Description: KiraDB is a simple, lightweight NoSQL-style embedded database Java API

If using Maven, add the repository and dependencies to your pom file:

            <repositories>
                    <repository>
                            <snapshots>
                                    <enabled>false</enabled>
                            </snapshots>
                            <id>bintray</id>
                            <name>bintray</name>
                            <url>https://dl.bintray.com/mrblog/bdt/</url>
                    </repository>
            </repositories>

Dependencies:

<dependency>
    <groupId>com.bdt</groupId>
    <artifactId>kira-db</artifactId>
    <version>1.1</version>
    <type>jar</type>
</dependency>

If you want to compile it yourself, here's how:

$ git clone https://github.com/mrblog/KiraDB.git
$ cd KiraDb
$ mvn install       # Requires maven, download from http://maven.apache.org/download.html

The pre-built jar is available at:

• kira-db-1.1.jar

You'll need to include versions of the dependencies yourself. You will need the following libraries (See the pom.xml for the more details):

• commons-io-1.4.jar
• commons-lang-2.4.jar
• ehcache-core-2.4.3.jar - optional, required for default caching
• jets3t-0.7.4.jar - only needed if using Amazon S3 backing store
• lucene-core-3.0.3.jar
• lucene-queries-3.0.3.jar
• slf4j-api-1.6.1.jar
• slf4j-jdk14-1.6.1.jar
• xpp3_min-1.1.4c.jar
• xstream-1.3.1.jar

You can view the javadocs for this project at: http://mrblog.github.com/KiraDB/apidocs/

Storing data with KiraDB is built around a Record interface. Each Record class describes the data model, as a POJO (Plain old Java Object) implementing the Record interface.

For example, let's say you're tracking high scores for a game. A GameScore class as follows implements the data model.

	public class GameScore {
	
		private String gameNumber;
		private int score;
		private String playerName;
		private String team;
		private Boolean cheatMode;
		
		public GameScore() {
		}

		public GameScore(String gameNumber, int score, String playerName, Boolean cheatMode) {
			setGameNumber(gameNumber);
			setScore(score);
			setPlayerName(playerName);
			setCheatMode(cheatMode);
		}
		public void setGameNumber(String gameNumber) {
			this.gameNumber = gameNumber;
		}
	
		public String getGameNumber() {
			return gameNumber;
		}
	
		public void setScore(int score) {
			this.score = score;
		}
	
		public int getScore() {
			return score;
		}
	
		public void setPlayerName(String playerName) {
			this.playerName = playerName;
		}
	
		public String getPlayerName() {
			return playerName;
		}

		public void setTeam(String team) {
			this.team = team;
		}

		public String getTeam() {
			return team;
		}

		public void setCheatMode(Boolean cheatMode) {
			this.cheatMode = cheatMode;
		}
	
		public Boolean getCheatMode() {
			return cheatMode;
		}
	
	}

In order to use this class with KiraDB, it would be augmented as follows to implement the Record interface:

	public class GameScore implements Record {
	
		private static final String RECORD_NAME = "scores";
		private static final String PRIMARY_KEY = "game";
		public static final String NAME = "name";
		public static final String SCORE = "score";
		public static final String TEAM = "team";

		...
	
		@Override
		public RecordDescriptor descriptor() {
			RecordDescriptor dr = new RecordDescriptor(RECORD_NAME);
			dr.setPrimaryKey(new Field(PRIMARY_KEY, FieldType.STRING, getGameNumber()));
			dr.addField(new Field(NAME, FieldType.STRING, getPlayerName()));
			dr.addField(new Field(SCORE, FieldType.NUMBER, getScore()));
			dr.addField(new Field(TEAM, FieldType.STRING, getTeam()));
			dr.setStoreMode(RecordDescriptor.STORE_MODE_INDEX);
			return dr;
		}
	
		@Override
		public String getRecordName() {
			return RECORD_NAME;
		}
	
		@Override
		public String getPrimaryKeyName() {
			return PRIMARY_KEY;
		}
	
	}

These methods inform KiraDB of the data model associated with the Object Class.

1. Define the name RECORD_NAME analogous to a SQL table name
2. Define the PRIMARY_KEY used for the record.
3. Define any additional index fields that may be needed for query/sort
4. Define the storage mode for the Object Class

Keys and Field names must be alphanumeric strings. Field values can be strings, numbers, dates, or full-text fields (more below). The Object itself can contain any kind of Java Object, Strings, Dates, Maps, Arrays, etc.

Each Record is an instance of a specific subclass with a class name and record name that you can use to distinguish different sorts of data. For example, we call the high score object a GameScore class with the record name "scores" (RECORD_NAME).

To create a new subclass, implement the Record interface. KiraDB will return instances of the new class for any Object with the specified record name

	KiraDb db = new KiraDb(new File("KiraDBIndex"));

Let's say you want to save the GameScore described above to KiraDB.

	GameScore p1 = new GameScore("Game157", 1337, "Kevin Blake", false);
	db.storeObject(p1);

If you have the primary key value, you can retrieve the Object using the retrieveObjectByPrimaryKey method:

	GameScore theScore = (GameScore) db.retrieveObjectByPrimaryKey(new GameScore(), "Game157");

We've already seen how you can retrieve a single Object from KiraDB. There are many other ways to retrieve data with the executeQuery method - you can retrieve many objects at once, put conditions on the objects you wish to retrieve, and more.

