streaming.py

# Tweepy
# Copyright 2009-2010 Joshua Roesslein
# See LICENSE for details.

# Appengine users: https://developers.google.com/appengine/docs/python/sockets/#making_httplib_use_sockets

from __future__ import absolute_import, print_function

import logging
import requests
from requests.exceptions import Timeout
from threading import Thread
from time import sleep

import six

import ssl

from tweepy.models import Status
from tweepy.api import API
from tweepy.error import TweepError

from tweepy.utils import import_simplejson, urlencode_noplus
json = import_simplejson()

STREAM_VERSION = '1.1'


class StreamListener(object):

    def __init__(self, api=None):
        self.api = api or API()

    def on_connect(self):
        """Called once connected to streaming server.

        This will be invoked once a successful response
        is received from the server. Allows the listener
        to perform some work prior to entering the read loop.
        """
        pass

    def on_data(self, raw_data):
        """Called when raw data is received from connection.

        Override this method if you wish to manually handle
        the stream data. Return False to stop stream and close connection.
        """
        data = json.loads(raw_data)

        if 'in_reply_to_status_id' in data:
            status = Status.parse(self.api, data)
            if self.on_status(status) is False:
                return False
        elif 'delete' in data:
            delete = data['delete']['status']
            if self.on_delete(delete['id'], delete['user_id']) is False:
                return False
        elif 'event' in data:
            status = Status.parse(self.api, data)
            if self.on_event(status) is False:
                return False
        elif 'direct_message' in data:
            status = Status.parse(self.api, data)
            if self.on_direct_message(status) is False:
                return False
        elif 'friends' in data:
            if self.on_friends(data['friends']) is False:
                return False
        elif 'limit' in data:
            if self.on_limit(data['limit']['track']) is False:
                return False
        elif 'disconnect' in data:
            if self.on_disconnect(data['disconnect']) is False:
                return False
        elif 'warning' in data:
            if self.on_warning(data['warning']) is False:
                return False
        else:
            logging.error("Unknown message type: " + str(raw_data))

    def on_status(self, status):
        """Called when a new status arrives"""
        return

    def on_exception(self, exception):
        """Called when an unhandled exception occurs."""
        return

    def on_delete(self, status_id, user_id):
        """Called when a delete notice arrives for a status"""
        return

    def on_event(self, status):
        """Called when a new event arrives"""
        return

    def on_direct_message(self, status):
        """Called when a new direct message arrives"""
        return

    def on_friends(self, friends):
        """Called when a friends list arrives.

        friends is a list that contains user_id
        """
        return

    def on_limit(self, track):
        """Called when a limitation notice arrives"""
        return

    def on_error(self, status_code):
        """Called when a non-200 status code is returned"""
        return False

    def on_timeout(self):
        """Called when stream connection times out"""
        return

    def on_disconnect(self, notice):
        """Called when twitter sends a disconnect notice

        Disconnect codes are listed here:
        https://dev.twitter.com/docs/streaming-apis/messages#Disconnect_messages_disconnect
        """
        return
    
    def on_warning(self, notice):
        """Called when a disconnection warning message arrives"""
        return


class ReadBuffer(object):
    """Buffer data from the response in a smarter way than httplib/requests can.

    Tweets are roughly in the 2-12kb range, averaging around 3kb.
    Requests/urllib3/httplib/socket all use socket.read, which blocks
    until enough data is returned. On some systems (eg google appengine), socket
    reads are quite slow. To combat this latency we can read big chunks,
    but the blocking part means we won't get results until enough tweets
    have arrived. That may not be a big deal for high throughput systems.
    For low throughput systems we don't want to sacrafice latency, so we
    use small chunks so it can read the length and the tweet in 2 read calls.
    """

    def __init__(self, stream, chunk_size):
        self._stream = stream
        self._buffer = u""
        self._chunk_size = chunk_size

    def read_len(self, length):
        while True:
            if len(self._buffer) >= length:
                return self._pop(length)
            read_len = max(self._chunk_size, length - len(self._buffer))
            self._buffer += self._stream.read(read_len).decode("ascii")

    def read_line(self, sep='\n'):
        start = 0
        while True:
            loc = self._buffer.find(sep, start)
            if loc >= 0:
                return self._pop(loc + len(sep))
            else:
                start = len(self._buffer)
            self._buffer += self._stream.read(self._chunk_size).decode("ascii")

