stride-align 0.6.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (190) hide show
  1. checksums.yaml +7 -0
  2. data/LICENSE +201 -0
  3. data/NOTICE +255 -0
  4. data/README.md +107 -0
  5. data/data/bmpm_data/gen_approx_any.txt +131 -0
  6. data/data/bmpm_data/gen_approx_arabic.txt +26 -0
  7. data/data/bmpm_data/gen_approx_common.txt +233 -0
  8. data/data/bmpm_data/gen_approx_cyrillic.txt +18 -0
  9. data/data/bmpm_data/gen_approx_czech.txt +18 -0
  10. data/data/bmpm_data/gen_approx_dutch.txt +18 -0
  11. data/data/bmpm_data/gen_approx_english.txt +47 -0
  12. data/data/bmpm_data/gen_approx_french.txt +25 -0
  13. data/data/bmpm_data/gen_approx_german.txt +73 -0
  14. data/data/bmpm_data/gen_approx_greek.txt +18 -0
  15. data/data/bmpm_data/gen_approx_greeklatin.txt +20 -0
  16. data/data/bmpm_data/gen_approx_hebrew.txt +18 -0
  17. data/data/bmpm_data/gen_approx_hungarian.txt +18 -0
  18. data/data/bmpm_data/gen_approx_italian.txt +18 -0
  19. data/data/bmpm_data/gen_approx_polish.txt +84 -0
  20. data/data/bmpm_data/gen_approx_portuguese.txt +18 -0
  21. data/data/bmpm_data/gen_approx_romanian.txt +18 -0
  22. data/data/bmpm_data/gen_approx_russian.txt +48 -0
  23. data/data/bmpm_data/gen_approx_spanish.txt +21 -0
  24. data/data/bmpm_data/gen_approx_turkish.txt +18 -0
  25. data/data/bmpm_data/gen_exact_any.txt +40 -0
  26. data/data/bmpm_data/gen_exact_approx_common.txt +79 -0
  27. data/data/bmpm_data/gen_exact_arabic.txt +18 -0
  28. data/data/bmpm_data/gen_exact_common.txt +32 -0
  29. data/data/bmpm_data/gen_exact_cyrillic.txt +18 -0
  30. data/data/bmpm_data/gen_exact_czech.txt +18 -0
  31. data/data/bmpm_data/gen_exact_dutch.txt +18 -0
  32. data/data/bmpm_data/gen_exact_english.txt +18 -0
  33. data/data/bmpm_data/gen_exact_french.txt +18 -0
  34. data/data/bmpm_data/gen_exact_german.txt +18 -0
  35. data/data/bmpm_data/gen_exact_greek.txt +18 -0
  36. data/data/bmpm_data/gen_exact_greeklatin.txt +18 -0
  37. data/data/bmpm_data/gen_exact_hebrew.txt +18 -0
  38. data/data/bmpm_data/gen_exact_hungarian.txt +18 -0
  39. data/data/bmpm_data/gen_exact_italian.txt +18 -0
  40. data/data/bmpm_data/gen_exact_polish.txt +23 -0
  41. data/data/bmpm_data/gen_exact_portuguese.txt +18 -0
  42. data/data/bmpm_data/gen_exact_romanian.txt +18 -0
  43. data/data/bmpm_data/gen_exact_russian.txt +19 -0
  44. data/data/bmpm_data/gen_exact_spanish.txt +19 -0
  45. data/data/bmpm_data/gen_exact_turkish.txt +18 -0
  46. data/data/bmpm_data/gen_hebrew_common.txt +113 -0
  47. data/data/bmpm_data/gen_lang.txt +295 -0
  48. data/data/bmpm_data/gen_languages.txt +36 -0
  49. data/data/bmpm_data/gen_rules_any.txt +367 -0
  50. data/data/bmpm_data/gen_rules_arabic.txt +76 -0
  51. data/data/bmpm_data/gen_rules_cyrillic.txt +99 -0
