Crawlee Tutorial: Easy Web Scraping and Browser Automation

Web scraping and browser automation have emerged as essential tools for businesses looking to stay competitive in the digital market. This easy tutorial covers everything you need to get started with Crawlee – a tool for web scraping and browser automation.

What is Crawlee?

Crawlee is a Node.JS package that offers a straightforward and adaptable interface for web scraping and browser automation. Users can retrieve web pages, apply CSS selectors to extract data from them, and navigate the DOM tree to follow links and scrape several sites.

Crawlee is a versatile tool that provides a uniform interface for web crawling via HTTP and headless browser approaches. It has an integrated persistent queue for handling URLs to crawl in either breadth-first or depth-first order.

Users can also benefit from integrated proxy rotation and session management, pluggable storage solutions for files and tabular data, and other features. Moreover, Crawlee offers hook-based customized lifecycles, programmable routing, error handling, and retries.

For speedy project setup, a CLI is accessible, and Dockerfiles are included to streamline deployment. Crawlee is a robust and effective tool for web scraping and crawling written in TypeScript using generics.

The benefits of using Crawlee for web scraping and browser automation

The following are some of the most common pros of using Crawlee for browser automation and web scraping:

• Single interface: Crawlee offers a single interface for headless browser crawling as well as HTTP crawling, making it simple to switch between the two based on your needs.

• Customizable lifecycles: Crawlee allows developers to alter their crawlers' lifecycles using hooks. These hooks can be used to carry out operations before or after specific events, such as before a request is made or after data is collected.

• Pluggable storage: Crawlee supports pluggable storage methods for both tabular data and files, making storing and managing the data you extract simple.

• Proxy rotation and session management: Crawlee has built-in support for these features, which can be used to manage complex web interactions and avoid IP blocking.

• Configurable request routing, error handling, and retries: Crawlee enables developers to control request routing, deal with errors, and retry requests as needed, making it simpler to handle edge cases and unexpected issues.

• Docker-capable: Crawlee comes with Dockerfiles that let developers rapidly and easily deploy their crawlers to production environments.

Crawlee web scraping tutorial

This section discusses the installation steps of Crawlee and explains how it works. Additionally, it contains a working example of scraping a website using Crawlee. 

Installation

To use Crawlee on your system, you must install Node.JS version 16.0 or above. Along with it, NPM should also be installed. 

The Crawlee CLI is the quickest and most efficient way to build new projects with Crawlee. The following command will create a new Crawlee project inside the “my-crawler” directory:

npx crawlee create my-crawler

The npx CLI tool runs the crawlee package locally without installing it globally on your machine. Running the above command will show you a prompt message to choose a template, as shown in the following snippet:

Since we are doing Node.JS development, we can select Getting Started Example (Javascript). This option will install all required dependencies and create a new directory named my-crawler in your current working directory. It will also add a package.json to this folder. Additionally, it will include example source code that you may use immediately. The image below shows the completed installation message:

Remember, the Crawlee project was created inside the my-crawler folder. You first need to change your current directory to this folder.

cd my-crawler

Next, run the following command to start the Crawlee project:

npm start

The start command will start the default crawler of the project. This crawler crawls the Crawlee website and outputs the titles of all the links on the website. 

Congratulations! You have successfully installed Crawlee and run one of the crawlers with it.

Crawlee working

Crawlee has three types of crawlers: CheerioCrawler, PuppeteerCrawler, and PlaywrightCrawler. They all share some basic characteristics.

Every crawler is designed to visit a webpage, carry out specific tasks, save the results, navigate to the next page, and repeat this cycle until the job is finished. Each crawler must therefore respond to two queries: Where should I go? What should I do there?

The crawler can start working when these queries are resolved since most other settings are pre-configured for the crawlers.

Crawlee web scraping example 

Our target for this web scraping demonstration is books.toscrape.com. We will be scraping the titles of the books listed on the website.

Open the main.js file from the src folder of your project and overwrite the following code in it:

import { PlaywrightCrawler } from 'crawlee';

const crawler = new PlaywrightCrawler({
    requestHandler: async ({ page }) => {
        
        // Waiting for book titles to load
        await page.waitForSelector('h3');
        
        // Execute a function in the browser that targets
        // the book title elements and allows their manipulation
        const bookTitles = await page.$$eval('h3', (els) => {
        
            // Extract text content from the titles list
            return els.map((el) => el.textContent);
        });
        
         bookTitles.forEach((text, i) => {
            console.log(`Book_${i + 1}: ${text}\n`);
        });
    },
});

await crawler.run(['https://books.toscrape.com/']);

This code illustrates how to scrape data from a website using Crawlee's PlaywrightCrawler class.

The code first imports the PlaywrightCrawler class from the crawlee package. Then it creates a new crawler of the PlaywrightCrawler type.

Instantiating the PlaywrightCrawler class requires an options object which includes a requestHandler function. The requestHandler function is an asynchronous function executed for every page visited by the crawler.

The requestHandler function first waits for the page's <h3> elements, representing the book titles, to be rendered using the page. waitForSelector() call. The page.$$eval() method is executed in the browser's context, extracting the text content from all of the <h3> elements on the page. Lastly, the crawler code logs the book titles to the console.

The last line of code initiates the crawling operation using the crawler.run() method.

The following command runs the project:

npm start

Here is what the code output when once you run the crawlee project: