# Tweepy # Copyright 2009-2023 Joshua Roesslein # See LICENSE for details. # Appengine users: https://developers.google.com/appengine/docs/python/sockets/#making_httplib_use_sockets from collections import namedtuple import json import logging from math import inf from platform import python_version import ssl import traceback from threading import Thread from time import sleep from typing import NamedTuple import requests import urllib3 import tweepy from tweepy.client import BaseClient, Response from tweepy.errors import TweepyException from tweepy.tweet import Tweet log = logging.getLogger(__name__) StreamResponse = namedtuple( "StreamResponse", ("data", "includes", "errors", "matching_rules") ) class BaseStream: def __init__(self, *, chunk_size=512, daemon=False, max_retries=inf, proxy=None, verify=True): self.chunk_size = chunk_size self.daemon = daemon self.max_retries = max_retries self.proxies = {"https": proxy} if proxy else {} self.verify = verify self.running = False self.session = requests.Session() self.thread = None self.user_agent = ( f"Python/{python_version()} " f"Requests/{requests.__version__} " f"Tweepy/{tweepy.__version__}" ) def _connect( self, method, url, auth=None, params=None, headers=None, body=None, timeout=21 ): self.running = True error_count = 0 # https://developer.twitter.com/en/docs/twitter-api/v1/tweets/filter-realtime/guides/connecting # https://developer.twitter.com/en/docs/twitter-api/tweets/filtered-stream/integrate/handling-disconnections # https://developer.twitter.com/en/docs/twitter-api/tweets/volume-streams/integrate/handling-disconnections network_error_wait = 0 network_error_wait_step = 0.25 network_error_wait_max = 16 http_error_wait = http_error_wait_start = 5 http_error_wait_max = 320 http_429_error_wait_start = 60 self.session.headers["User-Agent"] = self.user_agent try: while self.running and error_count <= self.max_retries: try: with self.session.request( method, url, params=params, headers=headers, data=body, timeout=timeout, stream=True, auth=auth, verify=self.verify, proxies=self.proxies ) as resp: if resp.status_code == 200: error_count = 0 http_error_wait = http_error_wait_start network_error_wait = 0 self.on_connect() if not self.running: break for line in resp.iter_lines( chunk_size=self.chunk_size ): if line: self.on_data(line) else: self.on_keep_alive() if not self.running: break if resp.raw.closed: self.on_closed(resp) else: self.on_request_error(resp.status_code) if not self.running: break # The error text is logged here instead of in # on_request_error to keep on_request_error # backwards-compatible. In a future version, the # Response should be passed to on_request_error. log.error( "HTTP error response text: %s", resp.text ) error_count += 1 if resp.status_code in (420, 429): if http_error_wait < http_429_error_wait_start: http_error_wait = http_429_error_wait_start sleep(http_error_wait) http_error_wait *= 2 if http_error_wait > http_error_wait_max: http_error_wait = http_error_wait_max except (requests.ConnectionError, requests.Timeout, requests.exceptions.ChunkedEncodingError, ssl.SSLError, urllib3.exceptions.ReadTimeoutError, urllib3.exceptions.ProtocolError) 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])): raise self.on_connection_error() if not self.running: break # The error text is logged here instead of in # on_connection_error to keep on_connection_error # backwards-compatible. In a future version, the error # should be passed to on_connection_error. log.error( "Connection error: %s", "".join( traceback.format_exception_only(type(exc), exc) ).rstrip() ) sleep(network_error_wait) network_error_wait += network_error_wait_step if network_error_wait > network_error_wait_max: network_error_wait = network_error_wait_max except Exception as exc: self.on_exception(exc) finally: self.session.close() self.running = False self.on_disconnect() def _threaded_connect(self, *args, **kwargs): self.thread = Thread(target=self._connect, name="Tweepy Stream", args=args, kwargs=kwargs, daemon=self.daemon) self.thread.start() return self.thread def disconnect(self): """Disconnect the stream""" self.running = False def on_closed(self, response): """This is called when the stream has been closed by Twitter. Parameters ---------- response : requests.Response The Response from Twitter """ log.error("Stream connection closed by Twitter") def on_connect(self): """This is called after successfully connecting to the streaming API. """ log.info("Stream connected") def on_connection_error(self): """This is called when the stream connection errors or times out.""" log.error("Stream connection has errored or timed