  52. data/data/bmpm_data/gen_rules_czech.txt +67 -0
  53. data/data/bmpm_data/gen_rules_dutch.txt +78 -0
  54. data/data/bmpm_data/gen_rules_english.txt +113 -0
  55. data/data/bmpm_data/gen_rules_french.txt +114 -0
  56. data/data/bmpm_data/gen_rules_german.txt +129 -0
  57. data/data/bmpm_data/gen_rules_greek.txt +97 -0
  58. data/data/bmpm_data/gen_rules_greeklatin.txt +118 -0
  59. data/data/bmpm_data/gen_rules_hebrew.txt +62 -0
  60. data/data/bmpm_data/gen_rules_hungarian.txt +83 -0
  61. data/data/bmpm_data/gen_rules_italian.txt +77 -0
  62. data/data/bmpm_data/gen_rules_polish.txt +185 -0
  63. data/data/bmpm_data/gen_rules_portuguese.txt +105 -0
  64. data/data/bmpm_data/gen_rules_romanian.txt +64 -0
  65. data/data/bmpm_data/gen_rules_russian.txt +142 -0
  66. data/data/bmpm_data/gen_rules_spanish.txt +85 -0
  67. data/data/bmpm_data/gen_rules_turkish.txt +50 -0
  68. data/data/keyboard_data/qwerty.npy +0 -0
  69. data/data/matrix_data/BLOSUM100 +31 -0
  70. data/data/matrix_data/BLOSUM30 +31 -0
  71. data/data/matrix_data/BLOSUM35 +31 -0
  72. data/data/matrix_data/BLOSUM40 +31 -0
  73. data/data/matrix_data/BLOSUM45 +25 -0
  74. data/data/matrix_data/BLOSUM50 +25 -0
  75. data/data/matrix_data/BLOSUM55 +31 -0
  76. data/data/matrix_data/BLOSUM60 +31 -0
  77. data/data/matrix_data/BLOSUM62 +25 -0
  78. data/data/matrix_data/BLOSUM65 +31 -0
  79. data/data/matrix_data/BLOSUM70 +31 -0
  80. data/data/matrix_data/BLOSUM75 +31 -0
  81. data/data/matrix_data/BLOSUM80 +25 -0
  82. data/data/matrix_data/BLOSUM85 +31 -0
  83. data/data/matrix_data/BLOSUM90 +25 -0
  84. data/data/matrix_data/NUC.4.4 +25 -0
  85. data/data/matrix_data/PAM10 +34 -0
  86. data/data/matrix_data/PAM100 +34 -0
  87. data/data/matrix_data/PAM110 +34 -0
  88. data/data/matrix_data/PAM120 +34 -0
  89. data/data/matrix_data/PAM130 +34 -0
  90. data/data/matrix_data/PAM140 +34 -0
  91. data/data/matrix_data/PAM150 +34 -0
  92. data/data/matrix_data/PAM160 +34 -0
  93. data/data/matrix_data/PAM170 +34 -0
  94. data/data/matrix_data/PAM180 +34 -0
  95. data/data/matrix_data/PAM190 +34 -0
  96. data/data/matrix_data/PAM20 +34 -0
  97. data/data/matrix_data/PAM200 +34 -0
  98. data/data/matrix_data/PAM210 +34 -0
  99. data/data/matrix_data/PAM220 +34 -0
  100. data/data/matrix_data/PAM230 +34 -0
  101. data/data/matrix_data/PAM240 +34 -0
  102. data/data/matrix_data/PAM250 +25 -0
  103. data/data/matrix_data/PAM260 +34 -0
  104. data/data/matrix_data/PAM270 +34 -0
  105. data/data/matrix_data/PAM280 +34 -0
  106. data/data/matrix_data/PAM290 +34 -0
  107. data/data/matrix_data/PAM30 +25 -0
  108. data/data/matrix_data/PAM300 +34 -0
  109. data/data/matrix_data/PAM310 +34 -0
  110. data/data/matrix_data/PAM320 +34 -0
  111. data/data/matrix_data/PAM330 +34 -0
  112. data/data/matrix_data/PAM340 +34 -0
