Repository navigation
Expand file tree
/
Copy pathalign-text.cc
More file actions
137 lines (117 loc) · 5.13 KB
/
Copy pathalign-text.cc
File metadata and controls
137 lines (117 loc) · 5.13 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
// bin/align-text.cc
// Copyright 2014 Guoguo Chen
// See ../../COPYING for clarification regarding multiple authors
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
// http://www.apache.org/licenses/LICENSE-2.0
//
// THIS CODE IS PROVIDED *AS IS* BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
// KIND, EITHER EXPRESS OR IMPLIED, INCLUDING WITHOUT LIMITATION ANY IMPLIED
// WARRANTIES OR CONDITIONS OF TITLE, FITNESS FOR A PARTICULAR PURPOSE,
// MERCHANTABLITY OR NON-INFRINGEMENT.
// See the Apache 2 License for the specific language governing permissions and
// limitations under the License.
#include "util/common-utils.h"
#include "util/parse-options.h"
#include "util/edit-distance.h"
#include <algorithm>
bool IsNotToken(const std::string &token) {
return ! kaldi::IsToken(token);
}
int main(int argc, char *argv[]) {
using namespace kaldi;
typedef kaldi::int32 int32;
try {
const char *usage =
"Computes alignment between two sentences with the same key in the\n"
"two given input text-rspecifiers. The current implementation uses\n"
"Levenshtein distance as the distance metric.\n"
"\n"
"The input text file looks like follows:\n"
" key1 a b c\n"
" key2 d e\n"
"\n"
"The output alignment file looks like follows:\n"
" key1 a a ; b <eps> ; c c \n"
" key2 d f ; e e \n"
"where the aligned pairs are separated by \";\"\n"
"\n"
"Usage: align-text [options] <text1-rspecifier> <text2-rspecifier> \\\n"
" <alignment-wspecifier>\n"
" e.g.: align-text ark:text1.txt ark:text2.txt ark,t:alignment.txt\n"
"See also: compute-wer,\n"
"Example scoring script: egs/wsj/s5/steps/score_kaldi.sh\n";
ParseOptions po(usage);
std::string special_symbol = "<eps>";
std::string separator = ";";
po.Register("special-symbol", &special_symbol, "Special symbol to be "
"aligned with the inserted or deleted words. Your sentences "
"should not contain this symbol.");
po.Register("separator", &separator, "Separator for each aligned pairs in "
"the output alignment file. Your sentences should not contain "
"this symbol.");
po.Read(argc, argv);
if (po.NumArgs() != 3) {
po.PrintUsage();
exit(1);
}
std::string text1_rspecifier = po.GetArg(1),
text2_rspecifier = po.GetArg(2),
align_wspecifier = po.GetArg(3);
SequentialTokenVectorReader text1_reader(text1_rspecifier);
RandomAccessTokenVectorReader text2_reader(text2_rspecifier);
TokenVectorWriter align_writer(align_wspecifier);
int32 n_done = 0;
int32 n_fail = 0;
for (; !text1_reader.Done(); text1_reader.Next()) {
std::string key = text1_reader.Key();
if (!text2_reader.HasKey(key)) {
KALDI_WARN << "Key " << key << " is in " << text1_rspecifier
<< ", but not in " << text2_rspecifier;
n_fail++;
continue;
}
const std::vector<std::string> &text1 = text1_reader.Value();
const std::vector<std::string> &text2 = text2_reader.Value(key);
// Checks if the special symbol and separator is in the string.
KALDI_ASSERT(std::find(text1.begin(),
text1.end(), special_symbol) == text1.end());
KALDI_ASSERT(std::find(text2.begin(),
text2.end(), special_symbol) == text2.end());
KALDI_ASSERT(std::find(text1.begin(),
text1.end(), separator) == text1.end());
KALDI_ASSERT(std::find(text2.begin(),
text2.end(), separator) == text2.end());
if (std::find_if(text1.begin(), text1.end(), IsNotToken) != text1.end()) {
KALDI_ERR << "In text1, the utterance " << key << " contains unprintable characters." \
<< "That means there is a problem with the text (such as incorrect encoding)." << std::endl;
return -1;
}
if (std::find_if(text2.begin(), text2.end(), IsNotToken) != text2.end()) {
KALDI_ERR << "In text2, the utterance " << key << " contains unprintable characters." \
<< "That means there is a problem with the text (such as incorrect encoding)." << std::endl;
return -1;
}
std::vector<std::pair<std::string, std::string> > aligned;
LevenshteinAlignment(text1, text2, special_symbol, &aligned);
std::vector<std::string> token_vec;
std::vector<std::pair<std::string, std::string> >::const_iterator iter;
for (iter = aligned.begin(); iter != aligned.end(); ++iter) {
token_vec.push_back(iter->first);
token_vec.push_back(iter->second);
if (aligned.end() - iter != 1)
token_vec.push_back(separator);
}
align_writer.Write(key, token_vec);
n_done++;
}
KALDI_LOG << "Done " << n_done << " sentences, failed for " << n_fail;
return (n_done != 0 ? 0 : 1);
} catch(const std::exception &e) {
std::cerr << e.what();
return -1;
}
}