Package edu.stanford.nlp.parser.lexparser

Examples of edu.stanford.nlp.parser.lexparser.TreebankLangParserParams

31
32
33
34
35
36
37
38
  final TreeBinarizer binarizer;
  final String tlppClass;

  public BinarizerAnnotator(String annotatorName, Properties props) {
    this.tlppClass = props.getProperty(annotatorName + ".tlppClass", DEFAULT_TLPP_CLASS);
    TreebankLangParserParams tlpp = ReflectionLoading.loadByReflection(tlppClass);
    this.binarizer = TreeBinarizer.simpleTreeBinarizer(tlpp.headFinder(), tlpp.treebankLanguagePack());
  }
View Full Code Here
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
      System.exit(-1);
    }
    File goldFile = new File(parsedArgs[0]);
    File guessFile = new File(parsedArgs[1]);
   
    final TreebankLangParserParams tlpp = Languages.getLanguageParams(LANGUAGE);
    final PrintWriter pwOut = tlpp.pw();

    final Treebank guessTreebank = tlpp.diskTreebank();
    guessTreebank.loadPath(guessFile);
    pwOut.println("GUESS TREEBANK:");
    pwOut.println(guessTreebank.textualSummary());

    final Treebank goldTreebank = tlpp.diskTreebank();
    goldTreebank.loadPath(goldFile);
    pwOut.println("GOLD TREEBANK:");
    pwOut.println(goldTreebank.textualSummary());

    final CollinsDepEval depEval = new CollinsDepEval("CollinsDep", true, tlpp.headFinder(), tlpp.treebankLanguagePack().startSymbol());

    final TreeTransformer tc = tlpp.collinizer();

    //PennTreeReader skips over null/malformed parses. So when the yields of the gold/guess trees
    //don't match, we need to keep looking for the next gold tree that matches.
    //The evalb ref implementation differs slightly as it expects one tree per line. It assigns
    //status as follows:
View Full Code Here
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
    if(args.length < minArgs) {
      System.out.println(usage.toString());
      System.exit(-1);
    }

    TreebankLangParserParams tlpp = new EnglishTreebankParserParams();
    int maxGoldYield = Integer.MAX_VALUE;
    int maxGuessYield = Integer.MAX_VALUE;
    boolean VERBOSE = false;
    boolean skipGuess = false;
    boolean tagMode = false;
    String guessFile = null;
    String goldFile = null;

    for(int i = 0; i < args.length; i++) {

      if(args[i].startsWith("-")) {

        switch (args[i]) {
          case "-l":
            Language lang = Language.valueOf(args[++i].trim());
            tlpp = Languages.getLanguageParams(lang);

            break;
          case "-y":
            maxGoldYield = Integer.parseInt(args[++i].trim());

            break;
          case "-t":
            tagMode = true;

            break;
          case "-v":
            VERBOSE = true;

            break;
          case "-g":
            maxGuessYield = Integer.parseInt(args[++i].trim());
            skipGuess = true;

            break;
          default:
            System.out.println(usage.toString());
            System.exit(-1);
        }

      } else {
        //Required parameters
        goldFile = args[i++];
        guessFile = args[i];
        break;
      }
    }

    final PrintWriter pwOut = tlpp.pw();

    final Treebank guessTreebank = tlpp.diskTreebank();
    guessTreebank.loadPath(guessFile);
    pwOut.println("GUESS TREEBANK:");
    pwOut.println(guessTreebank.textualSummary());

    final Treebank goldTreebank = tlpp.diskTreebank();
    goldTreebank.loadPath(goldFile);
    pwOut.println("GOLD TREEBANK:");
    pwOut.println(goldTreebank.textualSummary());

    final String evalName = (tagMode) ? "TsarfatyTAG" : "TsarfatySEG";
    final TsarfatyEval eval = new TsarfatyEval(evalName, tagMode);

    final TreeTransformer tc = tlpp.collinizer();

    //PennTreeReader skips over null/malformed parses. So when the yields of the gold/guess trees
    //don't match, we need to keep looking for the next gold tree that matches.
    //The evalb ref implementation differs slightly as it expects one tree per line. It assigns
    //status as follows:
View Full Code Here
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
      System.err.printf("Usage: java %s language filename features%n", TreebankFactoredLexiconStats.class.getName());
      System.exit(-1);
    }

    Language language = Language.valueOf(args[0]);
    TreebankLangParserParams tlpp = Languages.getLanguageParams(language);
    if (language.equals(Language.Arabic)) {
      String[] options = {"-arabicFactored"};
      tlpp.setOptionFlag(options, 0);
    } else {
      String[] options = {"-frenchFactored"};
      tlpp.setOptionFlag(options, 0);
    }
    Treebank tb = tlpp.diskTreebank();
    tb.loadPath(args[1]);

    MorphoFeatureSpecification morphoSpec = language.equals(Language.Arabic) ?
        new ArabicMorphoFeatureSpecification() : new FrenchMorphoFeatureSpecification();

    String[] features = args[2].trim().split(",");
    for (String feature : features) {
      morphoSpec.activate(MorphoFeatureType.valueOf(feature));
    }

