Welcome to mirror list, hosted at ThFree Co, Russian Federation.

DecodeStepTranslation.cpp « moses - github.com/moses-smt/mosesdecoder.git - Unnamed repository; edit this file 'description' to name the repository.
summaryrefslogtreecommitdiff
blob: c4a4fd0c46daaa985888ecd5dfe179df4e724c9c (plain)
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
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
// $Id$

/***********************************************************************
Moses - factored phrase-based language decoder
Copyright (C) 2006 University of Edinburgh

This library is free software; you can redistribute it and/or
modify it under the terms of the GNU Lesser General Public
License as published by the Free Software Foundation; either
version 2.1 of the License, or (at your option) any later version.

This library is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
Lesser General Public License for more details.

You should have received a copy of the GNU Lesser General Public
License along with this library; if not, write to the Free Software
Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA
***********************************************************************/

#include "DecodeStepTranslation.h"
#include "TranslationOption.h"
#include "TranslationOptionCollection.h"
#include "PartialTranslOptColl.h"
#include "FactorCollection.h"
#include "util/exception.hh"

using namespace std;

namespace Moses
{
DecodeStepTranslation::DecodeStepTranslation(PhraseDictionary* pdf,
    const DecodeStep* prev,
    const std::vector<FeatureFunction*> &features)
  : DecodeStep(pdf, prev, features)
{
  // don't apply feature functions that are from current phrase table.It should already have been
  // dont by the phrase table.
  const std::vector<FeatureFunction*> &pdfFeatures = pdf->GetFeaturesToApply();
  for (size_t i = 0; i < pdfFeatures.size(); ++i) {
    FeatureFunction *ff = pdfFeatures[i];
    RemoveFeature(ff);
  }
}

void DecodeStepTranslation::Process(const TranslationOption &inputPartialTranslOpt
                                    , const DecodeStep &decodeStep
                                    , PartialTranslOptColl &outputPartialTranslOptColl
                                    , TranslationOptionCollection *toc
                                    , bool adhereTableLimit
                                    , TargetPhraseCollection::shared_ptr phraseColl) const
{
  if (inputPartialTranslOpt.GetTargetPhrase().GetSize() == 0) {
    // word deletion
    outputPartialTranslOptColl.Add(new TranslationOption(inputPartialTranslOpt));
    return;
  }

  // normal trans step
  const Range &sourceWordsRange        = inputPartialTranslOpt.GetSourceWordsRange();
  const InputPath &inputPath = inputPartialTranslOpt.GetInputPath();
  const PhraseDictionary* phraseDictionary  =
    decodeStep.GetPhraseDictionaryFeature();
  const TargetPhrase &inPhrase = inputPartialTranslOpt.GetTargetPhrase();
  const size_t currSize = inPhrase.GetSize();
  const size_t tableLimit = phraseDictionary->GetTableLimit();

  if (phraseColl != NULL) {
    TargetPhraseCollection::const_iterator iterTargetPhrase, iterEnd;
    iterEnd = (!adhereTableLimit || tableLimit == 0 || phraseColl->GetSize() < tableLimit) ? phraseColl->end() : phraseColl->begin() + tableLimit;

    for (iterTargetPhrase = phraseColl->begin(); iterTargetPhrase != iterEnd; ++iterTargetPhrase) {
      const TargetPhrase& targetPhrase = **iterTargetPhrase;
      // const ScoreComponentCollection &transScores = targetPhrase.GetScoreBreakdown();
      // skip if the
      if (targetPhrase.GetSize() != currSize) continue;

      TargetPhrase outPhrase(inPhrase);

      if (IsFilteringStep()) {
        if (!inputPartialTranslOpt.IsCompatible(targetPhrase, m_conflictFactors))
          continue;
      }

      outPhrase.Merge(targetPhrase, m_newOutputFactors);
      outPhrase.EvaluateInIsolation(inputPath.GetPhrase(), m_featuresToApply); // need to do this as all non-transcores would be screwed up

