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 CHANNEL = "#plomlombot-test"
  21 PORT = 6667
  22 TIMEOUT = 240
  23 USERNAME = "plomlombot"
  24 NICKNAME = USERNAME
  25 TWTFILE = ""
  26 DBDIR = os.path.expanduser("~/plomlombot_db")
  27
  28
  29 def write_to_file(path, mode, text):
  30     f = open(path, mode)
  31     f.write(text)
  32     f.close()
  33
  34
  35 class ExceptionForRestart(Exception):
  36     pass
  37
  38
  39 class Line:
  40
  41     def __init__(self, line):
  42         self.line = line
  43         self.tokens = line.split(" ")
  44         self.sender = ""
  45         if self.tokens[0][0] == ":":
  46             for rune in self.tokens[0][1:]:
  47                 if rune in {"!", "@"}:
  48                     break
  49                 self.sender += rune
  50         self.receiver = ""
  51         if len(self.tokens) > 2:
  52             for rune in self.tokens[2]:
  53                 if rune in {"!", "@"}:
  54                     break
  55                 if rune != ":":
  56                     self.receiver += rune
  57
  58
  59 class Log:
  60
  61     def __init__(self, chandir, nickname, username, channel, rmlogs):
  62         self.nickname = nickname
  63         self.username = username
  64         self.channel = channel
  65         self.chandir = chandir
  66         self.rmlogcycle = rmlogs
  67         self.rawlogdir = chandir + "raw_logs/"
  68         self.logdir = chandir + "logs/"
  69         if not os.path.exists(self.logdir):
  70             os.makedirs(self.logdir)
  71         if not os.path.exists(self.rawlogdir):
  72             os.makedirs(self.rawlogdir)
  73
  74     def log(self, line, sent=False):
  75         identity = ""
  76         separator = " > "
  77         if sent:
  78             separator = " "
  79             line = Line("< " + line)
  80             line.sender = self.nickname
  81             identity = self.username + "@localhost"
  82         else:
  83             if type(line) == str:
  84                 line = Line(line)
  85         now = datetime.datetime.utcnow()
  86         form = "%Y-%m-%d %H:%M:%S UTC"
  87         write_to_file(self.rawlogdir + now.strftime("%Y-%m-%d") + ".txt", "a",
  88                       now.strftime(form) + separator + line.line + "\n")
  89         to_log = irclog.format_logline(line, self.channel, identity)
  90         if to_log != None:
  91             write_to_file(self.logdir + now.strftime("%Y-%m-%d") + ".txt", "a",
  92                           now.strftime(form) + " " + to_log + "\n")
  93
  94     def rmlogs(self):
  95         if self.rmlogcycle > 0:
  96             for f in os.listdir(self.logdir):
  97                 f = os.path.join(self.logdir, f)
  98                 if os.path.isfile(f) and \
  99                         os.stat(f).st_mtime < time.time() - self.rmlogcycle:
 100                     os.remove(f)
 101
 102     def separator_line(self):
 103         now = datetime.datetime.utcnow()
 104         write_to_file(self.logdir + now.strftime("%Y-%m-%d") + ".txt", "a",
 105                       "-----------------------\n")
 106
 107
 108 class IO:
 109
 110     def __init__(self, server, port, timeout):
 111         self.log = None
 112         self.timeout = timeout
 113         self.socket = socket.socket()
 114         try:
 115             self.socket.connect((server, port))
 116         except TimeoutError:
 117             raise ExceptionForRestart
 118         self.socket.setblocking(0)
 119         self.line_buffer = []
 120         self.rune_buffer = ""
 121         self.last_pong = time.time()
 122         line = self.recv_line(send_ping=False)
 123         if not line or len(line) < 1:
 124             raise ExceptionForRestart
 125         self.servername = line.split(" ")[0][1:]
 126
 127     def _pingtest(self, send_ping=True):
