Package ivory.regression.sigir2013.moses

Source Code of ivory.regression.sigir2013.moses.EnAr_TREC02

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
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
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
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
package ivory.regression.sigir2013.moses;

import ivory.core.eval.Qrels;
import ivory.regression.GroundTruth;
import ivory.regression.GroundTruth.Metric;
import ivory.smrf.retrieval.Accumulator;
import ivory.sqe.retrieval.QueryEngine;
import ivory.sqe.retrieval.RunQueryEngine;
import java.util.HashMap;
import java.util.Map;
import java.util.Set;
import junit.framework.JUnit4TestAdapter;
import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.fs.FileSystem;
import org.junit.Test;
import com.google.common.collect.Maps;
import edu.umd.cloud9.collection.DocnoMapping;
import edu.umd.cloud9.io.map.HMapSFW;

public class EnAr_TREC02 {
  private QueryEngine qe;
  private static String PATH = "en-ar.trec02";
  private static String LANGUAGE = "ar";
  private static int numTopics = 50;

  private static float expectedTokenMAP = 0.2714f;
  private static Map<Integer,float[]> expectedMAPs = new HashMap<Integer,float[]>();
 
  private static Map<Integer, String[]> grammar_AP = new HashMap<Integer, String[]>();
  private static Map<Integer, String[]> Nbest_AP = new HashMap<Integer, String[]>();
  private static Map<Integer, String[]> Onebest_AP = new HashMap<Integer, String[]>();
  private static Map<Integer, String[]> Interp_AP = new HashMap<Integer, String[]>();

