This is the Scrapy engine which controls the Scheduler, Downloader and Spiders.

For more information see docs/topics/architecture.rst

import logging
from time import time

from twisted.internet import defer, task
from twisted.python.failure import Failure

from scrapy import signals
from scrapy.core.scraper import Scraper
from scrapy.exceptions import DontCloseSpider
from scrapy.http import Response, Request
from scrapy.utils.misc import load_object
from scrapy.utils.reactor import CallLaterOnce
from scrapy.utils.log import logformatter_adapter, failure_to_exc_info

logger = logging.getLogger(__name__)

class Slot(object):

    def __init__(self, start_requests, close_if_idle, nextcall, scheduler):
        self.closing = False
        self.inprogress = set() # requests in progress
        self.start_requests = iter(start_requests)
        self.close_if_idle = close_if_idle
        self.nextcall = nextcall
        self.scheduler = scheduler
        self.heartbeat = task.LoopingCall(nextcall.schedule)

    def add_request(self, request):

    def remove_request(self, request):

    def close(self):
        self.closing = defer.Deferred()
        return self.closing

    def _maybe_fire_closing(self):
        if self.closing and not self.inprogress:
            if self.nextcall:
                if self.heartbeat.running:

class ExecutionEngine(object):

    def __init__(self, crawler, spider_closed_callback):
        self.crawler = crawler
        self.settings = crawler.settings
        self.signals = crawler.signals
        self.logformatter = crawler.logformatter
        self.slot = None
        self.spider = None
        self.running = False
        self.paused = False
        self.scheduler_cls = load_object(self.settings['SCHEDULER'])
        downloader_cls = load_object(self.settings['DOWNLOADER'])
        self.downloader = downloader_cls(crawler)
        self.scraper = Scraper(crawler)
        self._spider_closed_callback = spider_closed_callback

    def start(self):
        """Start the execution engine"""
        assert not self.running, "Engine already running"
        self.start_time = time()
        yield self.signals.send_catch_log_deferred(signal=signals.engine_started)
        self.running = True
        self._closewait = defer.Deferred()
        yield self._closewait

    def stop(self):
        """Stop the execution engine gracefully"""
        assert self.running, "Engine not running"
        self.running = False
        dfd = self._close_all_spiders()
        return dfd.addBoth(lambda _: self._finish_stopping_engine())

    def close(self):
        """Close the execution engine gracefully.

        If it has already been started, stop it. In all cases, close all spiders
        and the downloader.
        if self.running:
            # Will also close spiders and downloader
            return self.stop()
        elif self.open_spiders:
            # Will also close downloader
            return self._close_all_spiders()
            return defer.succeed(self.downloader.close())

    def pause(self):
        """Pause the execution engine"""
        self.paused = True

    def unpause(self):
        """Resume the execution engine"""
        self.paused = False

    def _next_request(self, spider):
        slot = self.slot
        if not slot:

        if self.paused:

        while not self._needs_backout(spider):
            if not self._next_request_from_scheduler(spider):

        if slot.start_requests and not self._needs_backout(spider):
                request = next(slot.start_requests)
            except StopIteration:
                slot.start_requests = None
            except Exception:
                slot.start_requests = None
                logger.error('Error while obtaining start requests',
                             exc_info=True, extra={'spider': spider})
                self.crawl(request, spider)

        if self.spider_is_idle(spider) and slot.close_if_idle:

    def _needs_backout(self, spider):
        slot = self.slot
        return not self.running \
            or slot.closing \
            or self.downloader.needs_backout() \
            or self.scraper.slot.needs_backout()

    def _next_request_from_scheduler(self, spider):
        slot = self.slot
        request = slot.scheduler.next_request()
        if not request:
        d = self._download(request, spider)
        d.addBoth(self._handle_downloader_output, request, spider)
        d.addErrback(lambda f: logger.info('Error while handling downloader output',
                                           extra={'spider': spider}))
        d.addBoth(lambda _: slot.remove_request(request))
        d.addErrback(lambda f: logger.info('Error while removing request from slot',
                                           extra={'spider': spider}))
        d.addBoth(lambda _: slot.nextcall.schedule())
        d.addErrback(lambda f: logger.info('Error while scheduling new request',
                                           extra={'spider': spider}))
        return d

    def _handle_downloader_output(self, response, request, spider):
        assert isinstance(response, (Request, Response, Failure)), response
        # downloader middleware can return requests (for example, redirects)
        if isinstance(response, Request):
            self.crawl(response, spider)
        # response is a Response or Failure
        d = self.scraper.enqueue_scrape(response, request, spider)
        d.addErrback(lambda f: logger.error('Error while enqueuing downloader output',
                                            extra={'spider': spider}))
        return d

    def spider_is_idle(self, spider):
        if not self.scraper.slot.is_idle():
            # scraper is not idle
            return False

        if self.downloader.active:
            # downloader has pending requests
            return False

