gajus / headless-crawler

A crawler implemented using a headless browser (Chrome).

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

headless-crawler 👻

Travis build status Coveralls NPM version Canonical Code Style Twitter Follow

A crawler implemented using a headless browser (Chrome).

Features

Usage

import puppeteer from 'puppeteer';
import {
  createHeadlessCrawler
} from 'headless-crawler';

const main = async () => {
  const browser = puppeteer.launch();

  // See Configuration documentation.
  const headlessCrawler = createHeadlessCrawler({
    onResult: (resource) => {
      console.log(resource.content.title);
    },
    browser
  });

  await headlessCrawler.queue('http://gajus.com/');
};

main();

Configuration

/**
 * @property browser Instance of [Puppeteer Browser](https://pptr.dev/#?product=Puppeteer&version=v1.11.0&show=api-class-browser).
 * @property concurrency Maximum concurrency. Defaults to 5.
 * @property extractContent A function [evaluted](https://pptr.dev/#?product=Puppeteer&version=v1.11.0&show=api-pageevaluatepagefunction-args) in the context of the browser. The result of the function is used to describe the contents of the website (see `ScrapeResultType#content` property).
 * @property filterLink Identifies which URLs to follow.
 * @property onPage Invoked when [Puppeteer Page](https://pptr.dev/#?product=Puppeteer&version=v1.11.0&show=api-class-page) instance is instantiated.
 * @property onResult Invoked after content is extracted from a new page. Must return a boolean value indicating whether the crawler should advance to the next URL.
 * @property sortQueuedLinks Sorts queued links.
 * @property waitFor Invoked before links are aggregated from the website and before `extractContent`.
 */
type HeadlessCrawlerUserConfigurationType = {|
  +browser: PuppeteerBrowserType,
  +concurrency: number,
  +extractContent?: ExtractContentHandlerType,
  +filterLink?: FilterLinkHandlerType,
  +onPage?: PageHandlerType,
  +onResult?: ResultHandlerType,
  +sortQueuedLinks?: SortQueuedLinksHandlerType,
  +waitFor?: WaitForHandlerType
|};

Default headlessCrawlerConfiguration.extractContent

The default extractContent function extracts page title.

(): ExtractContentHandlerType => {
  return `(() => {
    return {
      title: document.title
    };
  })();`;
};

Default headlessCrawlerConfiguration.filterLink

The default filterLink function includes all URLs allowed by robots.txt and does not visit previously scraped URLs.

type DefaultFilterLinkHandlerConfigurationType = {|
  +maxLinkDepth: number,
  +respectRobots: boolean
|};

type DefaultFilterLinkHandlerUserConfigurationType = {|
  +maxLinkDepth?: number,
  +respectRobots?: boolean
|};

(userConfiguration: DefaultFilterLinkHandlerUserConfigurationType): FilterLinkHandlerType => {
  const configuration: DefaultFilterLinkHandlerConfigurationType = {
    maxLinkDepth: 10,
    respectRobots: true,
    ...userConfiguration
  };

  let robotsAgent;

  if (configuration.respectRobots) {
    robotsAgent = createRobotsAgent();
  }

  return async (link, scrapedLinkHistory) => {
    if (link.linkDepth > configuration.maxLinkDepth) {
      return false;
    }

    if (configuration.respectRobots && robotsAgent.isRobotsAvailable(link.linkUrl) && !robotsAgent.isAllowed(link.linkUrl)) {
      return false;
    }

    for (const scrapedLink of scrapedLinkHistory) {
      if (scrapedLink.linkUrl === link.linkUrl) {
        return false;
      }
    }

    return true;
  };
};

Note: robots.txt support is implemented using robots-agent.

Default headlessCrawlerConfiguration.onError

(): ErrorHandlerType => {
  return (error) => {};
};

Default headlessCrawlerConfiguration.onResult

The default onResult logs the result and advances crawler to the next URL.