  public static void initialize() {
    expectedMAPs.put(0, new float[]{0.2725f, 0.2320f, 0.2489f, 0.2748f});   // "one2none" -> grammar,1best,10best,interp
    expectedMAPs.put(1, new float[]{0.2656f, 0.2492f, 0.2544f, 0.2735f});   // "one2one" -> grammar,1best,10best,interp
    expectedMAPs.put(2, new float[]{0.2742f, 0.2456f, 0.2636f, 0.2776f});   // "one2many" -> grammar,1best,10best,interp
    grammar_AP.put(0, new String[] {
        "35", "0.1804","36", "0.0021","33", "0.0299","34", "0.0062","39", "0.3566","37", "0.376","38", "0.0071","43", "0.0","42", "0.4193","41", "0.3308","40", "0.5147","67", "0.6256","66", "0.153","69", "0.5646","68", "0.2853","26", "0.0063","27", "0.3447","28", "0.0178","29", "0.5264","30", "0.7025","32", "0.0823","31", "0.0404","70", "0.5069","71", "0.2641","72", "0.0284","73", "0.282","74", "0.0305","75", "0.1662","59", "0.7346","58", "0.1665","57", "0.0961","56", "0.4687","55", "0.6265","64", "6.0E-4","65", "0.399","62", "0.0507","63", "0.1007","60", "0.735","61", "0.0802","49", "0.7932","48", "0.7049","45", "0.0297","44", "0.1628","47", "0.3341","46", "0.3137","51", "0.0539","52", "0.149","53", "0.0036","54", "0.0363","50", "0.7372"
    });
    grammar_AP.put(1, new String[] {
        "35", "0.1037","36", "0.0022","33", "0.0229","34", "0.0059","39", "0.3325","37", "0.3722","38", "0.0073","43", "0.0","42", "0.4189","41", "0.249","40", "0.3957","67", "0.6286","66", "0.2452","69", "0.5362","68", "0.3104","26", "0.0056","27", "0.3447","28", "0.02","29", "0.5354","30", "0.6487","32", "0.0259","31", "0.0472","70", "0.5058","71", "0.2728","72", "0.027","73", "0.289","74", "0.0406","75", "0.1742","59", "0.7391","58", "0.0146","57", "0.1097","56", "0.42","55", "0.6252","64", "6.0E-4","65", "0.5014","62", "0.0514","63", "0.1048","60", "0.6849","61", "0.044","49", "0.8038","48", "0.6961","45", "0.0868","44", "0.1631","47", "0.3459","46", "0.3115","51", "0.0501","52", "0.1545","53", "0.0055","54", "0.044","50", "0.755"
    });
    grammar_AP.put(2, new String[] {
        "35", "0.1934","36", "0.0022","33", "0.0311","34", "0.0062","39", "0.3568","37", "0.3762","38", "0.0070","43", "0.0","42", "0.4141","41", "0.1347","40", "0.5304","67", "0.6308","66", "0.1942","69", "0.5631","68", "0.2799","26", "0.0056","27", "0.3447","28", "0.021","29", "0.5279","30", "0.719","32", "0.0908","31", "0.0403","70", "0.5056","71", "0.2694","72", "0.0295","73", "0.2819","74", "0.0334","75", "0.1663","59", "0.7367","58", "0.1385","57", "0.0961","56", "0.488","55", "0.6263","64", "6.0E-4","65", "0.5259","62", "0.0494","63", "0.1007","60", "0.719","61", "0.0456","49", "0.797","48", "0.7042","45", "0.0623","44", "0.1639","47", "0.334","46", "0.3219","51", "0.0486","52", "0.1539","53", "0.0038","54", "0.0448","50", "0.7936"
    });
    Nbest_AP.put(0, new String[] {
        "35", "0.1857","36", "0.0032","33", "0.023","34", "0.0036","39", "0.3364","37", "0.2864","38", "0.0066","43", "0.0","42", "0.413","41", "0.035","40", "0.4676","67", "0.5442","66", "0.2495","69", "0.5353","68", "0.1407","26", "0.0153","27", "0.2883","28", "0.0218","29", "0.5233","30", "0.609","32", "0.0832","31", "0.0594","70", "0.5315","71", "0.0889","72", "0.0061","73", "0.1576","74", "0.0133","75", "0.1388","59", "0.7296","58", "0.1361","57", "0.1723","56", "0.4276","55", "0.5829","64", "6.0E-4","65", "0.3166","62", "0.0522","63", "0.0886","60", "0.7308","61", "0.0829","49", "0.7505","48", "0.6716","45", "0.0495","44", "0.1796","47", "0.3371","46", "0.3172","51", "0.0888","52", "0.0832","53", "0.0025","54", "0.0379","50", "0.8417"
    });
    Nbest_AP.put(1, new String[] {
        "35", "0.1857","36", "0.0032","33", "0.0204","34", "0.0036","39", "0.3159","37", "0.2864","38", "0.0066","43", "0.0","42", "0.413","41", "0.0285","40", "0.2389","67", "0.5924","66", "0.3667","69", "0.5353","68", "0.1407","26", "0.0050","27", "0.2883","28", "0.0251","29", "0.5233","30", "0.6284","32", "0.0635","31", "0.0594","70", "0.5315","71", "0.0889","72", "0.0228","73", "0.1757","74", "0.029","75", "0.1388","59", "0.7296","58", "0.1361","57", "0.1723","56", "0.3616","55", "0.5829","64", "6.0E-4","65", "0.5289","62", "0.0522","63", "0.0886","60", "0.7308","61", "0.0829","49", "0.8409","48", "0.6733","45", "0.3643","44", "0.1796","47", "0.3371","46", "0.3172","51", "0.0889","52", "0.079","53", "0.0025","54", "0.0332","50", "0.6194"
    });
    Nbest_AP.put(2, new String[] {
        "35", "0.1857","36", "0.0032","33", "0.0284","34", "0.0036","39", "0.3364","37", "0.2864","38", "0.0066","43", "0.0","42", "0.413","41", "0.0403","40", "0.4806","67", "0.5975","66", "0.3614","69", "0.5353","68", "0.1407","26", "0.0158","27", "0.2883","28", "0.0213","29", "0.5233","30", "0.6131","32", "0.0832","31", "0.0594","70", "0.5315","71", "0.0889","72", "0.0232","73", "0.1589","74", "0.0118","75", "0.1388","59", "0.7296","58", "0.1361","57", "0.1723","56", "0.408","55", "0.5829","64", "6.0E-4","65", "0.5386","62", "0.0522","63", "0.0886","60", "0.7308","61", "0.0829","49", "0.8021","48", "0.6724","45", "0.3457","44", "0.1796","47", "0.3371","46", "0.3172","51", "0.0888","52", "0.0796","53", "0.0025","54", "0.0348","50", "0.8186"
