Reranker.java
25.9 KB
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
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
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
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
package is2.parserR2;
import java.io.BufferedInputStream;
import java.io.BufferedOutputStream;
import java.io.DataInputStream;
import java.io.DataOutputStream;
import java.io.File;
import java.io.FileInputStream;
import java.io.FileNotFoundException;
import java.io.FileOutputStream;
import java.io.IOException;
import java.text.SimpleDateFormat;
import java.util.ArrayList;
import java.util.Date;
import java.util.HashMap;
import java.util.List;
import java.util.Map.Entry;
import java.util.zip.ZipEntry;
import java.util.zip.ZipInputStream;
import java.util.zip.ZipOutputStream;
import extractors.Extractor;
import extractors.ExtractorClusterStacked;
import extractors.ExtractorReranker;
import is2.data.Cluster;
import is2.data.DataF;
import is2.data.Edges;
import is2.data.F2SF;
import is2.data.FVR;
import is2.data.Instances;
import is2.data.Long2Int;
import is2.data.Long2IntInterface;
import is2.data.MFB;
import is2.data.Parse;
import is2.data.ParseNBest;
import is2.data.PipeGen;
import is2.data.SentenceData09;
import is2.io.CONLLReader09;
import is2.io.CONLLWriter09;
import is2.tools.Tool;
import is2.util.DB;
import is2.util.OptionsSuper;
import is2.util.ParserEvaluator;
public class Reranker implements Tool {
public static int THREADS = 4;
Long2IntInterface l2i;
// the parser models
ParametersFloat paramsParsers[];
// the re-ranker model
ParametersFloat parametersReranker;
PipeReranker pipeReranker;
Pipe pipe;
Options options;
HashMap<Integer, Integer> rank = new HashMap<Integer, Integer>();
int amongxbest = 0, amongxbest_ula = 0, nbest = 0, bestProj = 0, smallestErrorSum = 0, countAllNodes = 0;
static int NBest = 1000;
/**
* Initialize the parser
*
* @param options
*/
public Reranker(Options options) {
this.options = options;
}
/**
* @param modelFileName
* The file name of the parsing model
*/
public Reranker(String modelFileName) {
this(new Options(new String[] { "-model", modelFileName }));
}
public Reranker() {
// TODO Auto-generated constructor stub
}
public static void main(String[] args) throws Exception {
long start = System.currentTimeMillis();
Options options = new Options(args);
NBest = options.best;
DB.println("n-best " + NBest);
Runtime runtime = Runtime.getRuntime();
THREADS = runtime.availableProcessors();
if (options.cores < THREADS && options.cores > 0)
THREADS = options.cores;
DB.println("Found " + runtime.availableProcessors() + " cores use " + THREADS);
if (options.train) {
Reranker p = new Reranker();
p.options = options;
p.l2i = new Long2Int(options.hsize);
p.pipeReranker = new PipeReranker(options);
p.pipeReranker.extractor = new ExtractorReranker(p.l2i);
// initialize the parser
p.pipe = new Pipe(options);
// read parsing models
p.paramsParsers = new ParametersFloat[options.end + 1];
for (int m = 0; m <= options.end; m++) {
String name = options.prefix_model + m;
p.paramsParsers[m] = new ParametersFloat(0);
p.readModel(name, p.pipe, p.paramsParsers[m]);
}
// set up the reranker
p.parametersReranker = new ParametersFloat(p.l2i.size());
Instances[] iss = new Instances[options.end + 1];
for (int m = 0; m <= options.end; m++) {
String name = options.prefix_test + m;
iss[m] = new Instances();
DB.println("create instances of part " + name);
p.pipe.getInstances(name, iss[m]);
}
ExtractorReranker.initFeatures();
p.pipeReranker.extractor.init();
ExtractorReranker.initStat();
p.train(options, iss);
p.writeModell(options, p.parametersReranker, null, p.pipe.cl);
}
if (options.test) {
Reranker p = new Reranker();
p.options = options;
// set up the reranker
p.l2i = new Long2Int(options.hsize);
p.pipeReranker = new PipeReranker(options);
p.pipeReranker.extractor = new ExtractorReranker(p.l2i);
p.parametersReranker = new ParametersFloat(p.l2i.size());
// initialize the parser
p.pipe = new Pipe(options);
// read parsing models
p.paramsParsers = new ParametersFloat[options.end + 1];
String nbestName = "n-best+"
+ options.testfile.substring(options.testfile.length() - 12, options.testfile.length() - 1);
File fnbest = new File(nbestName);
int read = fnbest.exists() ? 2 : 1;
if (read != 2)
for (int m = 0; m <= options.end; m++) {
String name = options.prefix_model + m;
p.paramsParsers[m] = new ParametersFloat(0);
p.readModel(name, p.pipe, p.paramsParsers[m]);
}
p.readModel(options.modelName, p.pipeReranker, p.parametersReranker);
ExtractorReranker.initFeatures();
ExtractorReranker.initStat();
p.pipeReranker.extractor.init();
p.rerankedParses(options, p.pipe, p.parametersReranker, false, nbestName);
}
System.out.println();
if (options.eval) {
System.out.println("\nEVALUATION PERFORMANCE:");
ParserEvaluator.evaluate(options.goldfile, options.outfile);
}
long end = System.currentTimeMillis();
System.out.println("used time " + ((float) ((end - start) / 100) / 10));
Decoder.executerService.shutdown();
Pipe.executerService.shutdown();
System.out.println("end.");
}
/**
* Read the models and mapping
*
* @param options
* @param pipe
* @param prm
* @throws IOException
*/
public void readModel(String modelName, Pipe pipe, Parameters prm) throws IOException {
DB.println("Reading data started: " + modelName);
// prepare zipped reader
ZipInputStream zis = new ZipInputStream(new BufferedInputStream(new FileInputStream(modelName)));
zis.getNextEntry();
DataInputStream dis = new DataInputStream(new BufferedInputStream(zis));
pipe.mf.read(dis);
pipe.cl = new Cluster(dis);
prm.read(dis);
Long2Int l2i = new Long2Int(prm.size());
DB.println("li size " + l2i.size());
pipe.extractor = new ExtractorClusterStacked[THREADS];
for (int t = 0; t < THREADS; t++)
pipe.extractor[t] = new ExtractorClusterStacked(l2i);
ExtractorClusterStacked.initFeatures();
for (int t = 0; t < THREADS; t++) {
pipe.extractor[t].initStat();
pipe.extractor[t].init();
}
Edges.read(dis);
options.decodeProjective = dis.readBoolean();
ExtractorClusterStacked.maxForm = dis.readInt();
boolean foundInfo = false;
try {
String info = null;
int icnt = dis.readInt();
for (int i = 0; i < icnt; i++) {
info = dis.readUTF();
System.out.println(info);
}
} catch (Exception e) {
if (!foundInfo)
System.out.println("no info about training");
}
dis.close();
DB.println("Reading data finnished");
Decoder.NON_PROJECTIVITY_THRESHOLD = (float) options.decodeTH;
// ExtractorClusterStacked.initStat();
}
/**
* Read the models and mapping
*
* @param options
* @param pipe
* @param params
* @throws IOException
*/
public void readModel(String modelName, PipeReranker pipe, Parameters params) throws IOException {
DB.println("Reading data started: " + modelName);
// prepare zipped reader
ZipInputStream zis = new ZipInputStream(new BufferedInputStream(new FileInputStream(modelName)));
zis.getNextEntry();
DataInputStream dis = new DataInputStream(new BufferedInputStream(zis));
pipe.mf.read(dis);
// DB.println("reranker model "+pipe.mf.toString());
pipe.cl = new Cluster(dis);
params.read(dis);
this.l2i = new Long2Int(params.size());
DB.println("li size " + l2i.size());
pipe.extractor = new ExtractorReranker(l2i);
ExtractorReranker.initFeatures();
ExtractorReranker.initStat();
pipe.extractor.init();
Edges.read(dis);
options.decodeProjective = dis.readBoolean();
ExtractorClusterStacked.maxForm = dis.readInt();
boolean foundInfo = false;
try {
String info = null;
int icnt = dis.readInt();
for (int i = 0; i < icnt; i++) {
info = dis.readUTF();
System.out.println(info);
}
} catch (Exception e) {
if (!foundInfo)
System.out.println("no info about training");
}
dis.close();
DB.println("Reading data finnished");
Decoder.NON_PROJECTIVITY_THRESHOLD = (float) options.decodeTH;
// ExtractorClusterStacked.initStat();
}
/**
* Do the training
*
* @param instanceLengths
* @param options
* @param pipe
* @param parametersReranker
* @param is
* @param cluster
* @throws IOException
* @throws InterruptedException
* @throws ClassNotFoundException
*/
public void train(OptionsSuper options, Instances[] iss)
throws IOException, InterruptedException, ClassNotFoundException {
int read = 0; // 0 compute -- 1 compute and write -- 2 read parses
DB.println("Training Information ");
DB.println("-------------------- ");
ExtractorReranker.initStat();
pipeReranker.extractor.init();
for (Extractor e : this.pipe.extractor) {
e.init();
}
int numInstances = 0;
int maxLenInstances = 0;
// int maxLenSentence=1;
for (Instances is : iss) {
numInstances += is.size();
for (int i = 0; i < is.size(); i++)
if (maxLenInstances < is.length(i))
maxLenInstances = is.length(i);
}
DataF data = new DataF(maxLenInstances, pipe.mf.getFeatureCounter().get(PipeGen.REL).shortValue());
int iter = 0;
int del = 0;
float error = 0;
float f1 = 0;
double upd = (double) (options.count * options.numIters) + options.numIters * 10;
// float[][] = new float[this.NBest][3];
FVR act = new FVR();
FVR pred = new FVR();
FVR f = new FVR();
long[] vs = new long[ExtractorReranker._FC * maxLenInstances];
for (; iter < options.numIters; iter++) {
System.out.print("Iteration " + iter + ": ");
error = 0;
f1 = 0;
float las = 0, cnt = 0, averageScore = 0;
float firstBestTotalError = 0, totalError = 0;
long start = System.currentTimeMillis();
long last = System.currentTimeMillis();
long rerankTime = 0;
String nbest = "n-best";
File fnbest = new File(nbest);
read = fnbest.exists() ? 2 : 1;
DataInputStream dis = null;
DataOutputStream dos = null;
if (read == 1) {
DB.println("computing and writting nbest list to file: " + nbest);
ZipOutputStream zos = new ZipOutputStream(new BufferedOutputStream(new FileOutputStream(nbest)));
zos.putNextEntry(new ZipEntry("data"));
dos = new DataOutputStream(new BufferedOutputStream(zos));
}
// start reading again
if (read == 2) {
// DB.println("reading nbest list from file: "+nbest);
ZipInputStream zis = new ZipInputStream(new BufferedInputStream(new FileInputStream(nbest)));
zis.getNextEntry();
dis = new DataInputStream(new BufferedInputStream(zis));
}
HashMap<Integer, Integer> remapped = new HashMap<Integer, Integer>();
int round = 0, instance = 0, length = 0, count = 0, changes = 0;
for (Instances is : iss) {
F2SF fparser = this.paramsParsers[instance].getFV();
round++;
// go over the sentences in the instance
for (int n = 0; n < is.size(); n++) {
count += 1;
length += is.length(n);
upd--;
if (is.labels[n].length > options.maxLen)
continue;
List<ParseNBest> parses = null;
short pos[] = is.pposs[n];
// read or write nbest list
if (read == 0 || read == 1 && iter == 0) {
data = pipe.fillVector(fparser, is, n, data, pipe.cl, THREADS, l2i);
parses = Decoder.decode(pos, data, options.decodeProjective, pipe.extractor[0]);
if (parses.size() > NBest)
parses = parses.subList(0, NBest);
if (read == 1) {
// write the forest
dos.writeInt(parses.size());
for (int k = 0; k < parses.size(); k++) {
dos.writeUTF(parses.get(k).signature());
dos.writeFloat((float) parses.get(k).f1);
}
}
} else if (read == 2) {
parses = new ArrayList<ParseNBest>();
int parseCount = dis.readInt();
for (int k = 0; k < parseCount; k++) {
ParseNBest p = new ParseNBest(dis.readUTF(), dis.readFloat());
if (parses.size() < NBest)
parses.add(p);
}
}
int best = 0;
float bestScore = -100;
int goldBest = 0;
double goldError = Float.MAX_VALUE;
long startReranking = System.currentTimeMillis();
// score the n-best parses
for (int k = 0; k < parses.size(); k++) {
ParseNBest p = parses.get(k);
pipeReranker.extractor.extractFeatures(is, n, p, parses.indexOf(p), vs, pipe.cl);
int rank = 1 * ExtractorReranker.s_type;
f.clear();
for (long element : vs) {
if (element == Integer.MIN_VALUE)
break;
if (element > 0)
f.add(pipeReranker.extractor.li.l2i(element + rank));
}
f.add(pipeReranker.extractor.li.l2i(1 + rank), (float) p.f1);
float score = (parametersReranker.getScore(f));
if (score > bestScore) { // rankScore[k][2]>
bestScore = score;
best = k;
}
}
// get the best parse in the n-best list
for (int k = 0; k < parses.size(); k++) {
if (parses.get(k).heads.length != is.length(n)) {
DB.println("error " + n + " " + parses.get(k).heads.length + " " + is.length(n));
continue;
}
double errg = pipe.errors(is, n, parses.get(k));
if (goldError > errg) {
goldError = errg;
goldBest = k;
}
}
ParseNBest firstBest = parses.get(0);
ParseNBest predParse = parses.get(best);
ParseNBest goldBestParse = parses.get(goldBest);
double e = pipe.errors(is, n, predParse);
Integer ctb = remapped.get(best);
if (ctb == null)
remapped.put(best, 1);
else
remapped.put(best, ctb + 1);
String info = " 1best-error " + ((length - firstBestTotalError) / length) + " reranked "
+ ((length - totalError) / length) + " chd " + changes + " " + " ps las " + (las / cnt)
+ " avs " + (averageScore / count) + " ";
if ((n + 1) % 500 == 0)
del = PipeGen.outValueErr(count, Math.round(error * 1000) / 1000, f1 / count, del, last, upd,
info);
firstBestTotalError += Decoder.getError(firstBest, is, n, Decoder.LAS);
totalError += Decoder.getError(predParse, is, n, Decoder.LAS);
rerankTime += System.currentTimeMillis() - startReranking;
if (best != 0) {
changes++;
}
las += is.length(n) - Decoder.getError(goldBestParse, is, n, Decoder.LAS);
cnt += is.length(n);
averageScore += predParse.f1;
if (options.count < count)
break;
if (Decoder.getError(goldBestParse, is, n, Decoder.LAS) >= Decoder.getError(predParse, is, n,
Decoder.LAS))
continue;
// get predicted feature vector
pipeReranker.extractor.extractFeatures(is, n, predParse, parses.indexOf(predParse), vs, pipe.cl);
pred.clear();
int rank = 1 * ExtractorReranker.s_type;
for (long element : vs) {
if (element == Integer.MIN_VALUE)
break;
if (element > 0)
pred.add(pipeReranker.extractor.li.l2i(element + rank));
}
pred.add(pipeReranker.extractor.li.l2i(1 + rank), (float) predParse.f1);
error += 1;
pipeReranker.extractor.extractFeatures(is, n, goldBestParse, parses.indexOf(goldBestParse), vs,
pipe.cl);
act.clear();
rank = 1 * ExtractorReranker.s_type;
for (long element : vs) {
if (element == Integer.MIN_VALUE)
break;
if (element > 0)
act.add(pipeReranker.extractor.li.l2i(element + rank));
}
act.add(pipeReranker.extractor.li.l2i(1 + rank), (float) goldBestParse.f1);
float lam_dist = parametersReranker.getScore(act) - (parametersReranker.getScore(pred));
parametersReranker.update(act, pred, is, n, null, upd, e, lam_dist);
}
instance++;
}
String info = " td " + ((Decoder.timeDecotder) / 1000000F) + " tr " + ((Decoder.timeRearrange) / 1000000F)
+ " te " + ((Pipe.timeExtract) / 1000000F) + " nz " + parametersReranker.countNZ() + " 1best-error "
+ ((length - firstBestTotalError) / length) + " reranked-best " + ((length - totalError) / length)
+ " rds " + round + " " + " rerank-t " + (rerankTime / count) + " chd " + changes + " "
+ " ps las " + (las / cnt) + " avs " + (averageScore / count) + " ";
// DB.println("remapped "+remapped);
PipeGen.outValueErr(count, Math.round(error * 1000) / 1000, f1 / count, del, last, upd, info);
del = 0;
long end = System.currentTimeMillis();
System.out.println(" time:" + (end - start));
Decoder.timeDecotder = 0;
Decoder.timeRearrange = 0;
Pipe.timeExtract = 0;
if (dos != null)
dos.close();
if (dis != null)
dis.close();
}
DB.println("sb " + parametersReranker.parameters[this.pipeReranker.extractor.li
.l2i(4090378920L + 1 * ExtractorReranker.s_type)]);// 4090378266
parametersReranker.average(iter * numInstances);
}
/**
* Do the parsing
*
* @param options
* @param pipe
* @param params
* @throws IOException
*/
private void rerankedParses(OptionsSuper options, Pipe pipe, ParametersFloat params, boolean maxInfo,
String nbestName) throws Exception {
long start = System.currentTimeMillis();
ExtractorClusterStacked.initFeatures();
DataInputStream dis = null;
DataOutputStream dos = null;
float olas = 0, olcnt = 0;
File fnbest = new File(nbestName);
int read = fnbest.exists() ? 2 : 1;
if (read == 1) {
DB.println("computing and writting nbest list to file: " + nbestName);
ZipOutputStream zos = new ZipOutputStream(new BufferedOutputStream(new FileOutputStream(nbestName)));
zos.putNextEntry(new ZipEntry("data"));
dos = new DataOutputStream(new BufferedOutputStream(zos));
}
if (read == 2) {
// DB.println("reading nbest list from file: "+nbestName);
ZipInputStream zis = new ZipInputStream(new BufferedInputStream(new FileInputStream(nbestName)));
zis.getNextEntry();
dis = new DataInputStream(new BufferedInputStream(zis));
}
for (ParametersFloat paramsParser : this.paramsParsers) {
CONLLReader09 depReader = new CONLLReader09(options.testfile, options.formatTask);
CONLLWriter09 depWriter = new CONLLWriter09(options.outfile, options.formatTask);
float las = 0, lcnt = 0, averageScore = 0;
int cnt = 0;
int del = 0;
long last = System.currentTimeMillis();
String[] types = new String[pipe.mf.getFeatureCounter().get(PipeGen.REL)];
for (Entry<String, Integer> e : MFB.getFeatureSet().get(PipeGen.REL).entrySet())
types[e.getValue()] = e.getKey();
System.out.print("Processing Sentence: ");
FVR f = new FVR();
while (true) {
Instances is = new Instances();
is.init(1, new MFB(), options.formatTask);
SentenceData09 instance = pipe.nextInstance(is, depReader);
if (instance == null)
break;
cnt++;
String[] forms = instance.forms;
List<ParseNBest> parses = null;
// read or write nbest list
if (read == 0 || read == 1) {
parses = this.parseNBest(instance, paramsParser);
// data = pipe.fillVector(fparser, is, n, data,
// pipe.cl,THREADS);
// parses = Decoder.decode(pos, data,
// options.decodeProjective);
if (parses.size() > NBest)
parses = parses.subList(0, NBest);
if (read == 1) {
// write the forest
dos.writeInt(parses.size());
for (int k = 0; k < parses.size(); k++) {
dos.writeUTF(parses.get(k).signature());
dos.writeFloat((float) parses.get(k).f1);
}
}
} else if (read == 2) {
parses = new ArrayList<ParseNBest>();
int parseCount = dis.readInt();
for (int k = 0; k < parseCount; k++) {
ParseNBest p = new ParseNBest(dis.readUTF(), dis.readFloat());
if (parses.size() < NBest)
parses.add(p);
}
}
nbest += parses.size();
// List<ParseNBest> parses = this.parseNBest(instance,
// this.paramsParsers[m]);
long vs[] = new long[ExtractorReranker._FC * is.length(0)];
float bestScore = 0;
int best = 0;
for (int k = 0; k < parses.size(); k++) {
ParseNBest p = parses.get(k);
pipeReranker.extractor.extractFeatures(is, 0, p, k, vs, pipeReranker.cl);
int rank = 1 * ExtractorReranker.s_type;
f.clear();
for (long element : vs) {
if (element == Integer.MIN_VALUE)
break;
if (element > 0)
f.add(pipeReranker.extractor.li.l2i(element + rank));
}
f.add(pipeReranker.extractor.li.l2i(1 + rank), (float) p.f1);
float score = (parametersReranker.getScore(f));
if (score > bestScore) { // rankScore[k][2]>
bestScore = score;
best = k;
}
}
// change to best
ParseNBest d = parses.get(best);
las += (is.length(0) - 1) - Decoder.getError(d, is, 0, Decoder.LAS);
lcnt += is.length(0) - 1;
averageScore += d.f1;
SentenceData09 i09 = new SentenceData09(instance);
i09.createSemantic(instance);
for (int j = 0; j < forms.length - 1; j++) {
i09.plabels[j] = types[d.labels[j + 1]];
i09.pheads[j] = d.heads[j + 1];
}
depWriter.write(i09);
String info = "" + (averageScore / cnt) + " ";
if (cnt % 10 == 0)
del = PipeGen.outValueErr(cnt, lcnt - las, las / lcnt, del, last, 0, info);// outValue(cnt,
// del,last,
// info);
}
// pipe.close();
depWriter.finishWriting();
long end = System.currentTimeMillis();
DB.println("rank\n" + rank + "\n");
DB.println("x-best-las " + amongxbest + " x-best-ula " + amongxbest_ula + " cnt " + cnt + " x-best-las "
+ ((float) amongxbest / cnt) + " x-best-ula " + ((float) amongxbest_ula / cnt) + " nbest "
+ ((float) nbest / cnt) + " 1best " + ((float) (rank.get(0) == null ? 0 : rank.get(0)) / cnt)
+ " best-proj " + ((float) bestProj / cnt) + " Sum LAS "
+ ((float) this.smallestErrorSum / countAllNodes) + " " + "" + (las / lcnt));
// DB.println("errors "+error);
olas += las;
olcnt += lcnt;
rank.clear();
amongxbest = 0;
amongxbest_ula = 0;
cnt = 0;
nbest = 0;
bestProj = 0;
if (maxInfo)
System.out.println("Used time " + (end - start));
if (maxInfo)
System.out.println("forms count " + Instances.m_count + " unkown " + Instances.m_unkown);
}
if (dos != null) {
dos.flush();
dos.close();
}
if (dis != null)
dis.close();
DB.println("\n overall las " + (olas / olcnt));
}
public SentenceData09 parse(SentenceData09 instance, ParametersFloat params) {
String[] types = new String[pipe.mf.getFeatureCounter().get(PipeGen.REL)];
for (Entry<String, Integer> e : MFB.getFeatureSet().get(PipeGen.REL).entrySet())
types[e.getValue()] = e.getKey();
Instances is = new Instances();
is.init(1, new MFB(), options.formatTask);
new CONLLReader09().insert(is, instance);
String[] forms = instance.forms;
