diff --git a/bannedusernames.txt b/bannedusernames.txt new file mode 100644 index 0000000..5e6500b --- /dev/null +++ b/bannedusernames.txt @@ -0,0 +1 @@ +Бэдыч diff --git a/bannedwords.txt b/bannedwords.txt new file mode 100644 index 0000000..e69de29 diff --git a/helpers.py b/helpers.py new file mode 100644 index 0000000..1e347f5 --- /dev/null +++ b/helpers.py @@ -0,0 +1,10 @@ +class RID(object): + def __init__(self): + self._id = 0 + @property + def x(self): + self._id += 1 + return self._id + @x.setter + def x(self, value): + self._id = value diff --git a/modbot.py b/modbot.py new file mode 100644 index 0000000..bacb8fd --- /dev/null +++ b/modbot.py @@ -0,0 +1,592 @@ +# Tested on Python 3.10 +# Reqs: +# python -m pip install requests + +from datetime import datetime, timedelta +from threading import Thread +from time import sleep +import requests +import re +import base64 +import uuid +import random +import settings +import xmltodict +import queue +import os +from helpers import RID + +class ModBot: + + def __init__(self, instance, username, password, chatroom) -> None: + self.peertube_instance = instance + self.peertube_username = username + self.peertube_password = password + self.chat_room = chatroom + + self._prefixed_instance = f"https://{instance}" + self.rid = RID() + self.uid = RID() + self.users = {} + + self.msg_history = [] + self.msg_history_max_len = 500 + + self.restricted_mode = 0 + + # Timeout length, if !timeout command was send without 2nd argument + self.timeout_default = 300 + + # Rate limit settings + # Time window (in seconds) to measure rates + self.ratelimit_timewindow = 10 + # How many messages one user allowed to send in 'ratelimit_timewindow' seconds + self.ratelimit_maxmessages = 3 + # Length of timeout (in seconds) that will be applied to user who exceeded the limits + self.ratelimit_timeout = 60 + + self.resource_pref = "modbot.py" + self.bot_msg_prefix = "/me [ModBot]: " + self.last_msg = {"username": "", "count": 0} + + self.msg_queue = queue.Queue() + + self.banned_usernames = [] + self._bannedusernames_filename = "bannedusernames.txt" + self._bannedusernames_stamp = 0 + + self.banned_words = [] + self._bannedwords_filename = "bannedwords.txt" + self._bannedwords_stamp = 0 + + def _bosh_send(self, str): + req = requests.post(f'{self._prefixed_instance}{self.bosh_service_url}', + headers={'Content-Type': f'text/xml; charset=utf-8'}, timeout=60, data=str.encode('utf-8')) + #print(f"REQUEST: {str}\nRESPONSE: {req.text}\n") + return req.text + + def _generate_body_headers(self): + return f'rid="{self.rid.x}" sid="{self.authid}" xmlns="http://jabber.org/protocol/httpbind"' + + def update_user(self, presenses): + if isinstance(presenses, dict): + presenses = [presenses, ] + + for presense in presenses: + username = presense["@from"].split("/", 1)[1] + # If you are getting error HERE, your account... got banned from your room??? + x_items = presense["x"]["item"] + if isinstance(x_items, dict): + x_items = [x_items, ] + + for x_item in x_items: + + + userdata = { + "role": x_item["@role"], + "affiliation": x_item["@affiliation"], + "jid": x_item["@jid"].split("/", 1)[0], + } + + new_user = False + + if "@nick" in x_item: + for prev_username, prev_user in self.users.items(): + if prev_user["jid"] == userdata["jid"]: + newnick = x_item['@nick'] + self.users[newnick] = self.users[prev_username] + del self.users[prev_username] + user = self.users[newnick] + print(f"[{datetime.now().strftime('%H:%M:%S')}] ({user['uid']}|{prev_username}|{user['jid']} changed nickname to {newnick}", end="", flush=True) + + if self.users[newnick]["role"] == "visitor": + print(f" and was muted too)") + self.bosh_user_mute(newnick) + else: print(")") + break + else: + if username in self.users: + self.users[username]["role"] = userdata["role"] + self.users[username]["affiliation"] = userdata["affiliation"] + self.users[username]["jid"] = userdata["jid"] + else: + self.users[username] = userdata + self.users[username]["uid"] = self.uid.x + new_user = True + + if not userdata["role"] in ("visitor", "none"): + + if new_user: + if self.restricted_mode == 1 and "@anon." in userdata['jid']: + self.bosh_user_mute(username) + + if self.restricted_mode == 2 and "@anon." in userdata['jid']: + self.bosh_user_mute(username) + elif self.restricted_mode == 3 and userdata["role"] != "moderator": + self.bosh_user_mute(username) + + for banned_username in self.banned_usernames: + if banned_username in username: + print(f"User '{username}' was muted: contains banned username '{banned_username}'") + self.bosh_user_mute(username) + + def send_msg(self, msg, cmd=False): + if cmd == False: + someid = uuid.uuid4() + req_body = f'{self.bot_msg_prefix}{msg}' + return self._bosh_send(req_body) + else: + print(msg) + + def connect(self): + _pi = self._prefixed_instance + _room = self.chat_room + _inst = self.peertube_instance + + print(f'Peertube: Getting client tokens...') + req = requests.get(f'{_pi}/api/v1/oauth-clients/local') + client_tokens = req.json() + + print(f'Peertube: Authorizing and getting user token...') + req = requests.post(f'{_pi}/api/v1/users/token', data={ + "client_id": client_tokens["client_id"], + "client_secret": client_tokens["client_secret"], + "grant_type": "password", + "response_type": "code", + "username": self.peertube_username, + "password": self.peertube_password, + }) + token_resp = req.json() + if "access_token" in token_resp: + access_token = req.json()["access_token"] + else: + print(f"Error while logging in: {token_resp}") + exit(1) + + print(f"Webchat: Accessing room {_room} at instance {_inst}") + req = requests.get(f'{_pi}/plugins/livechat/router/webchat/room/{_room}') + + # Getting errors there? Does the room even exist? + self.livechat_version = re.search( + r"/livechat/(.*)/static/", req.text, re.MULTILINE).group(1) + print(f'Webchat: Livechat version: {self.livechat_version}') + + self.bosh_service_url = re.search(r"boshServiceUrl: '(.*)',", + req.text, re.MULTILINE).group(1) + self.authentication_url = re.search(r"authenticationUrl: '(.*)',", + req.text, re.MULTILINE).group(1).removeprefix(_pi) + + headers = {'authorization': f'Bearer {access_token}'} + req = requests.get(f'{_pi}{self.authentication_url}', headers=headers) + creds = {} + if (req.status_code == 200): + creds = req.json() + print(f"Webchat: Signed in as '{creds['nickname']}' ({creds['jid']})") + else: + print(f"Webchat: ERROR: UNAUTHORIZED (STATUS:{req.status_code})") + exit(1) + + self.creds = creds + + headers = {'Content-Type': f'text/xml; charset=utf-8'} + authb64 = base64.b64encode(f'\0{creds["nickname"]}\0{creds["password"]}'.encode('ASCII')).decode() + resource_id = f'{self.resource_pref}-{random.randint(10000000, 99999999)}' + self.resource_id = resource_id + + print(f'boshService: Fetching "authid"...') + req_body = f'' + req = self._bosh_send(req_body) + authid = xmltodict.parse(req)["body"]["@authid"] + + self.authid = authid + self._internal_user_addr = f'{creds["nickname"]}@{_inst}/{resource_id}' + self._internal_room_addr = f'{_room}@room.{_inst}' + + print(f'boshService: Authorizing...') + req_body = f'{authb64}' + req = self._bosh_send(req_body) + + print(f'boshService: Restarting datastream...') + req_body = f'' + req = self._bosh_send(req_body) + + print(f'boshService: Binding auth...') + req_body = f'{resource_id}' + req = self._bosh_send(req_body) + + print(f'boshService: Getting session...') + req_body = f'' + req = self._bosh_send(req_body) + + print(f'boshService: Joining room, sending presence...') + self.bosh_update_presenses() + + print(f'Done, check if there is ModBot message in chat.') + self.send_msg("Online!") + + def bosh_update_presenses(self): + req_body = f'' + req = self._bosh_send(req_body) + presenses = xmltodict.parse(req)["body"] + if "presence" in presenses: + presenses = presenses["presence"] + if isinstance(presenses, dict): + presenses = [presenses, ] + for presense in presenses: + self.update_user(presense) + + def bosh_user_unmute(self, username): + req_body = f'' + return self._bosh_send(req_body) + + def bosh_user_mute(self, username): + req_body = f'' + return self._bosh_send(req_body) + + def bosh_user_ban(self, jid): + req_body = f'' + return self._bosh_send(req_body) + + def bosh_retract_msg(self, msg_id): + req_body = f'' + return self._bosh_send(req_body) + + def wipe_user_msg(self, user_id): + newmsg = [] + for msg in self.msg_history: + if msg["user_id"] == user_id: + self.bosh_retract_msg(msg["msg_id"]) + else: + newmsg.append(msg) + self.msg_history = newmsg + + def process_command(self, body, cmd=False): + if body == "!help": + self.send_msg("List of commands:\n" + + "!users - Show list of online users\n" + + "!usersall - Show list of all known users\n" + + "!mute USERID - Mute user by ID\n" + + "!unmute USERID - Unmute user by ID\n" + "!timeout USERID SEC - Mute user by ID for SEC seconds\n" + "!wipe USERID - Retract user's recent messages\n" + "!ban USERID - Ban user by ID and retract his last messages\n" + "!mode 0 - Lift mutes, everyone can talk\n" + "!mode 1 - Mute anons-newcomers, existing and registered users can talk\n" + "!mode 2 - Mute all anons, registered users can talk\n" + "!mode 3 (or !shutup) - Siege mode: mute everyone except moderators.