Project

General

Profile

« Previous | Next » 

Revision 7572

Added by Jing Tao over 11 years ago

Add the necessary files for the solr home.

View differences:

metacat-index/src/main/resources/solr-home/conf/protwords.txt
1
# The ASF licenses this file to You under the Apache License, Version 2.0
2
# (the "License"); you may not use this file except in compliance with
3
# the License.  You may obtain a copy of the License at
4
#
5
#     http://www.apache.org/licenses/LICENSE-2.0
6
#
7
# Unless required by applicable law or agreed to in writing, software
8
# distributed under the License is distributed on an "AS IS" BASIS,
9
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
10
# See the License for the specific language governing permissions and
11
# limitations under the License.
12

  
13
#-----------------------------------------------------------------------
14
# Use a protected word file to protect against the stemmer reducing two
15
# unrelated words to the same base word.
16

  
17
# Some non-words that normally won't be encountered,
18
# just to test that they won't be stemmed.
19
dontstems
20
zwhacky
21

  
metacat-index/src/main/resources/solr-home/conf/stopwords.txt
1
# Licensed to the Apache Software Foundation (ASF) under one or more
2
# contributor license agreements.  See the NOTICE file distributed with
3
# this work for additional information regarding copyright ownership.
4
# The ASF licenses this file to You under the Apache License, Version 2.0
5
# (the "License"); you may not use this file except in compliance with
6
# the License.  You may obtain a copy of the License at
7
#
8
#     http://www.apache.org/licenses/LICENSE-2.0
9
#
10
# Unless required by applicable law or agreed to in writing, software
11
# distributed under the License is distributed on an "AS IS" BASIS,
12
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
13
# See the License for the specific language governing permissions and
14
# limitations under the License.
metacat-index/src/main/resources/solr-home/conf/elevate.xml
1
<?xml version="1.0" encoding="UTF-8" ?>
2
<!--
3
 Licensed to the Apache Software Foundation (ASF) under one or more
4
 contributor license agreements.  See the NOTICE file distributed with
5
 this work for additional information regarding copyright ownership.
6
 The ASF licenses this file to You under the Apache License, Version 2.0
7
 (the "License"); you may not use this file except in compliance with
8
 the License.  You may obtain a copy of the License at
9

  
10
     http://www.apache.org/licenses/LICENSE-2.0
11

  
12
 Unless required by applicable law or agreed to in writing, software
13
 distributed under the License is distributed on an "AS IS" BASIS,
14
 WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
15
 See the License for the specific language governing permissions and
16
 limitations under the License.
17
-->
18

  
19
<!-- If this file is found in the config directory, it will only be
20
     loaded once at startup.  If it is found in Solr's data
21
     directory, it will be re-loaded every commit.
22
-->
23

  
24
<elevate>
25
 <query text="foo bar">
26
  <doc id="1" />
27
  <doc id="2" />
28
  <doc id="3" />
29
 </query>
30
 
31
 <query text="ipod">
32
   <doc id="MA147LL/A" />  <!-- put the actual ipod at the top -->
33
   <doc id="IW-02" exclude="true" /> <!-- exclude this cable -->
34
 </query>
35
 
36
</elevate>
metacat-index/src/main/resources/solr-home/conf/stopwords_en.txt
1
# Licensed to the Apache Software Foundation (ASF) under one or more
2
# contributor license agreements.  See the NOTICE file distributed with
3
# this work for additional information regarding copyright ownership.
4
# The ASF licenses this file to You under the Apache License, Version 2.0
5
# (the "License"); you may not use this file except in compliance with
6
# the License.  You may obtain a copy of the License at
7
#
8
#     http://www.apache.org/licenses/LICENSE-2.0
9
#
10
# Unless required by applicable law or agreed to in writing, software
11
# distributed under the License is distributed on an "AS IS" BASIS,
12
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
13
# See the License for the specific language governing permissions and
14
# limitations under the License.
15

  
16
# a couple of test stopwords to test that the words are really being
17
# configured from this file:
18
stopworda
19
stopwordb
20

  
21
# Standard english stop words taken from Lucene's StopAnalyzer
22
a
23
an
24
and
25
are
26
as
27
at
28
be
29
but
30
by
31
for
32
if
33
in
34
into
35
is
36
it
37
no
38
not
39
of
40
on
41
or
42
such
43
that
44
the
45
their
46
then
47
there
48
these
49
they
50
this
51
to
52
was
53
will
54
with
metacat-index/src/main/resources/solr-home/conf/mapping-FoldToASCII.txt
1
# The ASF licenses this file to You under the Apache License, Version 2.0
2
# (the "License"); you may not use this file except in compliance with
3
# the License.  You may obtain a copy of the License at
4
#
5
#     http://www.apache.org/licenses/LICENSE-2.0
6
#
7
# Unless required by applicable law or agreed to in writing, software
8
# distributed under the License is distributed on an "AS IS" BASIS,
9
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
10
# See the License for the specific language governing permissions and
11
# limitations under the License.
12

  
13

  
14
# This map converts alphabetic, numeric, and symbolic Unicode characters
15
# which are not in the first 127 ASCII characters (the "Basic Latin" Unicode
16
# block) into their ASCII equivalents, if one exists.
17
#
18
# Characters from the following Unicode blocks are converted; however, only
19
# those characters with reasonable ASCII alternatives are converted:
20
#
21
# - C1 Controls and Latin-1 Supplement: http://www.unicode.org/charts/PDF/U0080.pdf
22
# - Latin Extended-A: http://www.unicode.org/charts/PDF/U0100.pdf
23
# - Latin Extended-B: http://www.unicode.org/charts/PDF/U0180.pdf
24
# - Latin Extended Additional: http://www.unicode.org/charts/PDF/U1E00.pdf
25
# - Latin Extended-C: http://www.unicode.org/charts/PDF/U2C60.pdf
26
# - Latin Extended-D: http://www.unicode.org/charts/PDF/UA720.pdf
27
# - IPA Extensions: http://www.unicode.org/charts/PDF/U0250.pdf
28
# - Phonetic Extensions: http://www.unicode.org/charts/PDF/U1D00.pdf
29
# - Phonetic Extensions Supplement: http://www.unicode.org/charts/PDF/U1D80.pdf
30
# - General Punctuation: http://www.unicode.org/charts/PDF/U2000.pdf
31
# - Superscripts and Subscripts: http://www.unicode.org/charts/PDF/U2070.pdf
32
# - Enclosed Alphanumerics: http://www.unicode.org/charts/PDF/U2460.pdf
33
# - Dingbats: http://www.unicode.org/charts/PDF/U2700.pdf
34
# - Supplemental Punctuation: http://www.unicode.org/charts/PDF/U2E00.pdf
35
# - Alphabetic Presentation Forms: http://www.unicode.org/charts/PDF/UFB00.pdf
36
# - Halfwidth and Fullwidth Forms: http://www.unicode.org/charts/PDF/UFF00.pdf
37
#  
38
# See: http://en.wikipedia.org/wiki/Latin_characters_in_Unicode
39
#
40
# The set of character conversions supported by this map is a superset of
41
# those supported by the map represented by mapping-ISOLatin1Accent.txt.
42
#
43
# See the bottom of this file for the Perl script used to generate the contents
44
# of this file (without this header) from ASCIIFoldingFilter.java.
45

  
46

  
47
# Syntax:
48
#   "source" => "target"
49
#     "source".length() > 0 (source cannot be empty.)
50
#     "target".length() >= 0 (target can be empty.)
51

