wantzel.py 17 KB
Newer Older
1 2
#encoding: utf-8
"""
Mindiell's avatar
Mindiell committed
3
Bot Wantzel from La Quadrature du Net.
4

Mindiell's avatar
Mindiell committed
5 6 7 8
License : AGPLv3
Doc     : https://wiki.laquadrature.net/Wantzel

TODO:
9 10 11 12 13
- Ajouter des commandes permettant de gérer le mediakit (moderators only)
    - Parser les urls et voir ce qu'on peut faire:
        - upload de vidéo
        - tag de vidéo
        - obtenir un lien vers une vidéo
14 15
"""

16
import feedparser
17
import HTMLParser
18
import importlib
19 20 21
from irc import IrcClientFactory
import MySQLdb
import re
22 23
import sqlite3
import time
24
from twisted.internet import reactor
25
from twitter import Twitter, OAuth
26 27 28 29 30
import urllib

import config
from messages import messages

Mindiell's avatar
Mindiell committed
31 32
def get_cursor():
    """
33
    This function connects to a MySQL database and returns a usable cursor.
Mindiell's avatar
Mindiell committed
34
    """
35
    connection = MySQLdb.connect(
Mindiell's avatar
Mindiell committed
36 37 38 39 40
        host=config.dbserver,
        user=config.dbuser,
        passwd=config.dbpassword,
        db=config.dbname
    )
41 42
    if connection:
        return connection.cursor()
Mindiell's avatar
Mindiell committed
43 44
    return None

45
def get_url(message):
Mindiell's avatar
Mindiell committed
46
    """
47
    Retrieve the url in the message.
Mindiell's avatar
Mindiell committed
48
    """
49 50
    # Let's get the url
    result = re.search("(https?[^ ]+)", message)
Mindiell's avatar
Mindiell committed
51 52
    if not result:
        return "http"
53
    url = result.group(1)
Mindiell's avatar
Mindiell committed
54 55 56 57 58 59 60 61
    # Removing anchor if needed
    result = re.search("^([^#]*)", url)
    if result:
        url = result.group(1)
    # Removing trackers
    url = re.sub("[?&](utm_medium|utm_source|utm_campaign|xtor)=[^&]*", "", url)
    return url

62 63 64
def get_title(message):
    title = ""
    website = ""
65 66 67 68
    exceptions = [
        "twitter.com",
        "github.com",
    ]
69
    try:
70 71 72 73
        url = get_url(message)
        for exception in exceptions:
            if exception in url:
                return ("", "")
74 75 76 77 78 79
        website = re.search("//([^/]*)", url).group(1)
        f = urllib.URLopener().open(url)
        content = f.read()
        title = re.search("<title>([^<]+)</title>", content).group(1)
    except:
        pass
80 81
    # Unescaping HTML entities and removing multiple lines
    title = HTMLParser.HTMLParser().unescape(re.sub("\n|\r", "", title))
82 83
    return (title, website)

84 85 86 87 88 89 90 91 92 93 94 95
def is_moderator(name):
    """
    This function verify if a user is a moderator.
    """
    connection = sqlite3.connect(config.sqlite_db)
    cursor = connection.cursor()
    cursor.execute("SELECT count(*) FROM moderator WHERE name=?", (name, ))
    if int(cursor.fetchone()[0])==1:
        return True
    return False


96 97 98 99 100 101 102 103
class Wantzel(object):
    """
    Wantzel bot.
    """
    def __init__(self):
        """
        Initialization of bot over IRC.
        """
104 105 106 107 108 109 110 111 112 113
        self.number = 0
        # default last_entry_published
        self.last_entry_published = time.strptime("2000-01-01", "%Y-%m-%d")
        # See if there is something in the db
        connection = sqlite3.connect(config.sqlite_db)
        for row in connection.execute("SELECT last_entry_published FROM tweets"):
            self.last_entry_published = time.strptime(
                row[0].encode("utf-8"),
                "%Y-%m-%d %H:%M:%S %Z"
            )
114 115 116
        self.irc = IrcClientFactory(config)
        self.irc.set_privmsg = self.set_privmsg
        reactor.connectTCP(config.server, config.port, self.irc)
117 118 119 120 121 122 123 124 125 126 127 128 129
        # Prepare timer
        reactor.callLater(config.timer, self.timer)

    def timer(self):
        """
        This method launches function regularly (see config.timer).
        """
        print("Timer called")
        self.rp_to_twitter("http://www.laquadrature.net/fr/revue-de-presse/feed")
        self.rp_to_twitter("http://www.laquadrature.net/en/press-review/feed")
        self.count_articles()
        # Recalling the timer
        reactor.callLater(config.timer, self.timer)
130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145