    // Counters
    Counter<String> wordTagCounter = new ClassicCounter<String>(30000);
    Counter<String> morphTagCounter = new ClassicCounter<String>(500);
//    Counter<String> signatureTagCounter = new ClassicCounter<String>();
    Counter<String> morphCounter = new ClassicCounter<String>(500);
    Counter<String> wordCounter = new ClassicCounter<String>(30000);
    Counter<String> tagCounter = new ClassicCounter<String>(300);

    Counter<String> lemmaCounter = new ClassicCounter<String>(25000);
    Counter<String> lemmaTagCounter = new ClassicCounter<String>(25000);

    Counter<String> richTagCounter = new ClassicCounter<String>(1000);

    Counter<String> reducedTagCounter = new ClassicCounter<String>(500);

    Counter<String> reducedTagLemmaCounter = new ClassicCounter<String>(500);

    Map<String,Set<String>> wordLemmaMap = Generics.newHashMap();

    TwoDimensionalIntCounter<String,String> lemmaReducedTagCounter = new TwoDimensionalIntCounter<String,String>(30000);
    TwoDimensionalIntCounter<String,String> reducedTagTagCounter = new TwoDimensionalIntCounter<String,String>(500);
    TwoDimensionalIntCounter<String,String> tagReducedTagCounter = new TwoDimensionalIntCounter<String,String>(300);

    int numTrees = 0;
    for (Tree tree : tb) {
      for (Tree subTree : tree) {
        if (!subTree.isLeaf()) {
          tlpp.transformTree(subTree, tree);
        }
      }
      List<Label> pretermList = tree.preTerminalYield();
      List<Label> yield = tree.yield();
      assert yield.size() == pretermList.size();
View Full Code Here
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
    if(args.length < minArgs) {
      System.out.println(usage.toString());
      System.exit(-1);
    }

    TreebankLangParserParams tlpp = new EnglishTreebankParserParams();
    DiskTreebank tb = null;
    String encoding = "UTF-8";
    Language lang = Language.English;

    for(int i = 0; i < args.length; i++) {
      if(args[i].startsWith("-")) {
        switch (args[i]) {
          case "-l":
            lang = Language.valueOf(args[++i].trim());
            tlpp = Languages.getLanguageParams(lang);

            break;
          case "-e":
            encoding = args[++i];

            break;
          default:
            System.out.println(usage.toString());
            System.exit(-1);
        }

      } else {
        if(tb == null) {
          if(tlpp == null) {
            System.out.println(usage.toString());
            System.exit(-1);
          } else {
            tlpp.setInputEncoding(encoding);
            tlpp.setOutputEncoding(encoding);
            tb = tlpp.diskTreebank();
          }
        }
        tb.loadPath(args[i]);
      }
    }

    PrintWriter pw = tlpp.pw();
    Options op = new Options();
    Options.LexOptions lexOptions = op.lexOptions;
    if(lang == Language.French) {
      lexOptions.useUnknownWordSignatures = 1;
      lexOptions.smartMutation = false;
      lexOptions.unknownSuffixSize = 2;
      lexOptions.unknownPrefixSize = 1;
    } else if(lang == Language.Arabic) {
      lexOptions.smartMutation = false;
      lexOptions.useUnknownWordSignatures = 9;
      lexOptions.unknownPrefixSize = 1;
      lexOptions.unknownSuffixSize = 1;
    }
    Index<String> wordIndex = new HashIndex<String>();
    Index<String> tagIndex = new HashIndex<String>();
    Lexicon lex = tlpp.lex(op, wordIndex, tagIndex);
   
    int computeAfter = (int) (0.50 * tb.size());
    Counter<String> vocab = new ClassicCounter<String>();
    Counter<String> unkCounter = new ClassicCounter<String>();
    int treeId = 0;
View Full Code Here
38
39
40
41
42
43
44
45
46
47
48
  public static TreebankLangParserParams getLanguageParams(String lang) {
    return getLanguageParams(Language.valueOf(lang));
  }

  public static TreebankLangParserParams getLanguageParams(Language lang) {
    TreebankLangParserParams tlpp; // initialized below
    switch(lang) {
    case Arabic:
      tlpp = new ArabicTreebankParserParams();
      break;
  
View Full Code Here
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
    boolean makeCopulaHead = props.getProperty("makeCopulaHead") != null;

