Example usage for org.apache.mahout.utils.vectors.io SequenceFileVectorWriter SequenceFileVectorWriter

Introduction

In this page you can find the example usage for org.apache.mahout.utils.vectors.io SequenceFileVectorWriter SequenceFileVectorWriter.

Prototype

public SequenceFileVectorWriter(SequenceFile.Writer writer) 

Source Link

Usage

From source file:com.grantingersoll.intell.clustering.KMeansClusteringEngine.java

License:Apache License

private static VectorWriter getSeqFileWriter(String outFile) throws IOException {
    Path path = new Path(outFile);
    Configuration conf = new Configuration();
    FileSystem fs = FileSystem.get(conf);
    // TODO: Make this parameter driven

    SequenceFile.Writer seqWriter = SequenceFile.createWriter(fs, conf, path, LongWritable.class,
            VectorWritable.class);

    return new SequenceFileVectorWriter(seqWriter);
}

---

From source file:com.tamingtext.mahout.VectorExamplesTest.java

License:Apache License

@Test
public void testProgrammatic() throws Exception {
    //<start id="vec.examples.programmatic"/>
    double[] vals = new double[] { 0.3, 1.8, 200.228 };
    Vector dense = new DenseVector(vals);//<co id="vec.exam.dense"/>
    assertTrue(dense.size() == 3);/*from   www. ja va 2  s .  c  o m*/
    Vector sparseSame = new SequentialAccessSparseVector(3);//<co id="vec.exam.sparse.same"/>
    Vector sparse = new SequentialAccessSparseVector(3000);//<co id="vec.exam.sparse"/>
    for (int i = 0; i < vals.length; i++) {//<co id="vec.exam.assign.sparse"/>
        sparseSame.set(i, vals[i]);
        sparse.set(i, vals[i]);
    }
    assertFalse(dense.equals(sparse));//<co id="vec.exam.notequals.d.s"/>
    assertEquals(dense, sparseSame);//<co id="vec.exam.equals.d.s"/>
    assertFalse(sparse.equals(sparseSame));
    /*
    <calloutlist>
    <callout arearefs="vec.exam.dense"><para>Create a <classname>DenseVector</classname> with a label of "my-dense" and 3 values.  The cardinality of this vector is 3 </para></callout>
    <callout arearefs="vec.exam.sparse.same"><para>Create a <classname>SparseVector</classname> with a label of my-sparse-same that has cardinality of 3</para></callout>
            
    <callout arearefs="vec.exam.sparse"><para>Create a <classname>SparseVector</classname> with a label of my-sparse and a cardinality of 3000.</para></callout>
    <callout arearefs="vec.exam.assign.sparse"><para>Set the values to the first 3 items in the sparse vectors.</para></callout>
    <callout arearefs="vec.exam.notequals.d.s"><para>The dense and the sparse <classname>Vector</classname>s are not equal because they have different cardinality.</para></callout>
    <callout arearefs="vec.exam.equals.d.s"><para>The dense and sparseSame <classname>Vector</classname>s are equal because they have the same values and cardinality</para></callout>
            
    </calloutlist>
    */
    //<end id="vec.examples.programmatic"/>
    //<start id="vec.examples.seq.file"/>
    File tmpDir = new File(System.getProperty("java.io.tmpdir"));
    File tmpLoc = new File(tmpDir, "sfvwt");
    tmpLoc.mkdirs();
    File tmpFile = File.createTempFile("sfvwt", ".dat", tmpLoc);

    Path path = new Path(tmpFile.getAbsolutePath());
    Configuration conf = new Configuration();//<co id="vec.examples.seq.conf"/>
    FileSystem fs = FileSystem.get(conf);
    SequenceFile.Writer seqWriter = SequenceFile.createWriter(fs, conf, path, LongWritable.class,
            VectorWritable.class);//<co id="vec.examples.seq.writer"/>
    VectorWriter vecWriter = new SequenceFileVectorWriter(seqWriter);//<co id="vec.examples.seq.vecwriter"/>
    List<Vector> vectors = new ArrayList<Vector>();
    vectors.add(sparse);
    vectors.add(sparseSame);
    vecWriter.write(vectors);//<co id="vec.examples.seq.write"/>
    vecWriter.close();
    /*
    <calloutlist>
    <callout arearefs="vec.examples.seq.conf"><para>Create a <classname>Configuration</classname> for Hadoop</para></callout>
    <callout arearefs="vec.examples.seq.writer"><para>Create a Hadoop <classname>SequenceFile.Writer</classname> to handle the job of physically writing out the vectors to a file in HDFS</para></callout>
    <callout arearefs="vec.examples.seq.vecwriter"><para>A <classname>VectorWriter</classname> processes the <classname>Vector</classname>s and invokes the underlying write methods on the <classname>SequenceFile.Writer</classname></para></callout>
    <callout arearefs="vec.examples.seq.write"><para>Do the work of writing out the files</para></callout>
            
    </calloutlist>
    */
    //<end id="vec.examples.seq.file"/>
}

---