    def set_privmsg(self):
        """
        This method set the methods to call for each callback received from IRC.
        """
        self.irc.client.privmsg = self.on_privmsg

    def send_message(self, channel, messages):
        """
        Sends a message on specified channel, cutting each line in a new message
        """
        for message in messages.splitlines():
            self.irc.client.msg(channel, message)

    def on_privmsg(self, user, channel, msg):
        """
146 147
        Wantzel can understand a lot of commands. Commands followed by a (*)
        are accessible only to moderators:
148 149 150 151 152
        - help
            Returns a message about how to use the bot.
            If a command is passed after help, the message explains how to use
            the command. 
        - rp(acp)
Mindiell's avatar
Mindiell committed
153
            Add an article in the database
154
        - stats
Mindiell's avatar
Mindiell committed
155
            Show some statistics about the RP
156
        - kill (*)
Mindiell's avatar
Mindiell committed
157
            Kill an article by giving it a score of -100
158 159 160 161 162 163
        - moderate list (*)
            List rights in private
        - moderate add (*)
            Add a new moderator to list
        - moderate remove (*)
            Remove a moderator from list
164 165
        """
        # Cleaning user name
Mindiell's avatar
Mindiell committed
166
        user = re.search("([^!]*)!", user).group(1)
Mindiell's avatar
Mindiell committed
167
        #print("Message received: %s %s %s" % (user, channel, msg))
168 169
        # Whatever is done, get the title of an existing url in a message
        title = ""
170
        if "http" in msg:
171
            title, website = get_title(msg)
Mindiell's avatar
Mindiell committed
172 173 174 175 176 177 178
        # Never answer to botself
        if user!=config.nickname:
            # If it's a query, bot should answer to the user as the channel
            if "#" not in channel:
                channel = user
            # Help command, specific
            if "wantzel" in msg and ("help" in msg or "aide" in msg):
179
                self.help(user, channel, msg)
Mindiell's avatar
Mindiell committed
180
            # Find known command
181
            command = re.search("!(rp[acp]*|kill|help|stats|admin)", msg)
Mindiell's avatar
Mindiell committed
182 183 184 185
            if command:
                command = command.group(1)
                print("Command: %s" % command)
                if command.startswith("rp"):
186
                    self.rp(command, user, channel, msg, title)
Mindiell's avatar
Mindiell committed
187 188 189 190 191 192
                elif command=="help":
                    self.help(user, channel, msg)
                elif command=="kill":
                    self.kill(user, channel, msg)
                elif command=="stats":
                    self.stats(user, channel, msg)
193 194
                elif command=="admin":
                    self.admin(user, channel, msg)
195 196
        if title and website:
            self.send_message(channel, messages["title"] % (title, website))
197 198 199 200 201 202 203 204

    def help(self, user, channel, msg):
        """
        Show global help.
        If a known command is behind the !help command, an adequate message is
        returned.
        """
        print("help command")
Mindiell's avatar
Mindiell committed
205
        # Searching for a command after help keyword
206
        command = re.search("!help (stats|rp|help|kill|admin)", msg)
Mindiell's avatar
Mindiell committed
207 208
        if command:
            command = command.group(1)
209
            self.send_message(channel, messages["help_"+command])            
Mindiell's avatar
Mindiell committed
210 211
        else:
            self.send_message(channel, messages["help"])
212

213
    def rp(self, command, user, channel, msg, title=""):
214 215 216 217 218 219
        """
        Adding the article in rp database.
        """
        print("rp command %s" % command)
        cite = 0
        note = 0
220
        url = get_url(msg)
221
        print("url: %s" % url)
Mindiell's avatar
Mindiell committed
222 223 224 225 226
        if url=="":
            return
        elif url=="http":
            self.send_message(channel, messages["rp_http"] % user)
            return
227

