X-Git-Url: http://www.chiark.greenend.org.uk/ucgi/~matthewv/git?p=irc.git;a=blobdiff_plain;f=commands.py;h=159bbb4e888dc67d28bff0e64eeb13d0e389d8af;hp=4bdc45ce01dcc632ae861a4194ced4a9f5c48aed;hb=e6da3880f361a431d5a04a146ea3062cc418cf6f;hpb=9a8c6f919c6e2df251c611630a1d4bdfb4205548

diff --git a/commands.py b/commands.py
index 4bdc45c..159bbb4 100644
--- a/commands.py
+++ b/commands.py
@@ -233,7 +233,7 @@ def googleq(bot, cmd, nick, conn, public):
 # Look up the definition of something using google
 def defineq(bot, cmd, nick, conn, public):
     cmdrest = string.join(cmd.split()[1:])
-    targ = ("http://www.google.com/search?q=define%%3A%s&ie=utf-8&oe=utf-8"
+    targ = ("http://www.google.co.uk/search?q=define%%3A%s&ie=utf-8&oe=utf-8"
             % urllib.quote_plus(cmdrest))
     try:
         # Just slurp everything into a string
@@ -242,16 +242,17 @@ def defineq(bot, cmd, nick, conn, public):
         # This is of course going to be a bit fragile. We first look for
         # 'Definitions of %s on the Web' -- if this isn't present we
         # assume we have the 'no definitions found page'.
-        # The first defn starts after the following <p> tag.
+        # The first defn starts after the following <p> tag, but as the
+        # first <li> in a <ul type="disc" class=std>
         # Following that we assume that each definition is all the non-markup
         # before a <br> tag. Currently we just dump out the first definition.
-        match = re.search(r"Definitions of <b>.*?</b> on the Web.*?<p>\s*([^>]*)<br>",defnpage,re.MULTILINE)
+        match = re.search(r"Definitions of <b>.*?</b> on the Web.*?<li>\s*([^>]*)((<br>)|(<li>))",defnpage,re.MULTILINE)
         if match == None:
            bot.automsg(public,nick,"Some things defy definition.")
         else:
            # We assume google has truncated the definition for us so this
            # won't flood the channel with text...
-           defn = " ".join(match.group(1).split("\n"));
+           defn = " ".join(match.group(1).split("\n"))
            bot.automsg(public,nick,defn)
     except IOError: # if the connection times out. This blocks. :(
          bot.automsg(public,nick,"The web's broken. Waah!")
@@ -280,8 +281,11 @@ def disconnq(bot, cmd, nick, conn, public):
         bot.disconnect(msg="Be right back.")
 
 ### list keys of a dictionary
-def listkeysq(bot, cmd, nick, conn, public, dict):
-    bot.automsg(public,nick,string.join(dict.keys()))
+def listkeysq(bot, cmd, nick, conn, public, dict, sort=False):
+    d=dict.keys()
+    if sort:
+        d.sort()
+    bot.automsg(public,nick,string.join(d))
 
 ### rot13 text (yes, I could have typed out the letters....)
 ### also "foo".encode('rot13') would have worked
@@ -290,3 +294,96 @@ def rot13q(bot, cmd, nick, conn, public):
     b=a[13:]+a[:13]
     trans=string.maketrans(a+a.upper(),b+b.upper())
     conn.notice(nick, string.join(cmd.split()[1:]).translate(trans))
+
+### URL-tracking stuff
+
+### return a easy-to-read approximation of a time period
+def nicetime(tempus):
+  if (tempus<120):
+    tm="%d seconds ago"%int(tempus)
+  elif (tempus<7200):
+    tm="%d minutes ago"%int(tempus/60)
+  if (tempus>7200):
+    tm="%d hours ago"%int(tempus/3600)
+  return tm
+
+### class to store URL data
+class UrlLog:
+    "contains meta-data about a URL seen on-channel"
+    def __init__(self,url,nick):
+        self.nick=nick
+        self.url=url
+        self.first=time.time()
+        self.count=1
+        self.lastseen=time.time()
+        self.lastasked=time.time()
+    def recenttime(self):
+        return max(self.lastseen,self.lastasked)
+    def firstmen(self):
+        return nicetime(time.time()-self.first)
+    def urltype(self):
+        z=min(len(urlcomplaints)-1, self.count-1)
+        return urlcomplaints[z]
+        
+urlre = re.compile("(https?://[^ ]+)( |$)")
+urlcomplaints = ["a contemporary","an interesting","a fascinating","an overused","a vastly overused"]
+
+### Deal with /msg bot url or ~url in channel
+def urlq(bot, cmd, nick, conn, public,urldb):
+  if (not urlre.search(cmd)):
+    bot.automsg(False,nick,"Please use 'url' only with http URLs")
+    return
+
+  urlstring=urlre.search(cmd).group(1)
+  url=canonical_url(urlstring)
+  if (url in urldb):
+    T = urldb[url]
+    complaint="That's %s URL that was first mentioned %s by %s" % \
+               (T.urltype(),T.firstmen(),T.nick)
+    if (public):
+      complaint=complaint+". Furthermore it defeats the point of this command to use it other than via /msg."
+      T.count+=1
+    bot.automsg(False,nick,complaint)
+    T.lastasked=time.time()
+  else:
+    if (public):
+      bot.automsg(False,nick,"That URL was unique. There is little point in using !url out loud; please use it via /msg")
+    else:
+      if urlstring != cmd.split()[1]: #first argument to URL was not the url
+        conn.privmsg(bot.channel,"%s remarks %s" % (nick," ".join(cmd.split()[1:])))
+      else:
+        conn.privmsg(bot.channel,"(via %s) %s"%(nick," ".join(cmd.split()[1:])))
+      bot.automsg(False,nick,"That URL was unique; I have posted it into IRC")
+    urldb[url]=UrlLog(url,nick)
+
+### Deal with URLs spotted in channel
+def dourl(bot,conn,nick,command,urldb):
+  urlstring=urlre.search(command).group(1)
+  urlstring=canonical_url(urlstring)
+
+  if urlstring in urldb:
+    T=urldb[urlstring]
+    message="observes %s URL, first mentioned %s by %s" % \
+             (T.urltype(),T.firstmen(),T.nick)
+    conn.action(bot.channel, message)
+    T.lastseen=time.time()
+    T.count+=1
+  else:
+    urldb[urlstring]=UrlLog(urlstring,nick)
+
+### Expire old urls
+def urlexpire(urldb,expire):
+    urls=urldb.keys()
+    for u in urls:
+        if time.time() - urldb[u].recenttime() > expire:
+            del urldb[u]
+
+# canonicalise BBC URLs (internal use only)
+def canonical_url(urlstring):
+  if (urlstring.find("news.bbc.co.uk") != -1):
+    for middle in ("/low/","/mobile/"):
+      x = urlstring.find(middle)
+      if (x != -1):
+        urlstring.replace(middle,"/hi/")
+  return urlstring
+