Package com.mozilla.hadoop.fs

Examples of com.mozilla.hadoop.fs.SequenceFileDirectoryReader

60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
    }
   
    public static Map<Integer, PriorityQueue<Pair<Double,String>>> getDocIds(Path docTopicsPath, int numDocs) {
        Map<Integer, PriorityQueue<Pair<Double,String>>> docIdMap = new HashMap<Integer, PriorityQueue<Pair<Double,String>>>();
        Map<Integer, Double> maxDocScores = new HashMap<Integer,Double>();
        SequenceFileDirectoryReader pointsReader = null;
        try {
            Text k = new Text();
            VectorWritable vw = new VectorWritable();
            pointsReader = new SequenceFileDirectoryReader(docTopicsPath);
            while (pointsReader.next(k, vw)) {
                String docId = k.toString();
                Vector normGamma = vw.get();
                Iterator<Element> iter = normGamma.iterateNonZero();
                double maxTopicScore = 0.0;
                int idx = 0;
                int topic = 0;
                while(iter.hasNext()) {
                    Element e = iter.next();
                    double score = e.get();
                    if (score > maxTopicScore) {
                        maxTopicScore = score;
                        topic = idx;
                    }
                   
                    idx++; 
                }
               
                PriorityQueue<Pair<Double,String>> docIdsForTopic = docIdMap.get(topic);
                if (docIdsForTopic == null) {
                    docIdsForTopic = new PriorityQueue<Pair<Double,String>>(numDocs);
                }
               
                Double maxDocScoreForTopic = maxDocScores.get(topic);
                if (maxDocScoreForTopic == null) {
                    maxDocScoreForTopic = 0.0;
                }
                if (maxTopicScore > maxDocScoreForTopic) {
                    maxDocScores.put(topic, maxTopicScore);
                }
               
                enqueue(docIdsForTopic, docId, maxTopicScore, numDocs);
                docIdMap.put(topic, docIdsForTopic);
            }
        } catch (IOException e) {
            LOG.error("IOException caught while reading clustered points", e);
        } finally {
            if (pointsReader != null) {
                pointsReader.close();
            }
        }

        for (Map.Entry<Integer, Double> entry : maxDocScores.entrySet()) {
            System.out.println("For topic: " + entry.getKey() + " max score: " + entry.getValue());
View Full Code Here
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
    }
   
    public static Map<Integer,PriorityQueue<Pair<Double,String>>> getTopWordsByTopic(String stateDirPath, Map<Integer,String> featureIndex, int numWordsToPrint) {
        Map<Integer,Double> expSums = new HashMap<Integer, Double>();
        Map<Integer,PriorityQueue<Pair<Double,String>>> queues = new HashMap<Integer,PriorityQueue<Pair<Double,String>>>();
        SequenceFileDirectoryReader reader = null;
        try {
            IntPairWritable k = new IntPairWritable();
            DoubleWritable v = new DoubleWritable();
            reader = new SequenceFileDirectoryReader(new Path(stateDirPath));
            while (reader.next(k, v)) {
                int topic = k.getFirst();
                int featureId = k.getSecond();
                if (featureId >= 0 && topic >= 0) {
                    double score = v.get();
                    Double curSum = expSums.get(topic);
                    if (curSum == null) {
                        curSum = 0.0;
                    }
                    expSums.put(topic, curSum + Math.exp(score));
                    String feature = featureIndex.get(featureId);
                   
                    PriorityQueue<Pair<Double,String>> q = queues.get(topic);
                    if (q == null) {
                        q = new PriorityQueue<Pair<Double,String>>(numWordsToPrint);
                    }
                    enqueue(q, feature, score, numWordsToPrint);
                    queues.put(topic, q);
                }
            }
        } catch (IOException e) {
            LOG.error("Error reading LDA state dir", e);
        } finally {
            if (reader != null) {
                reader.close();
            }
        }
       
        for (Map.Entry<Integer, PriorityQueue<Pair<Double,String>>> entry : queues.entrySet()) {
            int topic = entry.getKey();
View Full Code Here
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
    }
   
