Merge pull request #13 from kren1/sql_setup

Sql to weka

build.gradle

@@ -12,8 +12,10 @@ repositories {
 
 dependencies {
     compile group: 'nz.ac.waikato.cms.weka', name:'weka-stable', version: '3.6.13'
+    compile group: 'org.postgresql', name:'postgresql', version:  '9.4-1204-jdbc42'
     testCompile group: 'junit', name: 'junit', version: '4.11'
-
+    testCompile group: 'com.google.guava', name: 'guava', version: '15.0'
+    testCompile group: 'org.jmock', name: 'jmock-junit4', version: '2.8.1'
 }
 
 checkstyle {

sql/db_drop.sql

@@ -0,0 +1,3 @@
+DROP TABLE data;
+DROP TABLE batch;
+DROP TABLE feature;

sql/db_sample_data.sql

@@ -0,0 +1,15 @@
+INSERT INTO batch (added,description) VALUES (now(),'test');
+INSERT INTO feature(name) VALUES
+('f1'),
+('f2'),
+('f3')
+;
+INSERT INTO data (subjectID, featureID, batchID, featureValue,timeslice) VALUES
+(1,1,1,60,0),
+(1,2,1,2.22453,0),
+(1,3,1,0.36453,0),
+(2,3,1,5.32471,0),
+(2,2,1,4.72453,0),
+(2,1,1,70,0),
+(3,3,1,1.22451,0)
+;

sql/db_setup.sql

@@ -0,0 +1,26 @@
+CREATE TABLE feature
+(
+  featureID serial PRIMARY KEY,
+  name varchar(10) not null
+);
+
+CREATE TABLE batch
+(
+  batchID serial PRIMARY KEY,
+  added date not null,
+  description text not null
+);
+
+CREATE TABLE data
+(
+  subjectID bigserial,
+  featureID int REFERENCES feature,
+  batchID int REFERENCES batch,
+  featureValue numeric not null,
+  timeslice int not null,
+  PRIMARY KEY (subjectID, featureID, batchID, timeslice)
+
+);
+
+-- makes sure BPM always be feature with id 1
+INSERT INTO feature(name) VALUES ('BPM') ;

src/main/java/database/DataInstancesBuilder.java

@@ -0,0 +1,17 @@
+package database;
+
+import weka.core.Instances;
+
+import java.sql.Connection;
+import java.sql.Statement;
+
+/**
+ * Created by tk1713 on 02/11/15.
+ *  @param <T_batch> type of the batch selection input
+ *  @param <T_subject> type of the subject selection input
+ *  @param <T_feature> type of the feature selection input
+ */
+public interface DataInstancesBuilder<T_feature, T_subject, T_batch> {
+    DataInstancesBuilder<T_feature, T_subject, T_batch> withFeature(T_feature feature);
+    <T> DatabaseDataset<T> build(Connection connection);
+}

src/main/java/database/DatabaseConnection.java

@@ -0,0 +1,48 @@
+package database;
+
+import java.sql.Connection;
+import java.sql.DriverManager;
+import java.sql.SQLException;
+import java.util.Map;
+
+/**
+ * Created by Timotej on 31-Oct-15.
+ */
+public class DatabaseConnection {
+    /*
+     Gets the connection to our database. The information about the database is
+     stored in enviroment variables for 2 reasons. Firstly to not commit that data
+     to git and secondly so we can potentialy easily set up a database for travis.
+     It is easy to push encrypted enviroment variables to Travis.
+     */
+    public static Connection getDatabaseConnection() {
+        Connection connection;
+        Map<String, String> env = System.getenv();
+        if (!env.containsKey("PGHOST") || !env.containsKey("PGPORT") ||
+                !env.containsKey("PGDATABASE") || !env.containsKey("PGUSER") ||
+                !env.containsKey("PGPASS")) {
+            throw new RuntimeException("Cannot get db data from enviroment variables PGHOST, " +
+                    "PGPORT, PGDATABASE, PGUSER, PGPASS");
+        }
+
+        final String host = env.get("PGHOST");
+        final String port = env.get("PGPORT");
+        final String database = env.get("PGDATABASE");
+        final String username = env.get("PGUSER");
+        final String password = env.get("PGPASS");
+        try {
+            Class.forName("org.postgresql.Driver");
+            connection = DriverManager
+                    .getConnection("jdbc:postgresql://" + host + ":" + port + "/" + database +
+                                    //need this to fix SSL error
+                                    "?ssl=true&sslfactory=org.postgresql.ssl.NonValidatingFactory",
+                            username, password);
+        } catch (SQLException e) {
+            throw new RuntimeException("Cannot connect to your database " + database +
+                        " user: "  + username);
+        } catch (ClassNotFoundException e) {
+            throw new RuntimeException("JDBC not setup, build with gradle");
+        }
+        return connection;
+    }
+}

src/main/java/database/DatabaseDataset.java

@@ -0,0 +1,10 @@
+package database;
+
+/**
+ * Created by Timotej on 02-Nov-15.
+ * @param <T> return type in which this dataset should be returned
+ */
+public interface DatabaseDataset<T> {
+    public T getBatch(int batchID, int timeslice);
+    public T getBatchForSubject(int subjectID, int batchID, int timeslice);
+}

