forked from jhlau/acceptability_prediction
-
Notifications
You must be signed in to change notification settings - Fork 0
/
OutputHMMInducedClasses.java
168 lines (149 loc) · 5.86 KB
/
OutputHMMInducedClasses.java
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
/*
Author: Jey Han Lau
Date: May 14
Reads a BHMM model file and outputs the induced word classes of sentences of the training data.
*/
import tikka.bhmm.models.*;
import tikka.bhmm.model.base.*;
import tikka.bhmm.apps.*;
import org.apache.commons.cli.*;
import java.io.IOException;
import java.util.ArrayList;
import java.util.Arrays;
import java.util.Collections;
public class OutputHMMInducedClasses {
private static boolean debug = false;
public static void main(String[] args) {
CommandLineParser optparse = new PosixParser();
CommandLine cline = null;
Options options = setOptions();
String inputModel = "";
int numSent = 1000;
int maxSentLen = 10;
boolean conllFormat = false;
try {
cline = optparse.parse(options, args);
if (cline.hasOption('h')) {
HelpFormatter formatter = new HelpFormatter();
formatter.printHelp("java OutputHMMInducedClasses", options);
System.exit(0);
}
if (cline.hasOption("l")) {
inputModel = cline.getOptionValue("l");
}
if (cline.hasOption("xns")) {
numSent = Integer.parseInt(cline.getOptionValue("xns"));
}
if (cline.hasOption("xms")) {
maxSentLen = Integer.parseInt(cline.getOptionValue("xms"));
}
if (cline.hasOption("xconll")) {
conllFormat = true;
}
} catch (ParseException parseException) {
System.err.println("Error parsing command line arguments");
System.exit(0);
}
// load the model
HMMBase bhmm = null;
try {
CommandLineOptions modelOptions = new CommandLineOptions(cline);
SerializableModel serializableModel = new SerializableModel();
bhmm = serializableModel.loadModel(modelOptions,
modelOptions.getModelInputPath());
bhmm.resetTrainDataDir(); // reset the previous train directory (avoid IOException)
bhmm.initializeFromLoadedModel2(modelOptions);
} catch (IOException e) {
System.err.println("Error loading input model; path = " + inputModel);
System.exit(0);
}
// get the sentence states
int[] stateVector = bhmm.getStateVector();
int[] sentenceVector = bhmm.getSentenceVector();
int prevSent = 0;
ArrayList<Integer> currSent = new ArrayList<Integer>();
ArrayList<String[]> sents = new ArrayList<String[]>();
for (int i=0; i<stateVector.length; i++) {
int state = stateVector[i];
int stateSent = sentenceVector[i];
if ((stateSent != prevSent) || (i==(stateVector.length-1))) {
if (i==(stateVector.length-1)) {
currSent.add(state);
}
if ((maxSentLen == 0) || (currSent.size() <= maxSentLen)) {
sents.add(convToString(currSent));
}
currSent.clear();
if ((sents.size() % 10000) == 0) {
//System.err.println("Number of sentences processed = " + sents.size());
}
}
currSent.add(state);
prevSent = stateSent;
}
if (debug) {
System.out.println("Sentence Vector = " + Arrays.toString(sentenceVector));
System.out.println("State Vector = " + Arrays.toString(stateVector));
System.out.println("Pre-shuffle sentences = ");
for (String[] sent : sents) {
System.out.println(Arrays.toString(sent));
}
}
// shuffle the sents and select top-N
if (numSent != 0) {
System.err.println("Sorting...");
System.err.println("Number of items = " + sents.size());
Collections.shuffle(sents);
if (debug) {
System.out.println("Post-shuffle sentences = ");
for (String[] sent : sents) {
System.out.println(Arrays.toString(sent));
}
}
System.err.println("Done sorting...");
} else {
numSent = sents.size();
}
if (sents.size() < numSent) {
numSent = sents.size();
}
for (int i=0; i<numSent; i++) {
String[] outputSent = sents.get(i);
for (int j=0; j<outputSent.length; j++) {
System.out.print(outputSent[j]);
if (j == (outputSent.length-1)) {
if (!conllFormat) {
System.out.print("\n");
} else {
System.out.print("\n\n");
}
} else {
if (!conllFormat) {
System.out.print(" ");
} else {
System.out.print("\n");
}
}
}
}
}
private static String[] convToString(ArrayList<Integer> sent) {
String[] strList = new String[sent.size()];
for (int i=0; i<sent.size(); i++) {
strList[i] = "S" + sent.get(i);
}
return strList;
}
public static Options setOptions() {
Options options = new Options();
options.addOption("h", "help", false, "print help");
options.addOption("l", "input-model", true, "trained BHMM model");
options.addOption("xns", "num-sent", true, "number of sentences (0 = no restriction; " +
"default = 1000)");
options.addOption("xms", "max-sent-length", true, "maximum sentence length " +
"(0 = no restriction; default = 10)");
options.addOption("xconll", "conll-format", false, "use CONLL format (default = " +
"1 line per sentence)");
return options;
}
}