Merge branch 'jargonbot' of XPUB/federated-publishing-prototypes into master
onebigear
5 years ago
committed by
Gitea
commit
c72f94af62
@ -0,0 +1,65 @@
|
||||
#!/usr/bin/env python3
|
||||
import os, requests, urllib, subprocess
|
||||
from bs4 import BeautifulSoup
|
||||
|
||||
def pandoc (src, fro="markdown", to="html5"):
|
||||
# print ("[pandoc]", file=sys.stderr)
|
||||
p = subprocess.Popen(["pandoc", "--from", fro, "--to", to, "--section-divs"], stdin=subprocess.PIPE, stdout=subprocess.PIPE, stderr=subprocess.PIPE)
|
||||
stdout, stderr = p.communicate(src.encode("utf-8"))
|
||||
return stdout.decode("utf-8")
|
||||
|
||||
|
||||
def read_jargon_dictionary(url):
|
||||
data = urllib.request.urlopen(url).read() #bytes object
|
||||
# print('data', type(data), data)
|
||||
string = data.decode("utf-8")
|
||||
# print("string", type(string), string)
|
||||
# print ()
|
||||
# html = markdown(string)
|
||||
html = pandoc(string) # convert from markdown to html
|
||||
# print("html", html)
|
||||
|
||||
soup = BeautifulSoup(html, features="html5lib")
|
||||
# headers = soup.find_all('section', {"class": "level1"})
|
||||
# print("headers", headers)
|
||||
|
||||
defs = {} # creating empty dictionary
|
||||
for section in soup.find_all('section', {"class": "level1"}):
|
||||
h = section.find("h1")
|
||||
word = h.text # TCP
|
||||
d = section.find("p")
|
||||
if not d:
|
||||
d = section.find("ul")
|
||||
# print("found",h.text,d.text)
|
||||
#if d:
|
||||
#defition =
|
||||
#store the definition in the dict...
|
||||
# print ("i am in a loop with a variable header", type(header), header.text)
|
||||
defs[word.upper()] ={'name': word, 'text': d.text} # add to dictionary word(key): d.text (value)
|
||||
|
||||
|
||||
return defs
|
||||
|
||||
if __name__ == "__main__":
|
||||
# this only happens when you type python read.py
|
||||
# NOT when this file is imported
|
||||
url = "https://pad.xpub.nl/p/jargon-file.test/export/txt"
|
||||
defs = read_jargon_dictionary(url)
|
||||
print ("OUR JARGON DICTIONARY")
|
||||
from pprint import pprint
|
||||
pprint(defs)
|
||||
# print(defs["UDP"])
|
||||
|
||||
# cmd = '#TCP'
|
||||
# query = cmd.replace('#','')
|
||||
# for header in soup('h1'):
|
||||
# if query in header.string:
|
||||
# print(header.string)
|
||||
# print('---')
|
||||
|
||||
# cmd = '#UDP'
|
||||
# query = cmd.replace('#','')
|
||||
# for header in soup('h1'):
|
||||
# if query in header.string:
|
||||
# print(header.string)
|
||||
# print('---')
|
||||
@ -0,0 +1,190 @@
|
||||
#!/usr/bin/env python3
|
||||
# -*- coding: utf-8 -*-
|
||||
|
||||
"""
|
||||
Slixmpp: The Slick XMPP Library
|
||||
Copyright (C) 2010 Nathanael C. Fritz
|
||||
This file is part of Slixmpp.
|
||||
|
||||
See the file LICENSE for copying permission.
|
||||
"""
|
||||
|
||||
# Code source: https://git.poez.io/slixmpp/tree/examples/muc.py
|
||||
|
||||
# To run this bot:
|
||||
# $ python3 streambot.py --jid username@yourdomainname.ext --password password --room channel@groups.domainname.ext --nick nickname --output ./output/
|
||||
|
||||
|
||||
import logging
|
||||
from getpass import getpass
|
||||
from argparse import ArgumentParser
|
||||
|
||||
import slixmpp
|
||||
import ssl, os, requests, urllib, re
|
||||
|
||||
from jargon import *
|
||||
|
||||
class MUCBot(slixmpp.ClientXMPP):
|
||||
"""
|
||||
A simple Slixmpp bot that will save images
|
||||
and messages that are marked with #publish 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 streams 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):
|
||||
"""
|
||||
Process the session_start event.
