2
# Licensed to the Apache Software Foundation (ASF) under one or more
3
# contributor license agreements. See the NOTICE file distributed with
4
# this work for additional information regarding copyright ownership.
5
# The ASF licenses this file to You under the Apache License, Version 2.0
6
# (the "License"); you may not use this file except in compliance with
7
# the License. You may obtain a copy of the License at
9
# http://www.apache.org/licenses/LICENSE-2.0
11
# Unless required by applicable law or agreed to in writing, software
12
# distributed under the License is distributed on an "AS IS" BASIS,
13
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
14
# See the License for the specific language governing permissions and
15
# limitations under the License.
17
# Parses Lao text, with syllable as token.
19
# The definition of Lao syllable is based from:
21
# Syllabification of Lao Script for Line Breaking
22
# Phonpasit Phissamay, Valaxay Dalolay, Chitaphone Chanhsililath, Oulaiphone Silimasak,
23
# Sarmad Hussain, Nadir Durrani, Science Technology and Environment Agency, CRULP
24
# http://www.panl10n.net/english/final%20reports/pdf%20files/Laos/LAO06.pdf
25
# http://www.panl10n.net/Presentations/Cambodia/Phonpassit/LineBreakingAlgo.pdf
28
# There are some ambiguities in Lao syllabification without additional processing, as mentioned in the paper.
29
# For this reason, this RBBI grammar really only works with LaoBreakIterator, as it does this additional work.
31
# Syllable structure, where X is the nuclear consonant:
37
# +----+----+----+----+----+----+----+-----+
38
# | X0 | X1 | X | X6 | X7 | X8 | X9 | X10 |
39
# +----+----+----+----+----+----+----+-----+
45
# X0 represents a vowel which occurs before the nuclear consonant.
46
# It can always define the beginning of syllable.
47
$X0 = [\u0EC0-\u0EC4];
48
# X1 is a combination consonant which comes before the nuclear consonant,
49
# but only if nuclear consonant is one of {ງ ຍ ລ ວ ຼ ມ ນ ຣ}
51
# X represents the nuclear consonant.
52
$X = [\u0E81-\u0EAE\u0EDC\u0EDD];
53
# X2 is a combination consonant which comes after the nuclear consonant,
54
# which is placed under or next to the nuclear consonant.
55
$X2 = [\u0EBC\u0EA3\u0EA7\u0EA5];
56
# X3 represents a vowel which occurs under the nuclear consonant.
58
# X4 represents a vowel which occurs above the nuclear consonant.
59
$X4 = [\u0EB4-\u0EB7\u0ECD\u0EBB\u0EB1];
60
# X5 represents a tone mark which occurs above the nuclear consonant or upper vowel.
61
$X5 = [\u0EC8-\u0ECB];
62
# X6 represents a consonant vowel, which occurs after the nuclear consonant.
63
# It functions when the syllable doesn’t have any vowels. And it always exists with X8.
64
$X6 = [\u0EA7\u0EAD\u0EBD];
65
# X7 represents a final vowel.
66
# However X7_1 always represents the end of syllable and it never exists with tone mark.
67
$X7 = [\u0EB0\u0EB2\u0EB3];
68
# X8 represents an alternate consonant.
69
$X8 = [\u0E81\u0E87\u0E8D\u0E94\u0E99\u0EA1\u0E9A\u0EA7];
70
# X9 represents alternate consonants to pronounce foreign terms, it always exist with X10_3.
71
$X9 = [\u0E88\u0EAA\u0E8A\u0E9E\u0E9F\u0EA5];
