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 * http://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 package org.apache.commons.text.translate;
18
19 import java.util.Collections;
20 import java.util.HashMap;
21 import java.util.Map;
22 import java.util.Map.Entry;
23 import java.util.stream.Collectors;
24
25 /**
26 * Class holding various entity data for HTML and XML - generally for use with
27 * the LookupTranslator.
28 * All Maps are generated using {@code java.util.Collections.unmodifiableMap()}.
29 *
30 * @since 1.0
31 */
32 public class EntityArrays {
33
34 /**
35 * A Map<CharSequence, CharSequence> to escape
36 * <a href="https://secure.wikimedia.org/wikipedia/en/wiki/ISO/IEC_8859-1">ISO-8859-1</a>
37 * characters to their named HTML 3.x equivalents.
38 */
39 public static final Map<CharSequence, CharSequence> ISO8859_1_ESCAPE;
40 static {
41 final Map<CharSequence, CharSequence> initialMap = new HashMap<>();
42 initialMap.put("\u00A0", " "); // non-breaking space
43 initialMap.put("\u00A1", "¡"); // inverted exclamation mark
44 initialMap.put("\u00A2", "¢"); // cent sign
45 initialMap.put("\u00A3", "£"); // pound sign
46 initialMap.put("\u00A4", "¤"); // currency sign
47 initialMap.put("\u00A5", "¥"); // yen sign = yuan sign
48 initialMap.put("\u00A6", "¦"); // broken bar = broken vertical bar
49 initialMap.put("\u00A7", "§"); // section sign
50 initialMap.put("\u00A8", "¨"); // dieresis = spacing dieresis
51 initialMap.put("\u00A9", "©"); // © - copyright sign
52 initialMap.put("\u00AA", "ª"); // feminine ordinal indicator
53 initialMap.put("\u00AB", "«"); // left-pointing double angle quotation mark = left pointing guillemet
54 initialMap.put("\u00AC", "¬"); // not sign
55 initialMap.put("\u00AD", "­"); // soft hyphen = discretionary hyphen
56 initialMap.put("\u00AE", "®"); // ® - registered trademark sign
57 initialMap.put("\u00AF", "¯"); // macron = spacing macron = overline = APL overbar
58 initialMap.put("\u00B0", "°"); // degree sign
59 initialMap.put("\u00B1", "±"); // plus-minus sign = plus-or-minus sign
60 initialMap.put("\u00B2", "²"); // superscript two = superscript digit two = squared
61 initialMap.put("\u00B3", "³"); // superscript three = superscript digit three = cubed
62 initialMap.put("\u00B4", "´"); // acute accent = spacing acute
63 initialMap.put("\u00B5", "µ"); // micro sign
64 initialMap.put("\u00B6", "¶"); // pilcrow sign = paragraph sign
65 initialMap.put("\u00B7", "·"); // middle dot = Georgian comma = Greek middle dot
66 initialMap.put("\u00B8", "¸"); // cedilla = spacing cedilla
67 initialMap.put("\u00B9", "¹"); // superscript one = superscript digit one
68 initialMap.put("\u00BA", "º"); // masculine ordinal indicator
69 initialMap.put("\u00BB", "»"); // right-pointing double angle quotation mark = right pointing guillemet
70 initialMap.put("\u00BC", "¼"); // vulgar fraction one quarter = fraction one quarter
71 initialMap.put("\u00BD", "½"); // vulgar fraction one half = fraction one half
72 initialMap.put("\u00BE", "¾"); // vulgar fraction three quarters = fraction three quarters
73 initialMap.put("\u00BF", "¿"); // inverted question mark = turned question mark
74 initialMap.put("\u00C0", "À"); // Ã - uppercase A, grave accent
75 initialMap.put("\u00C1", "Á"); // Ã - uppercase A, acute accent
76 initialMap.put("\u00C2", "Â"); // Ã - uppercase A, circumflex accent
77 initialMap.put("\u00C3", "Ã"); // Ã - uppercase A, tilde
78 initialMap.put("\u00C4", "Ä"); // Ã - uppercase A, umlaut
79 initialMap.put("\u00C5", "Å"); // Ã - uppercase A, ring
80 initialMap.put("\u00C6", "Æ"); // Ã - uppercase AE
81 initialMap.put("\u00C7", "Ç"); // Ã - uppercase C, cedilla
82 initialMap.put("\u00C8", "È"); // Ã - uppercase E, grave accent