\n" + "!ratelimits M W S - Rewrite ratelimits: allow M messages per W seconds before S seconds timeout.\n", cmd) + + elif body == "!users": + usersstr = "List of online users:\n" + usersstr += f"ID) [name] (jid, role, affiliation)\n" + for username, user in self.users.items(): + if user['role'] != "none": + usersstr += f"{user['uid']}) [{username}] ({user['jid']}, {user['role']}, {user['affiliation']})\n" + self.send_msg(usersstr, cmd) + + elif body == "!usersall": + usersstr = "List of known users:\n" + usersstr += f"ID) [name] (jid, role, affiliation)\n" + for username, user in self.users.items(): + usersstr += f"{user['uid']}) [{username}] ({user['jid']}, {user['role']}, {user['affiliation']})\n" + self.send_msg(usersstr, cmd) + + elif body.startswith("!mute"): + _id = body[6:] + if _id.isdigit(): + _id = int(_id) + found = False + for username, user in self.users.items(): + if user["uid"] == _id: + found = True + self.bosh_user_mute(username) + self.send_msg(f"'{username}' muted", cmd) + if not found: + self.send_msg(f"Can't find user with ID={_id}", cmd) + else: + self.send_msg(f"Error: '{_id}' is not integer. Provide numeric user ID (you can list users with !users)", cmd) + + elif body.startswith("!unmute"): + _id = body[8:] + if _id.isdigit(): + _id = int(_id) + found = False + for username, user in self.users.items(): + if user["uid"] == _id: + found = True + self.bosh_user_unmute(username) + self.send_msg(f"'{username}' unmuted", cmd) + if not found: + self.send_msg(f"Can't find user with ID={_id}", cmd) + else: + self.send_msg(f"Error: '{_id}' is not integer. Provide numeric user ID (you can list users with !users)", cmd) + + elif body.startswith("!timeout"): + args = body[9:].split(" ") + _id = args[0] + _time = args[1] if len(args) > 1 else str(self.timeout_default) + if _id.isdigit() and _time.isdigit(): + _id = int(_id) + _time = int(_time) + found = False + for username, user in self.users.items(): + if user["uid"] == _id: + found = True + self.bosh_user_mute(username) + self.users[username]["timeout_until"] = datetime.now() + timedelta(seconds=_time) + self.send_msg(f"'{username}' muted for {_time} seconds", cmd) + if not found: + self.send_msg(f"Can't find user with ID={_id}", cmd) + else: + self.send_msg("Error: malformed command", cmd) + + elif body.startswith("!ratelimits"): + args = body[12:].split(" ") + if len(args) == 3: + _msgs = args[0] + _wnd = args[1] + _time = args[2] + if _msgs.isdigit() and _wnd.isdigit() and _time.isdigit(): + _msgs = int(_msgs) + _wnd = int(_wnd) + _time = int(_time) + self.ratelimit_maxmessages = _msgs + self.ratelimit_timewindow = _wnd + self.ratelimit_timeout = _time + self.send_msg(f"Ratelimits updated: users who send >={_msgs} messages in {_wnd} seconds will be timed out for {_time} seconds.", cmd) + else: + self.send_msg("Error: malformed command", cmd) + else: + self.send_msg("Error: malformed command", cmd) + + elif body == "!mode 0": + for username, user in self.users.items(): + self.bosh_user_unmute(username) + self.restricted_mode = 0 + self.send_msg("Code 🟩-0, everyone unmuted. Welcome back!", cmd) + + elif body == "!mode 1": + self.restricted_mode = 1 + self.send_msg("Code 🟨-0, newcomers will be muted.", cmd) + + elif body == "!mode 2": + for username, user in self.users.items(): + if "@anon." in user['jid']: + self.bosh_user_mute(username) + self.restricted_mode = 2 + self.send_msg("Code 🟨-1, anon users muted.", cmd) + + elif body == "!mode 3" or body == "!shutup": + for username, user in self.users.items(): + if user["role"] != "moderator": + self.bosh_user_mute(username) + self.restricted_mode = 3 + self.send_msg("Code 🟥-0, everyone muted. Stand by.", cmd) + + elif body.startswith("!wipe"): + _id = body[6:] + if _id.isdigit(): + _id = int(_id) + found = False + for username, user in self.users.items(): + if user["uid"] == _id: + found = True + self.wipe_user_msg(_id) + self.send_msg(f"'{username}' msgs wiped.", cmd) + if not found: + self.send_msg(f"Can't find user with ID={_id}", cmd) + else: + self.send_msg(f"Error: '{_id}' is not integer. Provide numeric user ID (you can list users with !users)", cmd) + + elif body.startswith("!ban"): + _id = body[5:] + if _id.isdigit(): + _id = int(_id) + found = False + for username, user in self.users.items(): + if user["uid"] == _id: + found = True + if user["jid"] == self.creds['jid']: + self.send_msg(f"Nope, won't ban myself ({self.creds['jid']})", cmd) + else: + self.bosh_user_ban(user["jid"]) + self.send_msg(f"'{username}' banned.", cmd) + self.wipe_user_msg(_id) + if not found: + self.send_msg(f"Can't find user with ID={_id}", cmd) + else: + self.send_msg(f"Error: '{_id}' is not integer. Provide numeric user ID (you can list users with !users)", cmd) + + def _receiver_loop(self): + while True: + req_body = f'' + resp = self._bosh_send(req_body) + self.msg_queue.put(resp) + + def start_receiver_loop(self): + t = Thread(target=self._receiver_loop) + t.daemon = True + t.start() + + def _reader_loop(self): + while True: + resp = self.msg_queue.get() + + #self.bosh_update_presenses() + + msg = xmltodict.parse(resp)["body"] + + #print(self.users) + #print(json.dumps(msg, indent=2) + "\n") + time_now = datetime.now() + + stamp = os.stat(self._bannedusernames_filename).st_mtime + if self._bannedusernames_stamp != stamp: + self._bannedusernames_stamp = stamp + with open(self._bannedusernames_filename, 'r', encoding='utf-8') as file: + self.banned_usernames = [line.strip() for line in file] + print(f"Banned usernames list updated (count: {len(self.banned_usernames)}).") + + stamp = os.stat(self._bannedwords_filename).st_mtime + if self._bannedwords_stamp != stamp: + self._bannedwords_stamp = stamp + with open(self._bannedwords_filename, 'r', encoding='utf-8') as file: + self.banned_words = [line.strip() for line in file] + print(f"Banned words list updated (count: {len(self.banned_words)}).") + + for username, user in self.users.items(): + if "timeout_until" in user: + if user["timeout_until"] < time_now: + self.send_msg(f"'{username}' unmuted after timeout.") + self.bosh_user_unmute(username) + del self.users[username]["timeout_until"] + + # should probably be removed since there is a bosh_update_presenses() call + if "presence" in msg: + self.update_user(msg["presence"]) + + if "message" in msg: + chatmsgs = msg["message"] + if isinstance(chatmsgs, dict): + chatmsgs = [chatmsgs, ] + + for chatmsg in chatmsgs: + + if "body" in chatmsg: + # If it's actually a message in the chat + + body = chatmsg["body"] + username = chatmsg["@from"].split("/", 1)[1] + msguser = self.users[username] + + self.msg_history.append({ + "user_id": msguser['uid'], + "body": body, + "msg_id": chatmsg["stanza-id"]["@id"] + }) + + if (len(self.msg_history) > self.msg_history_max_len): + self.msg_history.pop(0) + + if self.last_msg["username"] == username: + self.last_msg["count"] += 1 + else: + self.last_msg["username"] = username + self.last_msg["count"] = 1 + + msgtoprint = (body[:50] + '..') if len(body) > 50 else body + print(f"[{datetime.now().strftime('%H:%M:%S')}] ({msguser['uid']}|{username}|{msguser['jid']}|x{self.last_msg['count']}) {msgtoprint}") + + if self.users[username]["role"] == "moderator": + self.process_command(body) + + else: + + # Start of ratelimit maneuvers + user = self.users[username] + if not "msgtimes" in user: + self.users[username]["msgtimes"] = [] + + rl_tw = timedelta(seconds=self.ratelimit_timewindow) + for timestamp in user["msgtimes"]: + if timestamp < time_now - rl_tw: + self.users[username]["msgtimes"].remove(timestamp) + self.users[username]["msgtimes"].append(time_now) + + msg_duplicate = False + if "lastmsg" in self.users[username]: + if self.users[username]["lastmsg"] == hash(body): + msg_duplicate = True + + msgcount = 0 + for timestamp in self.users[username]["msgtimes"]: + if timestamp >= time_now - rl_tw: + if msg_duplicate: + self.users[username]["timeout_until"] = datetime.now() + timedelta(seconds=self.ratelimit_timeout) + self.send_msg(f"'{username}' muted for {self.ratelimit_timeout} seconds. Reason: spamming") + self.bosh_user_mute(username) + break + else: + msgcount += 1 + + if msgcount >= self.ratelimit_maxmessages: + self.users[username]["timeout_until"] = datetime.now() + timedelta(seconds=self.ratelimit_timeout) + self.send_msg(f"'{username}' muted for {self.ratelimit_timeout} seconds. Reason: ratelimits") + self.bosh_user_mute(username) + + self.users[username]["lastmsg"] = hash(body) + + # End of ratelimit maneuvers + + for banned_word in self.banned_words: + if banned_word in body.lower(): + self.users[username]["timeout_until"] = datetime.now() + timedelta(seconds=self.ratelimit_timeout) + self.send_msg(f"'{username}' muted for {self.ratelimit_timeout} seconds. Reason: banned word") + self.bosh_user_mute(username) + + #print(json.dumps(xmltodict.parse(req)["body"], indent=2) + "\n") + + # Put