Development | .NET Compact Framework

Invoking the Google Web Service on your SmartPhone
Written by W.G. Ryan  [author's bio]  [read 20265 times]
Edited by Derek

   

Page 1 

It's hard to imagine that any software developer hasn't heard of Google. What you may or may not be familiar with is their APIs. The first thing you need to do is register for the service. They'll give you 1,000 requests a day which is more than enough more for non-commercial application. If you need more I can't help you though b/c I wasn't able to find the pricing for it (by the time you're done reading though, maybe you can invoke a Google Search for their prices with your Smartphone). Registration is really easy and it's basically no different from every other sort of service you register for.

Ok, now I have to warn you about a few things. The Smartphone isn't the same thing as your desktop. The framework is called "Compact" for a good reason. This is obscured by the fact that in many instances, the code is identical to it's desktop counterpart. Don't lost sight of that fact though. You are in a resource constrained environment and there's no room for inefficient code. There aren't 50 events for every control nor are there a lot of controls in comparison to the desktop. Nor should there be, it just wouldn't make much sense.

I don't want to get into the guts of web services and how they work, but in a nutshell, it's pretty much the same as the desktop. Now, XML isn't the fastest way to access data and in most instances (particularly on a SmartPhone) it is pretty slow. So why am I using it?  Well, for one thing you don't have many alternatives on the Smartphone platform. For another it's Google's primary method of exposing its web service/API..

To set things up, I'm going to create a basic Smart Device Application and target it to the Smartphone platform. Once you do this, you should see a small form appear in your IDE. Its size and unique look clearly indicate you are targeting the Smartphone platform. Like I mentioned, you don't have all of the controls you are used to and the ones that do exist expose notably different functionality than you're probably used to dealing with. As such, I'm going to use a CheckBox control in place of where I'd typically use a CommandButton. When the user clicks on the checkbox, I'll examine its CheckState property and if it's checked, I'll fire my search.

Since querying Google wouldn't be of much value unless you can dynamically specify your search criteria, I'm going to include a TextBox control so the user can provide their search terms. Finally, once the service is invoked, I'll use the ComboBox control to load the results provided there are any. Once you have your controls loaded, you could easily expand this to do other things with your results. Air Time isn't cheap in most cases so this isn't economically the best way to hit data, but like I mentioned before, your data access options are limited on the Smartphone and invoking web services is one of your few options. Also note that you can easily load/read xml using System.Data which has a really strong implementation even on this platform. Anyway, let's go grab some data.

The first thing we need to do after we've registered for the service is to add a Web Reference to your project. In your Solution Explorer, right click References and choose the 'Add Web Reference' option. You can plug in http://api.google.com/GoogleSearch.wsdl to the URL box provided. At this point, Visual Studio .NET should create a proxy class for you provided it was able to see the .wsdl. It will probably call the proxy something like com.google.api. After this is created, the behavior you should see is virtually identical to any other class you would use or reference. If you look in the Object Browser ->  YourProject.com.google.api you - you'll see 4 classes (DirectoryCategory, DirectorySearchResult, DirectorySearchService & ResultElement). If you see those everything is good to go. Basically, to invoke and use Google's api, this is pretty much all there is to it

private void checkBox1_CheckStateChanged(object sender, System.EventArgs e)

{

          if(!checkBox1.Checked)return;

          com.google.api.GoogleSearchResult searchResult = new com.google.api.GoogleSearchResult();

          com.google.api.GoogleSearchService g = new com.google.api.GoogleSearchService();

          searchResult = g.doGoogleSearch("KeyHere", tbCriteria.Text , 0, 9, false,string.Empty,false, string.Empty, string.Empty, string.Empty);

     foreach(com.google.api.ResultElement re in searchResult.resultElements)

    {

      cboResults.Items.Add(re.URL);

     }

}

 

 

All that's really happening here is checking if the user actually wants to call the web service and if they do, I instantiated a new instance of a SearchResult, fire the service using my private key using doGoogleSearch and then iterate through the results. As I mentioned, there are 3 other classes you can manipulate w/ the API directly, but there's a TON of stuff within each of those. In addition, just like there's doGoogleSearch, there's an Async invocation called BegindoGoogleSearch following .NET's naming convention for async calls.

This is just a really basic intro into using the Google API but it lays the ground work for my next article, integrating the XML features of Infopath, XML Webservices and the Compact Framework

Back to .NET Compact Framework | [Article Index]