home · contact · privacy
Strip URLs from markov texts.
[plomlombot-irc.git] / plomlombot.py
index b5051927c58b11a772dc0f3597022eec5230e423..c8c11f45ebb03a3e0a73a9339ba82013813278b7 100755 (executable)
@@ -13,6 +13,8 @@ import hashlib
 import os
 import plomsearch
 
+URLREGEX = "(https?://[^\s>]+)"
+
 # Defaults, may be overwritten by command line arguments.
 SERVER = "irc.freenode.net"
 PORT = 6667
@@ -93,7 +95,7 @@ class IO:
         return line
 
 
-def handle_command(command, argument, notice, target):
+def handle_command(command, argument, notice, target, session):
     hash_string = hashlib.md5(target.encode("utf-8")).hexdigest()
     quotesfile_name = "quotes_" + hash_string
 
@@ -160,10 +162,82 @@ def handle_command(command, argument, notice, target):
             i = random.randrange(len(lines))
         notice("QUOTE #" + str(i + 1) + ": " + lines[i])
 
+    def markov():
+        from random import shuffle
+        select_length = 2
+        selections = []
+
+        def markov(snippet):
+            usable_selections = []
+            for i in range(select_length, 0, -1):
+                for selection in selections:
+                    add = True
+                    for j in range(i):
+                        if snippet[j] != selection[j]:
+                            add = False
+                            break
+                    if add:
+                        usable_selections += [selection]
+                if [] != usable_selections:
+                    break
+            if [] == usable_selections:
+                usable_selections = selections
+            shuffle(usable_selections)
+            return usable_selections[0][select_length]
+
+        def purge_undesired(tokens):
+            for token in tokens:
+                if None != re.match("^" + URLREGEX, token):
+                    del(tokens[tokens.index(token)])
+            for name in session.uses_in_chan:
+                while True:
+                    try:
+                        del(tokens[tokens.index(name)])
+                    except ValueError:
+                        break
+            return tokens
+
+        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.")
+            return
+        file = open(markovfeed_name, "r")
+        lines = file.readlines()
+        file.close()
+        tokens = []
+        for line in lines:
+            line = line.replace("\n", "")
+            tokens += line.split()
+        tokens = purge_undesired(tokens)
+        if len(tokens) <= select_length:
+            notice("NOT ENOUGH TEXT TO MARKOV.")
+            return
+        for i in range(len(tokens) - select_length):
+            token_list = []
+            for j in range(select_length + 1):
+                token_list += [tokens[i + j]]
+            selections += [token_list]
+        snippet = []
+        for i in range(select_length):
+            snippet += [""]
+        msg = ""
+        while 1:
+            new_end = markov(snippet)
+            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.")
+
     if "addquote" == command:
         addquote()
     elif "quote" == command:
         quote()
+    elif "markov" == command:
+        markov()
 
 
 def handle_url(url, notice, show_url=False):
@@ -189,7 +263,7 @@ def handle_url(url, notice, show_url=False):
         return
     if mobile_twitter_hack(url):
         return
-    title = bs4.BeautifulSoup(r.text).title
+    title = bs4.BeautifulSoup(r.text, "html.parser").title
     if title:
         prefix = "PAGE TITLE: "
         if show_url:
@@ -204,9 +278,11 @@ class Session:
     def __init__(self, io, username, nickname, channel):
         self.io = io
         self.nickname = nickname
+        self.channel = channel
+        self.uses_in_chan = []
         self.io.send_line("NICK " + self.nickname)
         self.io.send_line("USER " + username + " 0 * : ")
-        self.io.send_line("JOIN " + channel)
+        self.io.send_line("JOIN " + self.channel)
 
     def loop(self):
 
@@ -217,13 +293,19 @@ class Session:
                 def notice(msg):
                     self.io.send_line("NOTICE " + target + " :" + msg)
 
-                matches = re.findall("(https?://[^\s>]+)", msg)
+                matches = re.findall(URLREGEX, 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)
+                    handle_command(tokens[0], argument, notice, target, self)
+                    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]:
@@ -243,6 +325,14 @@ class Session:
             msg = str.join(" ", tokens[3:])[1:]
             handle_input(msg, target)
 
+        def name_from_join_or_part(tokens):
+            token = tokens[0][1:]
+            index_cut = token.find("@")
+            index_ex = token.find("!")
+            if index_ex > 0 and index_ex < index_cut:
+                index_cut = index_ex
+            return token[:index_cut]
+
         while True:
             line = self.io.recv_line()
             if not line:
@@ -253,7 +343,17 @@ class Session:
                     self.io.send_line("PONG " + tokens[1])
                 elif tokens[1] == "PRIVMSG":
                     handle_privmsg(tokens)
-
+                elif tokens[1] == "353":
+                    names = tokens[5:]
+                    names[0] = names[0][1:]
+                    self.uses_in_chan += names
+                elif tokens[1] == "JOIN":
+                    name = name_from_join_or_part(tokens)
+                    if name != self.nickname:
+                        self.uses_in_chan += [name]
+                elif tokens[1] == "PART":
+                    name = name_from_join_or_part(tokens)
+                    del(self.uses_in_chan[self.uses_in_chan.index(name)])
 
 def parse_command_line_arguments():
     parser = argparse.ArgumentParser()