Welcome to another post on the Crawlee blog; this time, we are going to compare Scrapy, one of the oldest and most popular web scraping libraries in the world, with Crawlee, a relative newcomer. This article will answer your questions about when to use Scrapy and help you decide when it would be better to use Crawlee instead. This article will be the first in a series comparing the various technical aspects of Crawlee with Scrapy.
Introduction:
Scrapy is an open-source Python-based web scraping framework that extracts data from websites. With Scrapy, you create spiders, which are autonomous scripts to download and process web content. The limitation of Scrapy is that it does not work very well with JavaScript rendered websites, as it was designed for static HTML pages. We will do a comparison later in the article about this.
Crawlee is also an open-source library that originated as Apify SDK. Crawlee has the advantage of being the latest library in the market, so it already has many features that Scrapy lacks, like autoscaling, headless browsing, working with JavaScript rendered websites without any plugins, and many more, which we are going to explain later on.
Feature comparison
We'll start comparing Scrapy and Crawlee by looking at language and development environments, and then features to make the scraping process easier for developers, like autoscaling, headless browsing, queue management, and more.
Language and development environments:
Scrapy is written in Python, making it easier for the data science community to integrate it with various tools. While Scrapy offers very detailed documentation, it can take a lot of work to get started with Scrapy. One of the reasons why it is considered not so beginner-friendly[1][2][3] is its complex architecture, which consists of various components like spiders, middleware, item pipelines, and settings. These can be challenging for beginners.
Crawlee is one of the few web scraping and automation libraries that supports JavaScript and TypeScript. Crawlee supports CLI just like Scrapy, but it also provides pre-built templates in TypeScript and JavaScript with support for Playwright and Puppeteer. These templates help beginners to quickly understand the file structure and how it works.
Headless browsing and JS rendering
Scrapy does not support headless browsers natively, but it supports them with its plugin system, similarly it does not support scraping JavaScript rendered websites, but the plugin system makes this possible. One of the best examples is its Playwright plugin.
Apify Store is a JavaScript rendered website, so we will scrape it in this example using the scrapy-playwright integration.
For installation and to make changes to [settings.py], please follow the instructions on the scrapy-playwrightrepository on GitHub.
Then, create a spider with this code to scrape the data:
In Crawlee, you can scrape JavaScript rendered websites using the built-in headless Puppeteer and Playwright browsers. It is important to note that, by default, Crawlee scrapes in headless mode. If you don't want headless, then just set headless: false.
Autoscaling refers to the capability of a library to automatically adjusting the number of concurrent tasks (such as browser instances, HTTP requests, etc.) based on the current load and system resources. This feature is particularly useful when handling web scraping and crawling tasks that may require dynamically scaled resources to optimize performance, manage system load, and handle rate limitations efficiently.
Scrapy does not have built-in autoscaling capabilities, but it can be done using external services like Scrapyd or deployed in a distributed manner with Scrapy Cluster.
Crawlee has built-in autoscaling with AutoscaledPool. It increases the number of requests that are processed concurrently within one crawler.
Queue Management
Scrapy supports both breadth-first and depth-first crawling strategies using a disk-based queuing system. By default, it uses the LIFO queue for the pending requests, which means it is using depth-first order, but if you want to use breadth-first order, you can do it by changing these settings:
Crawlee uses breadth-first by default and you can override it on a per-request basis by using the forefront: true argument in addRequest and its derivatives. If you use forefront: true for all requests, it becomes a depth-first process.
CLI Support
Scrapy has a powerful command-line interface that offers functionalities like starting a project, generating spiders, and controlling the crawling process.
Scrapy CLI comes with Scrapy. Just run this command, and you are good to go:
pip install scrapy
Crawlee also includes a CLI tool (crawlee-cli) that facilitates project setup, crawler creation and execution, streamlining the development process for users familiar with Node.js environments. The command for installation is:
npx crawlee create my-crawler
Proxy Rotation and Storage Management
Scrapy handles it via custom middleware. You have to install their scrapy-rotating-proxies package using pip.
pip install scrapy-rotating-proxies
Then in the settings.py file, add ROTATING_PROXY_LIST and the middleware to the DOWNLOADER_MIDDLEWARES and specify the list of proxy servers. For example:
DOWNLOADER_MIDDLEWARES={# Lower value means higher priority
'scrapy.downloadermiddlewares.retry.RetryMiddleware':90,'scrapy_rotating_proxies.middlewares.RotatingProxyMiddleware':610,'scrapy_rotating_proxies.middlewares.BanDetectionMiddleware':620,}ROTATING_PROXY_LIST=['proxy1.com:8000','proxy2.com:8031',# Add more proxies as needed
]
Now create a spider with the code you want to scrape any site and the ROTATING_PROXY_LIST in settings.py will manage which proxy to use for each request. Here middleware will treat each proxy initially as valid and then when a request is made, the middleware selects a proxy from the list of available proxies. The selection isn't purely sequential but is influenced by the recent history of proxy performance. The middleware has mechanisms to detect when a proxy might be banned or rendered ineffective. When such conditions are detected, the proxy is temporarily deactivated and put into a cooldown period. After the cooldown period expires, the proxy is reconsidered for use.