 128         if self.last_pong + self.timeout < time.time():
 129             print("SERVER NOT ANSWERING")
 130             raise ExceptionForRestart
 131         if send_ping:
 132             self.send_line("PING " + self.servername)
 133
 134     def send_line(self, msg):
 135         msg = msg.replace("\r", " ")
 136         msg = msg.replace("\n", " ")
 137         if len(msg.encode("utf-8")) > 510:
 138             print("NOT SENT LINE TO SERVER (too long): " + msg)
 139         print("LINE TO SERVER: "
 140               + str(datetime.datetime.now()) + ": " + msg)
 141         if self.log != None:
 142             self.log.log(msg, True)
 143         msg = msg + "\r\n"
 144         msg_len = len(msg)
 145         total_sent_len = 0
 146         while total_sent_len < msg_len:
 147             sent_len = self.socket.send(bytes(msg[total_sent_len:], "UTF-8"))
 148             if sent_len == 0:
 149                 print("SOCKET CONNECTION BROKEN")
 150                 raise ExceptionForRestart
 151             total_sent_len += sent_len
 152
 153     def _recv_line_wrapped(self, send_ping=True):
 154         if len(self.line_buffer) > 0:
 155             return self.line_buffer.pop(0)
 156         while True:
 157             ready = select.select([self.socket], [], [], int(self.timeout / 2))
 158             if not ready[0]:
 159                 self._pingtest(send_ping)
 160                 return None
 161             self.last_pong = time.time()
 162             received_bytes = self.socket.recv(1024)
 163             try:
 164                 received_runes = received_bytes.decode("UTF-8")
 165             except UnicodeDecodeError:
 166                 received_runes = received_bytes.decode("latin1")
 167             if len(received_runes) == 0:
 168                 print("SOCKET CONNECTION BROKEN")
 169                 raise ExceptionForRestart
 170             self.rune_buffer += received_runes
 171             lines_split = str.split(self.rune_buffer, "\r\n")
 172             self.line_buffer += lines_split[:-1]
 173             self.rune_buffer = lines_split[-1]
 174             if len(self.line_buffer) > 0:
 175                 return self.line_buffer.pop(0)
 176
 177     def recv_line(self, send_ping=True):
 178         line = self._recv_line_wrapped(send_ping)
 179         if line:
 180             if self.log != None:
 181                 self.log.log(line)
 182             print("LINE FROM SERVER " + str(datetime.datetime.now()) + ": " +
 183                   line)
 184         return line
 185
 186
 187 def handle_command(command, argument, notice, target, session):
 188
 189     def addquote():
 190         if not os.access(session.quotesfile, os.F_OK):
 191             write_to_file(session.quotesfile, "w",
 192                           "QUOTES FOR " + target + ":\n")
 193         write_to_file(session.quotesfile, "a", argument + "\n")
 194         quotesfile = open(session.quotesfile, "r")
 195         lines = quotesfile.readlines()
 196         quotesfile.close()
 197         notice("added quote #" + str(len(lines) - 1))
 198
 199     def quote():
 200
 201         def help():
 202             notice("syntax: !quote [int] OR !quote search QUERY "
 203                    "OR !quote offset-search [int] QUERY")
 204             notice("QUERY may be a boolean grouping of quoted or unquoted " +
 205                    "search terms, examples:")
 206             notice("!quote search foo")
 207             notice("!quote search foo AND (bar OR NOT baz)")
 208             notice("!quote search \"foo\\\"bar\" AND ('NOT\"' AND \"'foo'\"" +
 209                    " OR 'bar\\'baz')")
 210             notice("The offset-search int argument defines how many matches "
 211                    "to skip (useful if results are above maximum number to "
 212                    "display).")