  113. data/data/matrix_data/PAM350 +34 -0
  114. data/data/matrix_data/PAM360 +34 -0
  115. data/data/matrix_data/PAM370 +34 -0
  116. data/data/matrix_data/PAM380 +34 -0
  117. data/data/matrix_data/PAM390 +34 -0
  118. data/data/matrix_data/PAM40 +34 -0
  119. data/data/matrix_data/PAM400 +34 -0
  120. data/data/matrix_data/PAM410 +34 -0
  121. data/data/matrix_data/PAM420 +34 -0
  122. data/data/matrix_data/PAM430 +34 -0
  123. data/data/matrix_data/PAM440 +34 -0
  124. data/data/matrix_data/PAM450 +34 -0
  125. data/data/matrix_data/PAM460 +34 -0
  126. data/data/matrix_data/PAM470 +34 -0
  127. data/data/matrix_data/PAM480 +34 -0
  128. data/data/matrix_data/PAM490 +34 -0
  129. data/data/matrix_data/PAM50 +34 -0
  130. data/data/matrix_data/PAM500 +34 -0
  131. data/data/matrix_data/PAM60 +34 -0
  132. data/data/matrix_data/PAM70 +25 -0
  133. data/data/matrix_data/PAM80 +34 -0
  134. data/data/matrix_data/PAM90 +34 -0
  135. data/ext/stride_align/backend_avx2.cpp +2 -0
  136. data/ext/stride_align/backend_avx512bwvl.cpp +2 -0
  137. data/ext/stride_align/backend_generic.cpp +3 -0
  138. data/ext/stride_align/backend_impl.hpp +983 -0
  139. data/ext/stride_align/backend_lasx.cpp +2 -0
  140. data/ext/stride_align/backend_lsx.cpp +2 -0
  141. data/ext/stride_align/backend_neon.cpp +2 -0
  142. data/ext/stride_align/backend_rvv.cpp +2 -0
  143. data/ext/stride_align/backend_sse41.cpp +2 -0
  144. data/ext/stride_align/backend_sve.cpp +2 -0
  145. data/ext/stride_align/backend_sve2.cpp +2 -0
  146. data/ext/stride_align/backend_vsx.cpp +2 -0
  147. data/ext/stride_align/beider_morse_impl.cpp +5 -0
  148. data/ext/stride_align/cpu_detect.cpp +172 -0
  149. data/ext/stride_align/cpu_detect.hpp +6 -0
  150. data/ext/stride_align/extconf.rb +114 -0
  151. data/ext/stride_align/target_profile.hpp +82 -0
  152. data/ext/stride_align/vendor/beider_morse_impl.cpp +1467 -0
  153. data/ext/stride_align/vendor/stride_align/alignment.hpp +199 -0
  154. data/ext/stride_align/vendor/stride_align/batch.hpp +812 -0
  155. data/ext/stride_align/vendor/stride_align/beider_morse.hpp +121 -0
  156. data/ext/stride_align/vendor/stride_align/caverphone.hpp +222 -0
  157. data/ext/stride_align/vendor/stride_align/cologne_phonetic.hpp +202 -0
  158. data/ext/stride_align/vendor/stride_align/core.hpp +731 -0
  159. data/ext/stride_align/vendor/stride_align/daitch_mokotoff.hpp +631 -0
  160. data/ext/stride_align/vendor/stride_align/double_metaphone.hpp +796 -0
  161. data/ext/stride_align/vendor/stride_align/dtw.hpp +300 -0
  162. data/ext/stride_align/vendor/stride_align/encoded.hpp +235 -0
  163. data/ext/stride_align/vendor/stride_align/hamming.hpp +55 -0
  164. data/ext/stride_align/vendor/stride_align/indel.hpp +1200 -0