      TranslationOption *newTransOpt = new TranslationOption(sourceWordsRange, outPhrase);
      assert(newTransOpt != NULL);

      newTransOpt->SetInputPath(inputPath);

      outputPartialTranslOptColl.Add(newTransOpt );

    }
  } else if (sourceWordsRange.GetNumWordsCovered() == 1) {
    // unknown handler
    //toc->ProcessUnknownWord(sourceWordsRange.GetStartPos(), factorCollection);
  }
}

void
DecodeStepTranslation::
ProcessInitialTranslation(InputType const& source,
                          PartialTranslOptColl &outputPartialTranslOptColl,
                          size_t startPos, size_t endPos,
                          bool adhereTableLimit,
                          InputPath const& inputPath,
                          TargetPhraseCollection::shared_ptr phraseColl) const
{
  const PhraseDictionary* phraseDictionary = GetPhraseDictionaryFeature();
  const size_t tableLimit = phraseDictionary->GetTableLimit();

  const Range range(startPos, endPos);

  if (phraseColl != NULL) {
    IFVERBOSE(3) {
      if(source.GetType() == SentenceInput)
        TRACE_ERR("[" << source.GetSubString(range) << "; "
                  << startPos << "-" << endPos << "]\n");
      else
        TRACE_ERR("[" << startPos << "-" << endPos << "]" << std::endl);
    }

    TargetPhraseCollection::const_iterator iterTargetPhrase, iterEnd;
    iterEnd = (!adhereTableLimit || tableLimit == 0 || phraseColl->GetSize() < tableLimit) ? phraseColl->end() : phraseColl->begin() + tableLimit;

    for (iterTargetPhrase = phraseColl->begin() ; iterTargetPhrase != iterEnd ; ++iterTargetPhrase) {
      const TargetPhrase	&targetPhrase = **iterTargetPhrase;
      TranslationOption *transOpt = new TranslationOption(range, targetPhrase);

      transOpt->SetInputPath(inputPath);

      outputPartialTranslOptColl.Add (transOpt);

      VERBOSE(3,"\t" << targetPhrase << "\n");
    }
    VERBOSE(3,std::endl);
  }
}

void
DecodeStepTranslation::
ProcessInitialTransLEGACY(InputType const& source,
                          PartialTranslOptColl &outputPartialTranslOptColl,
                          size_t startPos, size_t endPos,
                          bool adhereTableLimit,
                          InputPathList const& inputPathList) const
{
  const PhraseDictionary* phraseDictionary = GetPhraseDictionaryFeature();
  const size_t tableLimit = phraseDictionary->GetTableLimit();

  const Range range(startPos, endPos);
  TargetPhraseCollectionWithSourcePhrase::shared_ptr phraseColl
  = phraseDictionary->GetTargetPhraseCollectionLEGACY(source,range);

  if (phraseColl != NULL) {
    IFVERBOSE(3) {
      if(source.GetType() == SentenceInput)
        TRACE_ERR("[" << source.GetSubString(range) << "; "
                  << startPos << "-" << endPos << "]\n");
      else
        TRACE_ERR("[" << startPos << "-" << endPos << "]" << std::endl);
    }

    const std::vector<Phrase> &sourcePhrases = phraseColl->GetSourcePhrases();

    TargetPhraseCollection::const_iterator iterTargetPhrase, iterEnd;
    std::vector<Phrase>::const_iterator iterSourcePhrase;
    iterEnd = (!adhereTableLimit || tableLimit == 0 || phraseColl->GetSize() < tableLimit) ? phraseColl->end() : phraseColl->begin() + tableLimit;

    for (iterTargetPhrase = phraseColl->begin(), iterSourcePhrase =  sourcePhrases.begin()
         ; iterTargetPhrase != iterEnd
         ; ++iterTargetPhrase, ++iterSourcePhrase) {
      assert(iterSourcePhrase != sourcePhrases.end());

      const TargetPhrase	&targetPhrase = **iterTargetPhrase;
      const Phrase			&sourcePhrase = *iterSourcePhrase;

      const InputPath &inputPath = GetInputPathLEGACY(targetPhrase, sourcePhrase, inputPathList);