    // TODO: if a parser is specified, load this from the parser
    // instead of ever loading it from this way
    String tLPP = props.getProperty("tLPP", "edu.stanford.nlp.parser.lexparser.EnglishTreebankParserParams");
    TreebankLangParserParams params = ReflectionLoading.loadByReflection(tLPP);
    if (makeCopulaHead) {
      // TODO: generalize and allow for more options
      String[] options = { "-makeCopulaHead" };
      params.setOptionFlag(options, 0);
    }

    if (sentFileName == null && (altDepReaderName == null || altDepReaderFilename == null) && treeFileName == null && conllXFileName == null && filter == null) {
      try {
        System.err.println("Usage: java GrammaticalStructure [options]* [-sentFile|-treeFile|-conllxFile file] [-testGraph]");
        System.err.println("  options: -basic, -collapsed, -CCprocessed [the default], -collapsedTree, -parseTree, -test, -parserFile file, -conllx, -keepPunct, -altprinter -altreader -altreaderfile");
        TreeReader tr = new PennTreeReader(new StringReader("((S (NP (NNP Sam)) (VP (VBD died) (NP-TMP (NN today)))))"));
        tb.add(tr.readTree());
      } catch (Exception e) {
        System.err.println("Horrible error: " + e);
        e.printStackTrace();
      }
    } else if (altDepReaderName != null && altDepReaderFilename != null) {
      DependencyReader altDepReader = loadAlternateDependencyReader(altDepReaderName);
      try {
        gsBank = altDepReader.readDependencies(altDepReaderFilename);
      } catch (IOException e) {
        System.err.println("Error reading " + altDepReaderFilename);
        return;
      }
    } else if (treeFileName != null) {
      tb.loadPath(treeFileName);
    } else if (filter != null) {
      tb.load(new BufferedReader(new InputStreamReader(System.in)));
    } else if (conllXFileName != null) {
      try {
        gsBank = params.readGrammaticalStructureFromFile(conllXFileName);
      } catch (RuntimeIOException e) {
        System.err.println("Error reading " + conllXFileName);
        return;
      }
    } else {
View Full Code Here
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
    if (fileName == null || fileName.equals("")) {
      System.out.println(usage);
      System.exit(-1);
    }
    Language language = PropertiesUtils.get(options, "l", Language.English, Language.class);
    TreebankLangParserParams tlpp = Languages.getLanguageParams(language);
    String encoding = options.getProperty("e", "UTF-8");
    tlpp.setInputEncoding(encoding);
    tlpp.setOutputEncoding(encoding);

    DiskTreebank tb = tlpp.diskTreebank();
    tb.loadPath(fileName);

    // Statistics
    Counter<String> binaryRuleTypes = new ClassicCounter<String>(20000);
    List<Integer> branchingFactors = new ArrayList<Integer>(20000);
    int nTrees = 0;
    int nUnaryRules = 0;
    int nBinaryRules = 0;
    int binaryBranchingFactors = 0;

    // Read the treebank
    PrintWriter pw = tlpp.pw();
    for (Tree tree : tb) {
      if (tree.value().equals("ROOT")) {
        tree = tree.firstChild();
      }
      ++nTrees;
View Full Code Here
29
30
31
32
33
34
35
36
37
38
39
40
    // the intermediate GrammaticalStructure instead
    SemanticGraph graph = SemanticGraphFactory.generateUncollapsedDependencies(tree);

    // Alternatively, this could have been the Chinese params or any
    // other language supported.  As of 2014, only English and Chinese
    TreebankLangParserParams params = new EnglishTreebankParserParams();
    GrammaticalStructureFactory gsf = params.treebankLanguagePack().grammaticalStructureFactory(params.treebankLanguagePack().punctuationWordRejectFilter(), params.typedDependencyHeadFinder());

    GrammaticalStructure gs = gsf.newGrammaticalStructure(tree);

    System.err.println(graph);
View Full Code Here
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
    if(!validateCommandLine(args)) {
      System.err.println(USAGE);
      System.exit(-1);
    }

    final TreebankLangParserParams tlpp = Languages.getLanguageParams(LANGUAGE);
    final PrintWriter pwOut = tlpp.pw();

    final Treebank guessTreebank = tlpp.diskTreebank();
    guessTreebank.loadPath(guessFile);
    pwOut.println("GUESS TREEBANK:");
    pwOut.println(guessTreebank.textualSummary());

    final Treebank goldTreebank = tlpp.diskTreebank();
    goldTreebank.loadPath(goldFile);
    pwOut.println("GOLD TREEBANK:");
    pwOut.println(goldTreebank.textualSummary());

    final LeafAncestorEval metric = new LeafAncestorEval("LeafAncestor");

    final TreeTransformer tc = tlpp.collinizer();

    //The evalb ref implementation assigns status for each tree pair as follows:
    //
    //   0 - Ok (yields match)
    //   1 - length mismatch
View Full Code Here
TOP

Related Classes of edu.stanford.nlp.parser.lexparser.TreebankLangParserParams

  • edu.stanford.nlp.international.Languages
  • edu.stanford.nlp.parser.eval.TreebankFactoredLexiconStats
  • edu.stanford.nlp.parser.eval.TreebankStats
  • edu.stanford.nlp.parser.eval.UNKPrinter
  • edu.stanford.nlp.parser.metrics.CollinsDepEval
  • edu.stanford.nlp.parser.metrics.Evalb
  • edu.stanford.nlp.parser.metrics.LeafAncestorEval
  • edu.stanford.nlp.parser.metrics.TaggingEval
  • edu.stanford.nlp.parser.metrics.TsarfatyEval
  • edu.stanford.nlp.parser.metrics.UnlabeledAttachmentEval
    • 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.