  
52

  
53
# À  [LATIN CAPITAL LETTER A WITH GRAVE]
54
"\u00C0" => "A"
55

  
56
# Á  [LATIN CAPITAL LETTER A WITH ACUTE]
57
"\u00C1" => "A"
58

  
59
# Â  [LATIN CAPITAL LETTER A WITH CIRCUMFLEX]
60
"\u00C2" => "A"
61

  
62
# Ã  [LATIN CAPITAL LETTER A WITH TILDE]
63
"\u00C3" => "A"
64

  
65
# Ä  [LATIN CAPITAL LETTER A WITH DIAERESIS]
66
"\u00C4" => "A"
67

  
68
# Å  [LATIN CAPITAL LETTER A WITH RING ABOVE]
69
"\u00C5" => "A"
70

  
71
# Ā  [LATIN CAPITAL LETTER A WITH MACRON]
72
"\u0100" => "A"
73

  
74
# Ă  [LATIN CAPITAL LETTER A WITH BREVE]
75
"\u0102" => "A"
76

  
77
# Ą  [LATIN CAPITAL LETTER A WITH OGONEK]
78
"\u0104" => "A"
79

  
80
# Ə  http://en.wikipedia.org/wiki/Schwa  [LATIN CAPITAL LETTER SCHWA]
81
"\u018F" => "A"
82