    });
    Onebest_AP.put(0, new String[] {
        "35", "0.1409","36", "0.0032","33", "0.0401","34", "0.0036","39", "0.2904","37", "0.2199","38", "0.0052","43", "0.0","42", "0.4017","41", "0.0","40", "0.4349","67", "0.5407","66", "0.342","69", "0.5263","68", "0.1377","26", "0.0176","27", "0.2883","28", "0.0015","29", "0.5233","30", "0.6088","32", "0.0832","31", "0.0743","70", "0.474","71", "0.0889","72", "0.0064","73", "0.1318","74", "0.0221","75", "0.1259","59", "0.7296","58", "0.1272","57", "0.0050","56", "0.4276","55", "0.6427","64", "6.0E-4","65", "0.5105","62", "0.0541","63", "0.0757","60", "0.7304","61", "0.0706","49", "0.704","48", "0.6716","45", "0.1433","44", "0.1796","47", "0.3341","46", "0.3358","51", "0.0825","52", "0.0832","53", "0.0021","54", "0.0339","50", "0.121"
    });
    Onebest_AP.put(1, new String[] {
        "35", "0.1409","36", "0.0032","33", "0.0193","34", "0.0036","39", "0.2904","37", "0.2199","38", "0.0052","43", "0.0","42", "0.4017","41", "0.0256","40", "0.4349","67", "0.5823","66", "0.342","69", "0.5263","68", "0.1377","26", "0.0176","27", "0.2883","28", "0.0544","29", "0.5233","30", "0.6088","32", "0.0484","31", "0.0743","70", "0.474","71", "0.0889","72", "0.0064","73", "0.1318","74", "0.0221","75", "0.1259","59", "0.7296","58", "0.1272","57", "0.0050","56", "0.2484","55", "0.6427","64", "6.0E-4","65", "0.5631","62", "0.0541","63", "0.0757","60", "0.7304","61", "0.0706","49", "0.704","48", "0.6716","45", "0.5262","44", "0.1796","47", "0.3341","46", "0.3358","51", "0.0825","52", "0.0832","53", "0.0021","54", "0.0339","50", "0.6625"
    });
    Onebest_AP.put(2, new String[] {
        "35", "0.1409","36", "0.0032","33", "0.0252","34", "0.0036","39", "0.2904","37", "0.2199","38", "0.0052","43", "0.0","42", "0.4017","41", "0.0295","40", "0.4349","67", "0.6033","66", "0.342","69", "0.5263","68", "0.1377","26", "0.0176","27", "0.2883","28", "0.0395","29", "0.5233","30", "0.6088","32", "0.0484","31", "0.0743","70", "0.474","71", "0.0889","72", "0.0064","73", "0.1318","74", "0.0221","75", "0.1259","59", "0.7296","58", "0.1272","57", "0.0050","56", "0.1912","55", "0.6427","64", "6.0E-4","65", "0.4456","62", "0.0541","63", "0.0757","60", "0.7304","61", "0.0706","49", "0.704","48", "0.6716","45", "0.513","44", "0.1796","47", "0.3341","46", "0.3358","51", "0.0825","52", "0.0832","53", "0.0021","54", "0.0339","50", "0.652"
    });
    Interp_AP.put(0, new String[] {
        "35", "0.1802","36", "0.0027","33", "0.0258","34", "0.0053","39", "0.3406","37", "0.3371","38", "0.0070","43", "0.0","42", "0.415","41", "0.3588","40", "0.5054","67", "0.6247","66", "0.2385","69", "0.5531","68", "0.2346","26", "0.0078","27", "0.311","28", "0.021","29", "0.5217","30", "0.6294","32", "0.2574","31", "0.0585","70", "0.532","71", "0.2315","72", "0.0266","73", "0.2417","74", "0.0253","75", "0.1621","59", "0.7341","58", "0.1366","57", "0.1264","56", "0.4036","55", "0.6272","64", "6.0E-4","65", "0.4858","62", "0.0506","63", "0.0848","60", "0.7417","61", "0.0891","49", "0.7783","48", "0.7015","45", "0.1264","44", "0.1696","47", "0.3355","46", "0.3179","51", "0.057","52", "0.1282","53", "0.0033","54", "0.0371","50", "0.7493"
    });
    Interp_AP.put(1, new String[] {
        "35", "0.1644","36", "0.0027","33", "0.0215","34", "0.0052","39", "0.3201","37", "0.3365","38", "0.0069","43", "0.0","42", "0.4157","41", "0.2666","40", "0.3948","67", "0.6229","66", "0.3132","69", "0.5314","68", "0.2389","26", "0.0071","27", "0.3111","28", "0.0221","29", "0.5288","30", "0.6816","32", "0.1236","31", "0.0602","70", "0.5315","71", "0.2376","72", "0.0257","73", "0.2471","74", "0.0324","75", "0.1614","59", "0.7345","58", "0.1099","57", "0.1269","56", "0.3877","55", "0.6295","64", "6.0E-4","65", "0.5192","62", "0.0511","63", "0.0832","60", "0.7306","61", "0.0899","49", "0.8235","48", "0.7013","45", "0.2919","44", "0.1701","47", "0.3379","46", "0.3166","51", "0.0575","52", "0.1312","53", "0.0035","54", "0.0393","50", "0.7298"
    });
    Interp_AP.put(2, new String[] {
        "35", "0.1799","36", "0.0027","33", "0.0278","34", "0.0052","39", "0.3408","37", "0.3385","38", "0.0070","43", "0.0","42", "0.4157","41", "0.1878","40", "0.5133","67", "0.6263","66", "0.2904","69", "0.5522","68", "0.2321","26", "0.0076","27", "0.311","28", "0.0194","29", "0.5242","30", "0.6404","32", "0.2063","31", "0.0585","70", "0.531","71", "0.2333","72", "0.0267","73", "0.2451","74", "0.0256","75", "0.1621","59", "0.7343","58", "0.135","57", "0.1264","56", "0.4115","55", "0.634","64", "6.0E-4","65", "0.5263","62", "0.0491","63", "0.0848","60", "0.7398","61", "0.0902","49", "0.7975","48", "0.702","45", "0.2794","44", "0.1701","47", "0.3348","46", "0.3215","51", "0.0571","52", "0.1295","53", "0.0033","54", "0.038","50", "0.8037"
    });
  }