  165. data/ext/stride_align/vendor/stride_align/jaro.hpp +517 -0
  166. data/ext/stride_align/vendor/stride_align/lcs.hpp +159 -0
  167. data/ext/stride_align/vendor/stride_align/levenshtein.hpp +1247 -0
  168. data/ext/stride_align/vendor/stride_align/levenshtein_prepared.hpp +193 -0
  169. data/ext/stride_align/vendor/stride_align/match_rating.hpp +168 -0
  170. data/ext/stride_align/vendor/stride_align/metaphone.hpp +291 -0
  171. data/ext/stride_align/vendor/stride_align/ngram.hpp +176 -0
  172. data/ext/stride_align/vendor/stride_align/nysiis.hpp +199 -0
  173. data/ext/stride_align/vendor/stride_align/pairwise_alignment.hpp +465 -0
  174. data/ext/stride_align/vendor/stride_align/partial_ratio.hpp +486 -0
  175. data/ext/stride_align/vendor/stride_align/ratcliff_obershelp.hpp +101 -0
  176. data/ext/stride_align/vendor/stride_align/soundex.hpp +108 -0
  177. data/ext/stride_align/vendor/stride_align/token_ratios.hpp +445 -0
  178. data/ext/stride_align/vendor/stride_align/types.hpp +16 -0
  179. data/ext/stride_align/vendor/stride_align/utf8.hpp +512 -0
  180. data/ext/stride_align/vendor/stride_align/wratio.hpp +363 -0
  181. data/lib/stride_align/algorithms.rb +296 -0
  182. data/lib/stride_align/alignment_path.rb +217 -0
  183. data/lib/stride_align/backend.rb +47 -0
  184. data/lib/stride_align/batch.rb +705 -0
  185. data/lib/stride_align/core.rb +180 -0
  186. data/lib/stride_align/keyboard.rb +200 -0
  187. data/lib/stride_align/matrices.rb +403 -0
  188. data/lib/stride_align/version.rb +5 -0
  189. data/lib/stride_align.rb +87 -0
  190. metadata +231 -0
@@ -0,0 +1,18 @@
1
+ /*
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
8
+ *
9
+ * https://www.apache.org/licenses/LICENSE-2.0
10
+ *
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.
16
+ */
17
+
18
+ #include gen_approx_french
@@ -0,0 +1,18 @@
1
+ /*
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
8
+ *
9
+ * https://www.apache.org/licenses/LICENSE-2.0
10
+ *
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.
16
+ */
17
+
18
+ #include gen_approx_french
@@ -0,0 +1,84 @@
1
+ /*
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
8
+ *
9
+ * https://www.apache.org/licenses/LICENSE-2.0
10
+ *
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.
16
+ */
17
+
18
+
19
+ "aiB" "" "[bp]" "(D|Dm)"
20
+ "oiB" "" "[bp]" "(D|Dm)"
21
+ "uiB" "" "[bp]" "(D|Dm)"
22
+ "eiB" "" "[bp]" "(D|Dm)"
23
+ "EiB" "" "[bp]" "(D|Dm)"
24
+ "iiB" "" "[bp]" "(D|Dm)"
25
+ "IiB" "" "[bp]" "(D|Dm)"
26
+
27
+ "aiB" "" "[dgkstvz]" "(D|Dn)"
28
+ "oiB" "" "[dgkstvz]" "(D|Dn)"
29
+ "uiB" "" "[dgkstvz]" "(D|Dn)"
30
+ "eiB" "" "[dgkstvz]" "(D|Dn)"
31
+ "EiB" "" "[dgkstvz]" "(D|Dn)"
32
+ "iiB" "" "[dgkstvz]" "(D|Dn)"
33
+ "IiB" "" "[dgkstvz]" "(D|Dn)"
34
+
35
+ "B" "" "[bp]" "(o|om|im)"
36
+ "B" "" "[dgkstvz]" "(o|on|in)"