In Crawlee, you can use your own proxy servers or proxy servers acquired from third-party providers. If you already have your proxy URLs, you can start using them like this:
Crawlee also has SessionPool, a built-in allocation system for proxies. It handles the rotation, creation, and persistence of user-like sessions. It creates a pool of session instances that are randomly rotated.
Data Storage
One of the most frequently required features when implementing scrapers is being able to store the scraped data as an "export file".
Scrapy provides this functionality out of the box with the Feed Exports, which allows it to generate feeds with the scraped items, using multiple serialization formats and storage backends. It supports CSV, JSON, JSON Lines, and XML.
To do this, you need to modify your settings.py file and enter:
# To store in CSV format
FEEDS={'data/crawl_data.csv':{'format':'csv','overwrite':True}}# OR to store in JSON format
FEEDS={'data/crawl_data.json':{'format':'json','overwrite':True}}
Crawlee's storage can be divided into two categories: Request Storage (Request Queue and Request List) and Results Storage (Datasets and Key Value Stores). Both are stored locally by default in the ./storage directory.
Also, remember that Crawlee, by default, clears its storages before starting a crawler run. This action is taken to prevent old data from interfering with new crawling sessions.
import{KeyValueStore}from'crawlee';//... Code to crawl the dataawaitKeyValueStore.setValue('key',{foo:'bar'});
Anti-blocking and fingerprints
In Scrapy, handling anti-blocking strategies like IP rotation, user-agent rotation, custom solutions via middleware, and plugins are needed.
Crawlee provides HTTP crawling and browser fingerprints with zero configuration necessary; fingerprints are enabled by default and available in PlaywrightCrawler and PuppeteerCrawler but also work with CheerioCrawler and the other HTTP Crawlers.
Error handling
Both libraries support error-handling practices like automatic retries, logging, and custom error handling.
In Scrapy, you can handle errors using middleware and signals. There are also exceptions like IgnoreRequest, which can be raised by Scheduler or any downloader middleware to indicate that the request should be ignored. Similarly, a spider callback can raise' CloseSpider' to close the spider.
Scrapy has built-in support for retrying failed requests. You can configure the retry policy (e.g., the number of retries, retrying on particular HTTP codes) via settings such as RETRY_TIMES, as shown in the example:
RETRY_ENABLED=TrueRETRY_TIMES=2# Number of retry attempts
RETRY_HTTP_CODES=[500,502,503,504,522,524]# HTTP error codes to retry
In Crawlee, you can also set up a custom error handler. For retries, maxRequestRetries controls how often Crawlee will retry a request before marking it as failed. To set it up, you just need to add the following line of code in your crawler.
constcrawler=newCheerioCrawler({maxRequestRetries:3// Crawler will retry three times.// ...})
There is also noRetry. If set to true then the request will not be automatically tried.
Crawlee also provides a built-in logging mechanism via log, allowing you to log warnings, errors, and other information effectively.
Deployment using Docker
Scrapy can be containerized using Docker, though it typically requires manual setup to create Dockerfiles and configure environments. While Crawlee includes ready-to-use Docker configurations, making deployment straightforward across various environments without additional configuration.
Community
Both projects are open source. Scrapy benefits from a large and well-established community. It has been around since 2008 and has attracted a lot of attention among developers, particularly those in the Python ecosystem.
Crawlee started its journey as Apify SDK in 2018. It now has more than 12K stars on GitHub.
CrawleeβA web scraping and browser automation library for Node.js to build reliable crawlers. In JavaScript and TypeScript. Extract data for AI, LLMs, RAG, or GPTs. Download HTML, PDF, JPG, PNG, and other files from websites. Works with Puppeteer, Playwright, Cheerio, JSDOM, and raw HTTP. Both headful and headless mode. With proxy rotation.
A web scraping and browser automation library
Crawlee covers your crawling and scraping end-to-end and helps you build reliable scrapers. Fast.
Your crawlers will appear human-like and fly under the radar of modern bot protections even with the default configuration. Crawlee gives you the tools to crawl the web for links, scrape data, and store it to disk or cloud while staying configurable to suit your project's needs.
This is the official developer community of Apify and Crawlee. | 8756 members
discord.com
It is used by the TypeScript and JavaScript community.
Conclusion
Both frameworks can handle a wide range of scraping tasks, and the best choice will depend on specific technical needs like language preference, project requirements, ease of use, etc.
If you are comfortable with Python and want to work only with it, go with Scrapy. It has very detailed documentation, and it is one of the oldest and most stable libraries in the space.
But if you want to explore or are comfortable working with TypeScript or JavaScript, our recommendation is Crawlee. With all the valuable features like a single interface for HTTP requests and headless browsing, making it work well with JavaScript rendered websites, autoscaling and fingerprint support, it is the best choice for scraping websites that can be complex, resource intensive, using JavaScript, or even have blocking methods.
As promised, this is just the first of the many articles comparing Scrapy and Crawlee. With the upcoming articles, you will learn more about every technical detail.