out") def on_disconnect(self): """This is called when the stream has disconnected.""" log.info("Stream disconnected") def on_exception(self, exception): """This is called when an unhandled exception occurs. Parameters ---------- exception : Exception The unhandled exception """ log.exception("Stream encountered an exception") def on_keep_alive(self): """This is called when a keep-alive signal is received.""" log.debug("Received keep-alive signal") def on_request_error(self, status_code): """This is called when a non-200 HTTP status code is encountered. Parameters ---------- status_code : int The HTTP status code encountered """ log.error("Stream encountered HTTP error: %d", status_code) class StreamingClient(BaseClient, BaseStream): """Filter and sample realtime Tweets with Twitter API v2 .. versionadded:: 4.6 Parameters ---------- bearer_token : str Twitter API Bearer Token return_type : type[dict | requests.Response | Response] Type to return from requests to the API wait_on_rate_limit : bool Whether or not to wait before retrying when a rate limit is encountered. This applies to requests besides those that connect to a stream (see ``max_retries``). chunk_size : int The default socket.read size. Default to 512, 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. daemon : bool Whether or not to use a daemon thread when using a thread to run the stream max_retries : int Max number of times to retry connecting the stream proxy : str | None URL of the proxy to use when connecting to the stream verify : bool | str Either a boolean, in which case it controls whether to verify the server’s TLS certificate, or a string, in which case it must be a path to a CA bundle to use. Attributes ---------- running : bool Whether there's currently a stream running session : :class:`requests.Session` Requests Session used to connect to the stream thread : :class:`threading.Thread` | None Thread used to run the stream user_agent : str User agent used when connecting to the stream """ def __init__(self, bearer_token, *, return_type=Response, wait_on_rate_limit=False, **kwargs): """__init__( \ bearer_token, *, return_type=Response, wait_on_rate_limit=False, \ chunk_size=512, daemon=False, max_retries=inf, proxy=None, \ verify=True \ ) """ BaseClient.__init__(self, bearer_token, return_type=return_type, wait_on_rate_limit=wait_on_rate_limit) BaseStream.__init__(self, **kwargs) def _connect(self, method, endpoint, **kwargs): self.session.headers["Authorization"] = f"Bearer {self.bearer_token}" url = f"https://api.twitter.com/2/tweets/{endpoint}/stream" super()._connect(method, url, **kwargs) def _process_data(self, data, data_type=None): if data_type is StreamRule: if isinstance(data, list): rules = [] for rule in data: if "tag" in rule: rules.append(StreamRule( value=rule["value"], id=rule["id"], tag=rule["tag"] )) else: rules.append(StreamRule(value=rule["value"], id=rule["id"])) return rules elif data is not None: if "tag" in data: return StreamRule(value=data["value"], id=data["id"], tag=data["tag"]) else: return StreamRule(value=data["value"], id=data["id"]) else: return super()._process_data(data, data_type=data_type) def add_rules(self, add, **params): """add_rules(add, *, dry_run) Add rules to filtered stream. Parameters ---------- add : list[StreamRule] | StreamRule Specifies the operation you want to perform on the rules. dry_run : bool Set to true to test the syntax of your rule without submitting it. This is useful if you want to check the syntax of a rule before removing one or more of your existing rules. Returns ------- dict | requests.Response | Response References ---------- https://developer.twitter.com/en/docs/twitter-api/tweets/filtered-stream/api-reference/post-tweets-search-stream-rules """ json = {"add": []} if isinstance(add, StreamRule): add = (add,) for rule in add: if rule.tag is not None: json["add"].append({"value": rule.value, "tag": rule.tag}) else: json["add"].append({"value": rule.value}) return self._make_request( "POST", f"/2/tweets/search/stream/rules", params=params, endpoint_parameters=("dry_run",), json=json, data_type=StreamRule ) def delete_rules(self, ids, **params): """delete_rules(ids, *, dry_run) Delete rules from filtered stream. Parameters ---------- ids : int | str | list[int | str | StreamRule] | StreamRule Array of rule IDs, each one representing a rule already active in your stream. IDs must be submitted as strings. dry_run : bool Set to true to test the syntax of your rule without submitting it. This is useful if you want to check the syntax of a rule before removing one or more of your existing rules. Returns ------- dict | requests.Response | Response References ---------- https://developer.twitter.com/en/docs/twitter-api/tweets/filtered-stream/api-reference/post-tweets-search-stream-rules """ json = {"delete": {"ids": []}} if isinstance(ids, (int, str, StreamRule)): ids = (ids,) for id in ids: if isinstance(id, StreamRule): json["delete"]["ids"].append(str(id.id)) else: json["delete"]["ids"].append(str(id)) return self._make_request( "POST", f"/2/tweets/search/stream/rules", params=params, endpoint_parameters=("dry_run",), json=json, data_type=StreamRule ) def filter(self, *, threaded=False, **params): """filter( \ *, backfill_minutes=None, expansions=None, media_fields=None, \ place_fields=None, poll_fields=None, tweet_fields=None, \ user_fields=None, threaded=False \ ) Streams Tweets in real-time based on a specific set of filter rules. If you are using the academic research product track, you can connect up to two `redundant connections `_ to maximize your streaming up-time. The Tweets returned by this endpoint count towards the Project-level `Tweet cap`_. Parameters ---------- backfill_minutes : int | None By passing this parameter, you can request up to five (5) minutes worth of streaming data that you might have missed during a disconnection to be delivered to you upon reconnection. The backfilled Tweets will automatically flow through the reconnected stream, with older Tweets generally being delivered before any newly matching Tweets. You must include a whole number between 1 and 5 as the value to this parameter. This feature will deliver duplicate Tweets, meaning that if you were disconnected for 90 seconds, and you requested two minutes of backfill, you will receive 30 seconds worth of duplicate Tweets. Due to this, you should make sure your system is tolerant of duplicate data. This feature is currently only available to the Academic Research product track. expansions : list[str] | str :ref:`expansions_parameter` media_fields : list[str] | str :ref:`media_fields_parameter` place_fields : list[str] | str :ref:`place_fields_parameter` poll_fields : list[str] | str :ref:`poll_fields_parameter` tweet_fields : list[str] | str :ref:`tweet_fields_parameter` user_fields : list[str] | str :ref:`user_fields_parameter` threaded : bool Whether or not to use a thread to run the stream Raises ------ TweepyException When the stream is already connected Returns ------- threading.Thread | None The thread if ``threaded`` is set to ``True``, else ``None`` References ---------- https://developer.twitter.com/en/docs/twitter-api/tweets/filtered-stream/api-reference/get-tweets-search-stream .. _filter redundant connections: https://developer.twitter.com/en/docs/twitter-api/tweets/filtered-stream/integrate/recovery-and-redundancy-features .. _Tweet cap: https://developer.twitter.com/en/docs/twitter-api/tweet-caps """ if self.running: raise TweepyException("Stream is already connected") method = "GET" endpoint = "search" params = self._process_params( params, endpoint_parameters=( "backfill_minutes", "expansions", "media.fields", "place.fields", "poll.fields", "tweet.fields", "user.fields" ) ) if threaded: return self._threaded_connect(method, endpoint, params=params) else: self._connect(method, endpoint, params=params) def get_rules(self, **params): """get_rules(*, ids) Return a list of rules currently active on the streaming endpoint, either as a list or individually. Parameters ---------- ids : list[str] | str Comma-separated list of rule IDs. If omitted, all rules are returned. Returns ------- dict | requests.Response | Response References ---------- https://developer.twitter.com/en/docs/twitter-api/tweets/filtered-stream/api-reference/get-tweets-search-stream-rules """ return self._make_request( "GET", f"/2/tweets/search/stream/rules", params=params, endpoint_parameters=("ids",), data_type=StreamRule ) def sample(self, *, threaded=False, **params): """sample( \ *, backfill_minutes=None, expansions=None, media_fields=None, \ place_fields=None, poll_fields=None, tweet_fields=None, \ user_fields=None, threaded=False \ ) Streams about 1% of all Tweets in real-time. If you are using the academic research product track, you can connect up to two `redundant connections `_ to maximize your streaming up-time. Parameters ---------- backfill_minutes : int | None By passing this parameter, you can request up to five (5) minutes worth of streaming data that you might have missed during a disconnection to be delivered to you upon reconnection. The backfilled Tweets will automatically flow through the reconnected stream, with older Tweets generally being delivered before any newly matching Tweets. You must include a whole