 213
 214         if "" == argument:
 215             tokens = []
 216         else:
 217             tokens = argument.split(" ")
 218         if (len(tokens) == 1 and not tokens[0].isdigit()) or \
 219            (len(tokens) > 1 and
 220             (tokens[0] not in {"search", "offset-search"} or
 221             (tokens[0] == "offset-search" and
 222              ((not len(tokens) > 2) or (not tokens[1].isdigit()))))):
 223             help()
 224             return
 225         if not os.access(session.quotesfile, os.F_OK):
 226             notice("no quotes available")
 227             return
 228         quotesfile = open(session.quotesfile, "r")
 229         lines = quotesfile.readlines()
 230         quotesfile.close()
 231         lines = lines[1:]
 232         if len(tokens) == 1:
 233             i = int(tokens[0])
 234             if i == 0 or i > len(lines):
 235                 notice("there's no quote of that index")
 236                 return
 237             i = i - 1
 238         elif len(tokens) > 1:
 239             to_skip = 0
 240             if tokens[0] == "search":
 241                 query = str.join(" ", tokens[1:])
 242             elif tokens[0] == "offset-search":
 243                 to_skip = int(tokens[1])
 244                 query = str.join(" ", tokens[2:])
 245             try:
 246                 results = plomsearch.search(query, lines)
 247             except plomsearch.LogicParserError as err:
 248                 notice("failed query parsing: " + str(err))
 249                 return
 250             if len(results) == 0:
 251                 notice("no quotes matching query")
 252             else:
 253                 if to_skip >= len(results):
 254                     notice("skipped all quotes matching query")
 255                 else:
 256                     notice("found %s matches, showing max. 3, skipping %s"
 257                            % (len(results), to_skip))
 258                 for i in range(len(results)):
 259                     if i >= to_skip and i < to_skip + 3:
 260                         result = results[i]
 261                         notice("quote #" + str(result[0] + 1) + ": "
 262                                + result[1][:-1])
 263             return
 264         else:
 265             i = random.randrange(len(lines))
 266         notice("quote #" + str(i + 1) + ": " + lines[i][:-1])
 267
 268     def markov():
 269
 270         def help():
 271             notice("syntax: !markov [integer from 1 to infinite]")
 272
 273         def markov(snippet):
 274             usable_selections = []
 275             for i in range(select_length, 0, -1):
 276                 for selection in selections:
 277                     add = True
 278                     for j in range(i):
 279                         j += 1
 280                         if snippet[-j] != selection[-(j+1)]:
 281                             add = False
 282                             break
 283                     if add:
 284                         usable_selections += [selection]
 285                 if [] != usable_selections:
 286                     break
 287             if [] == usable_selections:
 288                 usable_selections = selections
 289             selection = choice(usable_selections)
 290             return selection[select_length]
 291
 292         if "" == argument:
 293             tokens = []
 294         else:
 295             tokens = argument.split(" ")
 296         if (len(tokens) > 1 or (len(tokens) == 1 and not tokens[0].isdigit())):
 297             help()
 298             return
 299
 300         from random import choice, shuffle
 301         select_length = 2
 302         if len(tokens) == 1:
 303             n = int(tokens[0])
 304             if n > 0:
 305                 select_length = n
 306             else:
 307                 notice("bad value, using default: " + str(select_length))
 308         selections = []
 309
 310         if not os.access(session.markovfile, os.F_OK):
 311             notice("not enough text to markov for selection length")
 312             return
 313
 314         # Lowercase incoming lines, ensure they end in a sentence end mark.
 315         file = open(session.markovfile, "r")
 316         lines = file.readlines()
 317         file.close()
 318         tokens = []
 319         sentence_end_markers = ".!?)("
 320         for line in lines:
 321             line = line.lower().replace("\n", "")
 322             if line[-1] not in sentence_end_markers:
 323                 line += "."
 324             tokens += line.split()
 325         if len(tokens) - 1 <= select_length:
 326             notice("not enough text to markov")
 327             return
 328
 329         # Replace URLs with escape string for now, so that the Markov selector
 330         # won't see them as different strings. Stash replaced URLs in urls.
