Please advise how to scrape AJAX pages.
Overview:
All screen scraping first requires manual review of the page you want to extract resources from. When dealing with AJAX you usually just need to analyze a bit more than just simply the HTML.
When dealing with AJAX this just means that the value you want is not in the initial HTML document that you requested, but that javascript will be exectued which asks the server for the extra information you want.
You can therefore usually simply analyze the javascript and see which request the javascript makes and just call this URL instead from the start.
Example:
Take this as an example, assume the page you want to scrape from has the following script:
<script type="text/javascript">
function ajaxFunction()
{
var xmlHttp;
try
{
// Firefox, Opera 8.0+, Safari
xmlHttp=new XMLHttpRequest();
}
catch (e)
{
// Internet Explorer
try
{
xmlHttp=new ActiveXObject("Msxml2.XMLHTTP");
}
catch (e)
{
try
{
xmlHttp=new ActiveXObject("Microsoft.XMLHTTP");
}
catch (e)
{
alert("Your browser does not support AJAX!");
return false;
}
}
}
xmlHttp.onreadystatechange=function()
{
if(xmlHttp.readyState==4)
{
document.myForm.time.value=xmlHttp.responseText;
}
}
xmlHttp.open("GET","time.asp",true);
xmlHttp.send(null);
}
</script>
Then all you need to do is instead do an HTTP request to time.asp of the same server instead. Example from w3schools.
Advanced scraping with C++:
For complex usage, and if you're using C++ you could also consider using the firefox javascript engine SpiderMonkey to execute the javascript on a page.
Advanced scraping with Java:
For complex usage, and if you're using Java you could also consider using the firefox javascript engine for Java Rhino
Advanced scraping with .NET:
For complex usage, and if you're using .Net you could also consider using the Microsoft.vsa assembly. Recently replaced with ICodeCompiler/CodeDOM.
In my opinion the simpliest solution is to use Casperjs, a framework based on the WebKit headless browser phantomjs.
The whole page is loaded, and it's very easy to scrape any ajax-related data.
You can check this basic tutorial to learn Automating & Scraping with PhantomJS and CasperJS
You can also give a look at this example code, on how to scrape google suggests keywords :
/*global casper:true*/
var casper = require('casper').create();
var suggestions = [];
var word = casper.cli.get(0);
if (!word) {
casper.echo('please provide a word').exit(1);
}
casper.start('http://www.google.com/', function() {
this.sendKeys('input[name=q]', word);
});
casper.waitFor(function() {
return this.fetchText('.gsq_a table span').indexOf(word) === 0
}, function() {
suggestions = this.evaluate(function() {
var nodes = document.querySelectorAll('.gsq_a table span');
return [].map.call(nodes, function(node){
return node.textContent;
});
});
});
casper.run(function() {
this.echo(suggestions.join('\n')).exit();
});
If you can get at it, try examining the DOM tree. Selenium does this as a part of testing a page. It also has functions to click buttons and follow links, which may be useful.
The best way to scrape web pages using Ajax or in general pages using Javascript is with a browser itself or a headless browser (a browser without GUI). Currently phantomjs is a well promoted headless browser using WebKit. An alternative that I used with success is HtmlUnit (in Java or .NET via IKVM, which is a simulated browser. Another known alternative is using a web automation tool like Selenium.
I wrote many articles about this subject like web scraping Ajax and Javascript sites and automated browserless OAuth authentication for Twitter. At the end of the first article there are a lot of extra resources that I have been compiling since 2011.
I like PhearJS, but that might be partially because I built it.
That said, it's a service you run in the background that speaks HTTP(S) and renders pages as JSON for you, including any metadata you might need.
Depends on the ajax page. The first part of screen scraping is determining how the page works. Is there some sort of variable you can iterate through to request all the data from the page? Personally I've used Web Scraper Plus for a lot of screen scraping related tasks because it is cheap, not difficult to get started, non-programmers can get it working relatively quickly.
