通过链接进行JavaScript搜索

Question

I know from this that I can search an entire html page using

var position = document.documentElement.innerHTML.indexOf('dog');

However, I have a webpage with a ton of links and I want to make a simple script to visit and search through every linked page. Is that possible?

Answer 1

Because of same-origin policy and other challenges, this is not presently possible to do solely on the browser. However, there are server-side JavaScript projects using node.js that make this relatively straight-forward to do on a server or on your desktop (but not on your browser). Because it's nodejs, you can use the same kind of JavaScript (even jquery selectors) as you would in the browser. One nice example is node.io, although it's primarily built for use on Linux or OSX:

https://github.com/chriso/node.io

For a more WIndows-friendly approach, you could use nodejs along with Request ( https://github.com/mikeal/request ) and Cheerio, or else jsdom. There are a number of tutorials out there that use this approach (the first tutorial even extracts a list of URLs from a page):

• http://www.sitepoint.com/web-scraping-in-node-js
• http://okfnlabs.org/blog/2013/01/15/web-scraping-with-node-css-selectors.html
• http://blog.miguelgrinberg.com/post/easy-web-scraping-with-nodejs

If you need to work with browsers, you can originate the request in the browser, run it in the server, and then easily send the results to the browser via Ajax. But you'll need a server intermediary.

If you must do this in a browser, then you'll have to build it as some kind of browser extension or add-on.