 331         urls = []
 332         url_escape = "\nURL"
 333         url_starts = ["http://", "https://", "<http://", "<https://"]
 334         for i in range(len(tokens)):
 335             for url_start in url_starts:
 336                 if tokens[i][:len(url_start)] == url_start:
 337                     length = len(tokens[i])
 338                     if url_start[0] == "<":
 339                         try:
 340                             length = tokens[i].index(">") + 1
 341                         except ValueError:
 342                             pass
 343                     urls += [tokens[i][:length]]
 344                     tokens[i] = url_escape + tokens[i][length:]
 345                     break
 346
 347         # For each snippet of select_length, use markov() to find continuation
 348         # token from selections. Replace present users' names with malkovich.
 349         # Start snippets with the beginning of a sentence, if possible.
 350         for i in range(len(tokens) - select_length):
 351             token_list = []
 352             for j in range(select_length + 1):
 353                 token_list += [tokens[i + j]]
 354             selections += [token_list]
 355         snippet = []
 356         for i in range(select_length):
 357             snippet += [""]
 358         shuffle(selections)
 359         for i in range(len(selections)):
 360             if selections[i][0][-1] in sentence_end_markers:
 361                 for j in range(select_length):
 362                     snippet[j] = selections[j][j + 1]
 363                 break
 364         msg = ""
 365         malkovich = "malkovich"
 366         while 1:
 367             new_end = markov(snippet)
 368             for name in session.users_in_chan:
 369                 if new_end[:len(name)] == name.lower():
 370                     new_end = malkovich + new_end[len(name):]
 371                     break
 372             if len(msg) + len(new_end) > 200:
 373                 break
 374             msg += new_end + " "
 375             for i in range(select_length - 1):
 376                 snippet[i] = snippet[i + 1]
 377             snippet[select_length - 1] = new_end
 378
 379         # Replace occurences of url escape string with random choice from urls.
 380         while True:
 381             index = msg.find(url_escape)
 382             if index < 0:
 383                 break
 384             msg = msg.replace(url_escape, choice(urls), 1)
 385
 386         # More meaningful ways to randomly end sentences.
 387         notice(msg + malkovich + ".")
 388
 389     def twt():
 390         def try_open(mode):
 391             try:
 392                 twtfile = open(session.twtfile, mode)
 393             except (PermissionError, FileNotFoundError) as err:
 394                 notice("can't access or create twt file: " + str(err))
 395                 return None
 396             return twtfile
 397
 398         from datetime import datetime
 399         if not os.access(session.twtfile, os.F_OK):
 400             twtfile = try_open("w")
 401             if None == twtfile:
 402                 return
 403             twtfile.close()
 404         twtfile = try_open("a")
 405         if None == twtfile:
 406             return
 407         twtfile.write(datetime.utcnow().isoformat() + "\t" + argument + "\n")
 408         twtfile.close()
 409         notice("wrote twt.")