src/main/java/database/DatabaseInstancesFetcher.java

@@ -0,0 +1,106 @@
+package database;
+
+import weka.core.Attribute;
+import weka.core.FastVector;
+import weka.core.Instance;
+import weka.core.Instances;
+
+import java.sql.PreparedStatement;
+import java.sql.ResultSet;
+import java.sql.SQLException;
+import java.util.Map;
+
+/**
+ * Created by Timotej on 02-Nov-15.
+ */
+public class DatabaseInstancesFetcher implements DatabaseDataset<Instances>{
+    private final PreparedStatement stmt;
+    private final FastVector attributes;
+    private final Attribute subjectIdAttribute;
+    //Maps feautreID to attributes
+    private final Map<Integer, Attribute> attributeMap;
+    //Maps feautreID to the alias of the relation from where it's from
+    private final Map<Integer, String> features;
+
+    /*
+        Should not be constructed on it's own. Use DataInstancesBuilder
+    */
+    DatabaseInstancesFetcher(PreparedStatement stmt,
+                             FastVector attributes,
+                             Attribute subjectIdAttribute,
+                             Map<Integer, Attribute> attributeMap,
+                             Map<Integer, String> features) {
+        this.stmt = stmt;
+        this.attributes = attributes;
+        this.subjectIdAttribute = subjectIdAttribute;
+        this.attributeMap = attributeMap;
+        this.features = features;
+    }
+
+
+    public Instances getBatch(int batchID, int timeslice){
+        try {
+            stmt.setInt(1, batchID);
+            //don't care about that as we want all of them
+            stmt.setInt(2, 2342);
+            //we want all the subjects
+            stmt.setBoolean(3, true);
+            stmt.setInt(4, timeslice);
+            ResultSet rs = stmt.executeQuery();
+            return resultSetToInstances(rs);
+
+        } catch (SQLException e) {
+            e.printStackTrace();
+            throw new RuntimeException("SQL failed to fetch data");
+        }
+    }
+
+    public Instances getBatchForSubject(int subjectID, int batchID, int timeslice) {
+        try {
+            stmt.setInt(1, batchID);
+            //don't care about that as we want all of them
+            stmt.setInt(2, subjectID);
+            //we want all the subjects
+            stmt.setBoolean(3, false);
+            stmt.setInt(4, timeslice);
+            ResultSet rs = stmt.executeQuery();
+            return resultSetToInstances(rs);
+
+        } catch (SQLException e) {
+            e.printStackTrace();
+            throw new RuntimeException("SQL failed to fetch data");
+        }
+    }
+
+    public void close() {
+        try {
+            stmt.close();
+        } catch (SQLException e) {
+            e.printStackTrace();
+        }
+    }
+
+    private Instances resultSetToInstances(ResultSet rs) throws SQLException {
+        Instances dataset;
+        dataset = new Instances("data", attributes, 10);
+        Instance inst;
+        int numFeatures = attributes.size();
+        while (rs.next()) {
+            inst = new Instance(numFeatures);
+            inst.setValue(subjectIdAttribute, rs.getInt("subjectID"));
+            for (Map.Entry<Integer, String> feature : features.entrySet()) {
+
+                inst.setValue(attributeMap.get(feature.getKey()),
+                        rs.getDouble(feature.getValue() + "fv"));
+            }
+
+            dataset.add(inst);
+
+        }
+        //if BPM (featureID == 1) is one of the attributes, set is as the class
+        if (attributeMap.keySet().contains(1)) {
+            dataset.setClass(attributeMap.get(1));
+        }
+        return dataset;
+    }
+}

src/main/java/database/NameDataInstancesBuilder.java

@@ -0,0 +1,44 @@
+package database;
+
+import weka.core.Instances;
+
+import java.sql.Connection;
+import java.sql.PreparedStatement;
+import java.sql.ResultSet;
+import java.sql.SQLException;
+import java.util.HashSet;
+
+/**
+ * Created by tk1713 on 02/11/15.
+ */
+public class NameDataInstancesBuilder implements DataInstancesBuilder<String, Integer, Integer> {
+
+    private final DataInstancesBuilder<Integer, Integer, Integer> builder;
+    private final HashSet<String> features = new HashSet<String>();
+
+    public NameDataInstancesBuilder(DataInstancesBuilder<Integer, Integer, Integer> builder) {
+        this.builder = builder;
+    }
+    public NameDataInstancesBuilder withFeature(String featureName) {
+        features.add(featureName);
+        return this;
+    }
+
+    public DatabaseDataset<Instances> build(Connection conn) {
+        try {
+            PreparedStatement stmt = conn.prepareStatement("SELECT featureID, name FROM feature");
+            ResultSet rs = stmt.executeQuery();
+            while (rs.next()) {
+                if (features.contains(rs.getString("name"))) {
+                    builder.withFeature(rs.getInt("featureID"));
+                }
+            }
+            stmt.close();
+        } catch (SQLException e) {
+            e.printStackTrace();
+            throw new RuntimeException("SQL failed to fetch data");
+        }
+
+        return builder.build(conn);
+    }
+}