forum_bot.py

import os
import re
import time
import json
import hashlib
import calendar
import requests
import threading
import html2text

from datetime import datetime, date
from slackclient import SlackClient
from BeautifulSoup import BeautifulSoup as Soup
from BeautifulSoup import NavigableString

SLACK_BOT_TOKEN = 'CHANGE_THIS'
BOT_ID = 'CHANGE_THIS'

# constants
AT_BOT = "<@" + BOT_ID + ">"
COMMANDS = [ 'silence', 'listen', 'ping' ]
SILENCED = False

# instantiate Slack & Twilio clients
slack_client = SlackClient(SLACK_BOT_TOKEN)

def handle_command(command, channel):
    response = "No such command found. For a list of commands, use 'commands'."

    if command.startswith('commands') or command.startswith('help'):
        response = "Available commands: " + ', '.join(COMMANDS)
    elif command.startswith('silence'):
        SILENCED = True
        response = "Silence mode is now enabled."
    elif command.startswith('listen'):
        SILENCED = False
        response = "Silence mode is now disabled."
    elif command.startswith('ping'):
        response = "Pong."

    slack_client.api_call("chat.postMessage", channel=channel,
                          text=response, as_user=True)

def parse_slack_output(slack_rtm_output):
    output_list = slack_rtm_output
    if output_list and len(output_list) > 0:
        for output in output_list:
            if output and 'text' in output and AT_BOT in output['text']:
                # return text after the @ mention, whitespace removed
                return output['text'].split(AT_BOT)[1].strip().lower(), \
                       output['channel']
    return None, None

def strip_html(html):
    soup = Soup(html)
    [div.extract() for div in soup.findAll('div')]
    [span.extract() for span in soup.findAll('span')]
    [br.extract() for br in soup.findAll('br')]

    result = ''

    try:
        result = html2text.html2text(str(soup))
    except:
        result = str(soup)
    re.sub(r'(\r\n){2,}','\r\n', result)

    return result
 
def parse_forums():
    READ_FORUM_FEED_DELAY = 10
    PREVIOUS_POSTS = []
    FULLY_PARSED = False

    while True:
        req = requests.request('GET', 'http://forums.everybodyedits.com/extern.php?action=feed&type=xml')
        soup = Soup(req.content)

        posts = []
        for message in soup.findAll('topic'):
            post_id = message.attrs[0][1]
            title = message.find('title').text
            author_name = message.find('author').find('name').text
            author_profile = message.find('author').find('uri').text
            timestamp = datetime.strptime(message.find('posted').text, "%a, %d %b %Y %H:%M:%S +0000")
            content = strip_html(message.find('content').text)
            digest = hashlib.sha1(str.format('{}{}{}', post_id, author_profile, timestamp)).hexdigest()
            
            post = { 'post_id': post_id, 'title': title, 'author_name': author_name, 'author_profile': author_profile,
                     'timestamp': timestamp, 'content': content, 'digest': digest }
            posts.append(post)

        if FULLY_PARSED:
            found_posts = []

            for first in posts:
                found = False

                for second in PREVIOUS_POSTS:
                    if first['digest'] == second['digest']:
                        found = True
                        break

                if not found:
                    found_posts.append(first)

            for post in found_posts:
                print(post)

                channel_list = slack_client.api_call(
                    "channels.list",
                    exclude_archived=1
                )

                attachment = {
                    "fallback": str.format('[EEForums] {} replied to {}...', post['title'], 
                                str.format("http://forums.everybodyedits.com/viewtopic.php?id={}&action=new", post['post_id'])),
                    "color": "#36a64f",
                    "author_name": post['author_name'],
                    "author_link": post['author_profile'],
                    "author_icon": str.format("http://forums.everybodyedits.com/img/avatars/{}.png", post['author_profile'].split("?id=",1)[1]),
                    "title": post['title'],
                    "title_link": str.format("http://forums.everybodyedits.com/viewtopic.php?id={}&action=new", post['post_id']),
                    "text": post['content'][:150] + (post['content'][150:] and '..'),
                    "footer": "Everybody Edits Forums",
                    "footer_icon": "https://forums.everybodyedits.com/favicon.png",
                    "ts": int(calendar.timegm(post['timestamp'].timetuple()))
                }

                for channel in channel_list['channels']:
                    if channel['is_member'] and not SILENCED:
                            slack_client.api_call("chat.postMessage", channel=channel['id'], attachments=json.dumps( [attachment] ), as_user=True)

            PREVIOUS_POSTS = posts
        else:
            PREVIOUS_POSTS = posts
            FULLY_PARSED = True

        time.sleep(READ_FORUM_FEED_DELAY)

if __name__ == "__main__":
    READ_WEBSOCKET_DELAY = 1

    forum_parse_thread = threading.Thread(target=parse_forums, args=())
    forum_parse_thread.daemon = True
    forum_parse_thread.start()

    if slack_client.rtm_connect():
        print("BeefBot connected and running!")

        while True:
            command, channel = parse_slack_output(slack_client.rtm_read())
            if command and channel:
                handle_command(command, channel)
            time.sleep(READ_WEBSOCKET_DELAY)
    else:
        print("Connection failed. Invalid Slack token or bot ID?")