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