  
83
# Ǎ  [LATIN CAPITAL LETTER A WITH CARON]
84
"\u01CD" => "A"
85

  
86
# Ǟ  [LATIN CAPITAL LETTER A WITH DIAERESIS AND MACRON]
87
"\u01DE" => "A"
88

  
89
# Ǡ  [LATIN CAPITAL LETTER A WITH DOT ABOVE AND MACRON]
90
"\u01E0" => "A"
91

  
92
# Ǻ  [LATIN CAPITAL LETTER A WITH RING ABOVE AND ACUTE]
93
"\u01FA" => "A"
94

  
95
# Ȁ  [LATIN CAPITAL LETTER A WITH DOUBLE GRAVE]
96
"\u0200" => "A"
97

  
98
# Ȃ  [LATIN CAPITAL LETTER A WITH INVERTED BREVE]
99
"\u0202" => "A"
100

  
101
# Ȧ  [LATIN CAPITAL LETTER A WITH DOT ABOVE]
102
"\u0226" => "A"
103

  
104
# Ⱥ  [LATIN CAPITAL LETTER A WITH STROKE]
105
"\u023A" => "A"
106

  
107
# ᴀ  [LATIN LETTER SMALL CAPITAL A]
108
"\u1D00" => "A"
109

  
110
# Ḁ  [LATIN CAPITAL LETTER A WITH RING BELOW]
111
"\u1E00" => "A"
112

  
113
# Ạ  [LATIN CAPITAL LETTER A WITH DOT BELOW]
114
"\u1EA0" => "A"
115

  
116
# Ả  [LATIN CAPITAL LETTER A WITH HOOK ABOVE]
117
"\u1EA2" => "A"
118

  
119
# Ấ  [LATIN CAPITAL LETTER A WITH CIRCUMFLEX AND ACUTE]
120
"\u1EA4" => "A"
121

  
122
# Ầ  [LATIN CAPITAL LETTER A WITH CIRCUMFLEX AND GRAVE]
123
"\u1EA6" => "A"
124

  
125
# Ẩ  [LATIN CAPITAL LETTER A WITH CIRCUMFLEX AND HOOK ABOVE]
126
"\u1EA8" => "A"
127

  
128
# Ẫ  [LATIN CAPITAL LETTER A WITH CIRCUMFLEX AND TILDE]
129
"\u1EAA" => "A"
130

  
131
# Ậ  [LATIN CAPITAL LETTER A WITH CIRCUMFLEX AND DOT BELOW]
132
"\u1EAC" => "A"
133

  
134
# Ắ  [LATIN CAPITAL LETTER A WITH BREVE AND ACUTE]
135
"\u1EAE" => "A"
136

  
137
# Ằ  [LATIN CAPITAL LETTER A WITH BREVE AND GRAVE]
138
"\u1EB0" => "A"
139

  
140
# Ẳ  [LATIN CAPITAL LETTER A WITH BREVE AND HOOK ABOVE]
141
"\u1EB2" => "A"
142

  
143
# Ẵ  [LATIN CAPITAL LETTER A WITH BREVE AND TILDE]
144
"\u1EB4" => "A"
145

  
146
# Ặ  [LATIN CAPITAL LETTER A WITH BREVE AND DOT BELOW]
147
"\u1EB6" => "A"
148

  
149
# Ⓐ  [CIRCLED LATIN CAPITAL LETTER A]
150
"\u24B6" => "A"
151

  
152
# A  [FULLWIDTH LATIN CAPITAL LETTER A]
153
"\uFF21" => "A"
154

  
155
# à  [LATIN SMALL LETTER A WITH GRAVE]
156
"\u00E0" => "a"
157

  
158
# á  [LATIN SMALL LETTER A WITH ACUTE]
159
"\u00E1" => "a"
160

  
161
# â  [LATIN SMALL LETTER A WITH CIRCUMFLEX]
162
"\u00E2" => "a"
163

  
164
# ã  [LATIN SMALL LETTER A WITH TILDE]
165
"\u00E3" => "a"
166

  
167
# ä  [LATIN SMALL LETTER A WITH DIAERESIS]
168
"\u00E4" => "a"
169

  
170
# å  [LATIN SMALL LETTER A WITH RING ABOVE]
171
"\u00E5" => "a"
172

  
173
# ā  [LATIN SMALL LETTER A WITH MACRON]
174
"\u0101" => "a"
175

  
176
# ă  [LATIN SMALL LETTER A WITH BREVE]
177
"\u0103" => "a"
178

  
179
# ą  [LATIN SMALL LETTER A WITH OGONEK]
180
"\u0105" => "a"
181

  
182
# ǎ  [LATIN SMALL LETTER A WITH CARON]
183
"\u01CE" => "a"
184

  
185
# ǟ  [LATIN SMALL LETTER A WITH DIAERESIS AND MACRON]
186
"\u01DF" => "a"
187

  
188
# ǡ  [LATIN SMALL LETTER A WITH DOT ABOVE AND MACRON]
189
"\u01E1" => "a"
190

  
191
# ǻ  [LATIN SMALL LETTER A WITH RING ABOVE AND ACUTE]
192
"\u01FB" => "a"
193

  
194
# ȁ  [LATIN SMALL LETTER A WITH DOUBLE GRAVE]
195
"\u0201" => "a"
196

  
197
# ȃ  [LATIN SMALL LETTER A WITH INVERTED BREVE]
198
"\u0203" => "a"
199

  
200
# ȧ  [LATIN SMALL LETTER A WITH DOT ABOVE]
201
"\u0227" => "a"
202

  
203
# ɐ  [LATIN SMALL LETTER TURNED A]
204
"\u0250" => "a"
205

  
206
# ə  [LATIN SMALL LETTER SCHWA]
207
"\u0259" => "a"
208

  
209
# ɚ  [LATIN SMALL LETTER SCHWA WITH HOOK]
210
"\u025A" => "a"
211

  
212
# ᶏ  [LATIN SMALL LETTER A WITH RETROFLEX HOOK]
213
"\u1D8F" => "a"
214

  
215
# ᶕ  [LATIN SMALL LETTER SCHWA WITH RETROFLEX HOOK]
216
"\u1D95" => "a"
217

  
218
# ạ  [LATIN SMALL LETTER A WITH RING BELOW]
219
"\u1E01" => "a"
220

  
221
# ả  [LATIN SMALL LETTER A WITH RIGHT HALF RING]
222
"\u1E9A" => "a"
223

  
224
# ạ  [LATIN SMALL LETTER A WITH DOT BELOW]
225
"\u1EA1" => "a"
226

  
227
# ả  [LATIN SMALL LETTER A WITH HOOK ABOVE]
228
"\u1EA3" => "a"
229

  
230
# ấ  [LATIN SMALL LETTER A WITH CIRCUMFLEX AND ACUTE]
231
"\u1EA5" => "a"
232

  
233
# ầ  [LATIN SMALL LETTER A WITH CIRCUMFLEX AND GRAVE]
234
"\u1EA7" => "a"
235

  
236
# ẩ  [LATIN SMALL LETTER A WITH CIRCUMFLEX AND HOOK ABOVE]
237
"\u1EA9" => "a"
238

  
239
# ẫ  [LATIN SMALL LETTER A WITH CIRCUMFLEX AND TILDE]
240
"\u1EAB" => "a"
241

  
242
# ậ  [LATIN SMALL LETTER A WITH CIRCUMFLEX AND DOT BELOW]
243
"\u1EAD" => "a"
244

  
245
# ắ  [LATIN SMALL LETTER A WITH BREVE AND ACUTE]
246
"\u1EAF" => "a"
247

  
248
# ằ  [LATIN SMALL LETTER A WITH BREVE AND GRAVE]
249
"\u1EB1" => "a"
250

  
251
# ẳ  [LATIN SMALL LETTER A WITH BREVE AND HOOK ABOVE]
252
"\u1EB3" => "a"
253

  
254
# ẵ  [LATIN SMALL LETTER A WITH BREVE AND TILDE]
255
"\u1EB5" => "a"
256

  
257
# ặ  [LATIN SMALL LETTER A WITH BREVE AND DOT BELOW]
258
"\u1EB7" => "a"
259

  
260
# ₐ  [LATIN SUBSCRIPT SMALL LETTER A]
261
"\u2090" => "a"
262

  
263
# ₔ  [LATIN SUBSCRIPT SMALL LETTER SCHWA]
264
"\u2094" => "a"
265

  
266
# ⓐ  [CIRCLED LATIN SMALL LETTER A]
267
"\u24D0" => "a"
268

  
269
# ⱥ  [LATIN SMALL LETTER A WITH STROKE]
270
"\u2C65" => "a"
271

  
272
# Ɐ  [LATIN CAPITAL LETTER TURNED A]
273
"\u2C6F" => "a"
274

  