Mindiell's avatar
Mindiell committed
228 229
        # Looking for such an article in database
        cursor = get_cursor()
230 231 232 233 234 235 236 237 238 239 240 241
        cursor.execute("SELECT id, note, provenance FROM presse WHERE url = %s", (url, ))
        rows = cursor.fetchall()
        if not rows:
            # LQdN is quoted
            if "c" in command:
                cite += 2
            # the article speak about LQdN
            if command.count("p")>1:
                cite += 2
            # Archive this article
            if "a" in command:
                note -= 2
Mindiell's avatar
Mindiell committed
242
            print("Adding an article by %s: %s" % (user, url))
243
            result = cursor.execute(
244 245 246 247 248
                """INSERT INTO presse SET 
                url=%s, provenance=%s, cite=%s, note=%s, datec=NOW(), title=%s,
                lang='', published=0, nid=0, screenshot=0, fetched=0, seemscite=0
                """,
                (url, user, cite, note, title)
249
            )
Mindiell's avatar
Mindiell committed
250
            self.send_message(channel, messages["rp_new_article"] % user)
251 252
        else:
            if rows[0][2]!=user:
Mindiell's avatar
Mindiell committed
253
                print("Adding a point by %s on %s" % (user, rows[0][0]))
254 255 256 257 258
                result = cursor.execute(
                    "UPDATE presse SET note=note+1 WHERE id=%s",
                    (rows[0][0], )
                )
            if (rows[0][1]+1)<3:
Mindiell's avatar
Mindiell committed
259
                self.send_message(channel, messages["rp_known_article"] % user)
260
            else:
Mindiell's avatar
Mindiell committed
261
                self.send_message(channel, messages["rp_taken_article"] % user)
262 263
                # Update number of articles to do
                self.count_articles()
264 265 266

    def kill(self, user, channel, msg):
        """
Mindiell's avatar
Mindiell committed
267
        Kill an article by setting its score to -100.
268 269
        """
        print("kill command")
270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286
        if is_moderator(user):
            url = get_url(msg)
            print("url: %s" % url)
            if url=="":
                return
            elif url=="http":
                self.send_message(channel, messages["rp_http"] % user)
                return
            # Looking for such an article in database
            cursor = get_cursor()
            cursor.execute("SELECT id, note FROM presse WHERE url=%s", (url, ))
            rows = cursor.fetchall()
            if not rows:
                self.send_message(channel, messages["kill_none"] % url)
            else:
                cursor.execute("UPDATE presse SET note=-100 WHERE id=%s", (rows[0][0], ))
                self.send_message(channel, messages["kill_done"] % url)
Mindiell's avatar
Mindiell committed
287
        else:
288
            self.send_message(channel, messages["not_moderator"])
289 290 291 292 293 294

    def stats(self, user, channel, msg):
        """
        Returns stats on articles in press review.
        """
        print("stats command")
295
        cursor = get_cursor()
296 297 298 299 300 301 302
        periods = [1, 3, 7, 15]
        notes = [0, 3 ,4]
        notnull = 0
        somethingatall = 0
        result = ""
        for note in notes:
            notnull = 0
Mindiell's avatar
Mindiell committed
303
            result = result + "note>=%s: " % note
304 305 306 307 308 309 310 311 312 313 314 315
            for period in periods:
                cursor.execute(
                    """SELECT COUNT(id) AS cid FROM presse 
                    WHERE nid=0
                    AND datec>(NOW()-INTERVAL %s DAY)
                    AND note>=%s""",
                    (period, note)
                )
                rows = cursor.fetchall()
                if rows[0][0]>0:
                    result = result + "%sj:%s, " % (period, rows[0][0])
                    notnull = 1
Mindiell's avatar
Mindiell committed
316
                    somethingatall = 1
317 318
            if notnull:
                result = result[:-2] + "\n"
Mindiell's avatar
Mindiell committed
319 320
        if somethingatall==0:
            result = messages["stats_bravo"] % periods[-1]
321 322
        self.send_message(channel, result)

323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397
    def admin(self, user, channel, msg):
        """
        Manage moderation.
        A sub-command should be behind the !admin command.
        """
        print("admin command")
        # Searching for a command after admin keyword
        command = re.search("!admin (list|add|del)", msg)
        if command:
            command = command.group(1)
            if command=="list":
                self.admin_list(user, channel, msg)
            elif command=="add":
                self.admin_add(user, channel, msg)
            elif command=="del":
                self.admin_del(user, channel, msg)

    def admin_list(self, user, channel, msg):
        """
        List actual moderators.
        """
        print("admin_list command")
        if is_moderator(user):
            connection = sqlite3.connect(config.sqlite_db)
            names = []
            for row in connection.execute("SELECT name FROM moderator"):
                names.append(row[0].encode("utf-8"))
            self.send_message(channel, messages["admin_list"] % ", ".join(names))
        else:
            self.send_message(channel, messages["not_moderator"])