Side Note: Terms of Use is probably somewhere you might want to check before doing this. Depending on the site iterating through everything may raise some flags.
I think Brian R. Bondy's answer is useful when the source code is easy to read. I prefer an easy way using tools like Wireshark or HttpAnalyzer to capture the packet and get the url from the "Host" field and the "GET" field.
For example,I capture a packet like the following:
GET /hqzx/quote.aspx?type=3&market=1&sorttype=3&updown=up&page=1&count=8&time=164330
HTTP/1.1
Accept: */*
Referer: http://quote.hexun.com/stock/default.aspx
Accept-Language: zh-cn
Accept-Encoding: gzip, deflate
User-Agent: Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; SV1)
Host: quote.tool.hexun.com
Connection: Keep-Alive
Then the URL is :
http://quote.tool.hexun.com/hqzx/quote.aspx?type=3&market=1&sorttype=3&updown=up&page=1&count=8&time=164330
As a low cost solution you can also try SWExplorerAutomation (SWEA). The program creates an automation API for any Web application developed with HTML, DHTML or AJAX.
Selenium WebDriver is a good solution: you program a browser and you automate what needs to be done in the browser. Browsers (Chrome, Firefox, etc) provide their own drivers that work with Selenium. Since it works as an automated REAL browser, the pages (including javascript and Ajax) get loaded as they do with a human using that browser.
The downside is that it is slow (since you would most probably like to wait for all images and scripts to load before you do your scraping on that single page).
I have previously linked to MIT's solvent and EnvJS as my answers to scrape off Ajax pages. These projects seem no longer accessible.
Out of sheer necessity, I have invented another way to actually scrape off Ajax pages, and it has worked for tough sites like findthecompany which have methods to find headless javascript engines and show no data.
The technique is to use chrome extensions to do scraping. Chrome extensions are the best place to scrape off Ajax pages because they actually allow us access to javascript modified DOM. The technique is as follows, I will certainly open source the code in sometime. Create a chrome extension ( assuming you know how to create one, and its architecture and capabilities. This is easy to learn and practice as there are lots of samples),
Use content scripts to access the DOM, by using xpath. Pretty much get the entire list or table or dynamically rendered content using xpath into a variable as string HTML Nodes. ( Only content scripts can access DOM but they can't contact a URL using XMLHTTP )
From content script, using message passing, message the entire stripped DOM as string, to a background script. ( Background scripts can talk to URLs but can't touch the DOM ). We use message passing to get these to talk.
You can use various events to loop through web pages and pass each stripped HTML Node content to the background script.
Now use the background script, to talk to an external server (on localhost), a simple one created using Nodejs/python. Just send the entire HTML Nodes as string, to the server, where the server would just persist the content posted to it, into files, with appropriate variables to identify page numbers or URLs.
Now you have scraped AJAX content ( HTML Nodes as string ), but these are partial html nodes. Now you can use your favorite XPATH library to load these into memory and use XPATH to scrape information into Tables or text.
Please comment if you cant understand and I can write it better. ( first attempt ). Also, I am trying to release sample code as soon as possible.
Related
I'm trying to scrape data from a site in python, the payload is right and everything works but when I get the response of the site which would normally be the source code of the html page I instead, get just a script tag with some error written in it. See the response I get enclosed :
b'<script language="JavaScript">\nerr = "";\nlargeur = 1024;\nif (screen.width>largeur) { document.location.href="accueil.php?" +err;\t}\nelse { document.location.href="m.accueil.php?largeur=" +screen.width +\'&\' +err;\t}\n</script>'
Information :
after looking at the site it seems that it uses google analytics, I don't really know about what it is but maybe because of the preview things, it can't load the page since i'm not accessing it by a navigator.
What tool are you using to webscrape? Tools like beautiful soup parse pre-loaded HTML content. If a website uses client-side rendering and JavaScript to load content, often times HTML parsers will not function.