83 initialMap.put("\u00C9", "É"); // Ã - uppercase E, acute accent
84 initialMap.put("\u00CA", "Ê"); // Ã - uppercase E, circumflex accent
85 initialMap.put("\u00CB", "Ë"); // Ã - uppercase E, umlaut
86 initialMap.put("\u00CC", "Ì"); // Ã - uppercase I, grave accent
87 initialMap.put("\u00CD", "Í"); // Ã - uppercase I, acute accent
88 initialMap.put("\u00CE", "Î"); // Ã - uppercase I, circumflex accent
89 initialMap.put("\u00CF", "Ï"); // Ã - uppercase I, umlaut
90 initialMap.put("\u00D0", "Ð"); // Ã - uppercase Eth, Icelandic
91 initialMap.put("\u00D1", "Ñ"); // Ã - uppercase N, tilde
92 initialMap.put("\u00D2", "Ò"); // Ã - uppercase O, grave accent
93 initialMap.put("\u00D3", "Ó"); // Ã - uppercase O, acute accent
94 initialMap.put("\u00D4", "Ô"); // Ã - uppercase O, circumflex accent
95 initialMap.put("\u00D5", "Õ"); // Ã - uppercase O, tilde
96 initialMap.put("\u00D6", "Ö"); // Ã - uppercase O, umlaut
97 initialMap.put("\u00D7", "×"); // multiplication sign
98 initialMap.put("\u00D8", "Ø"); // Ã - uppercase O, slash
99 initialMap.put("\u00D9", "Ù"); // Ã - uppercase U, grave accent
100 initialMap.put("\u00DA", "Ú"); // Ã - uppercase U, acute accent
101 initialMap.put("\u00DB", "Û"); // Ã - uppercase U, circumflex accent
102 initialMap.put("\u00DC", "Ü"); // Ã - uppercase U, umlaut
103 initialMap.put("\u00DD", "Ý"); // Ã - uppercase Y, acute accent
104 initialMap.put("\u00DE", "Þ"); // Ã - uppercase THORN, Icelandic
105 initialMap.put("\u00DF", "ß"); // Ã - lowercase sharps, German
106 initialMap.put("\u00E0", "à"); // Ã - lowercase a, grave accent
107 initialMap.put("\u00E1", "á"); // á - lowercase a, acute accent
108 initialMap.put("\u00E2", "â"); // â - lowercase a, circumflex accent
109 initialMap.put("\u00E3", "ã"); // ã - lowercase a, tilde
110 initialMap.put("\u00E4", "ä"); // ä - lowercase a, umlaut
111 initialMap.put("\u00E5", "å"); // Ã¥ - lowercase a, ring
112 initialMap.put("\u00E6", "æ"); // æ - lowercase ae
113 initialMap.put("\u00E7", "ç"); // ç - lowercase c, cedilla
114 initialMap.put("\u00E8", "è"); // è - lowercase e, grave accent
115 initialMap.put("\u00E9", "é"); // é - lowercase e, acute accent
116 initialMap.put("\u00EA", "ê"); // ê - lowercase e, circumflex accent
117 initialMap.put("\u00EB", "ë"); // ë - lowercase e, umlaut
118 initialMap.put("\u00EC", "ì"); // ì - lowercase i, grave accent
119 initialMap.put("\u00ED", "í"); // Ã - lowercase i, acute accent
120 initialMap.put("\u00EE", "î"); // î - lowercase i, circumflex accent
121 initialMap.put("\u00EF", "ï"); // ï - lowercase i, umlaut
122 initialMap.put("\u00F0", "ð"); // ð - lowercase eth, Icelandic
123 initialMap.put("\u00F1", "ñ"); // ñ - lowercase n, tilde
124 initialMap.put("\u00F2", "ò"); // ò - lowercase o, grave accent
125 initialMap.put("\u00F3", "ó"); // ó - lowercase o, acute accent
126 initialMap.put("\u00F4", "ô"); // ô - lowercase o, circumflex accent
127 initialMap.put("\u00F5", "õ"); // õ - lowercase o, tilde
128 initialMap.put("\u00F6", "ö"); // ö - lowercase o, umlaut
129 initialMap.put("\u00F7", "÷"); // division sign
130 initialMap.put("\u00F8", "ø"); // ø - lowercase o, slash
131 initialMap.put("\u00F9", "ù"); // ù - lowercase u, grave accent
132 initialMap.put("\u00FA", "ú"); // ú - lowercase u, acute accent
133 initialMap.put("\u00FB", "û"); // û - lowercase u, circumflex accent
134 initialMap.put("\u00FC", "ü"); // ü - lowercase u, umlaut
135 initialMap.put("\u00FD", "ý"); // ý - lowercase y, acute accent
136 initialMap.put("\u00FE", "þ"); // þ - lowercase thorn, Icelandic
137 initialMap.put("\u00FF", "ÿ"); // ÿ - lowercase y, umlaut
138 ISO8859_1_ESCAPE = Collections.unmodifiableMap(initialMap);
139 }
140
141 /**
142 * Reverse of {@link #ISO8859_1_ESCAPE} for unescaping purposes.
