 argument = str.join(" ", tokens[1:])
 515                 handle_command(tokens[0], argument, notice, target, self)
 516                 return
 517             if self.markov_input:
 518                 write_to_file(self.markovfile, "a", msg + "\n")
 519
 520         while True:
 521             self.log.rmlogs()
 522             line = self.io.recv_line()
 523             if not line:
 524                 continue
 525             line = Line(line)
 526             if len(line.tokens) > 1:
 527                 if line.tokens[0] == "PING":
 528                     self.io.send_line("PONG " + line.tokens[1])
 529                 elif line.tokens[1] == "PRIVMSG":
 530                     handle_privmsg(line)
 531                 elif line.tokens[1] == "353":
 532                     names = line.tokens[5:]
 533                     names[0] = names[0][1:]
 534                     for i in range(len(names)):
 535                         names[i] = names[i].replace("@", "").replace("+", "")
 536                     self.users_in_chan += names
 537                 elif line.tokens[1] == "JOIN" and line.sender != self.nickname:
 538                     self.users_in_chan += [line.sender]
 539                 elif line.tokens[1] == "PART":
 540                     del(self.users_in_chan[self.users_in_chan.index(line.sender)])
 541                 elif line.tokens[1] == "NICK":
 542                     del(self.users_in_chan[self.users_in_chan.index(line.sender)])
 543                     self.users_in_chan += [line.receiver]
 544
 545
 546 def parse_command_line_arguments():
 547     parser = argparse.ArgumentParser()
 548     parser.add_argument("-s, --server", action="store", dest="server",
 549                         default=SERVER,
 550                         help="server or server net to connect to (default: "
 551                         + SERVER + ")")
 552     parser.add_argument("-p, --port", action="store", dest="port", type=int,
 553                         default=PORT, help="port to connect to (default : "
 554                         + str(PORT) + ")")
 555     parser.add_argument("-w, --wait", action="store", dest="timeout",
 556                         type=int, default=TIMEOUT,
 557                         help="timeout in seconds after which to attempt "
 558                         "reconnect (default: " + str(TIMEOUT) + ")")
 559     parser.add_argument("-u, --username", action="store", dest="username",
 560                         default=USERNAME, help="username to use (default: "
 561                         + USERNAME + ")")
 562     parser.add_argument("-n, --nickname", action="store", dest="nickname",
 563                         default=NICKNAME, help="nickname to use (default: "
 564                         + NICKNAME + ")")
 565     parser.add_argument("-t, --twtxtfile", action="store", dest="twtfile",
 566                         default=TWTFILE, help="twtxt file to use (default: "
 567                         + TWTFILE + ")")
 568     parser.add_argument("-d, --dbdir", action="store", dest="dbdir",
 569                         default=DBDIR, help="directory to store DB files in")
 570     parser.add_argument("-r, --rmlogs", action="store", dest="rmlogs",
 571                         type=int, default=0,
 572                         help="maximum age in seconds for logfiles in logs/ "
 573                         "(0 means: never delete, and is default)")
 574     parser.add_argument("-m, --markov_store", action="store_true",
 575                         dest="markov_store",
 576                         help="log channel discussions for !markov input")
 577     parser.add_argument("CHANNEL", action="store", help="channel to join")
 578     opts, unknown = parser.parse_known_args()
 579     return opts
 580
 581
 582 opts = parse_command_line_arguments()
 583 while True:
 584     try:
 585         io = IO(opts.server, opts.port, opts.timeout)
 586         hash_server = hashlib.md5(opts.server.encode("utf-8")).hexdigest()
 587         dbdir = opts.dbdir + "/" + hash_server
 588         session = Session(io, opts.username, opts.nickname, opts.CHANNEL,
 589                           opts.twtfile, dbdir, opts.rmlogs, opts.markov_store)
 590         session.loop()
 591     except ExceptionForRestart:
 592         io.socket.close()
 593         continue