|
||||
|
||||
Typical actions for the session_start event are
|
||||
requesting the roster and broadcasting an initial
|
||||
presence stanza.
|
||||
"""
|
||||
|
||||
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)
|
||||
|
||||
def muc_message(self, msg):
|
||||
"""
|
||||
Process incoming message stanzas from any chat room. Be aware
|
||||
that if you also have any handlers for the 'message' event,
|
||||
message stanzas may be processed by both handlers, so check
|
||||
the 'type' attribute when using a 'message' event handler.
|
||||
|
||||
Whenever the bot's nickname is mentioned, respond to
|
||||
the message.
|
||||
|
||||
IMPORTANT: Always check that a message is not from yourself,
|
||||
otherwise you will create an infinite loop responding
|
||||
to your own messages.
|
||||
|
||||
This handler will reply to messages that mention
|
||||
the bot's nickname.
|
||||
|
||||
Arguments:
|
||||
msg -- The received message stanza. See the documentation
|
||||
for stanza objects and the Message stanza to see
|
||||
how it may be used.
|
||||
"""
|
||||
|
||||
# Some inspection commands
|
||||
# print('Message:{}'.format(msg))
|
||||
# print('\nMessage TYPE:{}'.format(msg['type']))
|
||||
# print('\nMessage body:{}'.format(msg['body']))
|
||||
# print('\nMessage OOB:{}'.format(msg['oob']))
|
||||
# print('\nMessage OOB URL:{}'.format(msg['oob']['url']))
|
||||
# print('\nMessage MUCK NICK:{}'.format(msg['mucnick']))
|
||||
|
||||
# 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 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)
|
||||
|
||||
|
||||
# Include messages in the stream (only when '#publish' is used in the message)
|
||||
#if '#publish' in msg['body']:
|
||||
|
||||
# reply from the bot
|
||||
# print(defs.keys())
|
||||
|
||||
# body_list = re.findall('\w+', msg['body']) #.split(" ")
|
||||
# print(body_list)
|
||||
for key in defs.keys(): # for each word written by XMPP user
|
||||
if key in msg['body'].upper(): # check whether any word is a key of defs
|
||||
definition = defs[key]['text']
|
||||
name = defs[key]['name']
|
||||
print(name, definition)
|
||||
|
||||
self.send_message(mto=msg['from'].bare,
|
||||
mbody="{}: {}".format(name,definition),
|
||||
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", default="pe@please.undo.undo.it",
|
||||
help="JID to use")
|
||||
parser.add_argument("-p", "--password", dest="password", default="xpub",
|
||||
help="password to use")
|
||||
parser.add_argument("-r", "--room", dest="room", default="xpub@muc.vvvvvvaria.org",
|
||||
help="MUC room to join")
|
||||
parser.add_argument("-n", "--nick", dest="nick", default="jargonbot",
|
||||
help="MUC nickname")
|
||||
|
||||
# output folder for images
|
||||
parser.add_argument("-o", "--output", dest="output",
|
||||
help="output folder, this is where the files are stored",
|
||||
default="./output/", type=str)
|
||||
|
||||
args = parser.parse_args()
|
||||
|
||||
# get the jargon defs dictionary
|
||||
url = "https://pad.xpub.nl/p/jargon-file.test/export/txt"
|
||||
defs = read_jargon_dictionary(url)
|
||||
# Setup logging.
|
||||
logging.basicConfig(level=args.loglevel,
|
||||
format='%(levelname)-8s %(message)s')
|
||||
|
||||
if args.jid is None:
|
||||
args.jid = input("User: ")
|
||||
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()
|
||||
|
||||
Loading…
Reference in New Issue