275
# a  [FULLWIDTH LATIN SMALL LETTER A]
276
"\uFF41" => "a"
277

  
278
# Ꜳ  [LATIN CAPITAL LETTER AA]
279
"\uA732" => "AA"
280

  
281
# Æ  [LATIN CAPITAL LETTER AE]
282
"\u00C6" => "AE"
283

  
284
# Ǣ  [LATIN CAPITAL LETTER AE WITH MACRON]
285
"\u01E2" => "AE"
286

  
287
# Ǽ  [LATIN CAPITAL LETTER AE WITH ACUTE]
288
"\u01FC" => "AE"
289

  
290
# ᴁ  [LATIN LETTER SMALL CAPITAL AE]
291
"\u1D01" => "AE"
292

  
293
# Ꜵ  [LATIN CAPITAL LETTER AO]
294
"\uA734" => "AO"
295

  
296
# Ꜷ  [LATIN CAPITAL LETTER AU]
297
"\uA736" => "AU"
298

  
299
# Ꜹ  [LATIN CAPITAL LETTER AV]
300
"\uA738" => "AV"
301

  
302
# Ꜻ  [LATIN CAPITAL LETTER AV WITH HORIZONTAL BAR]
303
"\uA73A" => "AV"
304

  
305
# Ꜽ  [LATIN CAPITAL LETTER AY]
306
"\uA73C" => "AY"
307

  
308
# ⒜  [PARENTHESIZED LATIN SMALL LETTER A]
309
"\u249C" => "(a)"
310

  
311
# ꜳ  [LATIN SMALL LETTER AA]
312
"\uA733" => "aa"
313

  
314
# æ  [LATIN SMALL LETTER AE]
315
"\u00E6" => "ae"
316

  
317
# ǣ  [LATIN SMALL LETTER AE WITH MACRON]
318
"\u01E3" => "ae"
319

  
320
# ǽ  [LATIN SMALL LETTER AE WITH ACUTE]
321
"\u01FD" => "ae"
322

  
323
# ᴂ  [LATIN SMALL LETTER TURNED AE]
324
"\u1D02" => "ae"
325

  
326
# ꜵ  [LATIN SMALL LETTER AO]
327
"\uA735" => "ao"
328

  
329
# ꜷ  [LATIN SMALL LETTER AU]
330
"\uA737" => "au"
331

  
332
# ꜹ  [LATIN SMALL LETTER AV]
333
"\uA739" => "av"
334

  
335
# ꜻ  [LATIN SMALL LETTER AV WITH HORIZONTAL BAR]
336
"\uA73B" => "av"
337

  
338
# ꜽ  [LATIN SMALL LETTER AY]
339
"\uA73D" => "ay"
340

  
341
# Ɓ  [LATIN CAPITAL LETTER B WITH HOOK]
342
"\u0181" => "B"
343

  
344
# Ƃ  [LATIN CAPITAL LETTER B WITH TOPBAR]
345
"\u0182" => "B"
346

  
347
# Ƀ  [LATIN CAPITAL LETTER B WITH STROKE]
348
"\u0243" => "B"
349

  
350
# ʙ  [LATIN LETTER SMALL CAPITAL B]
351
"\u0299" => "B"
352

  
353
# ᴃ  [LATIN LETTER SMALL CAPITAL BARRED B]
354
"\u1D03" => "B"
355

  
356
# Ḃ  [LATIN CAPITAL LETTER B WITH DOT ABOVE]
357
"\u1E02" => "B"
358

  
359
# Ḅ  [LATIN CAPITAL LETTER B WITH DOT BELOW]
360
"\u1E04" => "B"
361

  
362
# Ḇ  [LATIN CAPITAL LETTER B WITH LINE BELOW]
363
"\u1E06" => "B"
364

  
365
# Ⓑ  [CIRCLED LATIN CAPITAL LETTER B]
366
"\u24B7" => "B"
367

  
368
# B  [FULLWIDTH LATIN CAPITAL LETTER B]
369
"\uFF22" => "B"
370

  
371
# ƀ  [LATIN SMALL LETTER B WITH STROKE]
372
"\u0180" => "b"
373

  
374
# ƃ  [LATIN SMALL LETTER B WITH TOPBAR]
375
"\u0183" => "b"
376

  
377
# ɓ  [LATIN SMALL LETTER B WITH HOOK]
378
"\u0253" => "b"
379

  
380
# ᵬ  [LATIN SMALL LETTER B WITH MIDDLE TILDE]
381
"\u1D6C" => "b"
382

  
383
# ᶀ  [LATIN SMALL LETTER B WITH PALATAL HOOK]
384
"\u1D80" => "b"
385

  
386
# ḃ  [LATIN SMALL LETTER B WITH DOT ABOVE]
387
"\u1E03" => "b"
388

  
389
# ḅ  [LATIN SMALL LETTER B WITH DOT BELOW]
390
"\u1E05" => "b"
391

  
392
# ḇ  [LATIN SMALL LETTER B WITH LINE BELOW]
393
"\u1E07" => "b"
394

  
395
# ⓑ  [CIRCLED LATIN SMALL LETTER B]
396
"\u24D1" => "b"
397

  
398
# b  [FULLWIDTH LATIN SMALL LETTER B]
399
"\uFF42" => "b"
400

  
401
# ⒝  [PARENTHESIZED LATIN SMALL LETTER B]
402
"\u249D" => "(b)"
403

  
404
# Ç  [LATIN CAPITAL LETTER C WITH CEDILLA]
405
"\u00C7" => "C"
406

  
407
# Ć  [LATIN CAPITAL LETTER C WITH ACUTE]
408
"\u0106" => "C"
409

  
410
# Ĉ  [LATIN CAPITAL LETTER C WITH CIRCUMFLEX]
411
"\u0108" => "C"
412

  
413
# Ċ  [LATIN CAPITAL LETTER C WITH DOT ABOVE]
414
"\u010A" => "C"
415

  
416
# Č  [LATIN CAPITAL LETTER C WITH CARON]
417
"\u010C" => "C"
418

  
419
# Ƈ  [LATIN CAPITAL LETTER C WITH HOOK]
420
"\u0187" => "C"
421

  
422
# Ȼ  [LATIN CAPITAL LETTER C WITH STROKE]
423
"\u023B" => "C"
424

  
425
# ʗ  [LATIN LETTER STRETCHED C]
426
"\u0297" => "C"
427

  
428
# ᴄ  [LATIN LETTER SMALL CAPITAL C]
429
"\u1D04" => "C"
430

  
431
# Ḉ  [LATIN CAPITAL LETTER C WITH CEDILLA AND ACUTE]
432
"\u1E08" => "C"
433

  
434
# Ⓒ  [CIRCLED LATIN CAPITAL LETTER C]
435
"\u24B8" => "C"
436

  
437
# C  [FULLWIDTH LATIN CAPITAL LETTER C]
438
"\uFF23" => "C"
439

  
440
# ç  [LATIN SMALL LETTER C WITH CEDILLA]
441
"\u00E7" => "c"
442

  
443
# ć  [LATIN SMALL LETTER C WITH ACUTE]
444
"\u0107" => "c"
445

  
446
# ĉ  [LATIN SMALL LETTER C WITH CIRCUMFLEX]
447
"\u0109" => "c"
448

  
449
# ċ  [LATIN SMALL LETTER C WITH DOT ABOVE]
450
"\u010B" => "c"
451

  
452
# č  [LATIN SMALL LETTER C WITH CARON]
453
"\u010D" => "c"
454

  
455
# ƈ  [LATIN SMALL LETTER C WITH HOOK]
456
"\u0188" => "c"
457

  
458
# ȼ  [LATIN SMALL LETTER C WITH STROKE]
459
"\u023C" => "c"
460

  
461
# ɕ  [LATIN SMALL LETTER C WITH CURL]
462
"\u0255" => "c"
463

  
464
# ḉ  [LATIN SMALL LETTER C WITH CEDILLA AND ACUTE]
465
"\u1E09" => "c"
466

  
467
# ↄ  [LATIN SMALL LETTER REVERSED C]
468
"\u2184" => "c"
469

  
470
# ⓒ  [CIRCLED LATIN SMALL LETTER C]
471
"\u24D2" => "c"
472

  
473
# Ꜿ  [LATIN CAPITAL LETTER REVERSED C WITH DOT]
474
"\uA73E" => "c"
475

  
476
# ꜿ  [LATIN SMALL LETTER REVERSED C WITH DOT]
477
"\uA73F" => "c"
478

  
479
# c  [FULLWIDTH LATIN SMALL LETTER C]
480
"\uFF43" => "c"
481

  
482
# ⒞  [PARENTHESIZED LATIN SMALL LETTER C]
483
"\u249E" => "(c)"
484

  
485