 410
 411     if "addquote" == command:
 412         addquote()
 413     elif "quote" == command:
 414         quote()
 415     elif "markov" == command:
 416         markov()
 417     elif "twt" == command:
 418         twt()
 419
 420
 421 def handle_url(url, notice, show_url=False):
 422
 423     def mobile_twitter_hack(url):
 424         re1 = 'https?://(mobile.twitter.com/)[^/]+(/status/)'
 425         re2 = 'https?://mobile.twitter.com/([^/]+)/status/([^\?/]+)'
 426         m = re.search(re1, url)
 427         if m and m.group(1) == 'mobile.twitter.com/' \
 428                 and m.group(2) == '/status/':
 429             m = re.search(re2, url)
 430             url = 'https://twitter.com/' + m.group(1) + '/status/' + m.group(2)
 431             handle_url(url, notice, True)
 432             return True
 433
 434     class TimeOut(Exception):
 435         pass
 436
 437     def timeout_handler(ignore1, ignore2):
 438         raise TimeOut("timeout")
 439
 440     signal.signal(signal.SIGALRM, timeout_handler)
 441     signal.alarm(15)
 442     try:
 443         r = requests.get(url, headers = {'User-Agent': 'plomlombot'}, stream=True)
 444         r.raw.decode_content = True
 445         text = r.raw.read(10000000+1)
 446         if len(text) > 10000000:
 447             raise ValueError('Too large a response')
 448     except (requests.exceptions.TooManyRedirects,
 449             requests.exceptions.ConnectionError,
 450             requests.exceptions.InvalidURL,
 451             TimeOut,
 452             UnicodeError,
 453             ValueError,
 454             requests.exceptions.InvalidSchema) as error:
 455         signal.alarm(0)
 456         notice("trouble following url: " + str(error))
 457         return False
 458     signal.alarm(0)
 459     if mobile_twitter_hack(url):
 460         return True
 461     title = bs4.BeautifulSoup(text, "html5lib").title
 462     if title and title.string:
 463         prefix = "page title: "
 464         if show_url:
 465             prefix = "page title for <" + url + ">: "
 466         notice(prefix + title.string.strip())
 467     else:
 468         notice("page has no title tag")
 469     return True
 470
 471
 472 class Session:
 473
 474     def __init__(self, io, username, nickname, sasl, channel, twtfile, dbdir, rmlogs,
 475                  markov_input, no_show_page_titles):
 476         import base64
 477         self.io = io
 478         self.nickname = nickname
 479         self.users_in_chan = []
 480         self.twtfile = twtfile
 481         hash_channel = hashlib.md5(channel.encode("utf-8")).hexdigest()
 482         chandir = dbdir + "/" + hash_channel + "/"
 483         self.markov_input = markov_input
 484         self.markovfile = chandir + "markovfeed"
 485         self.quotesfile = chandir + "quotes"
 486         self.log = Log(chandir, self.nickname, username, channel, rmlogs)
 487         if sasl:
 488             self.io.send_line("CAP REQ :sasl")
 489         self.io.send_line("NICK " + self.nickname)
 490         self.io.send_line("USER " + username + " 0 * : ")
 491         if sasl:
 492             self.io.send_line("AUTHENTICATE PLAIN")
 493             auth = username + '\0' + username + '\0' + password
 494             auth_encoded = base64.b64encode(auth.encode()).decode().rstrip()
 495             self.io.send_line("AUTHENTICATE " + auth_encoded)
 496             self.io.send_line("CAP END")
 497         self.io.send_line("JOIN " + channel)
 498         self.io.log = self.log
 499         self.log.separator_line()
 500         self.show_page_titles = not no_show_page_titles
 501
 502     def loop(self):
 503
 504         def handle_privmsg(line):
 505
 506             def notice(msg):
 507                 line = "NOTICE " + target + " :" + msg
 508                 self.io.send_line(line)
 509
 510             target = line.sender
 511             if line.receiver != self.nickname:
 512                 target = line.receiver
 513             msg = str.join(" ", line.tokens[3:])[1:]
 514             if self.show_page_titles:
 515                 matches = re.findall("(https?://[^\s>]+)", msg)
 516                 url_count = 0
 517                 for i in range(len(matches)):
 518                     if handle_url(matches[i], notice):
 519                         url_count += 1
 520                         if url_count == 3:
 521                             notice("maximum number of urls to parse per "
 522                                    "message reached")
 523                             break
 524             if "!" == msg[0] and len(msg) > 1:
 525                 tokens = msg[1:].split()
 526                 argument = str.join(" ", tokens[1:])
