|
1
|
|
|
/* |
|
2
|
|
|
* |
|
3
|
|
|
* Copyright 2014, Armenak Grigoryan, and individual contributors as indicated |
|
4
|
|
|
* by the @authors tag. See the copyright.txt in the distribution for a |
|
5
|
|
|
* full listing of individual contributors. |
|
6
|
|
|
* |
|
7
|
|
|
* This is free software; you can redistribute it and/or modify it |
|
8
|
|
|
* under the terms of the GNU Lesser General Public License as |
|
9
|
|
|
* published by the Free Software Foundation; either version 2.1 of |
|
10
|
|
|
* the License, or (at your option) any later version. |
|
11
|
|
|
* |
|
12
|
|
|
* This software is distributed in the hope that it will be useful, |
|
13
|
|
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of |
|
14
|
|
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
|
15
|
|
|
* Lesser General Public License for more details. |
|
16
|
|
|
*/ |
|
17
|
|
|
package com.strider.datadefender.discoverer; |
|
18
|
|
|
|
|
19
|
|
|
import com.strider.datadefender.DataDefenderException; |
|
20
|
|
|
import com.strider.datadefender.ModelDiscoveryConfig; |
|
21
|
|
|
import com.strider.datadefender.database.metadata.TableMetaData.ColumnMetaData; |
|
22
|
|
|
import com.strider.datadefender.requirement.file.Generator; |
|
23
|
|
|
|
|
24
|
|
|
import java.io.File; |
|
25
|
|
|
import java.io.IOException; |
|
26
|
|
|
import java.io.InputStream; |
|
27
|
|
|
import java.util.List; |
|
28
|
|
|
import java.util.Map; |
|
29
|
|
|
import java.util.stream.Collectors; |
|
30
|
|
|
|
|
31
|
|
|
import javax.xml.bind.JAXBException; |
|
32
|
|
|
|
|
33
|
|
|
import opennlp.tools.namefind.NameFinderME; |
|
34
|
|
|
import opennlp.tools.namefind.TokenNameFinderModel; |
|
35
|
|
|
import opennlp.tools.tokenize.TokenizerME; |
|
36
|
|
|
import opennlp.tools.tokenize.TokenizerModel; |
|
37
|
|
|
|
|
38
|
|
|
import lombok.Data; |
|
39
|
|
|
import lombok.extern.log4j.Log4j2; |
|
40
|
|
|
|
|
41
|
|
|
/** |
|
42
|
|
|
* Holds common logic for Discoverers. |
|
43
|
|
|
* @author Akira Matsuo |
|
44
|
|
|
*/ |
|
45
|
|
|
@Log4j2 |
|
46
|
|
|
public abstract class Discoverer { |
|
47
|
|
|
|
|
48
|
|
|
public static final String DEFAULT_TOKEN_MODEL = "en-token.bin"; |
|
49
|
|
|
public static final Map<String, String> BUILT_IN_MODELS = Map.of( |
|
50
|
|
|
"date", "en-ner-date.bin", |
|
51
|
|
|
"location", "en-ner-location.bin", |
|
52
|
|
|
"money", "en-ner-money.bin", |
|
53
|
|
|
"organization", "en-ner-organization.bin", |
|
54
|
|
|
"person", "en-ner-person.bin", |
|
55
|
|
|
"time", "en-ner-time.bin" |
|
56
|
|
|
); |
|
57
|
|
|
|
|
58
|
|
|
@Data |
|
59
|
|
|
public static class ColumnMatch { |
|
60
|
|
|
final private ColumnMetaData column; |
|
61
|
|
|
final private double averageProbability; |
|
62
|
|
|
final private String model; |
|
63
|
|
|
final private List<Probability> probabilityList; |
|
64
|
|
|
} |
|
65
|
|
|
|
|
66
|
|
|
protected List<ColumnMatch> matches; |
|
67
|
|
|
|
|
68
|
|
|
protected final ModelDiscoveryConfig config; |
|
69
|
|
|
protected final TokenizerME tokenizer; |
|
70
|
|
|
|
|
71
|
|
|
public Discoverer(ModelDiscoveryConfig config) throws IOException { |
|
72
|
|
|
this.config = config; |
|
73
|
|
|
if (config.getTokenModel() != null) { |
|
74
|
|
|
tokenizer = new TokenizerME(new TokenizerModel(config.getTokenModel())); |
|
75
|
|
|
} else { |
|
76
|
|
|
try (InputStream stream = Discoverer.class.getResourceAsStream(DEFAULT_TOKEN_MODEL)) { |
|
77
|
|
|
tokenizer = new TokenizerME(new TokenizerModel(stream)); |
|
78
|
|
|
} |
|
79
|
|
|
} |
|
80
|
|
|
} |
|
81
|
|
|
|
|
82
|
|
|
public double calculateAverage(final List<Probability> values) { |
|
83
|
|
|
Double sum = 0.0; |
|
84
|
|
|
|
|
85
|
|
|
if (!values.isEmpty()) { |
|
86
|
|
|
for (final Probability value : values) { |
|
87
|
|
|
sum += value.getProbabilityValue(); |
|
88
|
|
|
} |
|
89
|
|
|
|
|
90
|
|
|
return sum / values.size(); |
|
91
|
|
|
} |
|
92
|
|
|
|
|
93
|
|
|
return sum; |
|
94
|
|
|
} |
|
95
|
|
|
|
|
96
|
|
|
private Model createModelFrom(TokenNameFinderModel tnf, String modelName) { |
|
97
|
|
|
NameFinderME nameFinder = new NameFinderME(tnf); |
|
98
|
|
|
return new Model(tokenizer, nameFinder, modelName); |
|
99
|
|
|
} |
|
100
|
|
|
|
|
101
|
|
|
/** |
|
102
|
|
|
* Creates model POJO based on OpenNLP model file |
|
103
|
|
|
* |
|
104
|
|
|
* @param modelName |
|
105
|
|
|
* @return Model |
|
106
|
|
|
*/ |
|
107
|
|
|
public Model createModel(final File modelFile) throws IOException { |
|
108
|
|
|
return createModelFrom(new TokenNameFinderModel(modelFile), modelFile.getName()); |
|
109
|
|
|
} |
|
110
|
|
|
|
|
111
|
|
|
/** |
|
112
|
|
|
* Creates model POJO based on a built-in OpenNLP model |
|
113
|
|
|
* |
|
114
|
|
|
* @param modelName |
|
115
|
|
|
* @return Model |
|
116
|
|
|
*/ |
|
117
|
|
|
public Model createModel(final String modelName) throws IOException { |
|
118
|
|
|
try (InputStream stream = Discoverer.class.getResourceAsStream(BUILT_IN_MODELS.get(modelName))) { |
|
119
|
|
|
return createModelFrom(new TokenNameFinderModel(stream), modelName); |
|
120
|
|
|
} |
|
121
|
|
|
} |
|
122
|
|
|
} |
armenak /
DataDefender
