plomlombot.py

   1 #!/usr/bin/python3
   2
   3 import argparse
   4 import socket
   5 import datetime
   6 import select
   7 import time
   8 import re
   9 import requests
  10 import bs4
  11 import random
  12 import hashlib
  13 import os
  14 import plomsearch
  15 import irclog
  16
  17 # Defaults, may be overwritten by command line arguments.
  18 SERVER = "irc.freenode.net"
  19 PORT = 6667
  20 TIMEOUT = 240
  21 USERNAME = "plomlombot"
  22 NICKNAME = USERNAME
  23 TWTFILE = ""
  24 DBDIR = os.path.expanduser("~/plomlombot_db")
  25
  26
  27 class ExceptionForRestart(Exception):
  28     pass
  29
  30
  31 class Line:
  32
  33     def __init__(self, line):
  34         self.line = line
  35         self.tokens = line.split(" ")
  36         self.sender = ""
  37         if self.tokens[0][0] == ":":
  38             for rune in self.tokens[0][1:]:
  39                 if rune in {"!", "@"}:
  40                     break
  41                 self.sender += rune
  42         self.receiver = ""
  43         if len(self.tokens) > 2:
  44             for rune in self.tokens[2]:
  45                 if rune in {"!", "@"}:
  46                     break
  47                 if rune != ":":
  48                     self.receiver += rune
  49
  50
  51 class IO:
  52
  53     def __init__(self, server, port, timeout):
  54         self.timeout = timeout
  55         self.socket = socket.socket()
  56         self.socket.connect((server, port))
  57         self.socket.setblocking(0)
  58         self.line_buffer = []
  59         self.rune_buffer = ""
  60         self.last_pong = time.time()
  61         self.servername = self.recv_line(send_ping=False).split(" ")[0][1:]
  62
  63     def _pingtest(self, send_ping=True):
  64         if self.last_pong + self.timeout < time.time():
  65             print("SERVER NOT ANSWERING")
  66             raise ExceptionForRestart
  67         if send_ping:
  68             self.send_line("PING " + self.servername)
  69
  70     def send_line(self, msg):
  71         msg = msg.replace("\r", " ")
  72         msg = msg.replace("\n", " ")
  73         if len(msg.encode("utf-8")) > 510:
  74             print("NOT SENT LINE TO SERVER (too long): " + msg)
  75         print("LINE TO SERVER: "
  76               + str(datetime.datetime.now()) + ": " + msg)
  77         msg = msg + "\r\n"
  78         msg_len = len(msg)
  79         total_sent_len = 0
  80         while total_sent_len < msg_len:
  81             sent_len = self.socket.send(bytes(msg[total_sent_len:], "UTF-8"))
  82             if sent_len == 0:
  83                 print("SOCKET CONNECTION BROKEN")
  84                 raise ExceptionForRestart
  85             total_sent_len += sent_len
  86
  87     def _recv_line_wrapped(self, send_ping=True):
  88         if len(self.line_buffer) > 0:
  89             return self.line_buffer.pop(0)
  90         while True:
  91             ready = select.select([self.socket], [], [], int(self.timeout / 2))
  92             if not ready[0]:
  93                 self._pingtest(send_ping)
  94                 return None
  95             self.last_pong = time.time()
  96             received_bytes = self.socket.recv(1024)
  97             try:
  98                 received_runes = received_bytes.decode("UTF-8")
  99             except UnicodeDecodeError:
 100                 received_runes = received_bytes.decode("latin1")
 101             if len(received_runes) == 0:
 102                 print("SOCKET CONNECTION BROKEN")
 103                 raise ExceptionForRestart
 104             self.rune_buffer += received_runes
 105             lines_split = str.split(self.rune_buffer, "\r\n")
 106             self.line_buffer += lines_split[:-1]
 107             self.rune_buffer = lines_split[-1]
 108             if len(self.line_buffer) > 0:
 109                 return self.line_buffer.pop(0)
 110
 111     def recv_line(self, send_ping=True):
 112         line = self._recv_line_wrapped(send_ping)
 113         if line:
 114             print("LINE FROM SERVER " + str(datetime.datetime.now()) + ": " +
 115                   line)
 116         return line
 117
 118
 119 def handle_command(command, argument, notice, target, session):
 120
 121     def addquote():
 122         if not os.access(session.quotesfile, os.F_OK):
 123             quotesfile = open(session.quotesfile, "w")
 124             quotesfile.write("QUOTES FOR " + target + ":\n")
 125             quotesfile.close()
 126         quotesfile = open(session.quotesfile, "a")
 127         quotesfile.write(argument + "\n")
 128         quotesfile.close()
 129         quotesfile = open(session.quotesfile, "r")
 130         lines = quotesfile.readlines()
 131         quotesfile.close()
 132         notice("ADDED QUOTE #" + str(len(lines) - 1))
 133
 134     def quote():
 135
 136         def help():
 137             notice("SYNTAX: !quote [int] OR !quote search QUERY")
