.:] i.suggest [:. « beef.b10g-2.0.0

.:] i.suggest [:.

August 8th, 2007 – 13:01:57

Tagged as: java, live search, programming

Somewhat recently an enhancement came the wire at work for text input fields that auto-completed based off of a static set of data, specifically, e-mail domain names. The enhancement came to me and it was a fun thing to create, and something that could be potentially very useful for alot of different pieces of the application, so I thought that I would share the pattern that we came up with to solve this problem.

First we decided that we would like to make this an open-ended API, so in the future when they wanted to have other kinds of auto-completing fields, we could implement those using the same API and just write a new implementation to solve the problem at hand.

So the first thing we needed was a common interface to define what our auto-completer would need. We went with the term Suggester and the interface of ISuggest (clever eh?)

1   package isuggest;

2   

3   public interface ISuggest {

4      String [] getSuggestions( String searchText );

5   }

Then we will obviously need some type of implementation so I put together this quick dirty static text suggester implementation just for testing purposes. Obviously the implementation is where you would query your database, lucene index, or whatever datasource you choose to use. For this example, the datasource is a static string array containing guys and girls names.

1   package isuggest.impl;

2   

3   import java.util.List;

4   import java.util.ArrayList;

5   

6   import isuggest.ISuggest;

7   

8   public class StaticSuggester implements ISuggest {

9      private static String [] data = new String[] {

10        "Jacob",

11        "Michael",

12        "Joshua",

13        "Matthew",

14        "Ethan",

15        "Andrew",

16        "Daniel",

17        "Anthony",

18        "Christopher",

19        "Joseph",

20        "Emily",

21        "Emma",

22        "Madison",

23        "Abigail",

24        "Olivia",

25        "Isabella",

26        "Hannah",

27        "Samantha",

28        "Ava",

29        "Ashley",

30     };

31  

32     public String[] getSuggestions(String searchText) {

33        List results = new ArrayList();

34  

35        for ( int i = 0; i < data.length; i ++ ) {

36           if ( data[i].toLowerCase().startsWith( searchText.toLowerCase() ) ) {

37              results.add( data[i] );

38           }

39        }

40  

41        return (String []) results.toArray( new String[ results.size() ] );

42     }

43  

44  }

45

And there you have the basic setup for your suggester API. Now I like to take it one step further and give myself a factory implementation to get my suggester based of a specific type implementation, so I will create the SuggesterFactory next.

1   package isuggest;

2   

3   import java.util.HashMap;

4   import java.util.Map;

5   

6   public class SuggesterFactory {

7      public static ISuggest getSuggestor( SuggestType type ) throws SuggesterCreationException {

8         try {

9            return (ISuggest) type.getImplClass().newInstance();

10        } catch (Exception e) {

11           throw new SuggesterCreationException( "Unable to create suggester of type [" + type.getImplClassName() + "]", e );

12        }

13     }

14  }

15

You will notice that this factory takes a TypeSafe parameter called SuggestType which is outlined below

1   package isuggest;

2   

3   public class SuggestType {

4      public static final SuggestType STATIC_SUGGEST = new SuggestType( "isuggest.impl.StaticSuggester" );

5   

6      private String implClassName;

7   

8      public SuggestType( String implClassName ) {

9         this.implClassName = implClassName;

10     }

11  

12     public Class getImplClass() throws ClassNotFoundException {

13        return Class.forName( this.implClassName );

14     }

15  

16     public String getImplClassName() {

17        return this.implClassName;

18     }

19  }

20

And there you have it… A simple API to build the back-end to your auto-completer across your entire application.

The front end can be developed using Yahoo User Interface widgets ( Yahoo UI Library – AutoComplete ) or the fabulous Scriptaculous auto-completer ( script.aculo.us auto-completer demo ) or design your own as you need.

Posted by: beef | Comments: (121)