  private static String[] baseline_token_AP = new String[] {
    "35", "0.1661","36", "0.0022","33", "0.0198","34", "0.0059","39", "0.324","37", "0.3725","38", "0.0068","43", "0.0","42", "0.4007","41", "0.3236","40", "0.5037","67", "0.5584","66", "0.2819","69", "0.5405","68", "0.2965","26", "0.0037","27", "0.2963","28", "0.0021","29", "0.5275","30", "0.7243","32", "0.3655","31", "0.042","70", "0.5104","71", "0.2278","72", "0.0276","73", "0.248","74", "0.0159","75", "0.1714","59", "0.7359","58", "0.1171","57", "0.0971","56", "0.3585","55", "0.616","64", "6.0E-4","65", "0.5153","62", "0.0509","63", "0.0783","60", "0.7071","61", "0.0885","49", "0.7933","48", "0.7146","45", "0.0731","44", "0.171","47", "0.3368","46", "0.2876","51", "0.0186","52", "0.0617","53", "0.0038","54", "0.0409","50", "0.7403"
  };

  public EnAr_TREC02() {
    super();
    qe = new QueryEngine();
  }

  @Test
  public void runRegressions() throws Exception {
    initialize();
    runRegression(0);   // "one2none"
    runRegression(1);   // "one2one"
    runRegression(2);   // "one2many"
    verifyAllResults(qe.getModels(), qe.getAllResults(), qe.getDocnoMapping(),
        new Qrels("data/" + PATH + "/qrels." + PATH + ".txt"));
  }