You can instead use an automated browser that interacts with a website just as a regular user would. These automated browsers can operate with or without a GUI. Automated browsers when run without a GUI (also known as a headless browser) take up less time and resources than running them with a GUI. Here's a fairly exhaustive list of headless browsers you can use. Note that not all are compatible with Python.
As Buran mentioned in the comments Selenium is an option. Selenium is very well documented and has a large community following so it's easy to find helpful articles or tutorials. It's a multi-driver so it can run different types of browsers (firefox, chrome, etc.), both headless and with a GUI.
I am trying to scrape a web site using python and beautiful soup. I encountered that in some sites, the image links although seen on the browser is cannot be seen in the source code. However on using Chrome Inspect or Fiddler, we can see the the corresponding codes.
What I see in the source code is:
<div id="cntnt"></div>
But on Chrome Inspect, I can see a whole bunch of HTML\CSS code generated within this div class. Is there a way to load the generated content also within python? I am using the regular urllib in python and I am able to get the source but without the generated part.
I am not a web developer hence I am not able to express the behaviour in better terms. Please feel free to clarify if my question seems vague !
You need JavaScript Engine to parse and run JavaScript code inside the page.
There are a bunch of headless browsers that can help you
http://code.google.com/p/spynner/
http://phantomjs.org/
http://zombie.labnotes.org/
http://github.com/ryanpetrello/python-zombie
http://jeanphix.me/Ghost.py/
http://webscraping.com/blog/Scraping-JavaScript-webpages-with-webkit/
The Content of the website may be generated after load via javascript, In order to obtain the generated script via python refer to this answer
A regular scraper gets just the HTML document. To get any content generated by JavaScript logic, you rather need a Headless browser that would also generate the DOM, load and run the scripts like a regular browser would. The Wikipedia article and some other pages on the Net have lists of those and their capabilities.
Keep in mind when choosing that some previously major products of those are abandoned now.
TRY THIS FIRST!
Perhaps the data technically could be in the javascript itself and all this javascript engine business is needed. (Some GREAT links here!)
But from experience, my first guess is that the JS is pulling the data in via an ajax request. If you can get your program simulate that, you'll probably get everything you need handed right to you without any tedious parsing/executing/scraping involved!
It will take a little detective work though. I suggest turning on your network traffic logger (such as "Web Developer Toolbar" in Firefox) and then visiting the site. Focus your attention attention on any/all XmlHTTPRequests. The data you need should be found somewhere in one of these responses, probably in the middle of some JSON text.
Now, see if you can re-create that request and get the data directly. (NOTE: You may have to set the User-Agent of your request so the server thinks you're a "real" web browser.)
Can we use Scrapy for getting content from a web page which is loaded by Javascript?
I'm trying to scrape usage examples from this page,
but since they are loaded using Javascript as a JSON object I'm not able to get them with Scrapy.
Could you suggest what is the best way to deal with such issues?
Open your browser's developer tools and look at the Network tab. If you hit the "next" button on that page enough, it'll send out a new request:
After removing the JSONP paramter, the URL is pretty straightforward:
https://corpus.vocabulary.com/api/1.0/examples.json?query=unalienable&maxResults=24&startOffset=24&filter=0
By making the minimal number of requests, your spider will be fast.
If you want to just emulate a full browser and execute the JavaScript, you can use something like Selenium or Scrapinghub's Splash (and its corresponding Scrapy plugin).
I am trying to retrieve query results on sites based on ajax like www.snapbird.org using Python. Since it doesn't show in the page source, I am not sure how to proceed.
I am a Python newbie and hence it would be great if I could get a pointer in the right direction.
I am also open to some other approach to the task if that is easier
This is going to be complex but as a start, ppen firebug and find the URL that gets called when the AJAX request is handled. You can call that directly in your Python program and parse the output.
You could use Selenium's Python client driver to parse the page source. I usually use this in conjunction with PyQuery to make web scraping easier.
Here's the basic tutorial for Selenium's Python driver. Be sure to follow the instructions for Selenium version 2 instead of version 1 (unless you're using version 1 for some reason).