72
# X10 represents a sign mark.
73
# It always occurs at the end of a syllable, but mostly people keep it separate from syllable.
74
$X10 = [\u0EAF\u0EC6\u0ECC];
78
$X4_1_2 = [\u0EB4\u0EB5];
79
$X4_3_4 = [\u0EB6\u0EB7];
90
$Rule1_1 = $X0_1 ($X1)? $X ($X2)? ($X5)? ($X8)? ($X9 $X10_3)? ($X10_2)? ($X10_1)?;
91
$Rule1_2 = $X0_1 ($X1)? $X ($X2)? $X4_1_2 ($X5)? ($X8)? ($X9 $X10_3)? ($X10_2)? ($X10_1)?;
92
$Rule1_3 = $X0_1 ($X1)? $X ($X2)? $X4_3_4 ($X5)? $X6_2 ($X8)? ($X9 $X10_3)? ($X10_2)? ($X10_1)?;
93
$Rule1_4 = $X0_1 ($X1)? $X ($X2)? ($X7_2)? $X7_1;
94
$Rule1_5 = $X0_1 ($X1)? $X ($X2)? $X4_6 ($X5)? $X7_2 ($X9 $X10_3)? ($X10_2)? ($X10_1)?;
95
$Rule1_6 = $X0_1 ($X1)? $X ($X2)? $X4_7 ($X5)? $X8 ($X9 $X10_3)? ($X10_2)? ($X10_1)?;
96
$Rule1_7 = $X0_1 ($X1)? $X ($X2)? ($X4_7)? ($X5)? $X6_3 ($X9 $X10_3)? ($X10_2)? ($X10_1)?;
98
$Rule1 = ($Rule1_1 | $Rule1_2 | $Rule1_3 | $Rule1_4 | $Rule1_5 | $Rule1_6 | $Rule1_7);
103
$Rule2_1 = $X0_2 ($X1)? $X ($X2)? ($X5)? ($X6)? ($X8)? ($X9 $X10_3)? ($X10_2)? ($X10_1)?;
104
$Rule2_2 = $X0_2 ($X1)? $X ($X2)? $X7_1;
105
$Rule2_3 = $X0_2 ($X1)? $X ($X2)? $X4_7 ($X5)? $X8 ($X9 $X10_3)? ($X10_2)? ($X10_1)?;
107
$Rule2 = ($Rule2_1 | $Rule2_2 | $Rule2_3);
114
$Rule3_1 = $X0_3 ($X1)? $X ($X2)? ($X5)? ($X8)? ($X9 $X10_3)? ($X10_2)? ($X10_1)?;
115
$Rule3_2 = $X0_3 ($X1)? $X ($X2)? $X7_1;
116
$Rule3_3 = $X0_3 ($X1)? $X ($X2)? $X4_7 ($X5)? ($X8_3 | $X8_8);
118
$Rule3 = ($Rule3_1 | $Rule3_2 | $Rule3_3);
124
$Rule4 = $X0_4 ($X1)? $X ($X2)? ($X5)? ($X6_1)? ($X9 $X10_3)? ($X10_2)? ($X10_1)?;
129
$Rule5 = $X0_5 ($X1)? $X ($X2)? ($X5)? ($X6_1)? ($X9 $X10_3)? ($X10_2)? ($X10_1)?;
132
$Rule6 = ($X1)? $X ($X2)? $X3 ($X5)? ($X8)? ($X9 $X10_3)? ($X10_2)? ($X10_1)?;
135
$X4_1_4 = [\u0EB4-\u0EB7];
137
$Rule7 = ($X1)? $X ($X2)? $X4_1_4 ($X5)? ($X8)? ($X9 $X10_3)? ($X10_2)? ($X10_1)?;
142
$Rule8 = ($X1)? $X ($X2)? $X4_5 ($X5)? ($X7_2)? ($X9 $X10_3)? ($X10_2)? ($X10_1)?;
146
$Rule9_1 = ($X1)? $X ($X2)? $X4_6 ($X5)? $X8 ($X9 $X10_3)? ($X10_2)? ($X10_1)?;
147
$Rule9_2 = ($X1)? $X ($X2)? $X4_6 ($X5)? $X6_1 $X7_1;
149
$Rule9 = ($Rule9_1 | $Rule9_2);
152
$Rule10 = ($X1)? $X ($X2)? $X4_7 ($X5)? ($X6_1)? $X8 ($X9 $X10_3)? ($X10_2)? ($X10_1)?;
155
$Rule11 = ($X1)? $X ($X2)? ($X5)? $X6 $X8 ($X9 $X10_3)? ($X10_2)? ($X10_1)?;
158
$Rule12 = ($X1)? $X ($X2)? ($X5)? $X7_1;
161
$Rule13 = ($X1)? $X ($X2)? ($X5)? $X7_2 ($X8)? ($X9 $X10_3)? ($X10_2)? ($X10_1)?;
166
$Rule14 = ($X1)? $X ($X2)? ($X5)? $X7_3 ($X9 $X10_3)? ($X10_2)? ($X10_1)?;
168
$LaoSyllableEx = ($Rule1 | $Rule2 | $Rule3 | $Rule4 | $Rule5 | $Rule6 | $Rule7 | $Rule8 | $Rule9 | $Rule10 | $Rule11 | $Rule12 | $Rule13 | $Rule14);
170
$WordJoin = [:Line_Break=Word_Joiner:];
172
$LaoJoinedSyllableEx = $LaoSyllableEx ($WordJoin $LaoSyllableEx)*;
175
# default numerical definitions
177
$Extend = [\p{Word_Break = Extend}];
178
$Format = [\p{Word_Break = Format}];
179
$MidNumLet = [\p{Word_Break = MidNumLet}];
180
$MidNum = [\p{Word_Break = MidNum}];
181
$Numeric = [\p{Word_Break = Numeric}];
182
$ExtendNumLet = [\p{Word_Break = ExtendNumLet}];
183
$MidNumLetEx = $MidNumLet ($Extend | $Format)*;
184
$MidNumEx = $MidNum ($Extend | $Format)*;
185
$NumericEx = $Numeric ($Extend | $Format)*;
186
$ExtendNumLetEx = $ExtendNumLet ($Extend | $Format)*;
190
$LaoJoinedSyllableEx {200};
191
# default numeric rules
192
$NumericEx $ExtendNumLetEx? (($MidNumEx | $MidNumLetEx)? $NumericEx $ExtendNumLetEx?)* {100};