Beispiel #1
0
    def process_request(self, request, spider):
        """
        For Splash requests add 'cookies' key with current
        cookies to request.meta['splash']['args']
        """
        if 'splash' not in request.meta:
            return

        if request.meta.get('_splash_processed'):
            return

        splash_options = request.meta['splash']

        splash_args = splash_options.setdefault('args', {})
        if 'cookies' in splash_args:  # cookies already set
            return

        if 'session_id' not in splash_options:
            return

        jar = self.jars[splash_options['session_id']]

        cookies = self._get_request_cookies(request)
        har_to_jar(jar, cookies)

        splash_args['cookies'] = jar_to_har(jar)
    def process_request(self, request, spider):
        """
        For Splash requests add 'cookies' key with current
        cookies to ``request.meta['splash']['args']`` and remove cookie
        headers sent to Splash itself.
        """
        if 'splash' not in request.meta:
            return

        if request.meta.get('_splash_processed'):
            request.headers.pop('Cookie', None)
            return

        splash_options = request.meta['splash']

        splash_args = splash_options.setdefault('args', {})
        if 'cookies' in splash_args:  # cookies already set
            return

        if 'session_id' not in splash_options:
            return

        jar = self.jars[splash_options['session_id']]

        cookies = self._get_request_cookies(request)
        har_to_jar(jar, cookies)

        splash_args['cookies'] = jar_to_har(jar)
        self._debug_cookie(request, spider)
    def process_request(self, request, spider):
        """
        For Splash requests add 'cookies' key with current
        cookies to request.meta['splash']['args']
        """
        if 'splash' not in request.meta:
            return

        if request.meta.get('_splash_processed'):
            return

        splash_options = request.meta['splash']

        splash_args = splash_options.setdefault('args', {})
        if 'cookies' in splash_args:  # cookies already set
            return

        if 'session_id' not in splash_options:
            return

        jar = self.jars[splash_options['session_id']]

        cookies = self._get_request_cookies(request)
        har_to_jar(jar, cookies)

        splash_args['cookies'] = jar_to_har(jar)
        self._debug_cookie(request, spider)
    def process_request(self, request, spider):
        """
        For Splash requests add 'cookies' key with current
        cookies to request.meta['splash']['args']
        """
        if 'splash' not in request.meta:
            return

        if request.meta.get('_splash_processed'):
            return

        splash_options = request.meta['splash']

        splash_args = splash_options.setdefault('args', {})
        if 'cookies' in splash_args:  # cookies already set
            return

        if 'session_id' not in splash_options:
            return

        _jars = self.redis_conn.get(self.redis_key)
        if _jars is None:
            self.jars = defaultdict(CookieJar)
            self.redis_conn.set(self.redis_key, pickle.dumps(self.jars, protocol=-1))
        else:
            self.jars = pickle.loads(_jars)

        jar = self.jars[splash_options['session_id']]

        cookies = self._get_request_cookies(request)
        har_to_jar(jar, cookies)

        splash_args['cookies'] = jar_to_har(jar)
        self._debug_cookie(request, spider)