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