# Ð  [LATIN CAPITAL LETTER ETH]
486
"\u00D0" => "D"
487

  
488
# Ď  [LATIN CAPITAL LETTER D WITH CARON]
489
"\u010E" => "D"
490

  
491
# Đ  [LATIN CAPITAL LETTER D WITH STROKE]
492
"\u0110" => "D"
493

  
494
# Ɖ  [LATIN CAPITAL LETTER AFRICAN D]
495
"\u0189" => "D"
496

  
497
# Ɗ  [LATIN CAPITAL LETTER D WITH HOOK]
498
"\u018A" => "D"
499

  
500
# Ƌ  [LATIN CAPITAL LETTER D WITH TOPBAR]
501
"\u018B" => "D"
502

  
503
# ᴅ  [LATIN LETTER SMALL CAPITAL D]
504
"\u1D05" => "D"
505

  
506
# ᴆ  [LATIN LETTER SMALL CAPITAL ETH]
507
"\u1D06" => "D"
508

  
509
# Ḋ  [LATIN CAPITAL LETTER D WITH DOT ABOVE]
510
"\u1E0A" => "D"
511

  
512
# Ḍ  [LATIN CAPITAL LETTER D WITH DOT BELOW]
513
"\u1E0C" => "D"
514

  
515
# Ḏ  [LATIN CAPITAL LETTER D WITH LINE BELOW]
516
"\u1E0E" => "D"
517

  
518
# Ḑ  [LATIN CAPITAL LETTER D WITH CEDILLA]
519
"\u1E10" => "D"
520

  
521
# Ḓ  [LATIN CAPITAL LETTER D WITH CIRCUMFLEX BELOW]
522
"\u1E12" => "D"
523

  
524
# Ⓓ  [CIRCLED LATIN CAPITAL LETTER D]
525
"\u24B9" => "D"
526

  
527
# Ꝺ  [LATIN CAPITAL LETTER INSULAR D]
528
"\uA779" => "D"
529

  
530
# D  [FULLWIDTH LATIN CAPITAL LETTER D]
531
"\uFF24" => "D"
532

  
533
# ð  [LATIN SMALL LETTER ETH]
534
"\u00F0" => "d"
535

  
536
# ď  [LATIN SMALL LETTER D WITH CARON]
537
"\u010F" => "d"
538

  
539
# đ  [LATIN SMALL LETTER D WITH STROKE]
540
"\u0111" => "d"
541

  
542
# ƌ  [LATIN SMALL LETTER D WITH TOPBAR]
543
"\u018C" => "d"
544

  
545
# ȡ  [LATIN SMALL LETTER D WITH CURL]
546
"\u0221" => "d"
547

  
548
# ɖ  [LATIN SMALL LETTER D WITH TAIL]
549
"\u0256" => "d"
550

  
551
# ɗ  [LATIN SMALL LETTER D WITH HOOK]
552
"\u0257" => "d"
553

  
554
# ᵭ  [LATIN SMALL LETTER D WITH MIDDLE TILDE]
555
"\u1D6D" => "d"
556

  
557
# ᶁ  [LATIN SMALL LETTER D WITH PALATAL HOOK]
558
"\u1D81" => "d"
559

  
560
# ᶑ  [LATIN SMALL LETTER D WITH HOOK AND TAIL]
561
"\u1D91" => "d"
562

  
563
# ḋ  [LATIN SMALL LETTER D WITH DOT ABOVE]
564
"\u1E0B" => "d"
565

  
566
# ḍ  [LATIN SMALL LETTER D WITH DOT BELOW]
567
"\u1E0D" => "d"
568

  
569
# ḏ  [LATIN SMALL LETTER D WITH LINE BELOW]
570
"\u1E0F" => "d"
571

  
572
# ḑ  [LATIN SMALL LETTER D WITH CEDILLA]
573
"\u1E11" => "d"
574

  
575
# ḓ  [LATIN SMALL LETTER D WITH CIRCUMFLEX BELOW]
576
"\u1E13" => "d"
577

  
578
# ⓓ  [CIRCLED LATIN SMALL LETTER D]
579
"\u24D3" => "d"
580

  
581
# ꝺ  [LATIN SMALL LETTER INSULAR D]
582
"\uA77A" => "d"
583

  
584
# d  [FULLWIDTH LATIN SMALL LETTER D]
585
"\uFF44" => "d"
586

  
587
# DŽ  [LATIN CAPITAL LETTER DZ WITH CARON]
588
"\u01C4" => "DZ"
589

  
590
# DZ  [LATIN CAPITAL LETTER DZ]
591
"\u01F1" => "DZ"
592

  
593
# Dž  [LATIN CAPITAL LETTER D WITH SMALL LETTER Z WITH CARON]
594
"\u01C5" => "Dz"
595

  
596
# Dz  [LATIN CAPITAL LETTER D WITH SMALL LETTER Z]
597
"\u01F2" => "Dz"
598

  
599
# ⒟  [PARENTHESIZED LATIN SMALL LETTER D]
600
"\u249F" => "(d)"
601

  
602
# ȸ  [LATIN SMALL LETTER DB DIGRAPH]
603
"\u0238" => "db"
604

  
605
# dž  [LATIN SMALL LETTER DZ WITH CARON]
606
"\u01C6" => "dz"
607

  
608
# dz  [LATIN SMALL LETTER DZ]
609
"\u01F3" => "dz"
610

  
611
# ʣ  [LATIN SMALL LETTER DZ DIGRAPH]
612
"\u02A3" => "dz"
613

  
614
# ʥ  [LATIN SMALL LETTER DZ DIGRAPH WITH CURL]
615
"\u02A5" => "dz"
616

  
617
# È  [LATIN CAPITAL LETTER E WITH GRAVE]
618
"\u00C8" => "E"
619

  
620
# É  [LATIN CAPITAL LETTER E WITH ACUTE]
621
"\u00C9" => "E"
622

  
623
# Ê  [LATIN CAPITAL LETTER E WITH CIRCUMFLEX]
624
"\u00CA" => "E"
625

  
626
# Ë  [LATIN CAPITAL LETTER E WITH DIAERESIS]
627
"\u00CB" => "E"
628

  
629
# Ē  [LATIN CAPITAL LETTER E WITH MACRON]
630
"\u0112" => "E"
631

  
632
# Ĕ  [LATIN CAPITAL LETTER E WITH BREVE]
633
"\u0114" => "E"
634

  
635
# Ė  [LATIN CAPITAL LETTER E WITH DOT ABOVE]
636
"\u0116" => "E"
637

  
638
# Ę  [LATIN CAPITAL LETTER E WITH OGONEK]
639
"\u0118" => "E"
640

  
641
# Ě  [LATIN CAPITAL LETTER E WITH CARON]
642
"\u011A" => "E"
643

  
644
# Ǝ  [LATIN CAPITAL LETTER REVERSED E]
645
"\u018E" => "E"
646

  
647
# Ɛ  [LATIN CAPITAL LETTER OPEN E]
648
"\u0190" => "E"
649

  
650
# Ȅ  [LATIN CAPITAL LETTER E WITH DOUBLE GRAVE]
651
"\u0204" => "E"
652

  
653
# Ȇ  [LATIN CAPITAL LETTER E WITH INVERTED BREVE]
654
"\u0206" => "E"
655

  
656
# Ȩ  [LATIN CAPITAL LETTER E WITH CEDILLA]
657
"\u0228" => "E"
658

  
659
# Ɇ  [LATIN CAPITAL LETTER E WITH STROKE]
660
"\u0246" => "E"
661

  
662
# ᴇ  [LATIN LETTER SMALL CAPITAL E]
663
"\u1D07" => "E"
664

  
665
# Ḕ  [LATIN CAPITAL LETTER E WITH MACRON AND GRAVE]
666
"\u1E14" => "E"
667

  
668
# Ḗ  [LATIN CAPITAL LETTER E WITH MACRON AND ACUTE]
669
"\u1E16" => "E"
670

  
671
# Ḙ  [LATIN CAPITAL LETTER E WITH CIRCUMFLEX BELOW]
672
"\u1E18" => "E"
673

  
674
# Ḛ  [LATIN CAPITAL LETTER E WITH TILDE BELOW]
675
"\u1E1A" => "E"
676

  
677
# Ḝ  [LATIN CAPITAL LETTER E WITH CEDILLA AND BREVE]
678
"\u1E1C" => "E"
679

  
680
# Ẹ  [LATIN CAPITAL LETTER E WITH DOT BELOW]
681
"\u1EB8" => "E"
682

  
683
# Ẻ  [LATIN CAPITAL LETTER E WITH HOOK ABOVE]
684
"\u1EBA" => "E"
685

  
686
# Ẽ  [LATIN CAPITAL LETTER E WITH TILDE]
687
"\u1EBC" => "E"
688

  
689