You could also configure chrome/firefox to an HTTP proxy and then log/extract the necessary content with the proxy. I've tinkered with python proxies to save/log the requests/content based on content-type or URI globs.
For other projects I've written site-specific javascript bookmarklets which poll for new data and then POST it to my server (by dynamically creating both a form and iframe, and setting myform.target=myiframe;
Other javascript scripts/bookmarklets simulate a user interacting with sites, so instead of polling every few seconds the javascript automates clicking buttons and form submissions, etc. These scripts are always very site-specific of course but they've been hugely useful for me, especially when iterating over all the paginated results for a given search.
Here is a stripped down version of walking over a list of "paginated" results and preparing to send the data off to my server (which then further parses it with BeautifulSoup). In particular this was designed for Youtube's Sent/Inbox messages.
var tables = [];
function process_and_repeat(){
if(!(inbox && inbox.message_pane_ && inbox.message_pane_.innerHTML)){
alert("We've got no data!");
}
if(inbox.message_pane_.innerHTML.indexOf('<table') === 0)
{
tables.push(inbox.message_pane_.innerHTML);
inbox.next_page();
setTimeout("process_and_repeat()",3000);
}
else{
alert("Fininshed, [" + tables.length + " processed]");
document.write('<form action=http://curl.sente.cc method=POST><textarea name=sent.html>'+escape(tables.join('\n'))+'</textarea><input type=submit></form>')
}
}
process_and_repeat(); // now we wait and watch as all the paginated pages are viewed :)
This is a stripped down example without any fancy iframes/non-essentials which just add complexity.
Adding to what Liam said, Selenium is a great tool, too, which has aided in my various scraping needs. I'd be more than happy to help you out with this if you'd like.
One easy solution might be using a browser like Mechanize. So you can browse site, follow links, make searches and nearly everything that you can do with a browser with user interface.
But for a very sepcific job, you may not even need a such library, you can use urllib and urllib2 python libraries to make a connection and read response... You can use Firebug to see data structure of a search and response body. Then use urllib to make a request with relevant parameters...
With an example...
I made a search with joyvalencia and check the request url with firebug to see:
http://api.twitter.com/1/statuses/user_timeline.json?screen_name=joyvalencia&count=100&page=2&include_rts=true&callback=twitterlib1321017083330
So calling this url with urllib2.urlopen() will be the same with making the query on Snapbird. Response body is:
twitterlib1321017083330([{"id_str":"131548107799396357","place":null,"geo":null,"in_reply_to_user_id_str":null,"coordinates":.......
When you use urlopen() and read the response, the upper string is what you get... Then you can use json library of python to read the data and parse it to a pythonic data structure...
I have some html-page. There is a javascript which generates some content. I have to parse this content from python-script. I have saved copy of file on the computer. Are there any ways to work with 'already generated' html? Like I can see in the browser after opening page-file. As I understand, I have to work with DOM (maybe, xml2dom lib).
Have you saved "the file" (web page, I imagine) before or after Javascript has altered it?
If "after", then it doesn't matter any more that some of the HTML was done via Javascript -- you can just use popular parsers like lxml or BeautifulSoup to handle the HTML you have.
If "before", then first you need to let Javascript do its work by automating a real browser; for that task, I would recommend SeleniumRC -- which brings you back to the "after" case;-).
I think you may have a fundamental misunderstanding in regards to what runs where: At the time JavaScript generates the content (on client side), the server side processing of the document has already taken place. There is no direct way for a server side Python script to access HTML created by JavaScript. Basically, that HTML lives only "virtually" in the browser's DOM.
You would have to find a way to transmit that HTML to your Python script. Most likely using Ajax. You would take the HTML, and add it as a parameter to your Ajax call (Remember to use POST as the request method so you don't get size limitation problems.)
An example using jQuery's AJAX functions:
$.ajax({
url: "myscript.py",
type: "POST",
data: { html: your_html_content_here },
success: function(){
alert("sent HTML to python script!");
}});