37
+ "B" "" "" "o"
38
+
39
+ "aiF" "" "[bp]" "(D|Dm)"
40
+ "oiF" "" "[bp]" "(D|Dm)"
41
+ "uiF" "" "[bp]" "(D|Dm)"
42
+ "eiF" "" "[bp]" "(D|Dm)"
43
+ "EiF" "" "[bp]" "(D|Dm)"
44
+ "iiF" "" "[bp]" "(D|Dm)"
45
+ "IiF" "" "[bp]" "(D|Dm)"
46
+
47
+ "aiF" "" "[dgkstvz]" "(D|Dn)"
48
+ "oiF" "" "[dgkstvz]" "(D|Dn)"
49
+ "uiF" "" "[dgkstvz]" "(D|Dn)"
50
+ "eiF" "" "[dgkstvz]" "(D|Dn)"
51
+ "EiF" "" "[dgkstvz]" "(D|Dn)"
52
+ "iiF" "" "[dgkstvz]" "(D|Dn)"
53
+ "IiF" "" "[dgkstvz]" "(D|Dn)"
54
+
55
+ "F" "" "[bp]" "(i|im|om)"
56
+ "F" "" "[dgkstvz]" "(i|in|on)"
57
+ "F" "" "" "i"
58
+
59
+ "P" "" "" "(o|u)"
60
+
61
+ "I" "" "$" "i"
62
+ "I" "" "[^k]$" "i"
63
+ "Ik" "[lr]" "$" "(ik|Qk)"
64
+ "Ik" "" "$" "ik"
65
+ "sIts" "" "$" "(sits|sQts)"
66
+ "Its" "" "$" "its"
67
+ "I" "[aeiAEBFIou]" "" "i"
68
+ "I" "" "" "(i|Q)"
69
+
70
+ "au" "" "" "(D|a|u)"
71
+ "ou" "" "" "(D|o|u)"
72
+ "ai" "" "" "(D|a|i)"
73
+ "oi" "" "" "(D|o|i)"
74
+ "ui" "" "" "(D|u|i)"
75
+
76
+ "a" "" "" "(a|o)"
77
+ "e" "" "" "i"
78
+
79
+ "E" "" "[fklmnprst]$" "i"
80
+ "E" "" "ts$" "i"
81
+ "E" "" "$" "i"
82
+ "E" "[DaoiuQ]" "" "i"
83
+ "E" "" "[aoQ]" "i"
84
+ "E" "" "" "(Y|i)"
@@ -0,0 +1,18 @@
1
+ /*
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
8
+ *
9
+ * https://www.apache.org/licenses/LICENSE-2.0
10
+ *
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.
16
+ */
17
+
18
+ #include gen_approx_french
@@ -0,0 +1,18 @@
1
+ /*
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
8
+ *
9
+ * https://www.apache.org/licenses/LICENSE-2.0
10
+ *
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.
16
+ */
17
+
18
+ #include gen_approx_polish
@@ -0,0 +1,48 @@
1
+ /*
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
8
+ *
9
+ * https://www.apache.org/licenses/LICENSE-2.0
10
+ *
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.
16
+ */
17
+
18
+ // VOWELS
19
+ "I" "" "$" "i"
20
+ "I" "" "[^k]$" "i"
21
+ "Ik" "[lr]" "$" "(ik|Qk)"
22
+ "Ik" "" "$" "ik"
23
+ "sIts" "" "$" "(sits|sQts)"
24
+ "Its" "" "$" "its"
25
+ "I" "[aeiEIou]" "" "i"
26
+ "I" "" "" "(i|Q)"
27
+
28
+ "au" "" "" "(D|a|u)"
29
+ "ou" "" "" "(D|o|u)"
30
+ "ai" "" "" "(D|a|i)"
31
+ "oi" "" "" "(D|o|i)"
32
+ "ui" "" "" "(D|u|i)"
33
+
34
+ "om" "" "[bp]" "(om|im)"
35
+ "on" "" "[dgkstvz]" "(on|in)"
36
+ "em" "" "[bp]" "(im|om)"
37
+ "en" "" "[dgkstvz]" "(in|on)"
38
+ "Em" "" "[bp]" "(im|Ym|om)"
39
+ "En" "" "[dgkstvz]" "(in|Yn|on)"
40
+
41
+ "a" "" "" "(a|o)"
42
+ "e" "" "" "i"
43
+
44
+ "E" "" "[fklmnprsStv]$" "i"
45
+ "E" "" "ts$" "i"
46
+ "E" "[DaoiuQ]" "" "i"
47
+ "E" "" "[aoQ]" "i"
48
+ "E" "" "" "(Y|i)"
@@ -0,0 +1,21 @@
1
+ /*
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
8
+ *
9
+ * https://www.apache.org/licenses/LICENSE-2.0
10
+ *
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.
