I am using Selenium WebDriver for crawling a web site(only for example, I will be crawling other web sites too!) which has infinite scroll.
Problem statement:
Scroll down the infinite scroll page till the content stops loading using Selenium web driver.
My Approach: Currently I am doing this-
Step 1: Scroll to the page bottom
JavascriptExecutor js = (JavascriptExecutor) driver;
js.executeScript("javascript:window.onload=toBottom();"+
"function toBottom(){" +
"window.scrollTo(0,Math.max(document.documentElement.scrollHeight," +
"document.body.scrollHeight,document.documentElement.clientHeight));" +
"}");
Then I wait for some time to let the Ajax Request complete like this-
Step 2: Explicitly wait for Ajax request to be over
Thread.sleep(1000);
Then I give another java script to check if the page is scrollable
Step 3:Check if the page is scrollable
//Alternative to document.height is to be used which is document.body.clientHeight
//refer to https://developer.mozilla.org/en-US/docs/DOM/document.height
if((Long)js.executeScript("return " +
"(document.body.clientHeight-(window.pageYOffset + window.innerHeight))")>0)
If the above condition is true then I repeat the from Step 1 - 3, till condition in Step 3 is false.
The Problem:
I do not want to give the Thread.sleep(1000);
in step 2, rather I would like to check using Java Script if the background Ajax request is over and then scroll down further if the condition in Step 3 is true .
PS: I am not the developer of the page so I do not have access to the code running the page, I can just inject java scripts(as in Step 1 and 3) in the web page. And, I have to write a generic logic for any web site with Ajax requests during infinite scroll.
I will be grateful to some one could spare some time here!
EDIT : Ok, after struggling for 2 days, I have figured out that the pages which I am crawling through the Selenium WebDriver can have any of these JavaScript libraries and I will have to pool according to the different Library, for example, In case of the web application using jQuery api, I may be waiting for
(Long)((JavascriptExecutor)driver).executeScript("return jQuery.active")
to return a zero.
Likewise if the web application is using the Prototype JavaScript library I will have to wait for
(Long)((JavascriptExecutor)driver).executeScript("return Ajax.activeRequestCount")
to return a zero.
Now, the problem is how do I write a generic code which could handle most the JavaScript libraries available?
Problem I am facing in implementing this-
1. How do I find which JavaScript Library is being used in the Web Application(using Selenium WebDriver in Java), such that I can then write the corresponding wait methods? Currently, I am using this
2. This way I will have to write as many as 77 methods for separate JavaScript library so, I need a better way to handle this scenario as well.
In short, I need to figure out if the browser is making any call(Ajax or simple) with or without any JavaScript library through Selenium Web Driver's java implementation
PS: there are Add ons for Chorme's JavaScript Lib detector and Firefox's JavaScript Library detector which detect the JavaScript library being used.
For web pages with Ajax Response during the infinite scroll and using jQuery API(or other actions), before starting to opening the web page.
//Inject the pooling status variable
js.executeScript("window.status = 'fail';");
//Attach the Ajax call back method
js.executeScript( "$(document).ajaxComplete(function() {" +
"status = 'success';});");
Step 1: will remain the same as in the original question
Step 2 Pooling the following script(This is the one which removes the need of Thread.Sleep()
and makes the logic more dynamic)
String aStatus = (String)js.executeScript("return status;");
if(aStatus!=null && aStatus.equalsIgnoreCase("success")){
js.executeScript("status = 'fail';");
break poolingLoop;
}
Step 3: No need now!
Conclusion: No need to give blunt Thread.sleep(); again and again while using Selenium WebDriver!!
This approach works good only if there's jQuery api being used in the web application.
EDIT: As per the the link given by @jayati i injected the javascript-
Javascript one:
//XMLHttpRequest instrumentation/wrapping
var startTracing = function (onnew) {
var OldXHR = window.XMLHttpRequest;
// create a wrapper object that has the same interfaces as a regular XMLHttpRequest object
// see http://www.xulplanet.com/references/objref/XMLHttpRequest.html for reference on XHR object
var NewXHR = function() {
var self = this;
var actualXHR = new OldXHR();
// private callbacks (for UI):
// onopen, onsend, onsetrequestheader, onupdate, ...