(): ResultHandlerType => {
  return (scrapeResult) => {
    log.debug({
      scrapeResult
    }, 'new result');

    return true;
  };
};

Default headlessCrawlerConfiguration.sortQueuedLinks

(): SortQueuedLinksHandlerType => {
  return (links) => {
    return links;
  };
};

Default headlessCrawlerConfiguration.waitFor

(): WaitForHandlerType => {
  return (page) => {
    return page.waitForNavigation({
      waitUntil: 'networkidle2'
    });
  };
};

Create default handlers

You can import factory functions to create default handlers:

import {
  createDefaultExtractContentHandler,
  createDefaultFilterLinkHandler,
  createDefaultResultHandler,
  createDefaultSortQueuedLinksHandler,
  createDefaultWaitForHandler
} from 'headless-crawler';

This is useful for extending the default handlers, e.g.

const defaultFilterHandler = createDefaultFilterLinkHandler();

const myCustomFilterLinkHandler = (link, scrapedLinkHistory) => {
  if (link.linkUrl.startsWith('https://google.com/')) {
    return false;
  }

  return defaultFilterHandler(link, scrapedLinkHistory);
};

Recipes

Inject jQuery

Use extractContent to manipulate the Puppeteer Page object after it has been determined to be ready and create the function used to extract content from the website.

const main = async () => {
  const browser = await puppeteer.launch();

  const headlessCrawler = createHeadlessCrawler({
    browser,
    extractContent: async (page) => {
      await page.addScriptTag({
        url: 'https://code.jquery.com/jquery-3.3.1.min.js'
      });

      return `(() => {
        return $('title').text();
      })()`;
    }
  });
};

main();

Configure request parameters

Request parameters (such as geolocation, user-agent and viewport) can be configured using onPage handler, e.g.

const main = async () => {
  const browser = await puppeteer.launch();

  const onPage = async (page, scrapeConfiguration) => {
    await page.setGeolocation({
      latitude: 59.95,
      longitude: 30.31667
    });
    await page.setUserAgent('headless-crawler');
  };

  const headlessCrawler = createHeadlessCrawler({
    browser,
    onPage
  });
};

main();

Capture a screenshot

The extractContent method can capture the screenshot of the website as it was at the time just before the content-extraction function is executed, e.g.

const extractContent = async (page) => {
  await page.screenshot({
    fullPage: true,
    path: 'screenshot.png'
  });

  return `(() => {
    return {
      title: document.title
    };
  })()`;
};

Refer to Puppeteer page#screenshot documentation for other properties.

Configure a proxy

Note: These instructions are not specific headless-crawler; these are generic instructions for instructing Puppeteer to use HTTP proxy.

You must:

  1. Configure ignoreHTTPSErrors
  2. Configure --proxy-server

Example:

import puppeteer from 'puppeteer';
import {
  createHeadlessCrawler
} from 'headless-crawler';

const main = async () => {
  const browser = puppeteer.launch({
    args: [
      '--proxy-server=http://127.0.0.1:8080'
    ],
    ignoreHTTPSErrors: true
  });

  const headlessCrawler = createHeadlessCrawler({
    onResult: (resource) => {
      console.log(resource.content.title);
    },
    browser
  });

  await headlessCrawler.queue('http://gajus.com/');
};

main();

Types

This package is using Flow type annotations.

Refer to ./src/types.js for method parameter and result types.

Logging

This package is using roarr logger to log the program's state.

Export ROARR_LOG=true environment variable to enable log printing to stdout.

Use roarr-cli program to pretty-print the logs.

FAQ

What makes headless-crawler different from headless-chrome-crawler?

headless-chrome-crawler is the only other headless crawler in the Node.js ecosystem.

It appears that headless-chrome-crawler is no longer maintained. At the time of this writing, the author of headless-chrome-crawler has not made public contributions in over 6 months and the package includes bugs as a result of hardcoded dependency versions.

Maintenance issues aside, the headless-chrome-crawler is a feature-rich and configuration-driven framework. Meanwhile, headless-crawler provides a bare-bones framework for navigating the website and extracting the content. The consumer of the framework can extend the functionality using the provided handlers and directly consuming the Puppeteer API (e.g. see Capture a screenshot).

About

A crawler implemented using a headless browser (Chrome).

License:Other


Languages

Language:JavaScript 100.0%