X-Git-Url: http://www.chiark.greenend.org.uk/ucgi/~matthewv/git?p=irc.git;a=blobdiff_plain;f=commands.py;h=5a7cc94cdac1165910175f136cb1959afb9bd09b;hp=ec133f581869e488cd492a733582c24676997a75;hb=refs%2Fheads%2Fmaster;hpb=94aac51364d05d2f7c7521d781b6b6d381dbd0b0 diff --git a/commands.py b/commands.py index ec133f5..eec69e7 100755 --- a/commands.py +++ b/commands.py @@ -1,6 +1,13 @@ # Part of Acrobat. -import string, cPickle, random, urllib, sys, time, re, os, twitter, subprocess, datetime, urlparse +import string, cPickle, random, urllib, sys, time, re, os, twitter, subprocess, datetime, urlparse, hashlib +from collections import defaultdict from irclib import irc_lower, nm_to_n +import json + +try: + from blame_filter import bfd +except ImportError: + bfd = None # query karma def karmaq(bot, cmd, nick, conn, public, karma): @@ -47,6 +54,16 @@ def infoq(bot, cmd, nick, conn, public, karma): (bot.revision.split()[1], bot.channel, conn.get_nickname(), bot.owner, len(karma.keys()))) +class FishPond: + def __init__(fishpond): + fishpond.last=[] + fishpond.DoS=0 + fishpond.quotatime=0 + + def note_last(fishpond, msg, cfg): + fishpond.last.insert(0,(msg,cfg)) + fishpond.last = fishpond.last[0:10] + # Check on fish stocks def fish_quota(pond): if pond.DoS: @@ -87,7 +104,7 @@ def troutq(bot, cmd, nick, conn, public, cfg): return me = bot.connection.get_nickname() trout_msg = random.choice(fishlist) - fishpond.last=trout_msg + fishpond.note_last(trout_msg,cfg) # The bot won't trout or flirt with itself; if irc_lower(me) == irc_lower(target) or irc_lower(target) in synonyms: target = nick @@ -128,7 +145,7 @@ def slashq(bot, cmd, nick, conn, public, cfg): return me = bot.connection.get_nickname() slash_msg = random.choice(fishlist) - fishpond.last=slash_msg + fishpond.note_last(slash_msg,cfg) # The bot won't slash people with themselves if irc_lower(who[0]) == irc_lower(who[1]): conn.notice(nick, "oooooh no missus!") @@ -278,8 +295,12 @@ def __getcommitinfo(commit): return(err) ts,mes=out.split('|') + mes=mes.strip() + md5mes=hashlib.md5(mes).hexdigest() + if bfd and md5mes in bfd: + mes=bfd[md5mes] when=datetime.date.fromtimestamp(float(ts)) - return mes.strip(), when + return mes, when ###Return an array of commit messages and timestamps for lines in db that match what def __getcommits(db,keys,what): @@ -302,14 +323,30 @@ def __getall(tdb,tdbk,fdb,fdbk,sdb,sdbk,what): sans=__getcommits(sdb,sdbk,what) return tans+fans+sans -def blameq(bot,cmd,nick,conn,public,fish,tdb,tdbk,fdb,fdbk,sdb,sdbk): +def blameq(bot,cmd,nick,conn,public,fishpond,cfgs): + tdb,tdbk,x = cfgs[0][7] # urgh, magic, to support magic knowledge below + fdb,fdbk,x = cfgs[1][7] + sdb,sdbk,x = cfgs[2][7] clist=cmd.split() if len(clist) < 2: bot.automsg(public,nick,"Who or what do you want to blame?") return cwhat=' '.join(clist[2:]) + kindsfile = "fish?" if clist[1]=="#last": - ans=__getall(tdb,tdbk,fdb,fdbk,sdb,sdbk,fish.last) + try: + n = abs(int(clist[2]))-1 + if n < 0: raise ValueError + except IndexError: n = 0 + except ValueError: + bot.automsg(public,nick,"Huh?") + return + try: lmsg, lcfg = fishpond.last[n] + except IndexError: + bot.automsg(public,nick,"Nothing") + return + xdb,xdbk,kindsfile = lcfg[7] + ans=__getcommits(xdb,xdbk,lmsg) elif clist[1]=="#trouts" or clist[1]=="#trout": ans=__getcommits(tdb,tdbk,cwhat) elif clist[1]=="#flirts" or clist[1]=="#flirt": @@ -325,7 +362,7 @@ def blameq(bot,cmd,nick,conn,public,fish,tdb,tdbk,fdb,fdbk,sdb,sdbk): if len(ans[0])==1: bot.automsg(public,nick,ans[0]) else: - bot.automsg(public,nick,"Modified %s: %s" % (ans[0][2].isoformat(),ans[0][1])) + bot.automsg(public,nick,"Modified %s %s: %s" % (kindsfile, ans[0][2].isoformat(),ans[0][1])) elif len(ans)>4: bot.automsg(public,nick,"I found %d matches, which is too many. Please be more specific!" % (len(ans)) ) else: @@ -333,7 +370,7 @@ def blameq(bot,cmd,nick,conn,public,fish,tdb,tdbk,fdb,fdbk,sdb,sdbk): if len(a)==1: bot.automsg(public,nick,a) else: - bot.automsg(public,nick,"'%s' modified on %s: %s" % (a[0],a[2].isoformat(),a[1])) + bot.automsg(public,nick,"%s '%s' modified on %s: %s" % (kindsfile, a[0],a[2].isoformat(),a[1])) ### say to