number between 1 and 5 as the value to this parameter. This feature will deliver duplicate Tweets, meaning that if you were disconnected for 90 seconds, and you requested two minutes of backfill, you will receive 30 seconds worth of duplicate Tweets. Due to this, you should make sure your system is tolerant of duplicate data. This feature is currently only available to the Academic Research product track. expansions : list[str] | str :ref:`expansions_parameter` media_fields : list[str] | str :ref:`media_fields_parameter` place_fields : list[str] | str :ref:`place_fields_parameter` poll_fields : list[str] | str :ref:`poll_fields_parameter` tweet_fields : list[str] | str :ref:`tweet_fields_parameter` user_fields : list[str] | str :ref:`user_fields_parameter` threaded : bool Whether or not to use a thread to run the stream Raises ------ TweepyException When the stream is already connected Returns ------- threading.Thread | None The thread if ``threaded`` is set to ``True``, else ``None`` References ---------- https://developer.twitter.com/en/docs/twitter-api/tweets/volume-streams/api-reference/get-tweets-sample-stream .. _sample redundant connections: https://developer.twitter.com/en/docs/twitter-api/tweets/volume-streams/integrate/recovery-and-redundancy-features """ if self.running: raise TweepyException("Stream is already connected") method = "GET" endpoint = "sample" params = self._process_params( params, endpoint_parameters=( "backfill_minutes", "expansions", "media.fields", "place.fields", "poll.fields", "tweet.fields", "user.fields" ) ) if threaded: return self._threaded_connect(method, endpoint, params=params) else: self._connect(method, endpoint, params=params) def on_data(self, raw_data): """This is called when raw data is received from the stream. This method handles sending the data to other methods. Parameters ---------- raw_data : JSON The raw data from the stream References ---------- https://developer.twitter.com/en/docs/twitter-api/tweets/filtered-stream/integrate/consuming-streaming-data """ data = json.loads(raw_data) tweet = None includes = {} errors = [] matching_rules = [] if "data" in data: tweet = Tweet(data["data"]) self.on_tweet(tweet) if "includes" in data: includes = self._process_includes(data["includes"]) self.on_includes(includes) if "errors" in data: errors = data["errors"] self.on_errors(errors) if "matching_rules" in data: matching_rules = [ StreamRule(id=rule["id"], tag=rule["tag"]) for rule in data["matching_rules"] ] self.on_matching_rules(matching_rules) self.on_response( StreamResponse(tweet, includes, errors, matching_rules) ) def on_tweet(self, tweet): """This is called when a Tweet is received. Parameters ---------- tweet : Tweet The Tweet received """ pass def on_includes(self, includes): """This is called when includes are received. Parameters ---------- includes : dict The includes received """ pass def on_errors(self, errors): """This is called when errors are received. Parameters ---------- errors : dict The errors received """ log.error("Received errors: %s", errors) def on_matching_rules(self, matching_rules): """This is called when matching rules are received. Parameters ---------- matching_rules : list[StreamRule] The matching rules received """ pass def on_response(self, response): """This is called when a response is received. Parameters ---------- response : StreamResponse The response received """ log.debug("Received response: %s", response) class StreamRule(NamedTuple): """Rule for filtered stream .. versionadded:: 4.6 Parameters ---------- value : str | None The rule text. If you are using a `Standard Project`_ at the Basic `access level`_, you can use the basic set of `operators`_, can submit up to 25 concurrent rules, and can submit rules up to 512 characters long. If you are using an `Academic Research Project`_ at the Basic access level, you can use all available operators, can submit up to 1,000 concurrent rules, and can submit rules up to 1,024 characters long. tag : str | None The tag label. This is a free-form text you can use to identify the rules that matched a specific Tweet in the streaming response. Tags can be the same across rules. id : str | None Unique identifier of this rule. This is returned as a string. .. _Standard Project: https://developer.twitter.com/en/docs/projects .. _access level: https://developer.twitter.com/en/products/twitter-api/early-access/guide#na_1 .. _operators: https://developer.twitter.com/en/docs/twitter-api/tweets/search/integrate/build-a-query .. _Academic Research Project: https://developer.twitter.com/en/docs/projects """ value: str = None tag: str = None id: str = None