X-Git-Url: https://plomlompom.com/repos/?p=plomlombot-irc.git;a=blobdiff_plain;f=plomlombot.py;h=f46a0b1db619e1086188512d951838015c5eda47;hp=11a8206ab4691eaf500cedbb0873efeb532da561;hb=50a05c8c5789c22e60c7bd94e44a8cc4c29dbdd4;hpb=25fcd7f13104e84fe5726757632019a62afa0c74 diff --git a/plomlombot.py b/plomlombot.py index 11a8206..f46a0b1 100755 --- a/plomlombot.py +++ b/plomlombot.py @@ -11,7 +11,9 @@ import bs4 import random import hashlib import os +import signal import plomsearch +import irclog # Defaults, may be overwritten by command line arguments. SERVER = "irc.freenode.net" @@ -19,23 +21,107 @@ PORT = 6667 TIMEOUT = 240 USERNAME = "plomlombot" NICKNAME = USERNAME +TWTFILE = "" +DBDIR = os.path.expanduser("~/plomlombot_db") + + +def write_to_file(path, mode, text): + f = open(path, mode) + f.write(text) + f.close() class ExceptionForRestart(Exception): pass +class Line: + + def __init__(self, line): + self.line = line + self.tokens = line.split(" ") + self.sender = "" + if self.tokens[0][0] == ":": + for rune in self.tokens[0][1:]: + if rune in {"!", "@"}: + break + self.sender += rune + self.receiver = "" + if len(self.tokens) > 2: + for rune in self.tokens[2]: + if rune in {"!", "@"}: + break + if rune != ":": + self.receiver += rune + + +class Log: + + def __init__(self, chandir, nickname, username, channel, rmlogs): + self.nickname = nickname + self.username = username + self.channel = channel + self.chandir = chandir + self.rmlogcycle = rmlogs + self.rawlogdir = chandir + "raw_logs/" + self.logdir = chandir + "logs/" + if not os.path.exists(self.logdir): + os.makedirs(self.logdir) + if not os.path.exists(self.rawlogdir): + os.makedirs(self.rawlogdir) + + def log(self, line, sent=False): + identity = "" + separator = " > " + if sent: + separator = " " + line = Line("< " + line) + line.sender = self.nickname + identity = self.username + "@localhost" + else: + if type(line) == str: + line = Line(line) + now = datetime.datetime.utcnow() + form = "%Y-%m-%d %H:%M:%S UTC" + write_to_file(self.rawlogdir + now.strftime("%Y-%m-%d") + ".txt", "a", + now.strftime(form) + separator + line.line + "\n") + to_log = irclog.format_logline(line, self.channel, identity) + if to_log != None: + write_to_file(self.logdir + now.strftime("%Y-%m-%d") + ".txt", "a", + now.strftime(form) + " " + to_log + "\n") + + def rmlogs(self): + if self.rmlogcycle > 0: + for f in os.listdir(self.logdir): + f = os.path.join(self.logdir, f) + if os.path.isfile(f) and \ + os.stat(f).st_mtime < time.time() - self.rmlogcycle: + os.remove(f) + + def separator_line(self): + now = datetime.datetime.utcnow() + write_to_file(self.logdir + now.strftime("%Y-%m-%d") + ".txt", "a", + "-----------------------\n") + + class IO: def __init__(self, server, port, timeout): + self.log = None self.timeout = timeout self.socket = socket.socket() - self.socket.connect((server, port)) + try: + self.socket.connect((server, port)) + except TimeoutError: + raise ExceptionForRestart self.socket.setblocking(0) self.line_buffer = [] self.rune_buffer = "" self.last_pong = time.time() - self.servername = self.recv_line(send_ping=False).split(" ")[0][1:] + line = self.recv_line(send_ping=False) + if not line or len(line) < 1: + raise ExceptionForRestart + self.servername = line.split(" ")[0][1:] def _pingtest(self, send_ping=True): if self.last_pong + self.timeout < time.time(): @@ -51,6 +137,8 @@ class IO: print("NOT SENT LINE TO SERVER (too long): " + msg) print("LINE TO SERVER: " + str(datetime.datetime.now()) + ": " + msg) + if self.log != None: + self.log.log(msg, True) msg = msg + "\r\n" msg_len = len(msg) total_sent_len = 0 @@ -88,82 +176,98 @@ class IO: def recv_line(self, send_ping=True): line = self._recv_line_wrapped(send_ping) if