some rules or commands down there that apply for any user, regardless of role + else: + pass + #if not "urn:xmpp:hints" in req: + #print(json.dumps(xmltodict.parse(req)["body"], indent=2) + "\n") + else: + pass + #if not "urn:xmpp:hints" in req: + #print(json.dumps(xmltodict.parse(req)["body"], indent=2) + "\n") + + def start_reader_loop(self): + t = Thread(target=self._reader_loop) + t.daemon = True + t.start() + + def run(self): + self.connect() + self.start_receiver_loop() + self.start_reader_loop() + + while True: + command = input() + self.process_command(command, cmd=True) + +if __name__ == "__main__": + + bot = ModBot( + settings.instance, + settings.username, + settings.password, + settings.room + ) + + bot.timeout_default = settings.timeout_default + bot.ratelimit_timewindow = settings.ratelimit_timewindow + bot.ratelimit_maxmessages = settings.ratelimit_maxmessages + bot.ratelimit_timeout = settings.ratelimit_timeout + + + + #req = requests.get(f"https://xxivproduction.video/w/bcsWRyy1XgNdd9RdMriM9F") + #print(req.text) + + bot.run() diff --git a/settings.py b/settings.py new file mode 100644 index 0000000..238c55f --- /dev/null +++ b/settings.py @@ -0,0 +1,29 @@ +### Authorization settings ### + +# Peertube instance URL +instance = "xxivproduction.video" + +# Credentials of user with MODERATOR role on this instance or room +username = "USERNAME" +password = "PASSWORD" + +# Room UUID (from direct chat URL) +# To find one, click the chain icon in upper right corner in chat area on stream page. Example: +# Direct chat URL looks like this https://xxivproduction.video/plugins/livechat/router/webchat/room/4d393887-f850-eeee-bbbb-0178976ad016 +# You are looking for "4d393887-f850-eeee-bbbb-0178976ad016" part of it +# You can also find Room UUID in Share -> Embed menu +room = "f3e0a5fe-e0be-42c4-b3ab-9e42cf5ecc2e" + +### Moderation settings ### + +# Timeout length, if !timeout command was send without 2nd argument +timeout_default = 300 + +# Rate limit settings +# Time window (in seconds) to measure rates +ratelimit_timewindow = 10 +# How many messages one user allowed to send in 'ratelimit_timewindow' seconds +ratelimit_maxmessages = 4 +# Length of timeout (in seconds) that will be applied to user who exceeded the limits +ratelimit_timeout = 30 + diff --git a/xmltodict.py b/xmltodict.py new file mode 100644 index 0000000..6b5c28f --- /dev/null +++ b/xmltodict.py @@ -0,0 +1,549 @@ +#!/usr/bin/env python +"Makes working with XML feel like you are working with JSON" + +""" +Grabbed from https://github.com/martinblech/xmltodict + +Copyright (C) 2012 Martin Blech and individual contributors. +Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated documentation files (the "Software"), to deal in the Software without restriction, including without limitation the rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit persons to whom the Software is furnished to do so, subject to the following conditions: +The above copyright notice and this permission notice shall be included in all copies or substantial portions of the Software. +THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE. +""" + +try: + from defusedexpat import pyexpat as expat +except ImportError: + from xml.parsers import expat +from xml.sax.saxutils import XMLGenerator +from xml.sax.xmlreader import AttributesImpl +try: # pragma no cover + from cStringIO import StringIO +except ImportError: # pragma no cover + try: + from StringIO import StringIO + except ImportError: + from io import StringIO + +from collections import OrderedDict +from inspect import isgenerator + +try: # pragma no cover + _basestring = basestring +except NameError: # pragma no cover + _basestring = str +try: # pragma no cover + _unicode = unicode +except NameError: # pragma no cover + _unicode = str + +__author__ = 'Martin Blech' +__version__ = '0.12.0' +__license__ = 'MIT' + + +class ParsingInterrupted(Exception): + pass + + +class _DictSAXHandler(object): + def __init__(self, + item_depth=0, + item_callback=lambda *args: True, + xml_attribs=True, + attr_prefix='@', + cdata_key='#text', + force_cdata=False, + cdata_separator='', + postprocessor=None, + dict_constructor=OrderedDict, + strip_whitespace=True, + namespace_separator=':', + namespaces=None, + force_list=None, + comment_key='#comment'): + self.path = [] + self.stack = [] + self.data = [] + self.item = None + self.item_depth = item_depth + self.xml_attribs = xml_attribs + self.item_callback = item_callback + self.attr_prefix = attr_prefix + self.cdata_key = cdata_key + self.force_cdata = force_cdata + self.cdata_separator = cdata_separator + self.postprocessor = postprocessor + self.dict_constructor = dict_constructor + self.strip_whitespace = strip_whitespace + self.namespace_separator = namespace_separator + self.namespaces = namespaces + self.namespace_declarations = OrderedDict() + self.force_list = force_list + self.comment_key = comment_key + + def _build_name(self, full_name): + if self.namespaces is None: + return full_name + i = full_name.rfind(self.namespace_separator) + if i == -1: + return full_name + namespace, name = full_name[:i], full_name[i+1:] + try: + short_namespace = self.namespaces[namespace] + except KeyError: + short_namespace = namespace + if not short_namespace: + return name + else: + return self.namespace_separator.join((short_namespace, name)) + + def _attrs_to_dict(self, attrs): + if isinstance(attrs, dict): + return attrs + return self.dict_constructor(zip(attrs[0::2], attrs[1::2])) + + def startNamespaceDecl(self, prefix, uri): + self.namespace_declarations[prefix or ''] = uri + + def startElement(self, full_name, attrs): + name = self._build_name(full_name) + attrs = self._attrs_to_dict(attrs) + if attrs and self.namespace_declarations: + attrs['xmlns'] = self.namespace_declarations + self.namespace_declarations = OrderedDict() + self.path.append((name, attrs or None)) + if len(self.path) > self.item_depth: + self.stack.append((self.item, self.data)) + if self.xml_attribs: + attr_entries = [] + for key, value in attrs.items(): + key = self.attr_prefix+self._build_name(key) + if self.postprocessor: + entry = self.postprocessor(self.path, key, value) + else: + entry = (key, value) + if entry: + attr_entries.append(entry) + attrs = self.dict_constructor(attr_entries) + else: + attrs = None + self.item = attrs or None + self.data = [] + + def endElement(self, full_name): + name = self._build_name(full_name) + if len(self.path) == self.item_depth: + item = self.item + if item is None: + item = (None if not self.data + else self.cdata_separator.join(self.data)) + + should_continue = self.item_callback(self.path, item) + if not should_continue: + raise ParsingInterrupted() + if len(self.stack): + data = (None if not self.data + else self.cdata_separator.join(self.data)) + item = self.item + self.item, self.data = self.stack.pop() + if self.strip_whitespace and data: + data = data.strip() or None + if data and self.force_cdata and item is None: + item = self.dict_constructor() + if item is not None: + if data: + self.push_data(item, self.cdata_key, data) + self.item = self.push_data(self.item, name, item) + else: + self.item = self.push_data(self.item, name, data) + else: + self.item = None + self.data = [] + self.path.pop() + + def characters(self, data): + if not self.data: + self.data = [data] + else: + self.data.append(data) + + def comments(self, data): + if self.strip_whitespace: + data = data.strip() + self.item = self.push_data(self.item, self.comment_key, data) + + def push_data(self, item, key, data): + if self.postprocessor is not None: + result = self.postprocessor(self.path, key, data) + if result is None: + return item + key, data = result + if item is None: + item = self.dict_constructor() + try: + value = item[key] + if isinstance(value, list): + value.append(data) + else: + item[key] = [value, data] + except KeyError: + if self._should_force_list(key, data): + item[key] = [data] + else: + item[key] = data + return item + + def _should_force_list(self, key, value): + if not self.force_list: + return False + if isinstance(self.force_list, bool): + return self.force_list + try: + return key in self.force_list + except TypeError: + return self.force_list(self.path[:-1], key, value) + + +def parse(xml_input, encoding=None, expat=expat, process_namespaces=False, + namespace_separator=':', disable_entities=True, process_comments=False, **kwargs): + """Parse the given XML input and convert it into a dictionary. + + `xml_input` can either be a `string`, a file-like object, or a generator of strings. + + If `xml_attribs` is `True`, element attributes are put in the dictionary + among regular child elements, using `@` as a prefix to avoid collisions. If + set to `False`, they are just ignored. + + Simple example:: + + >>> import xmltodict + >>> doc = xmltodict.parse(\"\"\" + ... + ... 1 + ... 