Using Scrapy

tip

Our CLI now has native support for running Scrapy spiders on Apify! Check out the Scrapy migration guide for more information.

Scrapy is an open-source web scraping framework written in Python. It provides a complete set of tools for web scraping, including the ability to define how to extract data from websites, handle pagination and navigation.

Some of the key features of Scrapy for web scraping include:

• Request and response handling - Scrapy provides an easy-to-use interface for making HTTP requests and handling responses, allowing you to navigate through web pages and extract data.
• Robust Spider framework - Scrapy has a spider framework that allows you to define how to scrape data from websites, including how to follow links, how to handle pagination, and how to parse the data.
• Built-in data extraction - Scrapy includes built-in support for data extraction using XPath and CSS selectors, allowing you to easily extract data from HTML and XML documents.
• Integration with other tool - Scrapy can be integrated with other Python tools like BeautifulSoup and Selenium for more advanced scraping tasks.

Using Scrapy in Actors

To create Actors which use Scrapy, start from the Scrapy & Python Actor template.

This template already contains the structure and setup necessary to integrate Scrapy into your Actors, setting up the Scrapy settings, asyncio reactor, Actor logger and item pipeline as necessary to make Scrapy spiders run in Actors and save their outputs in Apify datasets.

Manual setup

If you don't want to use the template, there are several things you need to set up.

Event loop & reactor

Since the Actor class uses asyncio under the hood, Scrapy has to use the AsyncioSelectorReactor reactor. And to be able to run the Scrapy engine in an already running loop, you have to use the nest_asyncio package.

Item pipeline

To push the results into the Actor's default dataset, the engine has to use a custom ItemPipeline that calls Actor.push_data() on the scraped items.

Example Actor

This is a simple Actor that recursively scrapes titles from all linked websites, up to a maximum depth, starting from URLs in the Actor input.

It uses Scrapy download the pages, extract the results from each page, and continue recursively through the website pagination.

src/main.py

from urllib.parse import urljoin
import nest_asyncio
import scrapy
from itemadapter import ItemAdapter
from scrapy.crawler import CrawlerProcess
from scrapy.utils.project import get_project_settings
from scrapy.utils.reactor import install_reactor

from apify import Actor

# This is necessary so that twisted and asyncio work well together
install_reactor('twisted.internet.asyncioreactor.AsyncioSelectorReactor')
nest_asyncio.apply()

# Scrapes titles pages and enqueues all links it finds on the page
class TitleSpider(scrapy.Spider):
    name = 'title_spider'

    def __init__(self, start_urls, *args, **kwargs):
        super().__init__(*args, **kwargs)
        self.start_urls = start_urls

    def parse(self, response):
        yield {
            'url': response.url,
            'title': response.css('title::text').extract_first(),
        }
        for link_href in response.css('a::attr("href")'):
            link_url = urljoin(response.url, link_href.get())
            if link_url.startswith(('http://', 'https://')):
                yield scrapy.Request(link_url)

# Pushes the scraped items into the Actor's default dataset
class ActorDatasetPushPipeline:
    async def process_item(self, item, spider):
        item_dict = ItemAdapter(item).asdict()
        await Actor.push_data(item_dict)
        return item

async def main():
    async with Actor:
        actor_input = await Actor.get_input() or {}
        start_urls = actor_input.get('start_urls', [{ 'url': 'https://apify.com' }])
        start_urls = [start_url.get('url') for start_url in start_urls]

        settings = get_project_settings()
        settings['ITEM_PIPELINES'] = { ActorDatasetPushPipeline: 1 }
        settings['TWISTED_REACTOR'] = 'twisted.internet.asyncioreactor.AsyncioSelectorReactor'
        settings['DEPTH_LIMIT'] = actor_input.get('max_depth', 1)

        process = CrawlerProcess(settings)
        process.crawl(TitleSpider, start_urls=start_urls)
        process.start()