this.requestHeaders = "";
this.requestBody = "";
// emulate methods from regular XMLHttpRequest object
this.open = function(a, b, c, d, e) {
self.openMethod = a.toUpperCase();
self.openURL = b;
ajaxRequestStarted = 'open';
if (self.onopen != null && typeof(self.onopen) == "function") {
self.onopen(a,b,c,d,e); }
return actualXHR.open(a,b,c,d,e);
}
this.send = function(a) {
ajaxRequestStarted = 'send';
if (self.onsend != null && typeof(this.onsend) == "function") {
self.onsend(a); }
self.requestBody += a;
return actualXHR.send(a);
}
this.setRequestHeader = function(a, b) {
if (self.onsetrequestheader != null && typeof(self.onsetrequestheader) == "function") { self.onsetrequestheader(a, b); }
self.requestHeaders += a + ":" + b + "\r\n";
return actualXHR.setRequestHeader(a, b);
}
this.getRequestHeader = function() {
return actualXHR.getRequestHeader();
}
this.getResponseHeader = function(a) { return actualXHR.getResponseHeader(a); }
this.getAllResponseHeaders = function() { return actualXHR.getAllResponseHeaders(); }
this.abort = function() { return actualXHR.abort(); }
this.addEventListener = function(a, b, c) { return actualXHR.addEventListener(a, b, c); }
this.dispatchEvent = function(e) { return actualXHR.dispatchEvent(e); }
this.openRequest = function(a, b, c, d, e) { return actualXHR.openRequest(a, b, c, d, e); }
this.overrideMimeType = function(e) { return actualXHR.overrideMimeType(e); }
this.removeEventListener = function(a, b, c) { return actualXHR.removeEventListener(a, b, c); }
// copy the values from actualXHR back onto self
function copyState() {
// copy properties back from the actual XHR to the wrapper
try {
self.readyState = actualXHR.readyState;
} catch (e) {}
try {
self.status = actualXHR.status;
} catch (e) {}
try {
self.responseText = actualXHR.responseText;
} catch (e) {}
try {
self.statusText = actualXHR.statusText;
} catch (e) {}
try {
self.responseXML = actualXHR.responseXML;
} catch (e) {}
}
// emulate callbacks from regular XMLHttpRequest object
actualXHR.onreadystatechange = function() {
copyState();
try {
if (self.onupdate != null && typeof(self.onupdate) == "function") { self.onupdate(); }
} catch (e) {}
// onreadystatechange callback
if (self.onreadystatechange != null && typeof(self.onreadystatechange) == "function") { return self.onreadystatechange(); }
}
actualXHR.onerror = function(e) {
ajaxRequestComplete = 'err';
copyState();
try {
if (self.onupdate != null && typeof(self.onupdate) == "function") { self.onupdate(); }
} catch (e) {}
if (self.onerror != null && typeof(self.onerror) == "function") {
return self.onerror(e);
} else if (self.onreadystatechange != null && typeof(self.onreadystatechange) == "function") {
return self.onreadystatechange();
}
}
actualXHR.onload = function(e) {
ajaxRequestComplete = 'loaded';
copyState();
try {
if (self.onupdate != null && typeof(self.onupdate) == "function") { self.onupdate(); }
} catch (e) {}
if (self.onload != null && typeof(self.onload) == "function") {
return self.onload(e);
} else if (self.onreadystatechange != null && typeof(self.onreadystatechange) == "function") {
return self.onreadystatechange();
}
}
actualXHR.onprogress = function(e) {
copyState();
try {
if (self.onupdate != null && typeof(self.onupdate) == "function") { self.onupdate(); }
} catch (e) {}
if (self.onprogress != null && typeof(self.onprogress) == "function") {
return self.onprogress(e);
} else if (self.onreadystatechange != null && typeof(self.onreadystatechange) == "function") {
return self.onreadystatechange();
}
}
if (onnew && typeof(onnew) == "function") { onnew(this); }
}
window.XMLHttpRequest = NewXHR;
}
window.ajaxRequestComplete = 'no';//Make as a global javascript variable
window.ajaxRequestStarted = 'no';
startTracing();
Or Javascript Two:
var startTracing = function (onnew) {
window.ajaxRequestComplete = 'no';//Make as a global javascript variable
window.ajaxRequestStarted = 'no';
XMLHttpRequest.prototype.uniqueID = function() {
if (!this.uniqueIDMemo) {
this.uniqueIDMemo = Math.floor(Math.random() * 1000);
}
return this.uniqueIDMemo;
}
XMLHttpRequest.prototype.oldOpen = XMLHttpRequest.prototype.open;
var newOpen = function(method, url, async, user, password) {
ajaxRequestStarted = 'open';
/*alert(ajaxRequestStarted);*/
this.oldOpen(method, url, async, user, password);
}
XMLHttpRequest.prototype.open = newOpen;
XMLHttpRequest.prototype.oldSend = XMLHttpRequest.prototype.send;
var newSend = function(a) {
var xhr = this;
var onload = function() {
ajaxRequestComplete = 'loaded';
/*alert(ajaxRequestComplete);*/
};
var onerror = function( ) {
ajaxRequestComplete = 'Err';
/*alert(ajaxRequestComplete);*/
};
xhr.addEventListener("load", onload, false);
xhr.addEventListener("error", onerror, false);
xhr.oldSend(a);
}
XMLHttpRequest.prototype.send = newSend;
}
startTracing();
And checking the status of the status vars ajaxRequestStarted, ajaxRequestComplete
in the java code, one can determine if the ajax was started or completed.
Now I have a way to wait till an Ajax is complete, I can also find if the Ajax was triggered on some action