Java MySQLJDBCDataModel类代码示例

import org.apache.mahout.cf.taste.impl.model.jdbc.MySQLJDBCDataModel; //导入依赖的package包/类
/**
 * Builds a new Mahout Database model that connects to the MySQL
 * database using the credentials set-up in the WebDSL configuration.
 * @return The Mahoud JDBC Data Model instance
 */
private MySQLJDBCDataModel getMahoutDatabaseModel(){
    /*
     * Give Mahout this instance as the data source, Mahout will be able
     * to call the getConnection method to obtain the database jdbc connection
     * with the database that Hibernate is connected to. This has been implemented
     * like this so the H2 database could also be integrated. However, the function
     * used to obtain the datasource (Session.connection() of Hibernate is set to
     * deprecate in version 4. However, from what I read in their internal docs
     * they are still not convinced to remove it completely, because a lot of
     * customers need this function for their products.
     * Look at rev: 4809 for the version that used the properties of Hibernat
     * to set-up the database connection if you need to replace this later on.
     * ~ Siem
     */
    return new MySQLJDBCDataModel(this, "__"+this.name+"_Taste", "user_id", "item_id", "preference", null);
}
 

import org.apache.mahout.cf.taste.impl.model.jdbc.MySQLJDBCDataModel; //导入依赖的package包/类
public static void main(String[] args) {
    try {
        MysqlDataSource dataSource = new MysqlDataSource();
        dataSource.setServerName("localhost");
        dataSource.setUser("root");
        dataSource.setPassword("root");
        dataSource.setDatabaseName("rec");
        JDBCDataModel dm = new MySQLJDBCDataModel(dataSource,"ratings","userid","itemid","rating","");
        UserSimilarity similarity = new PearsonCorrelationSimilarity(dm);
        UserNeighborhood neighbor = new NearestNUserNeighborhood(2,similarity, dm);
        Recommender recommender = new GenericUserBasedRecommender(dm, neighbor, similarity);
        List<RecommendedItem> list = recommender.recommend(1, 3);// recommend
                                                                 // one item
                                                                 // to user
                                                                 // 1
        for (RecommendedItem ri : list) {
            System.out.println(ri);
        }
    } catch (Exception e) {
        e.printStackTrace();
    }

}
 

import org.apache.mahout.cf.taste.impl.model.jdbc.MySQLJDBCDataModel; //导入依赖的package包/类
public static DataModel loadFromDB() throws Exception {
	
	// Database-based DataModel - MySQLJDBCDataModel
	/*
	 * A JDBCDataModel backed by a PostgreSQL database and accessed via
	 * JDBC. It may work with other JDBC databases. By default, this class
	 * assumes that there is a DataSource available under the JNDI name
	 * "jdbc/taste", which gives access to a database with a
	 * "taste_preferences" table with the following schema: CREATE TABLE
	 * taste_preferences ( user_id BIGINT NOT NULL, item_id BIGINT NOT NULL,
	 * preference REAL NOT NULL, PRIMARY KEY (user_id, item_id) ) CREATE
	 * INDEX taste_preferences_user_id_index ON taste_preferences (user_id);
	 * CREATE INDEX taste_preferences_item_id_index ON taste_preferences
	 * (item_id);
	 */
	MysqlDataSource dbsource = new MysqlDataSource();
	dbsource.setUser("user");
	dbsource.setPassword("pass");
	dbsource.setServerName("localhost");
	dbsource.setDatabaseName("my_db");

	DataModel dataModelDB = new MySQLJDBCDataModel(dbsource,
			"taste_preferences", "user_id", "item_id", "preference",
			"timestamp");
	
	return dataModelDB;
}
 

import org.apache.mahout.cf.taste.impl.model.jdbc.MySQLJDBCDataModel; //导入依赖的package包/类
public static String getRecommenderItem(int id) throws IOException, TasteException {
		//DB연동 
		MysqlDataSource datasource = new MysqlDataSource();
		datasource.setServerName("localhost");
		datasource.setUser("root");
		datasource.setPassword("465651");
		datasource.setDatabaseName("tourOfAll2");
		
		
		DataModel model = new ReloadFromJDBCDataModel(new MySQLJDBCDataModel(datasource, "evaluations", "user_id", "item_id", "score", null));
		// DataModel model = new FileDataModel(
		// new
		// File("C:/Users/Administrator/git/RestfulMahoutRecommender/RestfulRecommenderApi/src/main/resources/ddd.csv"));
		
		//유사도 측정을 캐쉬로 저장
		UserSimilarity similarity = new CachingUserSimilarity(new EuclideanDistanceSimilarity(model),model);