    def _pop(self, length):
        r = self._buffer[:length]
        self._buffer = self._buffer[length:]
        return r


class Stream(object):

    host = 'stream.twitter.com'

    def __init__(self, auth, listener, **options):
        self.auth = auth
        self.listener = listener
        self.running = False
        self.timeout = options.get("timeout", 300.0)
        self.retry_count = options.get("retry_count")
        # values according to
        # https://dev.twitter.com/docs/streaming-apis/connecting#Reconnecting
        self.retry_time_start = options.get("retry_time", 5.0)
        self.retry_420_start = options.get("retry_420", 60.0)
        self.retry_time_cap = options.get("retry_time_cap", 320.0)
        self.snooze_time_step = options.get("snooze_time", 0.25)
        self.snooze_time_cap = options.get("snooze_time_cap", 16)

        # The default socket.read size. Default to less than half the size of
        # a tweet so that it reads tweets with the minimal latency of 2 reads
        # per tweet. Values higher than ~1kb will increase latency by waiting
        # for more data to arrive but may also increase throughput by doing
        # fewer socket read calls.
        self.chunk_size = options.get("chunk_size",  512)

        self.verify = options.get("verify", True)

        self.api = API()
        self.headers = options.get("headers") or {}
        self.new_session()
        self.body = None
        self.retry_time = self.retry_time_start
        self.snooze_time = self.snooze_time_step

    def new_session(self):
        self.session = requests.Session()
        self.session.headers = self.headers
        self.session.params = None

    def _run(self):
        # Authenticate
        url = "https://%s%s" % (self.host, self.url)

        # Connect and process the stream
        error_counter = 0
        resp = None
        exception = None
        while self.running:
            if self.retry_count is not None:
                if error_counter > self.retry_count:
                    # quit if error count greater than retry count
                    break
            try:
                auth = self.auth.apply_auth()
                resp = self.session.request('POST',
                                            url,
                                            data=self.body,
                                            timeout=self.timeout,
                                            stream=True,
                                            auth=auth,
                                            verify=self.verify)
                if resp.status_code != 200:
                    if self.listener.on_error(resp.status_code) is False:
                        break
                    error_counter += 1
                    if resp.status_code == 420:
                        self.retry_time = max(self.retry_420_start,
                                              self.retry_time)
                    sleep(self.retry_time)
                    self.retry_time = min(self.retry_time * 2,
                                          self.retry_time_cap)
                else:
                    error_counter = 0
                    self.retry_time = self.retry_time_start
                    self.snooze_time = self.snooze_time_step
                    self.listener.on_connect()
                    self._read_loop(resp)
            except (Timeout, ssl.SSLError) as exc:
                # This is still necessary, as a SSLError can actually be
                # thrown when using Requests
                # If it's not time out treat it like any other exception
                if isinstance(exc, ssl.SSLError):
                    if not (exc.args and 'timed out' in str(exc.args[0])):
                        exception = exc
                        break
                if self.listener.on_timeout() is False:
                    break
                if self.running is False:
                    break
                sleep(self.snooze_time)
                self.snooze_time = min(self.snooze_time + self.snooze_time_step,
                                       self.snooze_time_cap)
            except Exception as exc:
                exception = exc
                # any other exception is fatal, so kill loop
                break

        # cleanup
        self.running = False
        if resp:
            resp.close()

        self.new_session()

        if exception:
            # call a handler first so that the exception can be logged.
            self.listener.on_exception(exception)
            raise

    def _data(self, data):
        if self.listener.on_data(data) is False:
            self.running = False

    def _read_loop(self, resp):
        buf = ReadBuffer(resp.raw, self.chunk_size)

        while self.running:
            length = 0
            while True:
                line = buf.read_line().strip()
                if not line:
                    pass  # keep-alive new lines are expected
                elif line.isdigit():
                    length = int(line)
                    break
                else:
                    raise TweepError('Expecting length, unexpected value found')

            next_status_obj = buf.read_len(length)
            if self.running:
                self._data(next_status_obj)

