/*

  * Licensed to the Apache Software Foundation (ASF) under one

  * or more contributor license agreements.  See the NOTICE file

  * distributed with this work for additional information

  * regarding copyright ownership.  The ASF licenses this file

  * to you under the Apache License, Version 2.0 (the

  * "License"); you may not use this file except in compliance

  * with the License.  You may obtain a copy of the License at

  *

  *     http://www.apache.org/licenses/LICENSE-2.0

  *

  * Unless required by applicable law or agreed to in writing, software

  * distributed under the License is distributed on an "AS IS" BASIS,

  * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.

  * See the License for the specific language governing permissions and

  * limitations under the License.

  */

 package org.apache.giraph.examples;

 import org.apache.giraph.conf.ImmutableClassesGiraphConfigurable;

 import org.apache.giraph.conf.ImmutableClassesGiraphConfiguration;

 import org.apache.giraph.edge.Edge;

 import org.apache.giraph.edge.EdgeFactory;

 import org.apache.giraph.graph.Vertex;

 import org.apache.giraph.io.formats.TextVertexInputFormat;

 import org.apache.hadoop.io.DoubleWritable;

 import org.apache.hadoop.io.LongWritable;

 import org.apache.hadoop.mapreduce.InputSplit;

 import org.apache.hadoop.mapreduce.TaskAttemptContext;

 import com.google.common.collect.Lists;

 import java.io.IOException;

 import java.util.Collection;

 import java.util.List;

 import java.util.regex.Pattern;

 /**

  * Simple text-based {@link org.apache.giraph.io.VertexInputFormat} for

  * unweighted graphs with long ids. Each line consists of: vertex

  * neighbor1:weight1 neighbor2:weight2 ...

  */

 public class NormalizingLongDoubleDoubleTextInputFormat

     extends

     TextVertexInputFormat<LongWritable, DoubleWritable, DoubleWritable>

     implements ImmutableClassesGiraphConfigurable<LongWritable, DoubleWritable,

     DoubleWritable> {

   /** Configuration. */

   private ImmutableClassesGiraphConfiguration<LongWritable, DoubleWritable,

       DoubleWritable> conf;

   @Override

   public TextVertexReader createVertexReader(

       InputSplit split, TaskAttemptContext context) throws IOException {

     return new NormalizingLongDoubleDoubleDoubleVertexReader();

   }

   @Override

   public void setConf(ImmutableClassesGiraphConfiguration<LongWritable,

       DoubleWritable, DoubleWritable> configuration) {

     conf = configuration;

   }

   @Override

   public ImmutableClassesGiraphConfiguration<LongWritable, DoubleWritable,

       DoubleWritable> getConf() {

     return conf;

   }

   /**

    * Vertex reader associated with

    * {@link LongDoubleDoubleTextInputFormat}.

    */

   public class NormalizingLongDoubleDoubleDoubleVertexReader

       extends NormalizingLongDoubleDoubleTextInputFormat.TextVertexReader {

     /** Separator of the vertex and neighbors */

     private final Pattern edgeSeparator = Pattern.compile("\\s+");

     /** Separator of the edge id and edge weight */

     private final Pattern weightSeparator = Pattern.compile(":");

     @Override

     public Vertex<LongWritable, DoubleWritable, DoubleWritable>

     getCurrentVertex() throws IOException, InterruptedException {

       Vertex<LongWritable, DoubleWritable, DoubleWritable> vertex =

           conf.createVertex();

       String[] tokens = edgeSeparator.split(getRecordReader()

           .getCurrentValue().toString());

       List<Edge<LongWritable, DoubleWritable>> edges = Lists

           .newArrayListWithCapacity(tokens.length - 1);

       parse(tokens, edges);

       normalize(edges);

       LongWritable vertexId = new LongWritable(Long.parseLong(tokens[0]));

       vertex.initialize(vertexId, new DoubleWritable(), edges);

       return vertex;

     }

     /**

      * Parse a set of tokens into a map ID -> weight.

      * @param tokens The tokens to be parsed.

      * @param edges The map that will contain the result of the parsing.

      */

     void parse(String[] tokens,

                Collection<Edge<LongWritable, DoubleWritable>> edges) {

       for (int n = 1; n < tokens.length; n++) {

         String[] parts = weightSeparator.split(tokens[n]);

         edges.add(EdgeFactory.create(new LongWritable(Long.parseLong(parts[0])),

             new DoubleWritable(Double.parseDouble(parts[1]))));

       }

     }

     /**

      * Normalize the edges with L1 normalization.

      * @param edges The edges to be normalized.

      */

     void normalize(Collection<Edge<LongWritable, DoubleWritable>> edges) {

       if (edges == null || edges.size() == 0) {

         throw new IllegalArgumentException(

             "Cannot normalize an empy set of edges");

       }

       float normalizer = 0.0f;

       for (Edge<LongWritable, DoubleWritable> edge : edges) {

         normalizer += edge.getValue().get();

       }

       for (Edge<LongWritable, DoubleWritable> edge : edges) {

         edge.getValue().set(edge.getValue().get() / normalizer);

       }

     }

     @Override

     public boolean nextVertex() throws IOException, InterruptedException {

       return getRecordReader().nextKeyValue();

     }

   }

 }