line: + if self.log != None: + self.log.log(line) print("LINE FROM SERVER " + str(datetime.datetime.now()) + ": " + line) return line -def handle_command(command, argument, notice, target): - hash_string = hashlib.md5(target.encode("utf-8")).hexdigest() - quotesfile_name = "quotes_" + hash_string +def handle_command(command, argument, notice, target, session): def addquote(): - if not os.access(quotesfile_name, os.F_OK): - quotesfile = open(quotesfile_name, "w") - quotesfile.write("QUOTES FOR " + target + ":\n") - quotesfile.close() - quotesfile = open(quotesfile_name, "a") - quotesfile.write(argument + "\n") - quotesfile.close() - quotesfile = open(quotesfile_name, "r") + if not os.access(session.quotesfile, os.F_OK): + write_to_file(session.quotesfile, "w", + "QUOTES FOR " + target + ":\n") + write_to_file(session.quotesfile, "a", argument + "\n") + quotesfile = open(session.quotesfile, "r") lines = quotesfile.readlines() quotesfile.close() - notice("ADDED QUOTE #" + str(len(lines) - 1)) + notice("added quote #" + str(len(lines) - 1)) def quote(): def help(): - notice("SYNTAX: !quote [int] OR !quote search QUERY") + notice("syntax: !quote [int] OR !quote search QUERY " + "OR !quote offset-search [int] QUERY") notice("QUERY may be a boolean grouping of quoted or unquoted " + "search terms, examples:") notice("!quote search foo") notice("!quote search foo AND (bar OR NOT baz)") notice("!quote search \"foo\\\"bar\" AND ('NOT\"' AND \"'foo'\"" + " OR 'bar\\'baz')") + notice("The offset-search int argument defines how many matches " + "to skip (useful if results are above maximum number to " + "display).") if "" == argument: tokens = [] else: tokens = argument.split(" ") - if (len(tokens) > 1 and tokens[0] != "search") or \ - (len(tokens) == 1 and - (tokens[0] == "search" or not tokens[0].isdigit())): - help() + if (len(tokens) == 1 and not tokens[0].isdigit()) or \ + (tokens > 1 and + tokens[0] not in {"search", "offset-search"} or + (tokens[0] == "offset-search" and + ((not len(tokens) > 2) or (not tokens[1].isdigit())))): + help() + return + if not os.access(session.quotesfile, os.F_OK): + notice("no quotes available") return - if not os.access(quotesfile_name, os.F_OK): - notice("NO QUOTES AVAILABLE") - return - quotesfile = open(quotesfile_name, "r") + quotesfile = open(session.quotesfile, "r") lines = quotesfile.readlines() quotesfile.close() lines = lines[1:] if len(tokens) == 1: i = int(tokens[0]) if i == 0 or i > len(lines): - notice("THERE'S NO QUOTE OF THAT INDEX") + notice("there's no quote of that index") return i = i - 1 elif len(tokens) > 1: - query = str.join(" ", tokens[1:]) + to_skip = 0 + if tokens[0] == "search": + query = str.join(" ", tokens[1:]) + elif tokens[0] == "offset-search": + to_skip = int(tokens[1]) + query = str.join(" ", tokens[2:]) try: results = plomsearch.search(query, lines) except plomsearch.LogicParserError as err: - notice("FAILED QUERY PARSING: " + str(err)) + notice("failed query parsing: " + str(err)) return if len(results) == 0: - notice("NO QUOTES MATCHING QUERY") + notice("no quotes matching query") else: - for result in results: - notice("QUOTE #" + str(result[0] + 1) + " : " + result[1]) + if to_skip >= len(results): + notice("skipped all quotes matching query") + else: + notice("found %s matches, showing max. 3, skipping %s" + % (len(results), to_skip)) + for i in range(len(results)): + if i >= to_skip and i < to_skip + 3: + result = results[i] + notice("quote #" + str(result[0] + 1) + ": " + + result[1][:-1]) return else: i = random.randrange(len(lines)) - notice("QUOTE #" + str(i + 1) + ": " + lines[i]) + notice("quote #" + str(i + 1) + ": " + lines[i][:-1]) def markov(): - from random import shuffle - select_length = 2 - selections = [] + + def help(): + notice("syntax: !markov [integer from 1 to infinite]") def markov(snippet): usable_selections = [] @@ -171,7 +275,8 @@ def handle_command(command, argument, notice, target): for selection in selections: add = True for j in range(i): - if snippet[j] != selection[j]: + j += 1 + if snippet[-j] != selection[-(j+1)]: add = False break if add: @@ -180,24 +285,67 @@ def handle_command(command, argument, notice, target): break if [] == usable_selections: usable_selections = selections - shuffle(usable_selections) - return usable_selections[0][select_length] + selection = choice(usable_selections) + return selection[select_length] - hash_string = hashlib.md5(target.encode("utf-8")).hexdigest() - markovfeed_name = "markovfeed_" + hash_string - if not os.access(markovfeed_name, os.F_OK): - notice("NOT ENOUGH TEXT TO MARKOV.") + if "" == argument: + tokens = [] + else: + tokens = argument.split(" ") + if (len(tokens) > 1 or (len(tokens) == 1 and not tokens[0].isdigit())): + help() return - file = open(markovfeed_name, "r") + + from random import choice, shuffle + select_length = 2 + if len(tokens) == 1: + n = int(tokens[0]) + if n > 0: + select_length = n + else: + notice("bad value, using default: " + str(select_length)) + selections = [] + + if not os.access(session.markovfile, os.F_OK): + notice("not enough text to markov for selection length") + return + + # Lowercase incoming lines, ensure they end in a sentence end mark. + file = open(session.markovfile, "r") lines = file.readlines() file.close() tokens = [] + sentence_end_markers = ".!?)(" for line in lines: - line = line.replace("\n", "") + line = line.lower().replace("\n", "") + if line[-1] not in sentence_end_markers: + line += "." tokens += line.split() - if len(tokens) <= select_length: - notice("NOT ENOUGH TEXT TO MARKOV.") + if len(tokens) - 1 <= select_length: + notice("not enough text to markov") return + + # Replace URLs with escape string for now, so that the Markov selector + # won't see them as different strings. Stash replaced URLs in urls. + urls = [] + url_escape = "\nURL" + url_starts = ["http://", "https://", "") + 1 + except ValueError: + pass + urls += [tokens[i][:length]] + tokens[i] = url_escape + tokens[i][length:] + break + + # For each snippet of select_length, use markov() to find continuation + # token from selections. Replace present users' names with malkovich. + # Start snippets with the beginning of a sentence, if possible. for i in range(len(tokens) - select_length): token_list = [] for j in range(select_length + 1): @@ -206,16 +354,58 @@ def handle_command(command, argument, notice, target): snippet = [] for i in range(select_length): snippet += [""] + shuffle(selections) + for i in range(len(selections)): + if selections[i][0][-1] in sentence_end_markers: + for j in range(select_length): + snippet[j] = selections[j][j + 1] + break msg = "" + malkovich = "malkovich" while 1: new_end = markov(snippet) + for name in session.users_in_chan: + if new_end[:len(name)] == name.lower(): + new_end = malkovich + new_end[len(name):] + break if len(msg) + len(new_end) > 200: break msg += new_end + " " for i in range(select_length - 1): snippet[i] = snippet[i + 1] snippet[select_length - 1] = new_end - notice(msg.lower() + "malkovich.") + + # Replace occurences of url escape string with random choice from urls. + while True: + index = msg.find(url_escape) + if index < 0: + break + msg = msg.replace(url_escape, choice(urls), 1) + + # More meaningful ways to randomly end sentences. + notice(msg + malkovich + ".") + + def twt(): + def try_open(mode): + try: + twtfile = open(session.twtfile, mode) + except (PermissionError, FileNotFoundError) as err: + notice("can't access or create twt file: " + str(err)) + return None + return twtfile + + from datetime import datetime + if not os.access(session.twtfile, os.F_OK): + twtfile = try_open("w") + if None == twtfile: + return + twtfile.close() + twtfile = try_open("a") + if None == twtfile: + return + twtfile.write(datetime.utcnow().isoformat() + "\t" + argument + "\n") + twtfile.close() + notice("wrote twt.") if "addquote" == command: addquote() @@ -223,6 +413,8 @@ def handle_command(command, argument, notice, target): quote() elif "markov" == command: markov() + elif "twt" == command: + twt() def handle_url(url, notice, show_url=False): @@ -238,86 +430,117 @@ def handle_url(url, notice, show_url=False): handle_url(url, notice, True) return True + class TimeOut(Exception): + pass + + def timeout_handler(ignore1, ignore2): + raise TimeOut("timeout") + + signal.signal(signal.SIGALRM, timeout_handler) + signal.alarm(15) try: - r = requests.get(url, timeout=15) + r = requests.get(url, headers = {'User-Agent': 'plomlombot'}, stream=True) + r.raw.decode_content = True + text = r.raw.read(10000000+1) + if len(text) > 10000000: + raise ValueError('Too large a response') except (requests.exceptions.TooManyRedirects, requests.exceptions.ConnectionError, requests.exceptions.InvalidURL, + TimeOut, + UnicodeError, + ValueError, requests.exceptions.InvalidSchema) as error: - notice("TROUBLE FOLLOWING URL: " + str(error)) - return + signal.alarm(0) + notice("trouble following url: " + str(error)) + return False + signal.alarm(0) if mobile_twitter_hack(url): - return - title = bs4.BeautifulSoup(r.text).title - if title: - prefix = "PAGE TITLE: " + return True + title = bs4.BeautifulSoup(text, "html5lib").title + if title and title.string: + prefix = "page title: " if show_url: - prefix = "PAGE TITLE FOR <" + url + ">: " + prefix = "page title for <" + url + ">: " notice(prefix + title.string.strip()) else: - notice("PAGE HAS NO TITLE TAG") + notice("page has no title tag") + return True class Session: - def __init__(self, io, username, nickname, channel): + def __init__(self, io, username, nickname, channel, twtfile, dbdir, rmlogs, + markov_input): self.io = io self.nickname = nickname + self.users_in_chan = [] + self.twtfile = twtfile + hash_channel = hashlib.md5(channel.encode("utf-8")).hexdigest() + chandir = dbdir + "/" + hash_channel + "/" + self.markov_input = markov_input + self.markovfile = chandir + "markovfeed" + self.quotesfile = chandir + "quotes" + self.log = Log(chandir, self.nickname, username, channel, rmlogs) self.io.send_line("NICK " + self.nickname) self.io.send_line("USER " + username + " 0 * : ") self.io.send_line("JOIN " + channel) + self.io.log = self.log + self.log.separator_line() def loop(self): - def handle_privmsg(tokens): - - def handle_input(msg, target): - - def notice(msg): - self.io.send_line("NOTICE " + target + " :" + msg) - - matches = re.findall("(https?://[^\s>]+)", msg) - for i in range(len(matches)): - handle_url(matches[i], notice) - if "!" == msg[0]: - tokens = msg[1:].split() - argument = str.join(" ", tokens[1:]) - handle_command(tokens[0], argument, notice, target) - return - hash_string = hashlib.md5(target.encode("utf-8")).hexdigest() - markovfeed_name = "markovfeed_" + hash_string - file = open(markovfeed_name, "a") - file.write(msg + "\n") - file.close() - - sender = "" - for rune in tokens[0]: - if rune == "!": - break - if rune != ":": - sender += rune - receiver = "" - for rune in tokens[2]: - if rune == "!": - break - if rune != ":": - receiver += rune - target = sender - if receiver != self.nickname: - target = receiver - msg = str.join(" ", tokens[3:])[1:] - handle_input(msg, target) + def handle_privmsg(line): + + def notice(msg): + line = "NOTICE " + target + " :" + msg + self.io.send_line(line) + + target = line.sender + if line.receiver != self.nickname: + target = line.receiver + msg = str.join(" ", line.tokens[3:])[1:] + matches = re.findall("(https?://[^\s>]+)", msg) + url_count = 0 + for i in range(len(matches)): + if handle_url(matches[i], notice): + url_count += 1 + if url_count == 3: + notice("maximum number of urls to parse per message " + "reached") + break + if "!" == msg[0] and len(msg) > 1: + tokens = msg[1:].split() + argument = str.join(" ", tokens[1:]) + handle_command(tokens[0], argument, notice, target, self) + return + if self.markov_input: + write_to_file(self.markovfile, "a", msg + "\n") while True: + self.log.rmlogs() line = self.io.recv_line() if not line: continue - tokens = line.split(" ") - if len(tokens) > 1: - if tokens[0] == "PING": - self.io.send_line("PONG " + tokens[1]) - elif tokens[1] == "PRIVMSG": - handle_privmsg(tokens) + line = Line(line) + if len(line.tokens) > 1: + if line.tokens[0] == "PING": + self.io.send_line("PONG " + line.tokens[1]) + elif line.tokens[1] == "PRIVMSG": + handle_privmsg(line) + elif line.tokens[1] == "353": + names = line.tokens[5:] + names[0] = names[0][1:] + for i in range(len(names)): + names[i] = names[i].replace("@", "").replace("+", "") + self.users_in_chan += names + elif line.tokens[1] == "JOIN" and line.sender != self.nickname: + self.users_in_chan += [line.sender] + elif line.tokens[1] == "PART": + del(self.users_in_chan[self.users_in_chan.index(line.sender)]) + elif line.tokens[1] == "NICK": + del(self.users_in_chan[self.users_in_chan.index(line.sender)]) + self.users_in_chan += [line.receiver] def parse_command_line_arguments(): @@ -329,9 +552,9 @@ def parse_command_line_arguments(): parser.add_argument("-p, --port", action="store", dest="port", type=int, default=PORT, help="port to connect to (default : " + str(PORT) + ")") - parser.add_argument("-t, --timeout", action="store", dest="timeout", + parser.add_argument("-w, --wait", action="store", dest="timeout", type=int, default=TIMEOUT, - help="timeout in seconds after which to attempt " + + help="timeout in seconds after which to attempt " "reconnect (default: " + str(TIMEOUT) + ")") parser.add_argument("-u, --username", action="store", dest="username", default=USERNAME, help="username to use (default: " @@ -339,6 +562,18 @@ def parse_command_line_arguments(): parser.add_argument("-n, --nickname", action="store", dest="nickname", default=NICKNAME, help="nickname to use (default: " + NICKNAME + ")") + parser.add_argument("-t, --twtxtfile", action="store", dest="twtfile", + default=TWTFILE, help="twtxt file to use (default: " + + TWTFILE + ")") + parser.add_argument("-d, --dbdir", action="store", dest="dbdir", + default=DBDIR, help="directory to store DB files in") + parser.add_argument("-r, --rmlogs", action="store", dest="rmlogs", + type=int, default=0, + help="maximum age in seconds for logfiles in logs/ " + "(0 means: never delete, and is default)") + parser.add_argument("-m, --markov_store", action="store_true", + dest="markov_store", + help="log channel discussions for !markov input") parser.add_argument("CHANNEL", action="store", help="channel to join") opts, unknown = parser.parse_known_args() return opts @@ -348,7 +583,10 @@ opts = parse_command_line_arguments() while True: try: io = IO(opts.server, opts.port, opts.timeout) - session = Session(io, opts.username, opts.nickname, opts.CHANNEL) + hash_server = hashlib.md5(opts.server.encode("utf-8")).hexdigest() + dbdir = opts.dbdir + "/" + hash_server + session = Session(io, opts.username, opts.nickname, opts.CHANNEL, + opts.twtfile, dbdir, opts.rmlogs, opts.markov_store) session.loop() except ExceptionForRestart: io.socket.close()