16
+ */
17
+
18
+ #include gen_approx_french
19
+
20
+ "B" "" "" "(b|v)"
21
+ "V" "" "" "(b|v)"
@@ -0,0 +1,18 @@
1
+ /*
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
8
+ *
9
+ * https://www.apache.org/licenses/LICENSE-2.0
10
+ *
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.
16
+ */
17
+
18
+ #include gen_approx_french
@@ -0,0 +1,40 @@
1
+ /*
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
8
+ *
9
+ * https://www.apache.org/licenses/LICENSE-2.0
10
+ *
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.
16
+ */
17
+
18
+ // GENERAL
19
+ // A, E, I, O, P, U should create variants,
20
+ // EE = final "e" (english & french)
21
+ // V, B from Spanish
22
+ // but a, e, i, o, u should not create any new variant
23
+ "EE" "" "$" "e"
24
+
25
+ "A" "" "" "a"
26
+ "E" "" "" "e"
27
+ "I" "" "" "i"
28
+ "O" "" "" "o"
29
+ "P" "" "" "o"
30
+ "U" "" "" "u"
31
+
32
+ "B" "" "[fktSs]" "p"
33
+ "B" "" "p" ""
34
+ "B" "" "$" "p"
35
+ "V" "" "[pktSs]" "f"
36
+ "V" "" "f" ""
37
+ "V" "" "$" "f"
38
+
39
+ "B" "" "" "b"
40
+ "V" "" "" "v"
@@ -0,0 +1,79 @@
1
+ /*
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
8
+ *
9
+ * https://www.apache.org/licenses/LICENSE-2.0
10
+ *
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.
16
+ */
17
+
18
+ // GENERAL
19
+ "h" "" "$" ""
20
+
21
+ // VOICED - UNVOICED CONSONANTS
22
+ "b" "" "[fktSs]" "p"
23
+ "b" "" "p" ""
24
+ "b" "" "$" "p"
25
+ "p" "" "[vgdZz]" "b" // Ashk: "v" excluded (everythere)
26
+ "p" "" "b" ""
27
+
28
+ "v" "" "[pktSs]" "f"
29
+ "v" "" "f" ""
30
+ "v" "" "$" "f"
31
+ "f" "" "[vbgdZz]" "v"
32
+ "f" "" "v" ""
33
+
34
+ "g" "" "[pftSs]" "k"
35
+ "g" "" "k" ""
36
+ "g" "" "$" "k"
37
+ "k" "" "[vbdZz]" "g"
38
+ "k" "" "g" ""
39
+
40
+ "d" "" "[pfkSs]" "t"
41
+ "d" "" "t" ""
42
+ "d" "" "$" "t"
43
+ "t" "" "[vbgZz]" "d"
44
+ "t" "" "d" ""
45
+
46
+ "s" "" "dZ" ""
47
+ "s" "" "tS" ""
48
+
49
+ "z" "" "[pfkSt]" "s"
50
+ "z" "" "[sSzZ]" ""
51
+ "s" "" "[sSzZ]" ""
52
+ "Z" "" "[sSzZ]" ""
53
+ "S" "" "[sSzZ]" ""
54
+
55
+ // SIMPLIFICATION OF CONSONANT CLUSTERS
56
+ "jnm" "" "" "jm"
57
+
58
+ // DOUBLE --> SINGLE
59
+ "ji" "^" "" "i"
60
+ "jI" "^" "" "I"
61
+
62
+ "a" "" "[aA]" ""
63
+ "a" "A" "" ""
64
+ "A" "" "A" ""
65
+
66
+ "b" "" "b" ""
67
+ "d" "" "d" ""
68
+ "f" "" "f" ""
69
+ "g" "" "g" ""
70
+ "j" "" "j" ""
71
+ "k" "" "k" ""
72
+ "l" "" "l" ""
73
+ "m" "" "m" ""
74
+ "n" "" "n" ""
75
+ "p" "" "p" ""
76
+ "r" "" "r" ""
77
+ "t" "" "t" ""
78
+ "v" "" "v" ""
79
+ "z" "" "z" ""
@@ -0,0 +1,18 @@
1
+ /*
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
8
+ *
9
+ * https://www.apache.org/licenses/LICENSE-2.0
10
+ *
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.