    public static Map<Integer, PriorityQueue<Pair<Double,String>>> getTopDocIdsByTopic(Path docTopicsPath, int numDocs) {
        Map<Integer, PriorityQueue<Pair<Double,String>>> docIdMap = new HashMap<Integer, PriorityQueue<Pair<Double,String>>>();
        Map<Integer, Double> maxDocScores = new HashMap<Integer,Double>();
        SequenceFileDirectoryReader pointsReader = null;
        try {
            Text k = new Text();
            VectorWritable vw = new VectorWritable();
            pointsReader = new SequenceFileDirectoryReader(docTopicsPath);
            while (pointsReader.next(k, vw)) {
                String docId = k.toString();
                Vector normGamma = vw.get();
                Iterator<Element> iter = normGamma.iterateNonZero();
                double maxTopicScore = 0.0;
                int idx = 0;
                int topic = 0;
                while(iter.hasNext()) {
                    Element e = iter.next();
                    double score = e.get();
                    if (score > maxTopicScore) {
                        maxTopicScore = score;
                        topic = idx;
                    }
                   
                    idx++; 
                }
               
                PriorityQueue<Pair<Double,String>> docIdsForTopic = docIdMap.get(topic);
                if (docIdsForTopic == null) {
                    docIdsForTopic = new PriorityQueue<Pair<Double,String>>(numDocs);
                }
               
                Double maxDocScoreForTopic = maxDocScores.get(topic);
                if (maxDocScoreForTopic == null) {
                    maxDocScoreForTopic = 0.0;
                }
                if (maxTopicScore > maxDocScoreForTopic) {
                    maxDocScores.put(topic, maxTopicScore);
                }
               
                enqueue(docIdsForTopic, docId, maxTopicScore, numDocs);
                docIdMap.put(topic, docIdsForTopic);
            }
        } catch (IOException e) {
            LOG.error("IOException caught while reading clustered points", e);
        } finally {
            if (pointsReader != null) {
                pointsReader.close();
            }
        }

        for (Map.Entry<Integer, Double> entry : maxDocScores.entrySet()) {
            System.out.println("For topic: " + entry.getKey() + " max score: " + entry.getValue());
View Full Code Here
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
    }
   
    public static Map<Integer,PriorityQueue<Pair<Double,String>>> getTopWordsByTopics(String stateDirPath, Map<Integer,String> featureIndex, int numWordsToPrint) {
        Map<Integer,Double> expSums = new HashMap<Integer, Double>();
        Map<Integer,PriorityQueue<Pair<Double,String>>> queues = new HashMap<Integer,PriorityQueue<Pair<Double,String>>>();
        SequenceFileDirectoryReader reader = null;
        try {
            IntPairWritable k = new IntPairWritable();
            DoubleWritable v = new DoubleWritable();
            reader = new SequenceFileDirectoryReader(new Path(stateDirPath));
            while (reader.next(k, v)) {
                int topic = k.getFirst();
                int featureId = k.getSecond();
                if (featureId >= 0 && topic >= 0) {
                    double score = v.get();
                    Double curSum = expSums.get(topic);
                    if (curSum == null) {
                        curSum = 0.0;
                    }
                    expSums.put(topic, curSum + Math.exp(score));
                    String feature = featureIndex.get(featureId);
                   
                    PriorityQueue<Pair<Double,String>> q = queues.get(topic);
                    if (q == null) {
                        q = new PriorityQueue<Pair<Double,String>>(numWordsToPrint);
                    }
                    enqueue(q, feature, score, numWordsToPrint);
                    queues.put(topic, q);
                }
            }
        } catch (IOException e) {
            LOG.error("Error reading LDA state dir", e);
        } finally {
            if (reader != null) {
                reader.close();
            }
        }
       
        for (Map.Entry<Integer, PriorityQueue<Pair<Double,String>>> entry : queues.entrySet()) {
            int topic = entry.getKey();
View Full Code Here
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
    }
   
    public Map<Integer,Set<String>> getDocIds(double sampleRate) {
        Random rand = new Random();
        Map<Integer,Set<String>> docIdMap = new HashMap<Integer,Set<String>>();
        SequenceFileDirectoryReader pointsReader = null;
        try {
            IntWritable k = new IntWritable();
            WeightedVectorWritable wvw = new WeightedVectorWritable();
            pointsReader = new SequenceFileDirectoryReader(clusteredPointsPath);
            while (pointsReader.next(k, wvw)) {
                int clusterId = k.get();               
                Vector v = wvw.getVector();
                if (v instanceof NamedVector) {
                    if (rand.nextDouble() < sampleRate) {
                        NamedVector nv = (NamedVector)v;
                        nv.getName();
                        Set<String> curDocIds = docIdMap.get(clusterId);
                        if (curDocIds == null) {
                            curDocIds = new HashSet<String>();
                        }
                        curDocIds.add(nv.getName());
                        docIdMap.put(clusterId, curDocIds);
                    }
                }
            }
        } catch (IOException e) {
            LOG.error("IOException caught while reading clustered points", e);
        } finally {
            if (pointsReader != null) {
                pointsReader.close();
            }
        }
       