        if self.slot.start_requests is not None:
            # not all start requests are handled
            return False

        if self.slot.scheduler.has_pending_requests():
            # scheduler has pending requests
            return False

        return True

    def open_spiders(self):
        return [self.spider] if self.spider else []

    def has_capacity(self):
        """Does the engine have capacity to handle more spiders"""
        return not bool(self.slot)

    def crawl(self, request, spider):
        assert spider in self.open_spiders, \
            "Spider %r not opened when crawling: %s" % (spider.name, request)
        self.schedule(request, spider)

    def schedule(self, request, spider):
                request=request, spider=spider)
        if not self.slot.scheduler.enqueue_request(request):
                                        request=request, spider=spider)

    def download(self, request, spider):
        slot = self.slot
        d = self._download(request, spider)
        d.addBoth(self._downloaded, slot, request, spider)
        return d

    def _downloaded(self, response, slot, request, spider):
        return self.download(response, spider) \
                if isinstance(response, Request) else response

    def _download(self, request, spider):
        slot = self.slot
        def _on_success(response):
            assert isinstance(response, (Response, Request))
            if isinstance(response, Response):
                response.request = request # tie request to response received
                logkws = self.logformatter.crawled(request, response, spider)
                logger.log(*logformatter_adapter(logkws), extra={'spider': spider})
                self.signals.send_catch_log(signal=signals.response_received, \
                    response=response, request=request, spider=spider)
            return response

        def _on_complete(_):
            return _

        dwld = self.downloader.fetch(request, spider)
        return dwld

    def open_spider(self, spider, start_requests=(), close_if_idle=True):
        assert self.has_capacity(), "No free spider slot when opening %r" % \
        logger.info("Spider opened", extra={'spider': spider})
        nextcall = CallLaterOnce(self._next_request, spider)
        scheduler = self.scheduler_cls.from_crawler(self.crawler)
        start_requests = yield self.scraper.spidermw.process_start_requests(start_requests, spider)
        slot = Slot(start_requests, close_if_idle, nextcall, scheduler)
        self.slot = slot
        self.spider = spider
        yield scheduler.open(spider)
        yield self.scraper.open_spider(spider)
        yield self.signals.send_catch_log_deferred(signals.spider_opened, spider=spider)

    def _spider_idle(self, spider):
        """Called when a spider gets idle. This function is called when there
        are no remaining pages to download or schedule. It can be called
        multiple times. If some extension raises a DontCloseSpider exception
        (in the spider_idle signal handler) the spider is not closed until the
        next loop and this function is guaranteed to be called (at least) once
        again for this spider.
        res = self.signals.send_catch_log(signal=signals.spider_idle, \
            spider=spider, dont_log=DontCloseSpider)
        if any(isinstance(x, Failure) and isinstance(x.value, DontCloseSpider) \
                for _, x in res):

        if self.spider_is_idle(spider):
            self.close_spider(spider, reason='finished')

    def close_spider(self, spider, reason='cancelled'):
        """Close (cancel) spider and clear all its outstanding requests"""

        slot = self.slot
        if slot.closing:
            return slot.closing
        logger.info("Closing spider (%(reason)s)",
                    {'reason': reason},
                    extra={'spider': spider})

        dfd = slot.close()

        def log_failure(msg):
            def errback(failure):
                    extra={'spider': spider}
            return errback

        dfd.addBoth(lambda _: self.downloader.close())
        dfd.addErrback(log_failure('Downloader close failure'))

        dfd.addBoth(lambda _: self.scraper.close_spider(spider))
        dfd.addErrback(log_failure('Scraper close failure'))

        dfd.addBoth(lambda _: slot.scheduler.close(reason))
        dfd.addErrback(log_failure('Scheduler close failure'))

        dfd.addBoth(lambda _: self.signals.send_catch_log_deferred(
            signal=signals.spider_closed, spider=spider, reason=reason))
        dfd.addErrback(log_failure('Error while sending spider_close signal'))

        dfd.addBoth(lambda _: self.crawler.stats.close_spider(spider, reason=reason))
        dfd.addErrback(log_failure('Stats close failure'))

        dfd.addBoth(lambda _: logger.info("Spider closed (%(reason)s)",
                                          {'reason': reason},
                                          extra={'spider': spider}))

        dfd.addBoth(lambda _: setattr(self, 'slot', None))
        dfd.addErrback(log_failure('Error while unassigning slot'))

        dfd.addBoth(lambda _: setattr(self, 'spider', None))
        dfd.addErrback(log_failure('Error while unassigning spider'))

        dfd.addBoth(lambda _: self._spider_closed_callback(spider))

        return dfd

    def _close_all_spiders(self):
        dfds = [self.close_spider(s, reason='shutdown') for s in self.open_spiders]
        dlist = defer.DeferredList(dfds)
        return dlist

    def _finish_stopping_engine(self):
        yield self.signals.send_catch_log_deferred(signal=signals.engine_stopped)