# Ế  [LATIN CAPITAL LETTER E WITH CIRCUMFLEX AND ACUTE]
690
"\u1EBE" => "E"
691

  
692
# Ề  [LATIN CAPITAL LETTER E WITH CIRCUMFLEX AND GRAVE]
693
"\u1EC0" => "E"
694

  
695
# Ể  [LATIN CAPITAL LETTER E WITH CIRCUMFLEX AND HOOK ABOVE]
696
"\u1EC2" => "E"
697

  
698
# Ễ  [LATIN CAPITAL LETTER E WITH CIRCUMFLEX AND TILDE]
699
"\u1EC4" => "E"
700

  
701
# Ệ  [LATIN CAPITAL LETTER E WITH CIRCUMFLEX AND DOT BELOW]
702
"\u1EC6" => "E"
703

  
704
# Ⓔ  [CIRCLED LATIN CAPITAL LETTER E]
705
"\u24BA" => "E"
706

  
707
# ⱻ  [LATIN LETTER SMALL CAPITAL TURNED E]
708
"\u2C7B" => "E"
709

  
710
# E  [FULLWIDTH LATIN CAPITAL LETTER E]
711
"\uFF25" => "E"
712

  
713
# è  [LATIN SMALL LETTER E WITH GRAVE]
714
"\u00E8" => "e"
715

  
716
# é  [LATIN SMALL LETTER E WITH ACUTE]
717
"\u00E9" => "e"
718

  
719
# ê  [LATIN SMALL LETTER E WITH CIRCUMFLEX]
720
"\u00EA" => "e"
721

  
722
# ë  [LATIN SMALL LETTER E WITH DIAERESIS]
723
"\u00EB" => "e"
724

  
725
# ē  [LATIN SMALL LETTER E WITH MACRON]
726
"\u0113" => "e"
727

  
728
# ĕ  [LATIN SMALL LETTER E WITH BREVE]
729
"\u0115" => "e"
730

  
731
# ė  [LATIN SMALL LETTER E WITH DOT ABOVE]
732
"\u0117" => "e"
733

  
734
# ę  [LATIN SMALL LETTER E WITH OGONEK]
735
"\u0119" => "e"
736

  
737
# ě  [LATIN SMALL LETTER E WITH CARON]
738
"\u011B" => "e"
739

  
740
# ǝ  [LATIN SMALL LETTER TURNED E]
741
"\u01DD" => "e"
742

  
743
# ȅ  [LATIN SMALL LETTER E WITH DOUBLE GRAVE]
744
"\u0205" => "e"
745

  
746
# ȇ  [LATIN SMALL LETTER E WITH INVERTED BREVE]
747
"\u0207" => "e"
748

  
749
# ȩ  [LATIN SMALL LETTER E WITH CEDILLA]
750
"\u0229" => "e"
751

  
752
# ɇ  [LATIN SMALL LETTER E WITH STROKE]
753
"\u0247" => "e"
754

  
755
# ɘ  [LATIN SMALL LETTER REVERSED E]
756
"\u0258" => "e"
757

  
758
# ɛ  [LATIN SMALL LETTER OPEN E]
759
"\u025B" => "e"
760

  
761
# ɜ  [LATIN SMALL LETTER REVERSED OPEN E]
762
"\u025C" => "e"
763

  
764
# ɝ  [LATIN SMALL LETTER REVERSED OPEN E WITH HOOK]
765
"\u025D" => "e"
766

  
767
# ɞ  [LATIN SMALL LETTER CLOSED REVERSED OPEN E]
768
"\u025E" => "e"
769

  
770
# ʚ  [LATIN SMALL LETTER CLOSED OPEN E]
771
"\u029A" => "e"
772

  
773
# ᴈ  [LATIN SMALL LETTER TURNED OPEN E]
774
"\u1D08" => "e"
775

  
776
# ᶒ  [LATIN SMALL LETTER E WITH RETROFLEX HOOK]
777
"\u1D92" => "e"
778

  
779
# ᶓ  [LATIN SMALL LETTER OPEN E WITH RETROFLEX HOOK]
780
"\u1D93" => "e"
781

  
782
# ᶔ  [LATIN SMALL LETTER REVERSED OPEN E WITH RETROFLEX HOOK]
783
"\u1D94" => "e"
784

  
785
# ḕ  [LATIN SMALL LETTER E WITH MACRON AND GRAVE]
786
"\u1E15" => "e"
787

  
788
# ḗ  [LATIN SMALL LETTER E WITH MACRON AND ACUTE]
789
"\u1E17" => "e"
790

  
791
# ḙ  [LATIN SMALL LETTER E WITH CIRCUMFLEX BELOW]
792
"\u1E19" => "e"
793

  
794
# ḛ  [LATIN SMALL LETTER E WITH TILDE BELOW]
795
"\u1E1B" => "e"
796

  
797
# ḝ  [LATIN SMALL LETTER E WITH CEDILLA AND BREVE]
798
"\u1E1D" => "e"
799

  
800
# ẹ  [LATIN SMALL LETTER E WITH DOT BELOW]
801
"\u1EB9" => "e"
802

  
803
# ẻ  [LATIN SMALL LETTER E WITH HOOK ABOVE]
804
"\u1EBB" => "e"
805

  
806
# ẽ  [LATIN SMALL LETTER E WITH TILDE]
807
"\u1EBD" => "e"
808

  
809
# ế  [LATIN SMALL LETTER E WITH CIRCUMFLEX AND ACUTE]
810
"\u1EBF" => "e"
811

  
812
# ề  [LATIN SMALL LETTER E WITH CIRCUMFLEX AND GRAVE]
813
"\u1EC1" => "e"
814

  
815
# ể  [LATIN SMALL LETTER E WITH CIRCUMFLEX AND HOOK ABOVE]
816
"\u1EC3" => "e"
817

  
818
# ễ  [LATIN SMALL LETTER E WITH CIRCUMFLEX AND TILDE]
819
"\u1EC5" => "e"
820

  
821
# ệ  [LATIN SMALL LETTER E WITH CIRCUMFLEX AND DOT BELOW]
822
"\u1EC7" => "e"
823

  
824
# ₑ  [LATIN SUBSCRIPT SMALL LETTER E]
825
"\u2091" => "e"
826

  
827
# ⓔ  [CIRCLED LATIN SMALL LETTER E]
828
"\u24D4" => "e"
829

  
830
# ⱸ  [LATIN SMALL LETTER E WITH NOTCH]
831
"\u2C78" => "e"
832

  
833
# e  [FULLWIDTH LATIN SMALL LETTER E]
834
"\uFF45" => "e"
835

  
836
# ⒠  [PARENTHESIZED LATIN SMALL LETTER E]
837
"\u24A0" => "(e)"
838

  
839
# Ƒ  [LATIN CAPITAL LETTER F WITH HOOK]
840
"\u0191" => "F"
841

  
842
# Ḟ  [LATIN CAPITAL LETTER F WITH DOT ABOVE]
843
"\u1E1E" => "F"
844

  
845
# Ⓕ  [CIRCLED LATIN CAPITAL LETTER F]
846
"\u24BB" => "F"
847

  
848
# ꜰ  [LATIN LETTER SMALL CAPITAL F]
849
"\uA730" => "F"
850

  
851
# Ꝼ  [LATIN CAPITAL LETTER INSULAR F]
852
"\uA77B" => "F"
853

  
854
# ꟻ  [LATIN EPIGRAPHIC LETTER REVERSED F]
855
"\uA7FB" => "F"
856

  
857
# F  [FULLWIDTH LATIN CAPITAL LETTER F]
858
"\uFF26" => "F"
859

  
860
# ƒ  [LATIN SMALL LETTER F WITH HOOK]
861
"\u0192" => "f"
862

  
863
# ᵮ  [LATIN SMALL LETTER F WITH MIDDLE TILDE]
864
"\u1D6E" => "f"
865

  
866
# ᶂ  [LATIN SMALL LETTER F WITH PALATAL HOOK]
867
"\u1D82" => "f"
868

  
869
# ḟ  [LATIN SMALL LETTER F WITH DOT ABOVE]
870
"\u1E1F" => "f"
871

  
872
# ẛ  [LATIN SMALL LETTER LONG S WITH DOT ABOVE]
873
"\u1E9B" => "f"
874

  
875
# ⓕ  [CIRCLED LATIN SMALL LETTER F]
876
"\u24D5" => "f"
877

  
878
# ꝼ  [LATIN SMALL LETTER INSULAR F]
879
"\uA77C" => "f"
880

  
881
# f  [FULLWIDTH LATIN SMALL LETTER F]
882
"\uFF46" => "f"
883

  
884
# ⒡  [PARENTHESIZED LATIN SMALL LETTER F]
885
"\u24A1" => "(f)"
886

  
887