        return docIdMap;
    }
View Full Code Here
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
        this.invertedFeatureIndex = Dictionary.loadInvertedFeatureIndex(dictionaryPath);
    }
   
    public Map<Integer,Cloud> getClouds(Cloud template) {
        Map<Integer,Cloud> cloudMap = new HashMap<Integer,Cloud>();
        SequenceFileDirectoryReader pointsReader = null;
        try {
            IntWritable k = new IntWritable();
            WeightedVectorWritable wvw = new WeightedVectorWritable();
            pointsReader = new SequenceFileDirectoryReader(clusteredPointsPath);
            while (pointsReader.next(k, wvw)) {
                int clusterId = k.get();
                Cloud c = cloudMap.get(clusterId);
                if (c == null) {
                    c = new Cloud(template);
                }
                Iterator<Element> viter = wvw.getVector().iterateNonZero();
                while (viter.hasNext()) {
                    Element e = viter.next();
                    String feature = invertedFeatureIndex.get(e.index());
                    c.addTag(new Tag(feature, e.get()));
                }
               
                cloudMap.put(clusterId, c);
            }
        } catch (IOException e) {
            LOG.error("IOException caught while reading clustered points", e);
        } finally {
            if (pointsReader != null) {
                pointsReader.close();
            }
        }
       
        return cloudMap;
    }
View Full Code Here
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
    private static final Logger LOG = Logger.getLogger(DisplayKMeansBase.class);
   
    public List<Pair<Integer,Vector>> readClusteredPoints(Path clusteredPointsPath) {
        List<Pair<Integer,Vector>> clusteredPoints = new ArrayList<Pair<Integer,Vector>>();
        SequenceFileDirectoryReader pointsReader = null;
        try {
            IntWritable k = new IntWritable();
            WeightedVectorWritable wvw = new WeightedVectorWritable();
            pointsReader = new SequenceFileDirectoryReader(clusteredPointsPath);
            while (pointsReader.next(k, wvw)) {               
                clusteredPoints.add(new Pair<Integer,Vector>(k.get(), wvw.getVector()));
            }
        } catch (IOException e) {
            LOG.error("IOException caught while reading clustered points", e);
        } finally {
            if (pointsReader != null) {
                pointsReader.close();
            }
        }
       
        return clusteredPoints;
    }
View Full Code Here
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
        return clusteredPoints;
    }
   
    public List<Cluster> readClustersIteration(Path clusterIterationPath) {
        List<Cluster> clusters = new ArrayList<Cluster>();
        SequenceFileDirectoryReader iterationReader = null;
        try {
            Text k = new Text();
            Cluster c = new Cluster();
            iterationReader = new SequenceFileDirectoryReader(clusterIterationPath);
            while (iterationReader.next(k, c)) {
                clusters.add(c);
            }
        } catch (IOException e) {
            LOG.error("IOException caught while reading clustered points", e);
        } finally {
            if (iterationReader != null) {
                iterationReader.close();
            }
        }

        return clusters;
    }
View Full Code Here
TOP

Related Classes of com.mozilla.hadoop.fs.SequenceFileDirectoryReader

  • com.mozilla.grouperfish.mahout.clustering.display.kmeans.DisplayKMeansBase
  • com.mozilla.grouperfish.mahout.clustering.display.kmeans.OriginalText
  • com.mozilla.grouperfish.mahout.clustering.display.kmeans.WordCloud
  • com.mozilla.grouperfish.mahout.clustering.display.lda.DisplayLDABase
  • com.mozilla.grouperfish.mahout.clustering.display.lda.DisplayLDATopics
  • com.mozilla.grouperfish.mahout.clustering.display.lda.OriginalText
  • org.apache.hadoop.fs.Path
    • Copyright © 2018 www.massapicom. All rights reserved.
    All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.