    def admin_add(self, user, channel, msg):
        """
        Add some new moderators if not existing yet.
        """
        print("admin_add command")
        if is_moderator(user):
            try:
                names = []
                connection = sqlite3.connect(config.sqlite_db)
                result = re.search("!admin add (([^,]+, ?)+)?(.*)", msg)
                if result.group(1):
                    names = [name for name in result.group(1).split(", ") if name!=""]
                names.append(result.group(3))
                # Do not add actual moderators
                moderators = []
                for row in connection.execute("SELECT name FROM moderator"):
                    moderators.append(row[0].encode("utf-8"))
                names = set([name for name in names if name not in moderators])
                if names:
                    # Converting set in list of tuples
                    values = [(name,) for name in names]
                    connection.executemany("INSERT INTO moderator (name) VALUES (?)", values)
                    connection.commit()
                    self.send_message(channel, messages["admin_add"] % ", ".join(names))
                else:
                    self.send_message(channel, messages["admin_add_empty"])
            except:
                pass
        else:
            self.send_message(channel, messages["not_moderator"])

    def admin_del(self, user, channel, msg):
        """
        Delete a moderator from list.
        """
        print("admin_del command")
        if is_moderator(user):
            try:
                names = []
                result = re.search("!admin del (([^,]+, ?)+)?(.*)", msg)
                if result.group(1):
                    names = [name for name in result.group(1).split(", ") if name!=""]
                names.append(result.group(3))
                names = set(names)
Mindiell's avatar
Mindiell committed
398
                #print(names)
399 400 401 402 403 404 405 406 407 408 409 410 411 412 413
                connection = sqlite3.connect(config.sqlite_db)
                for name in names:
                    connection.execute("DELETE FROM moderator WHERE name=?", (name, ))
                connection.commit()
                self.send_message(channel, messages["admin_del"] % ", ".join(names))
            except:
                pass
        else:
            self.send_message(channel, messages["not_moderator"])

    def count_articles(self):
        """
        Count number of articles not done in RP and updates the topic of the
        press review channel if needed.
        """
Mindiell's avatar
Mindiell committed
414
        print("count_articles method")
415 416 417 418 419 420 421 422
        cursor = get_cursor()
        cursor.execute("""SELECT COUNT(*) FROM presse
            WHERE DATE_SUB(NOW(), INTERVAL 2 MONTH)<datec
            AND note > 2
            AND nid = 0""")
        rows = cursor.fetchall()
        number = int(rows[0][0])
        if self.number!=number:
Mindiell's avatar
Mindiell committed
423
            self.irc.client.topic("#lqdn-rp", messages["topic"] % number)
424 425 426 427 428 429 430 431 432 433 434 435 436
            pass
        self.number = number

    def rp_to_twitter(self, rss):
        """
        By parsing the RSS feed of the press-review, we know what to tweet.
        """
        print("rp_to_twitter method")
        now = time.localtime()
        today = time.strptime("%s-%s-%s %s" % (
                now.tm_year,
                now.tm_mon,
                now.tm_mday,
Mindiell's avatar
Mindiell committed
437
                time.tzname[0]
438 439 440
            ),
            "%Y-%m-%d %Z"
        )
Mindiell's avatar
Mindiell committed
441 442 443
        language = "fr"
        if "/en/" in rss:
            language = "en"
444 445 446 447 448 449 450
        entries = feedparser.parse(rss)['entries']
        entries.reverse()
        for entry in entries:
            # if date of publication is greater than today, midnight, and
            # lesser than future
            if today < entry.published_parsed < now:
                if self.last_entry_published < entry.published_parsed:
Mindiell's avatar
Mindiell committed
451
                    self.tweet(messages["tweet_rp_%s" % language] % (
452 453 454
                        entry.title.encode("utf-8"),
                        entry.link.encode("utf-8")
                    ))
Mindiell's avatar
Mindiell committed
455 456
                    #print(entry.published_parsed)
                    #print(entry.title)
457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484
                    # Save last_entry_published
                    self.last_entry_published = entry.published_parsed
                    last_entry_published = time.strftime(
                        "%Y-%m-%d %H:%M:%S %Z",
                        self.last_entry_published
                    )
                    connection = sqlite3.connect(config.sqlite_db)
                    connection.execute(
                        "UPDATE tweets SET last_entry_published=?",
                        (last_entry_published,)
                    )
                    connection.commit()
                    # Tweet only one message in order not to spam
                    return

    def tweet(self, message):
        """
        Tweet message on specified account
        """
        print("tweet method")
        auth = OAuth(
            config.TOKEN,
            config.TOKENSEC,
            config.CONSKEY,
            config.CONSSEC
        )
        twitter = Twitter(auth=auth)
        try:
Mindiell's avatar
Mindiell committed
485 486
            print("Tweeting: %s" % message)
            twitter.statuses.update(status=message)
487 488 489
        except:
            pass

490 491 492 493

if __name__ == '__main__':
    wantzel = Wantzel()
    reactor.run()