		// new SpearmanCorrelationSimilarity(model);
		
		//유저 이웃 계산 결과를 캐쉬로 저장
		UserNeighborhood neighborhood = new CachingUserNeighborhood(new ThresholdUserNeighborhood(0.75, similarity, model),model);
		
		
		// new NearestNUserNeighborhood(5,similarity,model);
		Recommender recommender = new GenericUserBasedRecommender(model, neighborhood, similarity);

		LoadEvaluator.runLoad(recommender);

		
		String json = "{" + "\"Items\"" + ":" + "[";

		IDRescorer testRescorer = new GenreRescorer(id);
		List<RecommendedItem> recommendations = recommender.recommend(id, 20, testRescorer);
		
//		List<RecommendedItem> recommendations = recommender.recommend(id, 10);
		// String Parsing 아이디값만 찾음
		Iterator<RecommendedItem> itr = recommendations.iterator();
		while (itr.hasNext()) {
			RecommendedItem item = itr.next();
			String str = item.toString();
			String ItemId = str.substring(str.indexOf(":") + 1, str.indexOf(","));
			String value = str.substring(str.indexOf("value:") + 6, str.indexOf("value:") + 9);
			getPlaceURL url = new getPlaceURL(Integer.parseInt(ItemId));
			GetPlaceTitle title = new GetPlaceTitle(Integer.parseInt(ItemId));
			if (itr.hasNext())
				json = json + "{" + "\"ID\"" + ":" + "\"" + ItemId + "\"" 
			            + ", " + "\"Value\"" + ":" + "\"" + value + "\""
			            + ", " + "\"URL\"" + ":" + "\"" + url.getURL() + "\""
			            + ", " + "\"Title\"" + ":" + "\"" + title.getTitle() + "\""
			            + "}" + ", ";
			else
				json = json + "{" + "\"ID\"" + ":" + "\"" + ItemId + "\""
						+ ", " + "\"Value\"" + ":" + "\"" + value + "\""
						+ ", " + "\"URL\"" + ":" + "\"" + url.getURL() + "\""
						+ ", " + "\"Title\"" + ":" + "\"" + title.getTitle() + "\""
						+ "}";
		}

		json = json + "]" + "}";
		return (json);
	}
 

import org.apache.mahout.cf.taste.impl.model.jdbc.MySQLJDBCDataModel; //导入依赖的package包/类
/**
 * This function will (re)construct all the recommendations using
 * Mahout.
 * NOTE: Be aware that this is a very expensive function in terms of
 * processing time, therefore this function is called on a regular
 * interval (say every x hours) to update the recommendation cache
 * according to the latest data available in the database.
 */
public void reconstructRecommendationCache(){
    long startTime = System.currentTimeMillis();

    try {
        /* Open the database and update the dictionary to Mahout style */
        convertUIDToBigIntInDB();

        /*
         * Database is converted, generate the recommendations seperately
         * from the currently cached recommendations if there are any.
         */
        MySQLJDBCDataModel model = getMahoutDatabaseModel();

        /*
         * Build the recommender itself.
         */
        Recommender userRecommender = null;
        Recommender itemRecommender = null;
        if(this.type == null || this.type.equalsIgnoreCase(T_USER) || this.type.equalsIgnoreCase(T_BOTH)){
            userRecommender = buildRecommenderInstance(model, RecFieldType.USER);
        }
        if(this.type == null || this.type.equalsIgnoreCase(T_ITEM) || this.type.equalsIgnoreCase(T_BOTH)){
            itemRecommender = buildRecommenderInstance(model, RecFieldType.ITEM);
        }

        /*
         * The recommendation process has finished, replace the cached
         * recommendations with the new updated version.
         */
        if(this.type == null || this.type.equalsIgnoreCase(T_USER) || this.type.equalsIgnoreCase(T_BOTH)){
            this.userRecommenderCache = new CachingRecommender(userRecommender);
        }
        if(this.type == null || this.type.equalsIgnoreCase(T_ITEM) || this.type.equalsIgnoreCase(T_BOTH)){
            this.itemRecommenderCache = new CachingRecommender(itemRecommender);
        }
        this.lastExecutionTime = System.currentTimeMillis() - startTime;
        org.webdsl.logging.Logger.info("Building the Mahout Index took: " + this.lastExecutionTime);
    } catch (Exception e){
        org.webdsl.logging.Logger.error("\n\n************ SEVERE ERROR ************\nThere was an exception while reconstructing the recommendations!\n" + e);
        org.webdsl.logging.Logger.error("EXCEPTION",e);
        org.webdsl.logging.Logger.error("************ END OF ERROR ************\n\n");
    }
}