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
|
package net.sourceforge.segment.srx;
import java.util.ArrayList;
import java.util.List;
import java.util.regex.Pattern;
/**
* Represents merged splitting pattern.
* Responsible for merging breaking rules into one large pattern and
* creating non breaking rules pattern.
* @author loomchild
*/
public class MergedPattern {
private Pattern breakingPattern;
private List<Pattern> nonBreakingPatternList;
public MergedPattern(List<LanguageRule> languageRuleList) {
this.nonBreakingPatternList = new ArrayList<Pattern>();
StringBuilder breakingPatternBuilder = new StringBuilder();
List<Rule> ruleList = extractRules(languageRuleList);
List<List<Rule>> ruleGroupList = groupRules(ruleList);
if (ruleGroupList.size() > 0) {
for (List<Rule> ruleGroup : ruleGroupList) {
if (breakingPatternBuilder.length() > 0) {
breakingPatternBuilder.append('|');
}
// All breaking rules need to be merged because segmentation
// need to be done in one pass when text is read from Reader.
// Breaking rule need to be inside capturing group so
// it is possible to recognize which breaking rule has been
// applied during the splitting and know which non-breaking
// rules to use.
// Breaking rule cannot contain capturing groups.
// Capturing groups are replaced with non-capturing groups
// inside create breaking pattern function.
String breakingGroupPattern = createBreakingPattern(ruleGroup);
breakingPatternBuilder.append("(" + breakingGroupPattern + ")");
// If first rule in the group is breaking then there are
// no non-breaking rules in the group. In this case null
// is appended to non breaking pattern list, because
// null does not match anything.
if (!ruleGroup.get(0).isBreaking()) {
Pattern nonBreakingGroupPattern =
Pattern.compile(createNonBreakingPattern(ruleGroup));
nonBreakingPatternList.add(nonBreakingGroupPattern);
} else {
nonBreakingPatternList.add(null);
}
}
this.breakingPattern = Pattern.compile(breakingPatternBuilder
.toString());
} else {
// null means that that pattern will not match anything
// (as empty pattern matches everything).
this.breakingPattern = null;
}
}
public Pattern getBreakingPattern() {
return breakingPattern;
}
public List<Pattern> getNonBreakingPatternList() {
return nonBreakingPatternList;
}
/**
* @param languageRuleList
* Language rule list.
* @return Returns merged list of rules form given language rules.
*/
private List<Rule> extractRules(List<LanguageRule> languageRuleList) {
List<Rule> ruleList = new ArrayList<Rule>();
for (LanguageRule languageRule : languageRuleList) {
ruleList.addAll(languageRule.getRuleList());
}
return ruleList;
}
/**
* Divides rules to groups where breaking and non-breaking rules cannot be
* interlaced.
*
* @param ruleList
* @return
*/
private List<List<Rule>> groupRules(List<Rule> ruleList) {
List<List<Rule>> ruleGroupList = new ArrayList<List<Rule>>();
List<Rule> ruleGroup = new ArrayList<Rule>();
boolean previousBreaking = false;
for (Rule rule : ruleList) {
if (rule.isBreaking() && !previousBreaking) {
ruleGroupList.add(ruleGroup);
} else if (!rule.isBreaking() && previousBreaking) {
ruleGroup = new ArrayList<Rule>();
}
ruleGroup.add(rule);
previousBreaking = rule.isBreaking();
}
return ruleGroupList;
}
/**
* Merges all breaking rules on list into one pattern.
*
* @param ruleList
* List of rules
* @return Returns pattern.
*/
private String createBreakingPattern(List<Rule> ruleList) {
StringBuilder patternBuilder = new StringBuilder();
for (Rule rule : ruleList) {
if (rule.isBreaking()) {
if (patternBuilder.length() > 0) {
patternBuilder.append('|');
}
// Capturing groups need to be removed from patterns as
// they will interfere with capturing group order
// which is used to recognize which breaking rule has been
// applied and decide which non-breaking rules to use.
// In addition as Java does not allow infinite length patterns
// in lookbehind, before pattern need to be shortened.
String beforePattern = Util.finitize(
Util.removeCapturingGroups(rule.getBeforePattern()));
String afterPattern = Util.removeCapturingGroups(rule
.getAfterPattern());
// Need to use lookbehind also in breaking rule because
// this way they become zero-length matches and I want
// to match shorter rules first, independent of occurrence order
// as in normal alternative.
// Example:
// Input: "aaa".
// Pattern "aa|a" matches "aa" first.
// Pattern "(?<=aa)|(?<=a)" matches "a" first.
if (beforePattern.length() > 0) {
patternBuilder.append("(?<=" + beforePattern + ")");
}
if (afterPattern.length() > 0) {
patternBuilder.append("(?=" + afterPattern + ")");
}
}
}
return patternBuilder.toString();
}
/**
* Creates non breaking pattern by merging given rules.
*
* @param ruleList
* Rule list.
* @return Non breaking pattern.
*/
private String createNonBreakingPattern(List<Rule> ruleList) {
StringBuilder patternBuilder = new StringBuilder();
for (Rule rule : ruleList) {
if (!rule.isBreaking()) {
if (patternBuilder.length() > 0) {
patternBuilder.append('|');
}
// As Java does not allow infinite length patterns
// in lookbehind, before pattern need to be shortened.
String beforePattern = Util.finitize(rule.getBeforePattern());
String afterPattern = rule.getAfterPattern();
if (beforePattern.length() > 0) {
patternBuilder.append("(?<=" + beforePattern + ")");
}
if (afterPattern.length() > 0) {
patternBuilder.append("(?=" + afterPattern + ")");
}
}
}
return patternBuilder.toString();
}
}
|