 527                 handle_command(tokens[0], argument, notice, target, self)
 528                 return
 529             if self.markov_input:
 530                 write_to_file(self.markovfile, "a", msg + "\n")
 531
 532         while True:
 533             self.log.rmlogs()
 534             line = self.io.recv_line()
 535             if not line:
 536                 continue
 537             line = Line(line)
 538             if len(line.tokens) > 1:
 539                 if line.tokens[0] == "PING":
 540                     self.io.send_line("PONG " + line.tokens[1])
 541                 elif line.tokens[1] == "PRIVMSG":
 542                     handle_privmsg(line)
 543                 elif line.tokens[1] == "353":
 544                     names = line.tokens[5:]
 545                     names[0] = names[0][1:]
 546                     for i in range(len(names)):
 547                         names[i] = names[i].replace("@", "").replace("+", "")
 548                     self.users_in_chan += names
 549                 elif line.tokens[1] == "JOIN" and line.sender != self.nickname:
 550                     self.users_in_chan += [line.sender]
 551                 elif line.tokens[1] == "PART":
 552                     del(self.users_in_chan[self.users_in_chan.index(line.sender)])
 553                 elif line.tokens[1] == "NICK":
 554                     del(self.users_in_chan[self.users_in_chan.index(line.sender)])
 555                     self.users_in_chan += [line.receiver]
 556
 557
 558 def parse_command_line_arguments():
 559     parser = argparse.ArgumentParser()
 560     parser.add_argument("-s, --server", action="store", dest="server",
 561                         default=SERVER,
 562                         help="server or server net to connect to (default: "
 563                         + SERVER + ")")
 564     parser.add_argument("-p, --port", action="store", dest="port", type=int,
 565                         default=PORT, help="port to connect to (default : "
 566                         + str(PORT) + ")")
 567     parser.add_argument("-c, --channel", action="store", dest="channel",
 568                         default=SERVER, help="channel to join")
 569     parser.add_argument("-w, --wait", action="store", dest="timeout",
 570                         type=int, default=TIMEOUT,
 571                         help="timeout in seconds after which to attempt "
 572                         "reconnect (default: " + str(TIMEOUT) + ")")
 573     parser.add_argument("-u, --username", action="store", dest="username",
 574                         default=USERNAME, help="username to use (default: "
 575                         + USERNAME + ")")
 576     parser.add_argument("-n, --nickname", action="store", dest="nickname",
 577                         default=NICKNAME, help="nickname to use (default: "
 578                         + NICKNAME + ")")
 579     parser.add_argument("-a, --authenticate", action="store", dest="sasl",
 580                         default=None, help="SASL password (default: none)")
 581     parser.add_argument("-t, --twtxtfile", action="store", dest="twtfile",
 582                         default=TWTFILE, help="twtxt file to use (default: "
 583                         + TWTFILE + ")")
 584     parser.add_argument("-d, --dbdir", action="store", dest="dbdir",
 585                         default=DBDIR, help="directory to store DB files in")
 586     parser.add_argument("-r, --rmlogs", action="store", dest="rmlogs",
 587                         type=int, default=0,
 588                         help="maximum age in seconds for logfiles in logs/ "
 589                         "(0 means: never delete, and is default)")
 590     parser.add_argument("-m, --markov_store", action="store_true",
 591                         dest="markov_store",
 592                         help="log channel discussions for !markov input")
 593     parser.add_argument("--no-show-page-titles", action="store_true",
 594                         dest="no_show_page_titles",
 595                         help="do not show page titles")
 596     opts, unknown = parser.parse_known_args()
 597     return opts
 598
 599
 600 opts = parse_command_line_arguments()
 601 while True:
 602     try:
 603         io = IO(opts.server, opts.port, opts.timeout)
 604         hash_server = hashlib.md5(opts.server.encode("utf-8")).hexdigest()
 605         dbdir = opts.dbdir + "/" + hash_server
 606         session = Session(io, opts.username, opts.nickname, opts.sasl, opts.channel,
 607                           opts.twtfile, dbdir, opts.rmlogs, opts.markov_store,
 608                           opts.no_show_page_titles)
 609         session.loop()
 610     except ExceptionForRestart:
 611         io.socket.close()
 612         continue