  public void runRegression(int heuristic) throws Exception {
    /////// baseline-token
    Configuration conf = RunQueryEngine.parseArgs(new String[] {
        "--xml", "data/" + PATH + "/run_en-" + LANGUAGE + ".token.xml",
        "--queries_path", "data/" + PATH + "/moses/title_en-" + LANGUAGE + "-trans10-filtered.xml", "--one2many", heuristic + "", "--is_stemming", "--is_doc_stemmed",
        "--unknown", "data/" + PATH + "/moses/10.unk"});
    FileSystem fs = FileSystem.getLocal(conf);
       
    // no need to repeat token-based case for other heuristics
    if (heuristic == 0) {
      qe.init(conf, fs);
      qe.runQueries(conf);
    }
    /////// 1-best

    conf = RunQueryEngine.parseArgs(new String[] {
        "--xml", "data/" + PATH + "/run_en-" + LANGUAGE + ".1best.xml",
        "--queries_path", "data/" + PATH + "/moses/title_en-" + LANGUAGE + "-trans1-filtered.xml", "--one2many", heuristic + "", "--is_stemming", "--is_doc_stemmed",
        "--unknown", "data/" + PATH + "/moses/1.unk"});

    qe.init(conf, fs);
    qe.runQueries(conf);

    /////// grammar

    conf = RunQueryEngine.parseArgs(new String[] {
        "--xml", "data/" + PATH + "/run_en-" + LANGUAGE + ".grammar.xml",
        "--queries_path", "data/" + PATH + "/moses/title_en-" + LANGUAGE + "-trans10-filtered.xml", "--one2many", heuristic + "", "--is_stemming", "--is_doc_stemmed",
        "--unknown", "data/" + PATH + "/moses/10.unk"});

    qe.init(conf, fs);
    qe.runQueries(conf);

    /////// 10-best

    conf = RunQueryEngine.parseArgs(new String[] {
        "--xml", "data/" + PATH + "/run_en-" + LANGUAGE + ".10best.xml",
        "--queries_path", "data/" + PATH + "/moses/title_en-" + LANGUAGE + "-trans10-filtered.xml", "--one2many", heuristic + "", "--is_stemming", "--is_doc_stemmed",
        "--unknown", "data/" + PATH + "/moses/10.unk"});

    qe.init(conf, fs);
    qe.runQueries(conf);

    /////// interp

    conf = RunQueryEngine.parseArgs(new String[] {
        "--xml", "data/"+ PATH + "/run_en-" + LANGUAGE + ".interp.xml",
        "--queries_path", "data/" + PATH + "/moses/title_en-" + LANGUAGE + "-trans10-filtered.xml", "--one2many", heuristic + "", "--is_stemming", "--is_doc_stemmed",
        "--unknown", "data/" + PATH + "/moses/10.unk"});

    qe.init(conf, fs);
    qe.runQueries(conf);
  }

  public static void verifyAllResults(Set<String> models,
      Map<String, Map<String, Accumulator[]>> results, DocnoMapping mapping, Qrels qrels) {

    Map<String, GroundTruth> g = Maps.newHashMap();

    g.put("en-" + LANGUAGE + ".token_10-0-100-100_0", new GroundTruth(Metric.AP, numTopics, baseline_token_AP, expectedTokenMAP));

