logbot/logbot.py

#!/usr/bin/env python3
# -*- coding: utf-8 -*-

# To run this bot:
# $ python3 logbot.py
# The output folder of this bot currently is: /var/www/logs/digital-autonomy

import logging
from getpass import getpass
from argparse import ArgumentParser

import slixmpp
import ssl, os, requests, urllib

class MUCBot(slixmpp.ClientXMPP):
	"""
	A simple Slixmpp bot that will save images
	and messages that are marked with @bot to a folder.
	"""

	def __init__(self, jid, password, room, nick, output):
		slixmpp.ClientXMPP.__init__(self, jid, password)

		self.room = room
		self.nick = nick
		self.output = output

		# The session_start event will be triggered when
		# the bot establishes its connection with the server
		# and the XML logs are ready for use. We want to
		# listen for this event so that we we can initialize
		# our roster.
		self.add_event_handler("session_start", self.start)

		# The groupchat_message event is triggered whenever a message
		# stanza is received from any chat room. If you also also
		# register a handler for the 'message' event, MUC messages
		# will be processed by both handlers.
		self.add_event_handler("groupchat_message", self.muc_message)


	def start(self, event):
		self.get_roster()
		self.send_presence()

		# https://xmpp.org/extensions/xep-0045.html
		self.plugin['xep_0045'].join_muc(self.room,
										 self.nick,
										 # If a room password is needed, use:
										 # password=the_room_password,
										 wait=True)

		# Send a message to the room
		self.send_message(mto=self.room, mbody='Hello! Logbot here. I\'m here to log all the images that are send to this group. You can also log text messages, by including @bot in your message. Happy logging! PS. you can access the logs at https://vvvvvvaria.org/logs/', mtype='groupchat')

	def muc_message(self, msg):
		# Some inspection commands
		#print('Message: {}'.format(msg))

		# Always check that a message is not the bot itself, otherwise you will create an infinite loop responding to your own messages.
		if msg['mucnick'] != self.nick:

			# Check if output folder exists
			if not os.path.exists(self.output):
				os.mkdir(self.output)

			# Check if an OOB URL is included in the stanza (which is how an image is sent)
			# (OOB object - https://xmpp.org/extensions/xep-0066.html#x-oob)
			if len(msg['oob']['url']) > 0:

				# Send a reply
				self.send_message(mto=self.room,
								  mbody="Super, our log is growing. Your image is added!",
								  mtype='groupchat')

				# Save the image to the output folder
				url = msg['oob']['url'] # grep the url in the message
				filename = os.path.basename(url) # grep the filename in the url
				output_path = os.path.join(self.output, filename)
				u = urllib.request.urlopen(url) # read the image data
				f = open(output_path, 'wb') # open the output file
				f.write(u.read()) # write image to file
				f.close() # close the output file

				# Add image to log
				img = '<img class="image" src="{}">'.format(msg['oob']['url'])
				log = 'log.html'
				log_path = os.path.join(self.output, log)
				f = open(log_path, 'a+')
				f.write(img+'\n')
				f.close()


			# Include messages in the log (only when '#publish' is used in the message)
			if '@bot' in msg['body']:

				# reply from the bot
				self.send_message(mto=self.room,
								  mbody="Noted! And added to the log. Thanks {}!".format(msg['mucnick']),
								  mtype='groupchat')

				# Add message to log
				message = '<p class="message">{}</p>'.format(msg['body'].replace('@bot',''))
				log = 'log.html'
				log_path = os.path.join(self.output, log)
				f = open(log_path, 'a+')
				f.write(message+'\n')
				f.close()

				if '/book' in msg['body']: # Check if this is a book

					self.send_message(mto=self.room,
									  mbody="Oh a book, that's cool! Thanks {}!".format(msg['mucnick']),
									  mtype='groupchat')


				from bs4 import BeautifulSoup
				import requests
				import re

				book = msg['body'].replace('@bot', '').replace('/book', '')
				book = re.sub(' +', ' ', book) # remove double spaces
				book = book.lstrip().rstrip() # remove spaces at the beginning and at the end
				book = book.replace(' ', '+').lower() # turn space into + and lowercase

				page_link = 'https://www.worldcat.org/search?q={}&qt=results_page'.format(book)

				page_response = requests.get(page_link, timeout=5)

				page_content = BeautifulSoup(page_response.content, "html.parser")

				try:
					book_title = page_content.findAll("div", {"class": "name"})[0].text
					book_author = page_content.findAll("div", {"class": "author"})[0].text
					book_publisher = page_content.findAll("div", {"class": "publisher"})[0].text

					response = '<b>BOOK</b>: ' + book_title + ' ' + book_author + ' ' + book_publisher

					book_found = True

				except IndexError:

					book_found = False

				if book_found:

					# Add message to log
					message = '<b>BOOK</b>: ' + book_title + ' ' + book_author + ' ' + book_publisher
					log = 'log.html'
					log_path = os.path.join(self.output, log)
					f = open(log_path, 'a+')
					f.write(message+'\n')
					f.close()

					self.send_message(mto=self.room, mbody='Hope this was the book you were looking for: ' + book_title + ' ' + book_author + ' ' + book_publisher, mtype='groupchat')

				else:

					self.send_message(mto=self.room, mbody='Sorry, no book found!', mtype='groupchat')




if __name__ == '__main__':
	# Setup the command line arguments.
	parser = ArgumentParser()

	# output verbosity options.
	parser.add_argument("-q", "--quiet", help="set logging to ERROR",
						action="store_const", dest="loglevel",
						const=logging.ERROR, default=logging.INFO)
	parser.add_argument("-d", "--debug", help="set logging to DEBUG",
						action="store_const", dest="loglevel",
						const=logging.DEBUG, default=logging.INFO)

	# JID and password options.
	parser.add_argument("-j", "--jid", dest="jid",
						help="JID to use")
	parser.add_argument("-p", "--password", dest="password",
						help="password to use")
	parser.add_argument("-r", "--room", dest="room",
						help="MUC room to join")
	parser.add_argument("-n", "--nick", dest="nick",
						help="MUC nickname")

	# output folder for images
	parser.add_argument("-o", "--output", dest="output",
						help="output folder, this is where the files are stored",
						type=str)

	args = parser.parse_args()

	# Setup logging.
	logging.basicConfig(level=args.loglevel,
						format='%(levelname)-8s %(message)s')

	if args.jid is None:
		args.jid = input("XMPP address: ")
	if args.password is None:
		args.password = getpass("Password: ")
	if args.room is None:
		args.room = input("MUC room: ")
	if args.nick is None:
		args.nick = input("MUC nickname: ")
	if args.output is None:
		args.output = input("Output folder: ")

	# Setup the MUCBot and register plugins. Note that while plugins may
	# have interdependencies, the order in which you register them does
	# not matter.
	xmpp = MUCBot(args.jid, args.password, args.room, args.nick, args.output)
	xmpp.register_plugin('xep_0030') # Service Discovery
	xmpp.register_plugin('xep_0045') # Multi-User Chat
	xmpp.register_plugin('xep_0199') # XMPP Ping
	xmpp.register_plugin('xep_0066') # Process URI's (files, images)

	# Connect to the XMPP server and start processing XMPP stanzas.
	xmpp.connect()
	xmpp.process()