// use for the training ppos
DataF d2;
try {
d2 = pipe.fillVector(params.getFV(), is, 0, null, pipe.cl, THREADS, l2i);// cnt-1
} catch (Exception e) {
e.printStackTrace();
return null;
}
short[] pos = is.pposs[0];
List<ParseNBest> parses = null;
Parse d = null;
try {
parses = Decoder.decode(pos, d2, options.decodeProjective, pipe.extractor[0]); // cnt-1
d = parses.get(0);
} catch (Exception e) {
e.printStackTrace();
}
if (parses.size() > NBest)
parses = parses.subList(0, NBest);
int g_las = Decoder.getGoldRank(parses, is, 0, Decoder.LAS);
int g_ula = Decoder.getGoldRank(parses, is, 0, !Decoder.LAS);
int smallest = Decoder.getSmallestError(parses, is, 0, !Decoder.LAS);
smallestErrorSum += is.length(0) - smallest;
countAllNodes += is.length(0);
if (g_las >= 0)
amongxbest++;
if (g_ula >= 0)
amongxbest_ula++;
nbest += parses.size();
Integer r = rank.get(g_las);
if (r == null)
rank.put(g_las, 1);
else
rank.put(g_las, r + 1);
this.pipe.errors(is, 0, d);
float errBestProj = (float) this.pipe.errors(is, 0, Decoder.bestProj);
if (errBestProj == 0)
bestProj++;
SentenceData09 i09 = new SentenceData09(instance);
i09.createSemantic(instance);
for (int j = 0; j < forms.length - 1; j++) {
i09.plabels[j] = types[d.labels[j + 1]];
i09.pheads[j] = d.heads[j + 1];
}
return i09;
}
public List<ParseNBest> parseNBest(SentenceData09 instance, ParametersFloat params) {
Instances is = new Instances();
is.init(1, new MFB(), options.formatTask);
new CONLLReader09().insert(is, instance);
// use for the training ppos
DataF d2;
try {
d2 = pipe.fillVector(params.getFV(), is, 0, null, pipe.cl, THREADS, l2i);// cnt-1
} catch (Exception e) {
e.printStackTrace();
return null;
}
short[] pos = is.pposs[0];
List<ParseNBest> parses = null;
try {
parses = Decoder.decode(pos, d2, options.decodeProjective, pipe.extractor[0]); // cnt-1
} catch (Exception e) {
e.printStackTrace();
}
if (parses.size() > NBest)
parses = parses.subList(0, NBest);
return parses;
}
/*
* (non-Javadoc)
*
* @see is2.tools.Tool#apply(is2.data.SentenceData09)
*/
@Override
public SentenceData09 apply(SentenceData09 snt09) {
try {
parse(snt09, this.parametersReranker);
} catch (Exception e) {
e.printStackTrace();
}
Decoder.executerService.shutdown();
Pipe.executerService.shutdown();
return snt09;
}
/**
* Write the parsing model
*
* @param options
* @param params
* @param extension
* @throws FileNotFoundException
* @throws IOException
*/
private void writeModell(OptionsSuper options, ParametersFloat params, String extension, Cluster cs)
throws FileNotFoundException, IOException {
String name = extension == null ? options.modelName : options.modelName + extension;
DB.println("Writting model: " + name);
ZipOutputStream zos = new ZipOutputStream(new BufferedOutputStream(new FileOutputStream(name)));
zos.putNextEntry(new ZipEntry("data"));
DataOutputStream dos = new DataOutputStream(new BufferedOutputStream(zos));
MFB.writeData(dos);
cs.write(dos);
params.write(dos);
Edges.write(dos);
dos.writeBoolean(options.decodeProjective);
dos.writeInt(ExtractorClusterStacked.maxForm);
dos.writeInt(5); // Info count
dos.writeUTF("Used parser " + Reranker.class.toString());
dos.writeUTF("Creation date " + (new SimpleDateFormat("yyyy.MM.dd HH:mm:ss")).format(new Date()));
dos.writeUTF("Training data " + options.trainfile);
dos.writeUTF("Iterations " + options.numIters + " Used sentences " + options.count);
dos.writeUTF("Cluster " + options.clusterFile);
dos.flush();
dos.close();
}
}