    for (int heuristic=0; heuristic <= 2; heuristic++) {
      g.put("en-" + LANGUAGE + ".grammar_10-0-0-100_" + heuristic, new GroundTruth(Metric.AP, numTopics, grammar_AP.get(heuristic), expectedMAPs.get(heuristic)[0]));
      g.put("en-" + LANGUAGE + ".1best_1-0-0-100_" + heuristic, new GroundTruth(Metric.AP, numTopics, Onebest_AP.get(heuristic), expectedMAPs.get(heuristic)[1]));
      g.put("en-" + LANGUAGE + ".10best_10-100-0-100_" + heuristic, new GroundTruth(Metric.AP, numTopics, Nbest_AP.get(heuristic), expectedMAPs.get(heuristic)[2]));
      g.put("en-" + LANGUAGE + ".interp_10-30-30-100_" + heuristic, new GroundTruth(Metric.AP, numTopics, Interp_AP.get(heuristic), expectedMAPs.get(heuristic)[3]));
    }

    for (String model : models) {
      System.err.println("Verifying results of model \"" + model + "\"");

      GroundTruth groundTruth = g.get(model);
      Map<String, Accumulator[]> result = results.get(model);
      groundTruth.verify(result, mapping, qrels);

      System.err.println("Done!");
    }
  }
 
  public static junit.framework.Test suite() {
    return new JUnit4TestAdapter(EnAr_TREC02.class);
  }

  public static void main(String[] args) {
    //    HMapSFW gridAPMap = array2Map(Interp_AP);
    HMapSFW tenbestAPMap = array2Map(Nbest_AP.get(2));
    HMapSFW onebestAPMap = array2Map(Onebest_AP.get(1));
    HMapSFW grammarAPMap = array2Map(grammar_AP.get(0));
    HMapSFW tokenAPMap = array2Map(baseline_token_AP);
    //    System.out.println(countNumberOfImprovedTopics(tokenAPMap, gridAPMap));
    System.out.println(countNumberOfImprovedTopics(tokenAPMap, tenbestAPMap));
    System.out.println(countNumberOfImprovedTopics(tokenAPMap, onebestAPMap));
    System.out.println(countNumberOfImprovedTopics(tokenAPMap, grammarAPMap));
    System.out.println(countNumberOfImprovedTopics(tokenAPMap, tokenAPMap));
    //    System.out.println(countNumberOfNegligibleTopics(tokenAPMap, gridAPMap));
    System.out.println(countNumberOfNegligibleTopics(tokenAPMap, tenbestAPMap));
    System.out.println(countNumberOfNegligibleTopics(tokenAPMap, onebestAPMap));
    System.out.println(countNumberOfNegligibleTopics(tokenAPMap, grammarAPMap));
    System.out.println(countNumberOfNegligibleTopics(tokenAPMap, tokenAPMap));
  }

  private static int countNumberOfImprovedTopics(HMapSFW tokenAPMap, HMapSFW gridAPMap) {
    int cnt = 0;
    for (String key : tokenAPMap.keySet()) {
      float difference = gridAPMap.get(key) - tokenAPMap.get(key);
      if ( difference > 0.001 ) {
        cnt++;
      }
    }
    return cnt;
  }

  private static int countNumberOfNegligibleTopics(HMapSFW tokenAPMap, HMapSFW gridAPMap) {
    int cnt = 0;
    for (String key : tokenAPMap.keySet()) {
      float difference = gridAPMap.get(key) - tokenAPMap.get(key);
      if ( difference > -0.001 && difference < 0.001 ) {
        cnt++;
      }
    }
    return cnt;
  }

  private static HMapSFW array2Map(String[] array) {
    HMapSFW map = new HMapSFW();
    for ( int i = 0; i < array.length; i += 2 ) {
      map.put(array[i], Float.parseFloat(array[i+1]));
    }
    return map;
  }

}
TOP

Related Classes of ivory.regression.sigir2013.moses.EnAr_TREC02

  • edu.umd.cloud9.io.map.HMapSFW
  • ivory.core.eval.Qrels
  • ivory.regression.GroundTruth
  • ivory.sqe.retrieval.QueryEngine
  • junit.framework.JUnit4TestAdapter
  • org.apache.hadoop.conf.Configuration
  • org.apache.hadoop.fs.FileSystem
    • TOP
    Copyright © 2018 www.massapi.com. 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.