2 + ... + ... \"\"\") + >>> doc['a']['@prop'] + u'x' + >>> doc['a']['b'] + [u'1', u'2'] + + If `item_depth` is `0`, the function returns a dictionary for the root + element (default behavior). Otherwise, it calls `item_callback` every time + an item at the specified depth is found and returns `None` in the end + (streaming mode). + + The callback function receives two parameters: the `path` from the document + root to the item (name-attribs pairs), and the `item` (dict). If the + callback's return value is false-ish, parsing will be stopped with the + :class:`ParsingInterrupted` exception. + + Streaming example:: + + >>> def handle(path, item): + ... print('path:%s item:%s' % (path, item)) + ... return True + ... + >>> xmltodict.parse(\"\"\" + ... + ... 1 + ... 2 + ... \"\"\", item_depth=2, item_callback=handle) + path:[(u'a', {u'prop': u'x'}), (u'b', None)] item:1 + path:[(u'a', {u'prop': u'x'}), (u'b', None)] item:2 + + The optional argument `postprocessor` is a function that takes `path`, + `key` and `value` as positional arguments and returns a new `(key, value)` + pair where both `key` and `value` may have changed. Usage example:: + + >>> def postprocessor(path, key, value): + ... try: + ... return key + ':int', int(value) + ... except (ValueError, TypeError): + ... return key, value + >>> xmltodict.parse('12x', + ... postprocessor=postprocessor) + OrderedDict([(u'a', OrderedDict([(u'b:int', [1, 2]), (u'b', u'x')]))]) + + You can pass an alternate version of `expat` (such as `defusedexpat`) by + using the `expat` parameter. E.g: + + >>> import defusedexpat + >>> xmltodict.parse('hello', expat=defusedexpat.pyexpat) + OrderedDict([(u'a', u'hello')]) + + You can use the force_list argument to force lists to be created even + when there is only a single child of a given level of hierarchy. The + force_list argument is a tuple of keys. If the key for a given level + of hierarchy is in the force_list argument, that level of hierarchy + will have a list as a child (even if there is only one sub-element). + The index_keys operation takes precedence over this. This is applied + after any user-supplied postprocessor has already run. + + For example, given this input: + + + host1 + Linux + + + em0 + 10.0.0.1 + + + + + + If called with force_list=('interface',), it will produce + this dictionary: + {'servers': + {'server': + {'name': 'host1', + 'os': 'Linux'}, + 'interfaces': + {'interface': + [ {'name': 'em0', 'ip_address': '10.0.0.1' } ] } } } + + `force_list` can also be a callable that receives `path`, `key` and + `value`. This is helpful in cases where the logic that decides whether + a list should be forced is more complex. + + + If `process_comment` is `True` then comment will be added with comment_key + (default=`'#comment'`) to then tag which contains comment + + For example, given this input: + + + + + + 1 + + 2 + + + + If called with process_comment=True, it will produce + this dictionary: + 'a': { + 'b': { + '#comment': 'b comment', + 'c': { + + '#comment': 'c comment', + '#text': '1', + }, + 'd': '2', + }, + } + """ + handler = _DictSAXHandler(namespace_separator=namespace_separator, + **kwargs) + if isinstance(xml_input, _unicode): + if not encoding: + encoding = 'utf-8' + xml_input = xml_input.encode(encoding) + if not process_namespaces: + namespace_separator = None + parser = expat.ParserCreate( + encoding, + namespace_separator + ) + try: + parser.ordered_attributes = True + except AttributeError: + # Jython's expat does not support ordered_attributes + pass + parser.StartNamespaceDeclHandler = handler.startNamespaceDecl + parser.StartElementHandler = handler.startElement + parser.EndElementHandler = handler.endElement + parser.CharacterDataHandler = handler.characters + if process_comments: + parser.CommentHandler = handler.comments + parser.buffer_text = True + if disable_entities: + try: + # Attempt to disable DTD in Jython's expat parser (Xerces-J). + feature = "http://apache.org/xml/features/disallow-doctype-decl" + parser._reader.setFeature(feature, True) + except AttributeError: + # For CPython / expat parser. + # Anything not handled ends up here and entities aren't expanded. + parser.DefaultHandler = lambda x: None + # Expects an integer return; zero means failure -> expat.ExpatError. + parser.ExternalEntityRefHandler = lambda *x: 1 + if hasattr(xml_input, 'read'): + parser.ParseFile(xml_input) + elif isgenerator(xml_input): + for chunk in xml_input: + parser.Parse(chunk,False) + parser.Parse(b'',True) + else: + parser.Parse(xml_input, True) + return handler.item + + +def _process_namespace(name, namespaces, ns_sep=':', attr_prefix='@'): + if not namespaces: + return name + try: + ns, name = name.rsplit(ns_sep, 1) + except ValueError: + pass + else: + ns_res = namespaces.get(ns.strip(attr_prefix)) + name = '{}{}{}{}'.format( + attr_prefix if ns.startswith(attr_prefix) else '', + ns_res, ns_sep, name) if ns_res else name + return name + + +def _emit(key, value, content_handler, + attr_prefix='@', + cdata_key='#text', + depth=0, + preprocessor=None, + pretty=False, + newl='\n', + indent='\t', + namespace_separator=':', + namespaces=None, + full_document=True, + expand_iter=None): + key = _process_namespace(key, namespaces, namespace_separator, attr_prefix) + if preprocessor is not None: + result = preprocessor(key, value) + if result is None: + return + key, value = result + if (not hasattr(value, '__iter__') + or isinstance(value, _basestring) + or isinstance(value, dict)): + value = [value] + for index, v in enumerate(value): + if full_document and depth == 0 and index > 0: + raise ValueError('document with multiple roots') + if v is None: + v = OrderedDict() + elif isinstance(v, bool): + if v: + v = _unicode('true') + else: + v = _unicode('false') + elif not isinstance(v, dict): + if expand_iter and hasattr(v, '__iter__') and not isinstance(v, _basestring): + v = OrderedDict(((expand_iter, v),)) + else: + v = _unicode(v) + if isinstance(v, _basestring): + v = OrderedDict(((cdata_key, v),)) + cdata = None + attrs = OrderedDict() + children = [] + for ik, iv in v.items(): + if ik == cdata_key: + cdata = iv + continue + if ik.startswith(attr_prefix): + ik = _process_namespace(ik, namespaces, namespace_separator, + attr_prefix) + if ik == '@xmlns' and isinstance(iv, dict): + for k, v in iv.items(): + attr = 'xmlns{}'.format(':{}'.format(k) if k else '') + attrs[attr] = _unicode(v) + continue + if not isinstance(iv, _unicode): + iv = _unicode(iv) + attrs[ik[len(attr_prefix):]] = iv + continue + children.append((ik, iv)) + if pretty: + content_handler.ignorableWhitespace(depth * indent) + content_handler.startElement(key, AttributesImpl(attrs)) + if pretty and children: + content_handler.ignorableWhitespace(newl) + for child_key, child_value in children: + _emit(child_key, child_value, content_handler, + attr_prefix, cdata_key, depth+1, preprocessor, + pretty, newl, indent, namespaces=namespaces, + namespace_separator=namespace_separator, + expand_iter=expand_iter) + if cdata is not None: + content_handler.characters(cdata) + if pretty and children: + content_handler.ignorableWhitespace(depth * indent) + content_handler.endElement(key) + if pretty and depth: + content_handler.ignorableWhitespace(newl) + + +def unparse(input_dict, output=None, encoding='utf-8', full_document=True, + short_empty_elements=False, + **kwargs): + """Emit an XML document for the given `input_dict` (reverse of `parse`). + + The resulting XML document is returned as a string, but if `output` (a + file-like object) is specified, it is written there instead. + + Dictionary keys prefixed with `attr_prefix` (default=`'@'`) are interpreted + as XML node attributes, whereas keys equal to `cdata_key` + (default=`'#text'`) are treated as character data. + + The `pretty` parameter (default=`False`) enables pretty-printing. In this + mode, lines are terminated with `'\n'` and indented with `'\t'`, but this + can be customized with the `newl` and `indent` parameters. + + """ + if full_document and len(input_dict) != 1: + raise ValueError('Document must have exactly one root.') + must_return = False + if output is None: + output = StringIO() + must_return = True + if short_empty_elements: + content_handler = XMLGenerator(output, encoding, True) + else: + content_handler = XMLGenerator(output, encoding) + if full_document: + content_handler.startDocument() + for key, value in input_dict.items(): + _emit(key, value, content_handler, full_document=full_document, + **kwargs) + if full_document: + content_handler.endDocument() + if must_return: + value = output.getvalue() + try: # pragma no cover + value = value.decode(encoding) + except AttributeError: # pragma no cover + pass + return value + + +if __name__ == '__main__': # pragma: no cover + import sys + import marshal + try: + stdin = sys.stdin.buffer + stdout = sys.stdout.buffer + except AttributeError: + stdin = sys.stdin + stdout = sys.stdout + + (item_depth,) = sys.argv[1:] + item_depth = int(item_depth) + + def handle_item(path, item): + marshal.dump((path, item), stdout) + return True + + try: + root = parse(stdin, + item_depth=item_depth, + item_callback=handle_item, + dict_constructor=dict) + if item_depth == 0: + handle_item([], root) + except KeyboardInterrupt: + pass