msg/channel def sayq(bot, cmd, nick, conn, public): @@ -392,23 +429,33 @@ class UrlLog: self.nick=nick self.url=url self.first=time.time() + self.localfirst=time.localtime(self.first) 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) + n=time.localtime(time.time()) + s="%02d:%02d" % (self.localfirst.tm_hour,self.localfirst.tm_min) + if n.tm_yday != self.localfirst.tm_yday: + s+=time.strftime(" on %d %B", self.localfirst) + return s def urltype(self): - z=min(len(urlcomplaints)-1, self.count-1) - return urlcomplaints[z] + z=min(len(urlinfos)-1, self.count-1) + return urlinfos[z] #(?:) is a regexp that doesn't group urlre = re.compile(r"((?:(?:http)|(?:nsfw))s?://[^ ]+)( |$)") hturlre= re.compile(r"(http)(s?://[^ ]+)( |$)") #matches \bre\:?\s+ before a regexp; (?i)==case insensitive match shibboleth = re.compile(r"(?i)\bre\:?\s+((?:(?:http)|(?:nsfw))s?://[^ ]+)( |$)") -urlcomplaints = ["a contemporary","an interesting","a fascinating","an overused","a vastly overused"] +#How long (in s) to wait since the most recent mention before commenting +url_repeat_time = 300 +urlinfos = ["a new", + "a fascinating", + "an interesting", + "a popular"] ### Deal with /msg bot url or ~url in channel def urlq(bot, cmd, nick, conn, public,urldb): @@ -420,12 +467,12 @@ def urlq(bot, cmd, nick, conn, public,urldb): 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) + comment="I saw that URL in scrool, first mentioned by %s at %s" % \ + (T.nick,T.firstmen()) if (public): - complaint=complaint+". Furthermore it defeats the point of this command to use it other than via /msg." + comment=comment+". Furthermore it defeats the point of this command to use it other than via /msg." T.count+=1 - bot.automsg(False,nick,complaint) + bot.automsg(False,nick,comment) T.lastasked=time.time() #URL suppressed, so mention in #urls if urlstring != cmd.split()[1]: #first argument to URL was not the url @@ -449,9 +496,10 @@ def dourl(bot,conn,nick,command,urldb): if urlstring in urldb: T=urldb[urlstring] - message="observes %s URL, first mentioned %s by %s" % \ - (T.urltype(),T.firstmen(),T.nick) - if shibboleth.search(command)==None: + message="saw that URL in scrool, first mentioned by %s at %s" % \ + (T.nick,T.firstmen()) + if shibboleth.search(command)==None and \ + time.time() - T.lastseen > url_repeat_time: conn.action(bot.channel, message) T.lastseen=time.time() T.count+=1 @@ -497,47 +545,136 @@ def twitterq(bot,cmd,nick,conn,public,twitapi): urlstring = urlre.search(cmd).group(1) if (urlstring.find("twitter.com") !=-1): - stringout = getTweet(urlstring,twitapi) - bot.automsg(public, nick, stringout) + stringsout = getTweet(urlstring,twitapi) + for stringout in stringsout: + bot.automsg(public, nick, stringout) -def getTweet(urlstring,twitapi): - parts = string.split(urlstring,'/') - tweetID = parts[-1] +def getTweet(urlstring,twitapi,inclusion=False,recurlvl=0): + unobfuscate_urls=True + expand_included_tweets=True + stringsout=[] + + path = urlparse.urlparse(urlstring).path + tweetID = path.split('/')[-1] try: status = twitapi.GetStatus(tweetID) if status == {}: return "twitapi.GetStatus returned nothing :-(" if status.user == None and status.text == None: return "Empty status object returned :(" + if status.retweeted_status and status.retweeted_status.text: + status = status.retweeted_status if status.user is not None: - tweeter_screen = status.user.screen_name.encode('UTF-8', 'replace') - tweeter_name = status.user.name.encode('UTF-8', 'replace') + tweeter_screen = status.user.screen_name #.encode('UTF-8', 'replace') + tweeter_name = status.user.name #.encode('UTF-8', 'replace') else: tweeter_screen = "[not returned]" ; tweeter_name = "[not returned]" - tweetText = status.text.encode('UTF-8', 'replace') - tweetText = tweetText.replace('\n',' ') + tweeter_name = tweeter_name + " RTing " + status.user.name #.encode('UTF-8', 'replace') + tweetText = status.full_text + if status.media: + replacements = defaultdict(list) + + for medium in status.media: + replacements[medium.url].append(medium.media_url_https) + + # The twitter-api 'conveniently' parses this for you and + # throws away the actual video URLs, so we have to take the + # JSON and reparse it :sadpanda: + # This is particularly annoying because we don't know + # for sure that status.media and the JSON 'media' entry + # have the same elements in the same order. Probably they + # do but maybe twitter-api randomly reorganised things or + # filtered the list or something. So instead we go through + # the JSON and handle the media urls, discarding whatever + # unfortunate thing we have put in replacements already. + parsed_tweet = json.loads(status.AsJsonString()) + for medium in parsed_tweet.get('media', []): + if medium['type'] == 'video': + best = { 'bitrate': -1 } + for vt in medium['video_info']['variants']: + if (vt.get('content_type') == 'video/mp4' and + vt.get('bitrate', -1) > best['bitrate']): + best = vt + if 'url' in best: + video_url = best['url'].split('?',1)[0] + duration = medium['video_info']['duration_millis'] + # ^ duration_millis is a string + duration = "%.1f" % (float(duration)/1000.) + video_desc = "%s (%ss)" % (video_url, duration) + replacements[medium['url']] = [video_desc] + + for k,v in replacements.items(): + if len(v) > 1: + replacementstring = "[" + " ; ".join(v) +"]" + else: + replacementstring = v[0] + tweetText = tweetText.replace(k, replacementstring) - for medium in status.media: - if "media_url_https" in medium: - tweetText = tweetText.replace(medium["url"], medium["media_url_https"]) for url in status.urls: toReplace = url.expanded_url + if unobfuscate_urls: + import urllib + rv = urlparse.urlparse(toReplace) + if rv.hostname in { + # sourced from http://bit.do/list-of-url-shorteners.php + "bit.do", "t.co", "lnkd.in", "db.tt", "qr.ae", "adf.ly", + "goo.gl", "bitly.com", "cur.lv", "tinyurl.com", "ow.ly", + "bit.ly", "adcrun.ch", "ity.im", "q.gs", "viralurl.com", + "is.gd", "po.st", "vur.me", "bc.vc", "twitthis.com", "u.to", + "j.mp", "buzurl.com", "cutt.us", "u.bb", "yourls.org", + "crisco.com", "x.co", "prettylinkpro.com", "viralurl.biz", + "adcraft.co", "virl.ws", "scrnch.me", "filoops.info", "vurl.bz", + "vzturl.com", "lemde.fr", "qr.net", "1url.com", "tweez.me", + "7vd.cn", "v.gd", "dft.ba", "aka.gr", "tr.im", + # added by ASB: + "trib.al", "dlvr.it" + }: + #expand list as needed. + response = urllib.urlopen('http://urlex.org/txt/' + toReplace) + resptext = response.read() + if resptext.startswith('http'): # ie it looks urlish (http or https) + if resptext != toReplace: + toReplace = resptext + # maybe make a note of the domain of the original URL to compile list of shortenable domains? + # remove tracking utm_ query parameters, for privacy and brevity # code snippet from https://gist.github.com/lepture/5997883 rv = urlparse.urlparse(toReplace) if rv.query: query = re.sub(r'utm_\w+=[^&]+&?', '', rv.query) - toReplace = '%s://%s%s?%s' % (rv.scheme, rv.hostname, rv.path, query) + if query: + toReplace = '%s://%s%s?%s' % (rv.scheme, rv.hostname, rv.path, query) + else: + toReplace = '%s://%s%s' % (rv.scheme, rv.hostname, rv.path) # leave off the final '?' + + if expand_included_tweets: + if rv.hostname == 'twitter.com' and re.search(r'status/\d+',rv.path): + if recurlvl > 2: + stringsout = [ "{{ Recursion level too high }}" ] + stringsout + else: + quotedtweet = getTweet(toReplace, twitapi, inclusion=True, recurlvl=recurlvl+1) # inclusion parameter limits recursion. + if not quotedtweet: + quotedtweet = [""] + quotedtweet[0] = "Q{ " + quotedtweet[0] + quotedtweet[-1] += " }" + stringsout = quotedtweet + stringsout tweetText = tweetText.replace(url.url, toReplace) tweetText = tweetText.replace(">",">") tweetText = tweetText.replace("<","<") tweetText = tweetText.replace("&","&") - + tweetText = tweetText.replace("\n"," ") stringout = "tweet by %s (%s): %s" %(tweeter_screen,tweeter_name,tweetText) except twitter.TwitterError: terror = sys.exc_info() stringout = "Twitter error: %s" % terror[1].__str__() - return stringout + except Exception: + terror = sys.exc_info() + stringout = "Error: %s" % terror[1].__str__() + stringsout = [stringout] + stringsout + if inclusion: + return stringsout # don't want to double-encode it, so just pass it on for now and encode later + + return map(lambda x: x.encode('UTF-8', 'replace'), stringsout)