Got Scraping

Got Scraping is a small but powerful got extension that allows you to send browser-like requests with only a little configuration seamlessly and with a single function call. Got is a widely use and powerful client that provides extensibility and customization. Sending browser-like requests is essential in the web scraping industry to blend in with the website traffic. You can read more about the mimics here.

Motivation

With the increasing popularity of web scraping, it is becoming more important to blend in with the typical internet traffic. This package implements multiple mimics to make GET requests similar to browsers. Thanks to the included header-generator package, you can choose various browsers from different operating systems and devices. The header-generator package generates all the headers for you so that you can focus on the important stuff.

But the got-scraping doesn't stop here. The got-scraping package also tries to mimic the full request fingerprint, including the correct HTTP protocol and TLS ciphers.

Proxies are essential in the web scraping industry. Another goal of this package is to simplify the usage of HTTP and HTTPS proxies. All you have to do is pass the proxyUrl option.

There is one more good news for loyal got users. This package is modified got instance using handlers so that you can use the got interface as always. We've added just a few extras.

• Simulating browser-like requests
• Proxies
• Installation
• Examples
  • Simple GET request
  • GET request with proxy
  • Overriding request headers
  • Customizing Header generator options
  • Get JSON
• API reference
  • Got scraping default values
  • Got scraping extra options
  • Error recovery

Simulating browser-like requests

The Got Scraping package's primary goal is to make the document GET request look like a real browser made it. The crucial part of this process is to send browser-like headers. This process is outsourced to the header-generator package, allowing you to customize browsers, devices, languages, and operating systems to generate the request headers. The generation of the request headers is based on real-world data, and the algorithm is periodically updated. For more information, see the header-generator README.

This package can only generate all the standard attributes. There still might be site-specific attributes that you need to add. One of those attributes you might want to add is the referer header. Referer header might add more credibility to your request and hence reduce the blocking. Check out more info about the referer header here. Please bear in mind that these headers are made for GET requests for HTML documents. If you want to make POST requests or GET requests for any other content type, you should alter these headers according to your needs. You can do so by passing a headers option or writing a custom got handler.

The second colossal factor is using the same HTTP version as browsers. Most modern browsers use HTTP v2, so using it when the server supports it takes you one more step further to look like a browser. Luckily you don't have to care about the server HTTP version support since got automatically handles ALPN protocol negotiation and uses the HTTP v2 only if it is supported. The last step is to have a browser-like TLS suite and ciphers. According to our research, the cipher TLS_AES_256_GCM_SHA384 is used among all modern browsers. We use this cipher as a default one. However, feel free to change it.

This package should provide a solid start for your browser request emulation process. All websites are built differently, and some of them might require some additional special care.

Proxies

Got Scraping package makes using proxies with your requests ridiculously easy. It supports multiple variations of connections through HTTP1 and HTTP2 proxies to HTTP1 and HTTP2 servers. You can see the supported variations below. All you have to do is pass a valid proxyUrl option.

Proxy type	Proxy HTTP version	Agents
https	http2	http2, https, http
https	http1	http2, https
http	http1	http2, https, http

The proxy type and proxy HTTP version is a type of connection to a proxy. The agents are supported connections from the proxy to the target, let's say, a website.

Installation

We strongly recommend using Node 12+ because of the compatible TLS ciphers we use to emulate the browser. However, this package is compatible with Node 10+.

1$ npm install got-scraping

Examples

These examples should help you to grasp the concept of the got-scraping package quickly.

Simple GET request

1const gotScraping = require('got-scraping');
2
3gotScraping
4    .get('https://apify.com')
5    .then( ({ body }) => console.log(body))

GET request with proxy

1const gotScraping = require('got-scraping');
2
3gotScraping({
4    url: 'https://apify.com',
5    proxyUrl: 'http://myproxy.com:1234',
6})
7    .then(({ body }) => console.log(body));
8

Overriding request headers

By passing options:

1const response = await gotScraping({
2    url: 'https://apify.com/',
3    headers: {
4        'user-agent': 'test',
5    },
6});

You can check out how to override request headers using handlers in the original got repo under the instances section.

Customizing Header generator options

1const response = await gotScraping({
2    url: 'https://api.apify.com/v2/browser-info',
3    headerGeneratorOptions:{
4        browsers:[
5            {
6                name: 'chrome',
7                minVersion: 87,
8                maxVersion: 89
9            }
10        ],
11        devices: ['desktop'],
12        languages: ['de-DE', 'en-US'],
13        operatingSystems: ['windows', 'linux'] 
14    }
15});

Get JSON

You can get JSON body with this package too, but please bear in mind that the request header generation is done specifically for HTML content type. You might want to alter the generated headers to match the browser ones.

1const response = await gotScraping({
2    responseType: 'json',
3    url: 'https://api.apify.com/v2/browser-info',
4    ciphers: undefined,
5});

API reference

Got scraping package is built using got.extend functionality and supports all of the got API. On top of that, it adds few more options to the got once and then passes them by a handler to the got context object. This package also alters got defaults to be more scraping friendly.

Got scraping default values

1const SCRAPING_DEFAULT_OPTIONS = {
2    // Most of the new browsers use HTTP2
3    http2: true,
4    https: {
5        // We usually don't want to fail because of SSL errors.
6        // We want the content.
7        rejectUnauthorized: false,
8    },
9    // This would fail all of 404, 403 responses.
10    // We usually don't want to consider these as errors.
11    // We want to take some action after this.
12    throwHttpErrors: false,
13    // Node js uses different TLS ciphers by default.
14    ciphers: getCiphersBasedOnNode(),
15    // We need to have browser-like headers to blend in.
16    useHeaderGenerator: true,
17};

Got scraping extra options

proxyUrl - Url of HTTPS or HTTP based proxy. HTTP2 proxies are supported.

useHeaderGenerator - Turns on the generation of the browser-like header. By default, it is set to true.

headerGeneratorOptions - See the @TODO: proper link (HeaderGeneratroOptions)[https://github.com/apify/header-generator]

Error recovery

This section covers possible errors that might happen due to different site implementations.

RequestError: Client network socket disconnected before secure TLS connection was established - Try changing the ciphers parameter to either undefined or a custom value.