 138             notice("QUERY may be a boolean grouping of quoted or unquoted " +
 139                    "search terms, examples:")
 140             notice("!quote search foo")
 141             notice("!quote search foo AND (bar OR NOT baz)")
 142             notice("!quote search \"foo\\\"bar\" AND ('NOT\"' AND \"'foo'\"" +
 143                    " OR 'bar\\'baz')")
 144
 145         if "" == argument:
 146             tokens = []
 147         else:
 148             tokens = argument.split(" ")
 149         if (len(tokens) > 1 and tokens[0] != "search") or \
 150             (len(tokens) == 1 and
 151                 (tokens[0] == "search" or not tokens[0].isdigit())):
 152             help()
 153             return
 154         if not os.access(session.quotesfile, os.F_OK):
 155             notice("NO QUOTES AVAILABLE")
 156             return
 157         quotesfile = open(session.quotesfile, "r")
 158         lines = quotesfile.readlines()
 159         quotesfile.close()
 160         lines = lines[1:]
 161         if len(tokens) == 1:
 162             i = int(tokens[0])
 163             if i == 0 or i > len(lines):
 164                 notice("THERE'S NO QUOTE OF THAT INDEX")
 165                 return
 166             i = i - 1
 167         elif len(tokens) > 1:
 168             query = str.join(" ", tokens[1:])
 169             try:
 170                 results = plomsearch.search(query, lines)
 171             except plomsearch.LogicParserError as err:
 172                 notice("FAILED QUERY PARSING: " + str(err))
 173                 return
 174             if len(results) == 0:
 175                 notice("NO QUOTES MATCHING QUERY")
 176             else:
 177                 for result in results:
 178                     notice("QUOTE #" + str(result[0] + 1) + " : "
 179                            + result[1][-1])
 180             return
 181         else:
 182             i = random.randrange(len(lines))
 183         notice("QUOTE #" + str(i + 1) + ": " + lines[i][:-1])
 184
 185     def markov():
 186         from random import choice, shuffle
 187         select_length = 2
 188         selections = []
 189
 190         def markov(snippet):
 191             usable_selections = []
 192             for i in range(select_length, 0, -1):
 193                 for selection in selections:
 194                     add = True
 195                     for j in range(i):
 196                         j += 1
 197                         if snippet[-j] != selection[-(j+1)]:
 198                             add = False
 199                             break
 200                     if add:
 201                         usable_selections += [selection]
 202                 if [] != usable_selections:
 203                     break
 204             if [] == usable_selections:
 205                 usable_selections = selections
 206             selection = choice(usable_selections)
 207             return selection[select_length]
 208
 209         if not os.access(session.markovfile, os.F_OK):
 210             notice("NOT ENOUGH TEXT TO MARKOV.")
 211             return
 212
 213         # Lowercase incoming lines, ensure they end in a sentence end mark.
 214         file = open(session.markovfile, "r")
 215         lines = file.readlines()
 216         file.close()
 217         tokens = []
 218         sentence_end_markers = ".!?)("
 219         for line in lines:
 220             line = line.lower().replace("\n", "")
 221             if line[-1] not in sentence_end_markers:
 222                 line += "."
 223             tokens += line.split()
 224         if len(tokens) <= select_length:
 225             notice("NOT ENOUGH TEXT TO MARKOV.")
 226             return
 227
 228         # Replace URLs with escape string for now, so that the Markov selector
 229         # won't see them as different strings. Stash replaced URLs in urls.
 230         urls = []
 231         url_escape = "\nURL"
 232         url_starts = ["http://", "https://", "<http://", "<https://"]
 233         for i in range(len(tokens)):
 234             for url_start in url_starts:
 235                 if tokens[i][:len(url_start)] == url_start:
 236                     length = len(tokens[i])
 237                     if url_start[0] == "<":
 238                         try:
 239                             length = tokens[i].index(">") + 1
 240                         except ValueError:
 241                             pass
 242                     urls += [tokens[i][:length]]
 243                     tokens[i] = url_escape + tokens[i][length:]
 244                     break
 245
 246         # For each snippet of select_length, use markov() to find continuation
 247         # token from selections. Replace present users' names with malkovich.
 248         # Start snippets with the beginning of a sentence, if possible.
 249         for i in range(len(tokens) - select_length):
 250             token_list = []
 251             for j in range(select_length + 1):
 252                 token_list += [tokens[i + j]]
 253             selections += [token_list]
 254         snippet = []
 255         for i in range(select_length):
 256             snippet += [""]
 257         shuffle(selections)
 258         for i in range(len(selections)):
 259             if selections[i][0][-1] in sentence_end_markers:
 260                 for i in range(select_length):
 261                     snippet[i] = selections[i][i + 1]
 262                 break
 263         msg = ""
 264         malkovich = "malkovich"
 265         while 1:
 266             new_end = markov(snippet)
 267             for name in session.users_in_chan:
 268                 if new_end[:len(name)] == name.lower():
 269                     new_end = malkovich + new_end[len(name):]
 270                     break
 271             if len(msg) + len(new_end) > 200:
 272                 break
 273             msg += new_end + " "
 274             for i in range(select_length - 1):
 275                 snippet[i] = snippet[i + 1]
 276             snippet[select_length - 1] = new_end
 277
 278         # Replace occurences of url escape string with random choice from urls.
 279         while True:
 280             index = msg.find(url_escape)
 281             if index < 0:
 282                 break
 283             msg = msg.replace(url_escape, choice(urls), 1)
 284
 285         # More meaningful ways to randomly end sentences.
 286         notice(msg + malkovich + ".")
 287
 288     def twt():
 289         def try_open(mode):
 290             try:
 291                 twtfile = open(session.twtfile, mode)
 292             except (PermissionError, FileNotFoundError) as err:
 293                 notice("CAN'T ACCESS OR CREATE TWT FILE: " + str(err))
 294                 return None
 295             return twtfile
 296
 297         from datetime import datetime
 298         if not os.access(session.twtfile, os.F_OK):
 299             twtfile = try_open("w")
 300             if None == twtfile:
 301                 return
 302             twtfile.close()
 303         twtfile = try_open("a")
 304         if None == twtfile:
 305             return
 306         twtfile.write(datetime.utcnow().isoformat() + "\t" + argument + "\n")
 307         twtfile.close()
 308         notice("WROTE TWT.")
 309
 310     if "addquote" == command:
 311         addquote()
 312     elif "quote" == command:
 313         quote()
 314     elif "markov" == command:
 315         markov()
 316     elif "twt" == command:
 317         twt()
 318
 319
 320 def handle_url(url, notice, show_url=False):
 321
 322     def mobile_twitter_hack(url):
 323         re1 = 'https?://(mobile.twitter.com/)[^/]+(/status/)'
 324         re2 = 'https?://mobile.twitter.com/([^/]+)/status/([^\?/]+)'
 325         m = re.search(re1, url)
 326         if m and m.group(1) == 'mobile.twitter.com/' \
 327                 and m.group(2) == '/status/':
 328             m = re.search(re2, url)
 329             url = 'https://twitter.com/' + m.group(1) + '/status/' + m.group(2)
 330             handle_url(url, notice, True)
 331             return True
 332
 333     try:
 334         r = requests.get(url, timeout=15)
 335     except (requests.exceptions.TooManyRedirects,
 336             requests.exceptions.ConnectionError,
 337             requests.exceptions.InvalidURL,
 338             UnicodeError,
 339             requests.exceptions.InvalidSchema) as error:
 340         notice("TROUBLE FOLLOWING URL: " + str(error))
 341         return
 342     if mobile_twitter_hack(url):
 343         return
 344     title = bs4.BeautifulSoup(r.text, "html5lib").title
 345     if title and title.string:
 346         prefix = "PAGE TITLE: "
 347         if show_url:
 348             prefix = "PAGE TITLE FOR <" + url + ">: "
 349         notice(prefix + title.string.strip())
 350     else:
 351         notice("PAGE HAS NO TITLE TAG")
 352
 353
 354 class Session:
 355
 356     def __init__(self, io, username, nickname, channel, twtfile, dbdir):
 357         self.io = io
 358         self.nickname = nickname
 359         self.username = username
 360         self.channel = channel
 361         self.users_in_chan = []
 362         self.twtfile = twtfile
 363         self.dbdir = dbdir
 364         self.io.send_line("NICK " + self.nickname)
 365         self.io.send_line("USER " + self.username + " 0 * : ")
 366         self.io.send_line("JOIN " + self.channel)
 367         hash_channel = hashlib.md5(self.channel.encode("utf-8")).hexdigest()
 368         self.chandir = self.dbdir + "/" + hash_channel + "/"
 369         self.rawlogdir = self.chandir + "raw_logs/"
 370         self.logdir = self.chandir + "logs/"
 371         if not os.path.exists(self.logdir):
 372             os.makedirs(self.logdir)
 373         if not os.path.exists(self.rawlogdir):
 374             os.makedirs(self.rawlogdir)
 375         self.markovfile = self.chandir + "markovfeed"
 376         self.quotesfile = self.chandir + "quotes"
 377
 378     def loop(self):
 379
 380         def log(line):
 381             if type(line) == str:
 382                 line = Line(":" + self.nickname + "!~" + self.username +
 383                             "@localhost" + " " + line)
 384             now = datetime.datetime.utcnow()
 385             logfile = open(self.rawlogdir + now.strftime("%Y-%m-%d") + ".txt", "a")
 386             form = "%Y-%m-%d %H:%M:%S UTC\t"
 387             logfile.write(now.strftime(form) + " " + line.line + "\n")
 388             logfile.close()
 389             to_log = irclog.format_logline(line, self.channel)
 390             if to_log != None:
 391                 logfile = open(self.logdir + now.strftime("%Y-%m-%d") + ".txt", "a")
 392                 logfile.write(now.strftime(form) + " " + to_log + "\n")
 393                 logfile.close()
 394
 395         def handle_privmsg(line):
 396
 397             def notice(msg):
 398                 line = "NOTICE " + target + " :" + msg
 399                 self.io.send_line(line)
 400                 log(line)
 401
 402             target = line.sender
 403             if line.receiver != self.nickname:
 404                 target = line.receiver
 405             msg = str.join(" ", line.tokens[3:])[1:]
 406             matches = re.findall("(https?://[^\s>]+)", msg)
 407             for i in range(len(matches)):
 408                 handle_url(matches[i], notice)
 409             if "!" == msg[0]:
 410                 tokens = msg[1:].split()
 411                 argument = str.join(" ", tokens[1:])
 412                 handle_command(tokens[0], argument, notice, target, self)
 413                 return
 414             file = open(self.markovfile, "a")
 415             file.write(msg + "\n")
 416             file.close()
 417
 418         while True:
 419             line = self.io.recv_line()
 420             if not line:
 421                 continue
 422             line = Line(line)
 423             log(line)
 424             if len(line.tokens) > 1:
 425                 if line.tokens[0] == "PING":
 426                     self.io.send_line("PONG " + line.tokens[1])
 427                 elif line.tokens[1] == "PRIVMSG":
 428                     handle_privmsg(line)
 429                 elif line.tokens[1] == "353":
 430                     names = line.tokens[5:]
 431                     names[0] = names[0][1:]
 432                     for i in range(len(names)):
 433                         names[i] = names[i].replace("@", "").replace("+", "")
 434                     self.users_in_chan += names
 435                 elif line.tokens[1] == "JOIN" and line.sender != self.nickname:
 436                     self.users_in_chan += [line.sender]
 437                 elif line.tokens[1] == "PART":
 438                     del(self.users_in_chan[self.users_in_chan.index(line.sender)])
 439                 elif line.tokens[1] == "NICK":
 440                     del(self.users_in_chan[self.users_in_chan.index(line.sender)])
 441                     self.users_in_chan += [line.receiver]
 442
 443
 444 def parse_command_line_arguments():
 445     parser = argparse.ArgumentParser()
 446     parser.add_argument("-s, --server", action="store", dest="server",
 447                         default=SERVER,
 448                         help="server or server net to connect to (default: "
 449                         + SERVER + ")")
 450     parser.add_argument("-p, --port", action="store", dest="port", type=int,
 451                         default=PORT, help="port to connect to (default : "
 452                         + str(PORT) + ")")
 453     parser.add_argument("-w, --wait", action="store", dest="timeout",
 454                         type=int, default=TIMEOUT,
 455                         help="timeout in seconds after which to attempt " +
 456                         "reconnect (default: " + str(TIMEOUT) + ")")
 457     parser.add_argument("-u, --username", action="store", dest="username",
 458                         default=USERNAME, help="username to use (default: "
 459                         + USERNAME + ")")
 460     parser.add_argument("-n, --nickname", action="store", dest="nickname",
 461                         default=NICKNAME, help="nickname to use (default: "
 462                         + NICKNAME + ")")
 463     parser.add_argument("-t, --twtxtfile", action="store", dest="twtfile",
 464                         default=TWTFILE, help="twtxt file to use (default: "
 465                         + TWTFILE + ")")
 466     parser.add_argument("-d, --dbdir", action="store", dest="dbdir",
 467                         default=DBDIR, help="directory to store DB files in")
 468     parser.add_argument("CHANNEL", action="store", help="channel to join")
 469     opts, unknown = parser.parse_known_args()
 470     return opts
 471
 472
 473 opts = parse_command_line_arguments()
 474 while True:
 475     try:
 476         io = IO(opts.server, opts.port, opts.timeout)
 477         hash_server = hashlib.md5(opts.server.encode("utf-8")).hexdigest()
 478         dbdir = opts.dbdir + "/" + hash_server
 479         session = Session(io, opts.username, opts.nickname, opts.CHANNEL,
 480             opts.twtfile, dbdir)
 481         session.loop()
 482     except ExceptionForRestart:
 483         io.socket.close()
 484         continue