# ff  [LATIN SMALL LIGATURE FF]
888
"\uFB00" => "ff"
889

  
890
# ffi  [LATIN SMALL LIGATURE FFI]
891
"\uFB03" => "ffi"
892

  
893
# ffl  [LATIN SMALL LIGATURE FFL]
894
"\uFB04" => "ffl"
895

  
896
# fi  [LATIN SMALL LIGATURE FI]
897
"\uFB01" => "fi"
898

  
899
# fl  [LATIN SMALL LIGATURE FL]
900
"\uFB02" => "fl"
901

  
902
# Ĝ  [LATIN CAPITAL LETTER G WITH CIRCUMFLEX]
903
"\u011C" => "G"
904

  
905
# Ğ  [LATIN CAPITAL LETTER G WITH BREVE]
906
"\u011E" => "G"
907

  
908
# Ġ  [LATIN CAPITAL LETTER G WITH DOT ABOVE]
909
"\u0120" => "G"
910

  
911
# Ģ  [LATIN CAPITAL LETTER G WITH CEDILLA]
912
"\u0122" => "G"
913

  
914
# Ɠ  [LATIN CAPITAL LETTER G WITH HOOK]
915
"\u0193" => "G"
916

  
917
# Ǥ  [LATIN CAPITAL LETTER G WITH STROKE]
918
"\u01E4" => "G"
919

  
920
# ǥ  [LATIN SMALL LETTER G WITH STROKE]
921
"\u01E5" => "G"
922

  
923
# Ǧ  [LATIN CAPITAL LETTER G WITH CARON]
924
"\u01E6" => "G"
925

  
926
# ǧ  [LATIN SMALL LETTER G WITH CARON]
927
"\u01E7" => "G"
928

  
929
# Ǵ  [LATIN CAPITAL LETTER G WITH ACUTE]
930
"\u01F4" => "G"
931

  
932
# ɢ  [LATIN LETTER SMALL CAPITAL G]
933
"\u0262" => "G"
934

  
935
# ʛ  [LATIN LETTER SMALL CAPITAL G WITH HOOK]
936
"\u029B" => "G"
937

  
938
# Ḡ  [LATIN CAPITAL LETTER G WITH MACRON]
939
"\u1E20" => "G"
940

  
941
# Ⓖ  [CIRCLED LATIN CAPITAL LETTER G]
942
"\u24BC" => "G"
943

  
944
# Ᵹ  [LATIN CAPITAL LETTER INSULAR G]
945
"\uA77D" => "G"
946

  
947
# Ꝿ  [LATIN CAPITAL LETTER TURNED INSULAR G]
948
"\uA77E" => "G"
949

  
950
# G  [FULLWIDTH LATIN CAPITAL LETTER G]
951
"\uFF27" => "G"
952

  
953
# ĝ  [LATIN SMALL LETTER G WITH CIRCUMFLEX]
954
"\u011D" => "g"
955

  
956
# ğ  [LATIN SMALL LETTER G WITH BREVE]
957
"\u011F" => "g"
958

  
959
# ġ  [LATIN SMALL LETTER G WITH DOT ABOVE]
960
"\u0121" => "g"
961

  
962
# ģ  [LATIN SMALL LETTER G WITH CEDILLA]
963
"\u0123" => "g"
964

  
965
# ǵ  [LATIN SMALL LETTER G WITH ACUTE]
966
"\u01F5" => "g"
967

  
968
# ɠ  [LATIN SMALL LETTER G WITH HOOK]
969
"\u0260" => "g"
970

  
971
# ɡ  [LATIN SMALL LETTER SCRIPT G]
972
"\u0261" => "g"
973

  
974
# ᵷ  [LATIN SMALL LETTER TURNED G]
975
"\u1D77" => "g"
976

  
977
# ᵹ  [LATIN SMALL LETTER INSULAR G]
978
"\u1D79" => "g"
979

  
980
# ᶃ  [LATIN SMALL LETTER G WITH PALATAL HOOK]
981
"\u1D83" => "g"
982

  
983
# ḡ  [LATIN SMALL LETTER G WITH MACRON]
984
"\u1E21" => "g"
985

  
986
# ⓖ  [CIRCLED LATIN SMALL LETTER G]
987
"\u24D6" => "g"
988

  
989
# ꝿ  [LATIN SMALL LETTER TURNED INSULAR G]
990
"\uA77F" => "g"
991

  
992
# g  [FULLWIDTH LATIN SMALL LETTER G]
993
"\uFF47" => "g"
994

  
995
# ⒢  [PARENTHESIZED LATIN SMALL LETTER G]
996
"\u24A2" => "(g)"
997

  
998
# Ĥ  [LATIN CAPITAL LETTER H WITH CIRCUMFLEX]
999
"\u0124" => "H"
1000

  
1001
# Ħ  [LATIN CAPITAL LETTER H WITH STROKE]
1002
"\u0126" => "H"
1003

  
1004
# Ȟ  [LATIN CAPITAL LETTER H WITH CARON]
1005
"\u021E" => "H"
1006

  
1007
# ʜ  [LATIN LETTER SMALL CAPITAL H]
1008
"\u029C" => "H"
1009

  
1010
# Ḣ  [LATIN CAPITAL LETTER H WITH DOT ABOVE]
1011
"\u1E22" => "H"
1012

  
1013
# Ḥ  [LATIN CAPITAL LETTER H WITH DOT BELOW]
1014
"\u1E24" => "H"
1015

  
1016
# Ḧ  [LATIN CAPITAL LETTER H WITH DIAERESIS]
1017
"\u1E26" => "H"
1018

  
1019
# Ḩ  [LATIN CAPITAL LETTER H WITH CEDILLA]
1020
"\u1E28" => "H"
1021

  
1022
# Ḫ  [LATIN CAPITAL LETTER H WITH BREVE BELOW]
1023
"\u1E2A" => "H"
1024

  
1025
# Ⓗ  [CIRCLED LATIN CAPITAL LETTER H]
1026
"\u24BD" => "H"
1027

  
1028
# Ⱨ  [LATIN CAPITAL LETTER H WITH DESCENDER]
1029
"\u2C67" => "H"
1030

  
1031
# Ⱶ  [LATIN CAPITAL LETTER HALF H]
1032
"\u2C75" => "H"
1033

  
1034
# H  [FULLWIDTH LATIN CAPITAL LETTER H]
1035
"\uFF28" => "H"
1036

  
1037
# ĥ  [LATIN SMALL LETTER H WITH CIRCUMFLEX]
1038
"\u0125" => "h"
1039

  
1040
# ħ  [LATIN SMALL LETTER H WITH STROKE]
1041
"\u0127" => "h"
1042

  
1043
# ȟ  [LATIN SMALL LETTER H WITH CARON]
1044
"\u021F" => "h"
1045

  
1046
# ɥ  [LATIN SMALL LETTER TURNED H]
1047
"\u0265" => "h"
1048

  
1049
# ɦ  [LATIN SMALL LETTER H WITH HOOK]
1050
"\u0266" => "h"
1051

  
1052
# ʮ  [LATIN SMALL LETTER TURNED H WITH FISHHOOK]
1053
"\u02AE" => "h"
1054

  
1055
# ʯ  [LATIN SMALL LETTER TURNED H WITH FISHHOOK AND TAIL]
1056
"\u02AF" => "h"
1057

  
1058
# ḣ  [LATIN SMALL LETTER H WITH DOT ABOVE]
1059
"\u1E23" => "h"
1060

  
1061
# ḥ  [LATIN SMALL LETTER H WITH DOT BELOW]
1062
"\u1E25" => "h"
1063

  
1064
# ḧ  [LATIN SMALL LETTER H WITH DIAERESIS]
1065
"\u1E27" => "h"
1066

  
1067
# ḩ  [LATIN SMALL LETTER H WITH CEDILLA]
1068
"\u1E29" => "h"
1069

  
1070
# ḫ  [LATIN SMALL LETTER H WITH BREVE BELOW]
1071
"\u1E2B" => "h"
1072

  
1073
# ẖ  [LATIN SMALL LETTER H WITH LINE BELOW]
1074
"\u1E96" => "h"
1075

  
1076
# ⓗ  [CIRCLED LATIN SMALL LETTER H]
1077
"\u24D7" => "h"
1078

  
1079
# ⱨ  [LATIN SMALL LETTER H WITH DESCENDER]
1080
"\u2C68" => "h"
1081

  
1082
# ⱶ  [LATIN SMALL LETTER HALF H]
1083
"\u2C76" => "h"
1084

  
1085
# h  [FULLWIDTH LATIN SMALL LETTER H]
1086
"\uFF48" => "h"
1087

  
1088
# Ƕ  http://en.wikipedia.org/wiki/Hwair  [LATIN CAPITAL LETTER HWAIR]
1089
"\u01F6" => "HV"
1090