The general approach is to create a query, put conditions on it, and then retrieve a List of matching Objects using executeQuery. For example, retrieve the scores associated with a particular team by constraining the value for a key:

        List<GameScore> qResults = db.executeQuery(new GameScore(), GameScore.TEAM, "shaggy", 10, 0, GameScore.SCORE, true);

This returns scores belonging to team "shaggy" returning 10 results at a time, skipping 0 results (i.e. starting at 0), sorting by SCORE in reverse sort order (highest score first).

There are several ways to put constraints on the objects retrieved from executeQuery . You can give multiple constraints, and objects will only be in the results if they match all of the constraints. In other words, it's like an AND of constraints.

    Query *query = new Query(new GameScorre());
	query.whereMatches(GameScore.NAME, "Sam Bidwell");
	query.whereMatches(GameScore.TEAM, "pilots");
	List<GameScore> qResults = db.executeQuery(query);
	
   

Find records all players named "Sam Bidwell" that are also members of team "pilots".

You can limit the number of results with setLimit. By default, results are limited to 100, but anything from 1 to Integer.MAX_VALUE is a valid limit:

    query.setLimit(10); // limit to at most 10 results
	

You can skip the first results with setStart. This can be useful for pagination:

    query.setStart(10); // skip the first 10 results
	

You can control the order in which results are returned:

    
    setSortField(GameScore.SCORE);
	

And reverse the order (ascending is the default sort order):

    
    setSortField(GameScore.SCORE, true);
	

If you just need to count how many objects match a query, but you do not need to retrieve all the objects that match, you can use getTotalHits(). For example, to count how many games have been played by a particular player:

        List<GameScore> qResults = db.executeQuery(new GameScore(), GameScore.NAME, "Kevin Blake", 1, 0, GameScore.SCORE, true);
	int count = db.getTotalHits();

To page through query results, use the hitsPerPage and skipDocs parameters:

       List<GameScore> qResults = db.executeQuery(new GameScore(), GameScore.TEAM, "shaggy", PER_PAGE, (page-1)*PER_PAGE, GameScore.SCORE, true);

where PER_PAGE is the number of objects per page and page in this case is the page number (starting from 1).

Primary Keys in KiraDB must be String fields (FieldType.STRING) which are simple case-sensitive string values. The Primary Key is unique across all records of this class.

The GameScore example adds a field for the user's score. This field is to be treated as a NUMBER by KiraDB. This effects searching and sorting based on the field.

The supported field types are:

• STRING - simple case-sensitive string
• NUMBER - integer
• DATE - java.util.Date object
• FULLTEXT - a string field with full-text search capabilities

Record, Primary Key, and Field Names are case-sensitive and must consist of letters and digits only.

KiraDB can operate in multiple modes:

1. STORE_MODE_NONE Indexing (supporting searching and sorting operations) only (no storage of the full object)
2. STORE_MODE_INDEX Storing Objects in the KiraDB index
3. STORE_MODE_BACKING Storing Objects in a separate Backing Store (BackingStore)

More details on these modes is provided in the sections that follow.

In this mode, KiraDB is only performing indexing (searching and sorting) for the user. Object storage and retrieval is entirely the responsibility of the user. Record objects returned by KiraDB contain only the primary key and the fields associated with the Record class.

In this mode, KiraDB will store user objects with the index. KiraDB will return the full object class implementing the Record interface as shown in the examples above for the GameScore object class. However, in this mode, there is no authoritative data store other than the index, which reduces redundancy. If the index should become corrupt, user data could be lost.

In this mode, KiraDB works with a connected BackingStore for the authoritative object repository. If the index should become corrupt, it could be rebuilt from the backing store.

KiraDB includes two included backing store implementations, a simple File-based object store (FileBackingStore) and an Amazon S3 based backing store (S3BackingStore)

Users can provide their own custom backing store by implementing their own BackingStore class.

Convenience classes S3KiraDB and FileSystemKiraDb are provided for constructing a Core KiraDB instance configured with the corresponding backing store. Refer to their Javadocs for information on their use.

KiraDb supports full-text searching including an English-language stemmer. Let's say you have a field in your Record class for a title. Add this field as a FULLTEXT field and you can then perform full-text searches on that field:

dr.addField(new Field(TITLE, FieldType.FULLTEXT, getTitle()));

Perform a full-text search on this field with the executeQuery() method:

   List<MyClass> qResults = db.executeQuery(new MyClass(), MyClass.TITLE, "logic", null, false);

KiraDb provides a feature to locate Objects that are similar to a specified test string. This can be used to identify other Objects that are similar to a given Object. For example:

    String[] fieldNames = { TextDocument.BODY };
    List<String> matches = db.relatedObjects(new TextDocument(), doc.getBody(), fieldNames, 5, doc.getId());

This returns a list of up to 5 matching document Id's (the primary field for the TextDocument Object class), excluding the input "reference" object (doc).

You can also find all documents that match any given reference string, as in:

    String[] fieldNames = { TextDocument.BODY };
    List<String> matches = db.relatedObjects(new TextDocument(), testStr, fieldNames, 5, null);

This can be useful, for instance, to assist users in avoiding providing duplicate content.