16
+ */
17
+
18
+ "1" "" "" ""
@@ -0,0 +1,32 @@
1
+ /*
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
8
+ *
9
+ * https://www.apache.org/licenses/LICENSE-2.0
10
+ *
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.
16
+ */
17
+
18
+ #include gen_exact_approx_common
19
+
20
+ "H" "" "" ""
21
+
22
+ // VOICED - UNVOICED CONSONANTS
23
+ "s" "[^t]" "[bgZd]" "z"
24
+ "Z" "" "[pfkst]" "S"
25
+ "Z" "" "$" "S"
26
+ "S" "" "[bgzd]" "Z"
27
+ "z" "" "$" "s"
28
+
29
+ "ji" "[aAoOeEiIuU]" "" "j"
30
+ "jI" "[aAoOeEiIuU]" "" "j"
31
+ "je" "[aAoOeEiIuU]" "" "j"
32
+ "jE" "[aAoOeEiIuU]" "" "j"
@@ -0,0 +1,18 @@
1
+ /*
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
8
+ *
9
+ * https://www.apache.org/licenses/LICENSE-2.0
10
+ *
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.
16
+ */
17
+
18
+ #include gen_exact_russian
@@ -0,0 +1,18 @@
1
+ /*
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
8
+ *
9
+ * https://www.apache.org/licenses/LICENSE-2.0
10
+ *
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.
16
+ */
17
+
18
+ #include gen_exact_russian
@@ -0,0 +1,18 @@
1
+ /*
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
8
+ *
9
+ * https://www.apache.org/licenses/LICENSE-2.0
10
+ *
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.
16
+ */
17
+
18
+ // empty
@@ -0,0 +1,18 @@
1
+ /*
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
8
+ *
9
+ * https://www.apache.org/licenses/LICENSE-2.0
10
+ *
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.
16
+ */
17
+
18
+ #include gen_exact_russian
@@ -0,0 +1,18 @@
1
+ /*
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
8
+ *
9
+ * https://www.apache.org/licenses/LICENSE-2.0
10
+ *
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.
16
+ */
17
+
18
+ // empty
@@ -0,0 +1,18 @@
1
+ /*
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
8
+ *
9
+ * https://www.apache.org/licenses/LICENSE-2.0
10
+ *
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.
16
+ */
17
+
18
+ #include gen_exact_any
@@ -0,0 +1,18 @@
1
+ /*
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
8
+ *
9
+ * https://www.apache.org/licenses/LICENSE-2.0
10
+ *
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.
16
+ */
17
+
18
+ // empty
@@ -0,0 +1,18 @@
1
+ /*
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
8
+ *
9
+ * https://www.apache.org/licenses/LICENSE-2.0
10
+ *
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.
16
+ */
17
+
18
+ "N" "" "" "n"
@@ -0,0 +1,18 @@
1
+ /*
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
8
+ *
9
+ * https://www.apache.org/licenses/LICENSE-2.0
10
+ *
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.
16
+ */
17
+
18
+ // empty
@@ -0,0 +1,18 @@
1
+ /*
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
8
+ *
9
+ * https://www.apache.org/licenses/LICENSE-2.0
10
+ *
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.
16
+ */
17
+
18
+ // empty
@@ -0,0 +1,18 @@
1
+ /*
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
8
+ *
9
+ * https://www.apache.org/licenses/LICENSE-2.0
10
+ *
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.
16
+ */
17
+
18
+ // empty