  
1091
# ⒣  [PARENTHESIZED LATIN SMALL LETTER H]
1092
"\u24A3" => "(h)"
1093

  
1094
# ƕ  [LATIN SMALL LETTER HV]
1095
"\u0195" => "hv"
1096

  
1097
# Ì  [LATIN CAPITAL LETTER I WITH GRAVE]
1098
"\u00CC" => "I"
1099

  
1100
# Í  [LATIN CAPITAL LETTER I WITH ACUTE]
1101
"\u00CD" => "I"
1102

  
1103
# Î  [LATIN CAPITAL LETTER I WITH CIRCUMFLEX]
1104
"\u00CE" => "I"
1105

  
1106
# Ï  [LATIN CAPITAL LETTER I WITH DIAERESIS]
1107
"\u00CF" => "I"
1108

  
1109
# Ĩ  [LATIN CAPITAL LETTER I WITH TILDE]
1110
"\u0128" => "I"
1111

  
1112
# Ī  [LATIN CAPITAL LETTER I WITH MACRON]
1113
"\u012A" => "I"
1114

  
1115
# Ĭ  [LATIN CAPITAL LETTER I WITH BREVE]
1116
"\u012C" => "I"
1117

  
1118
# Į  [LATIN CAPITAL LETTER I WITH OGONEK]
1119
"\u012E" => "I"
1120

  
1121
# İ  [LATIN CAPITAL LETTER I WITH DOT ABOVE]
1122
"\u0130" => "I"
1123

  
1124
# Ɩ  [LATIN CAPITAL LETTER IOTA]
1125
"\u0196" => "I"
1126

  
1127
# Ɨ  [LATIN CAPITAL LETTER I WITH STROKE]
1128
"\u0197" => "I"
1129

  
1130
# Ǐ  [LATIN CAPITAL LETTER I WITH CARON]
1131
"\u01CF" => "I"
1132

  
1133
# Ȉ  [LATIN CAPITAL LETTER I WITH DOUBLE GRAVE]
1134
"\u0208" => "I"
1135

  
1136
# Ȋ  [LATIN CAPITAL LETTER I WITH INVERTED BREVE]
1137
"\u020A" => "I"
1138

  
1139
# ɪ  [LATIN LETTER SMALL CAPITAL I]
1140
"\u026A" => "I"
1141

  
1142
# ᵻ  [LATIN SMALL CAPITAL LETTER I WITH STROKE]
1143
"\u1D7B" => "I"
1144

  
1145
# Ḭ  [LATIN CAPITAL LETTER I WITH TILDE BELOW]
1146
"\u1E2C" => "I"
1147

  
1148
# Ḯ  [LATIN CAPITAL LETTER I WITH DIAERESIS AND ACUTE]
1149
"\u1E2E" => "I"
1150

  
1151
# Ỉ  [LATIN CAPITAL LETTER I WITH HOOK ABOVE]
1152
"\u1EC8" => "I"
1153

  
1154
# Ị  [LATIN CAPITAL LETTER I WITH DOT BELOW]
1155
"\u1ECA" => "I"
1156

  
1157
# Ⓘ  [CIRCLED LATIN CAPITAL LETTER I]
1158
"\u24BE" => "I"
1159

  
1160
# ꟾ  [LATIN EPIGRAPHIC LETTER I LONGA]
1161
"\uA7FE" => "I"
1162

  
1163
# I  [FULLWIDTH LATIN CAPITAL LETTER I]
1164
"\uFF29" => "I"
1165

  
1166
# ì  [LATIN SMALL LETTER I WITH GRAVE]
1167
"\u00EC" => "i"
1168

  
1169
# í  [LATIN SMALL LETTER I WITH ACUTE]
1170
"\u00ED" => "i"
1171

  
1172
# î  [LATIN SMALL LETTER I WITH CIRCUMFLEX]
1173
"\u00EE" => "i"
1174

  
1175
# ï  [LATIN SMALL LETTER I WITH DIAERESIS]
1176
"\u00EF" => "i"
1177

  
1178
# ĩ  [LATIN SMALL LETTER I WITH TILDE]
1179
"\u0129" => "i"
1180

  
1181
# ī  [LATIN SMALL LETTER I WITH MACRON]
1182
"\u012B" => "i"
1183

  
1184
# ĭ  [LATIN SMALL LETTER I WITH BREVE]
1185
"\u012D" => "i"
1186

  
1187
# į  [LATIN SMALL LETTER I WITH OGONEK]
1188
"\u012F" => "i"
1189

  
1190
# ı  [LATIN SMALL LETTER DOTLESS I]
1191
"\u0131" => "i"
1192

  
1193
# ǐ  [LATIN SMALL LETTER I WITH CARON]
1194
"\u01D0" => "i"
1195

  
1196
# ȉ  [LATIN SMALL LETTER I WITH DOUBLE GRAVE]
1197
"\u0209" => "i"
1198

  
1199
# ȋ  [LATIN SMALL LETTER I WITH INVERTED BREVE]
1200
"\u020B" => "i"
1201

  
1202
# ɨ  [LATIN SMALL LETTER I WITH STROKE]
1203
"\u0268" => "i"
1204

  
1205
# ᴉ  [LATIN SMALL LETTER TURNED I]
1206
"\u1D09" => "i"
1207

  
1208
# ᵢ  [LATIN SUBSCRIPT SMALL LETTER I]
1209
"\u1D62" => "i"
1210

  
1211
# ᵼ  [LATIN SMALL LETTER IOTA WITH STROKE]
1212
"\u1D7C" => "i"
1213

  
1214
# ᶖ  [LATIN SMALL LETTER I WITH RETROFLEX HOOK]
1215
"\u1D96" => "i"
1216

  
1217
# ḭ  [LATIN SMALL LETTER I WITH TILDE BELOW]
1218
"\u1E2D" => "i"
1219

  
1220
# ḯ  [LATIN SMALL LETTER I WITH DIAERESIS AND ACUTE]
1221
"\u1E2F" => "i"
1222

  
1223
# ỉ  [LATIN SMALL LETTER I WITH HOOK ABOVE]
1224
"\u1EC9" => "i"
1225

  
1226
# ị  [LATIN SMALL LETTER I WITH DOT BELOW]
1227
"\u1ECB" => "i"
1228

  
1229
# ⁱ  [SUPERSCRIPT LATIN SMALL LETTER I]
1230
"\u2071" => "i"
1231

  
1232
# ⓘ  [CIRCLED LATIN SMALL LETTER I]
1233
"\u24D8" => "i"
1234

  
1235
# i  [FULLWIDTH LATIN SMALL LETTER I]
1236
"\uFF49" => "i"
1237

  
1238
# IJ  [LATIN CAPITAL LIGATURE IJ]
1239
"\u0132" => "IJ"
1240

  
1241
# ⒤  [PARENTHESIZED LATIN SMALL LETTER I]
1242
"\u24A4" => "(i)"
1243

  
1244
# ij  [LATIN SMALL LIGATURE IJ]
1245
"\u0133" => "ij"
1246

  
1247
# Ĵ  [LATIN CAPITAL LETTER J WITH CIRCUMFLEX]
1248
"\u0134" => "J"
1249

  
1250
# Ɉ  [LATIN CAPITAL LETTER J WITH STROKE]
1251
"\u0248" => "J"
1252

  
1253
# ᴊ  [LATIN LETTER SMALL CAPITAL J]
1254
"\u1D0A" => "J"
1255

  
1256
# Ⓙ  [CIRCLED LATIN CAPITAL LETTER J]
1257
"\u24BF" => "J"
1258

  
1259
# J  [FULLWIDTH LATIN CAPITAL LETTER J]
1260
"\uFF2A" => "J"
1261

  
1262
# ĵ  [LATIN SMALL LETTER J WITH CIRCUMFLEX]
1263
"\u0135" => "j"
1264

  
1265
# ǰ  [LATIN SMALL LETTER J WITH CARON]
1266
"\u01F0" => "j"
1267

  
1268
# ȷ  [LATIN SMALL LETTER DOTLESS J]
1269
"\u0237" => "j"
1270

  
1271
# ɉ  [LATIN SMALL LETTER J WITH STROKE]
1272
"\u0249" => "j"
1273

  
1274
# ɟ  [LATIN SMALL LETTER DOTLESS J WITH STROKE]
1275
"\u025F" => "j"
1276

  
1277
# ʄ  [LATIN SMALL LETTER DOTLESS J WITH STROKE AND HOOK]
1278
"\u0284" => "j"
1279

  
1280
# ʝ  [LATIN SMALL LETTER J WITH CROSSED-TAIL]
1281
"\u029D" => "j"
1282

  
1283
# ⓙ  [CIRCLED LATIN SMALL LETTER J]
1284
"\u24D9" => "j"
1285

  
1286
# ⱼ  [LATIN SUBSCRIPT SMALL LETTER J]
1287
"\u2C7C" => "j"
1288

  
1289
# j  [FULLWIDTH LATIN SMALL LETTER J]
1290
"\uFF4A" => "j"
1291

  
1292
# ⒥  [PARENTHESIZED LATIN SMALL LETTER J]
1293
"\u24A5" => "(j)"
1294

  
1295
# Ķ  [LATIN CAPITAL LETTER K WITH CEDILLA]
1296
"\u0136" => "K"
1297

  
1298
# Ƙ  [LATIN CAPITAL LETTER K WITH HOOK]
1299
"\u0198" => "K"
1300

  
1301
# Ǩ  [LATIN CAPITAL LETTER K WITH CARON]
1302
"\u01E8" => "K"
1303

  
1304
# ᴋ  [LATIN LETTER SMALL CAPITAL K]
1305
"\u1D0B" => "K"
1306

  
1307
# Ḱ  [LATIN CAPITAL LETTER K WITH ACUTE]
1308
"\u1E30" => "K"
1309

  
1310
# Ḳ  [LATIN CAPITAL LETTER K WITH DOT BELOW]
1311
"\u1E32" => "K"
1312

  
1313
# Ḵ  [LATIN CAPITAL LETTER K WITH LINE BELOW]
1314
"\u1E34" => "K"
1315

  
1316
# Ⓚ  [CIRCLED LATIN CAPITAL LETTER K]
1317
"\u24C0" => "K"
1318

  
1319
# Ⱪ  [LATIN CAPITAL LETTER K WITH DESCENDER]
1320
"\u2C69" => "K"
1321

  
1322
# Ꝁ  [LATIN CAPITAL LETTER K WITH STROKE]
1323
"\uA740" => "K"
1324

  
1325
# Ꝃ  [LATIN CAPITAL LETTER K WITH DIAGONAL STROKE]
1326
"\uA742" => "K"
1327

  
1328
# Ꝅ  [LATIN CAPITAL LETTER K WITH STROKE AND DIAGONAL STROKE]
1329
"\uA744" => "K"
1330

  
1331
# K  [FULLWIDTH LATIN CAPITAL LETTER K]
1332
"\uFF2B" => "K"
1333

  
1334
# ķ  [LATIN SMALL LETTER K WITH CEDILLA]
1335
"\u0137" => "k"
1336

  
1337
# ƙ  [LATIN SMALL LETTER K WITH HOOK]
1338
"\u0199" => "k"
1339

  
1340
# ǩ  [LATIN SMALL LETTER K WITH CARON]
1341
"\u01E9" => "k"
1342

  
1343
# ʞ  [LATIN SMALL LETTER TURNED K]
1344
"\u029E" => "k"
1345

  
1346
# ᶄ  [LATIN SMALL LETTER K WITH PALATAL HOOK]
1347
"\u1D84" => "k"
1348

  
1349
# ḱ  [LATIN SMALL LETTER K WITH ACUTE]
1350
"\u1E31" => "k"
1351

  
... This diff was truncated because it exceeds the maximum size that can be displayed.

Also available in: Unified diff