      TranslationOption *transOpt = new TranslationOption(range, targetPhrase);
      transOpt->SetInputPath(inputPath);

      outputPartialTranslOptColl.Add (transOpt);

      VERBOSE(3,"\t" << targetPhrase << "\n");
    }
    VERBOSE(3,std::endl);
  }
}

const InputPath &DecodeStepTranslation::GetInputPathLEGACY(
  const TargetPhrase targetPhrase,
  const Phrase sourcePhrase,
  const InputPathList &inputPathList) const
{
  const Word &wordFromPt =  sourcePhrase.GetWord(0);

  InputPathList::const_iterator iter;
  for (iter = inputPathList.begin(); iter != inputPathList.end(); ++iter) {
    const InputPath &inputPath = **iter;
    const Phrase &phraseFromIP = inputPath.GetPhrase();

    const Word *wordIP = NULL;
    for (size_t i = 0; i < phraseFromIP.GetSize(); ++i) {
      const Word &tempWord =  phraseFromIP.GetWord(i);
      if (!tempWord.IsEpsilon()) {
        wordIP = &tempWord;
        break;
      }
    }

    // const Range &range = inputPath.GetWordsRange();

    if (wordIP && *wordIP == wordFromPt) {
      return inputPath;
    }
  }

  UTIL_THROW(util::Exception, "Input path not found");
}

void
DecodeStepTranslation::
ProcessLEGACY(TranslationOption const& in,
              DecodeStep const& decodeStep,
              PartialTranslOptColl &out,
              TranslationOptionCollection *toc,
              bool adhereTableLimit) const
{
  if (in.GetTargetPhrase().GetSize() == 0) {
    // word deletion
    out.Add(new TranslationOption(in));
    return;
  }

  // normal trans step
  Range const& srcRange = in.GetSourceWordsRange();
  InputPath const& inputPath = in.GetInputPath();
  PhraseDictionary const* pdict  = decodeStep.GetPhraseDictionaryFeature();
  TargetPhrase const& inPhrase = in.GetTargetPhrase();
  size_t const currSize = inPhrase.GetSize();
  size_t const tableLimit = pdict->GetTableLimit();

  TargetPhraseCollectionWithSourcePhrase::shared_ptr phraseColl
  = pdict->GetTargetPhraseCollectionLEGACY(toc->GetSource(),srcRange);

  if (phraseColl != NULL) {
    TargetPhraseCollection::const_iterator iterTargetPhrase, iterEnd;
    iterEnd = ((adhereTableLimit && tableLimit && phraseColl->GetSize() >= tableLimit)
               ? phraseColl->begin() + tableLimit : phraseColl->end());

    for (iterTargetPhrase = phraseColl->begin();
         iterTargetPhrase != iterEnd;
         ++iterTargetPhrase) {
      TargetPhrase const& targetPhrase = **iterTargetPhrase;
      if (targetPhrase.GetSize() != currSize ||
          (IsFilteringStep() && !in.IsCompatible(targetPhrase, m_conflictFactors)))
        continue;

      TargetPhrase outPhrase(inPhrase);
      outPhrase.Merge(targetPhrase, m_newOutputFactors);
      outPhrase.EvaluateInIsolation(inputPath.GetPhrase(), m_featuresToApply); // need to do this as all non-transcores would be screwed up

      TranslationOption *newTransOpt = new TranslationOption(srcRange, outPhrase);
      assert(newTransOpt != NULL);

      newTransOpt->SetInputPath(inputPath);

      out.Add(newTransOpt);

    }
  }
}
}