            # # Note: keep-alive newlines might be inserted before each length value.
            # # read until we get a digit...
            # c = b'\n'
            # for c in resp.iter_content(decode_unicode=True):
            #     if c == b'\n':
            #         continue
            #     break
            #
            # delimited_string = c
            #
            # # read rest of delimiter length..
            # d = b''
            # for d in resp.iter_content(decode_unicode=True):
            #     if d != b'\n':
            #         delimited_string += d
            #         continue
            #     break
            #
            # # read the next twitter status object
            # if delimited_string.decode('utf-8').strip().isdigit():
            #     status_id = int(delimited_string)
            #     next_status_obj = resp.raw.read(status_id)
            #     if self.running:
            #         self._data(next_status_obj.decode('utf-8'))


        if resp.raw._fp.isclosed():
            self.on_closed(resp)

    def _start(self, async):
        self.running = True
        if async:
            self._thread = Thread(target=self._run)
            self._thread.start()
        else:
            self._run()

    def on_closed(self, resp):
        """ Called when the response has been closed by Twitter """
        pass

    def userstream(self,
                   stall_warnings=False,
                   _with=None,
                   replies=None,
                   track=None,
                   locations=None,
                   async=False,
                   encoding='utf8'):
        self.session.params = {'delimited': 'length'}
        if self.running:
            raise TweepError('Stream object already connected!')
        self.url = '/%s/user.json' % STREAM_VERSION
        self.host = 'userstream.twitter.com'
        if stall_warnings:
            self.session.params['stall_warnings'] = stall_warnings
        if _with:
            self.session.params['with'] = _with
        if replies:
            self.session.params['replies'] = replies
        if locations and len(locations) > 0:
            if len(locations) % 4 != 0:
                raise TweepError("Wrong number of locations points, "
                                 "it has to be a multiple of 4")
            self.session.params['locations'] = ','.join(['%.2f' % l for l in locations])
        if track:
            self.session.params['track'] = u','.join(track).encode(encoding)

        self._start(async)

    def firehose(self, count=None, async=False):
        self.session.params = {'delimited': 'length'}
        if self.running:
            raise TweepError('Stream object already connected!')
        self.url = '/%s/statuses/firehose.json' % STREAM_VERSION
        if count:
            self.url += '&count=%s' % count
        self._start(async)

    def retweet(self, async=False):
        self.session.params = {'delimited': 'length'}
        if self.running:
            raise TweepError('Stream object already connected!')
        self.url = '/%s/statuses/retweet.json' % STREAM_VERSION
        self._start(async)

    def sample(self, async=False, languages=None):
        self.session.params = {'delimited': 'length'}
        if self.running:
            raise TweepError('Stream object already connected!')
        self.url = '/%s/statuses/sample.json' % STREAM_VERSION
        if languages:
            self.session.params['language'] = ','.join(map(str, languages))
        self._start(async)

    def filter(self, follow=None, track=None, async=False, locations=None,
               stall_warnings=False, languages=None, encoding='utf8'):
        self.session.params = {}
        self.session.headers['Content-type'] = "application/x-www-form-urlencoded"
        if self.running:
            raise TweepError('Stream object already connected!')
        self.url = '/%s/statuses/filter.json' % STREAM_VERSION
        if follow:
            self.session.params['follow'] = u','.join(follow).encode(encoding)
        if track:
            self.session.params['track'] = u','.join(track).encode(encoding)
        if locations and len(locations) > 0:
            if len(locations) % 4 != 0:
                raise TweepError("Wrong number of locations points, "
                                 "it has to be a multiple of 4")
            self.session.params['locations'] = u','.join(['%.4f' % l for l in locations])
        if stall_warnings:
            self.session.params['stall_warnings'] = stall_warnings
        if languages:
            self.session.params['language'] = u','.join(map(str, languages))
        self.body = urlencode_noplus(self.session.params)
        self.session.params = {'delimited': 'length'}
        self.host = 'stream.twitter.com'
        self._start(async)

    def sitestream(self, follow, stall_warnings=False,
                   with_='user', replies=False, async=False):
        self.parameters = {}
        if self.running:
            raise TweepError('Stream object already connected!')
        self.url = '/%s/site.json' % STREAM_VERSION
        self.parameters['follow'] = u','.join(map(six.text_type, follow))
        self.parameters['delimited'] = 'length'
        if stall_warnings:
            self.parameters['stall_warnings'] = stall_warnings
        if with_:
            self.parameters['with'] = with_
        if replies:
            self.parameters['replies'] = replies
        self.body = urlencode_noplus(self.parameters)
        self._start(async)

    def disconnect(self):
        if self.running is False:
            return
        self.running = False