143 */
144 public static final Map<CharSequence, CharSequence> ISO8859_1_UNESCAPE;
145
146 static {
147 ISO8859_1_UNESCAPE = Collections.unmodifiableMap(invert(ISO8859_1_ESCAPE));
148 }
149
150 /**
151 * A Map<CharSequence, CharSequence> to escape additional
152 * <a href="http://www.w3.org/TR/REC-html40/sgml/entities.html">character entity
153 * references</a>. Note that this must be used with {@link #ISO8859_1_ESCAPE} to get the full list of
154 * HTML 4.0 character entities.
155 */
156 public static final Map<CharSequence, CharSequence> HTML40_EXTENDED_ESCAPE;
157
158 static {
159 final Map<CharSequence, CharSequence> initialMap = new HashMap<>();
160 // <!-- Latin Extended-B -->
161 initialMap.put("\u0192", "ƒ"); // latin small f with hook = function= florin, U+0192 ISOtech -->
162 // <!-- Greek -->
163 initialMap.put("\u0391", "Α"); // greek capital letter alpha, U+0391 -->
164 initialMap.put("\u0392", "Β"); // greek capital letter beta, U+0392 -->
165 initialMap.put("\u0393", "Γ"); // greek capital letter gamma,U+0393 ISOgrk3 -->
166 initialMap.put("\u0394", "Δ"); // greek capital letter delta,U+0394 ISOgrk3 -->
167 initialMap.put("\u0395", "Ε"); // greek capital letter epsilon, U+0395 -->
168 initialMap.put("\u0396", "Ζ"); // greek capital letter zeta, U+0396 -->
169 initialMap.put("\u0397", "Η"); // greek capital letter eta, U+0397 -->
170 initialMap.put("\u0398", "Θ"); // greek capital letter theta,U+0398 ISOgrk3 -->
171 initialMap.put("\u0399", "Ι"); // greek capital letter iota, U+0399 -->
172 initialMap.put("\u039A", "Κ"); // greek capital letter kappa, U+039A -->
173 initialMap.put("\u039B", "Λ"); // greek capital letter lambda,U+039B ISOgrk3 -->
174 initialMap.put("\u039C", "Μ"); // greek capital letter mu, U+039C -->
175 initialMap.put("\u039D", "Ν"); // greek capital letter nu, U+039D -->
176 initialMap.put("\u039E", "Ξ"); // greek capital letter xi, U+039E ISOgrk3 -->
177 initialMap.put("\u039F", "Ο"); // greek capital letter omicron, U+039F -->
178 initialMap.put("\u03A0", "Π"); // greek capital letter pi, U+03A0 ISOgrk3 -->
179 initialMap.put("\u03A1", "Ρ"); // greek capital letter rho, U+03A1 -->
180 // <!-- there is no Sigmaf, and no U+03A2 character either -->
181 initialMap.put("\u03A3", "Σ"); // greek capital letter sigma,U+03A3 ISOgrk3 -->
182 initialMap.put("\u03A4", "Τ"); // greek capital letter tau, U+03A4 -->
183 initialMap.put("\u03A5", "Υ"); // greek capital letter upsilon,U+03A5 ISOgrk3 -->
184 initialMap.put("\u03A6", "Φ"); // greek capital letter phi,U+03A6 ISOgrk3 -->
185 initialMap.put("\u03A7", "Χ"); // greek capital letter chi, U+03A7 -->
186 initialMap.put("\u03A8", "Ψ"); // greek capital letter psi,U+03A8 ISOgrk3 -->
187 initialMap.put("\u03A9", "Ω"); // greek capital letter omega,U+03A9 ISOgrk3 -->
188 initialMap.put("\u03B1", "α"); // greek small letter alpha,U+03B1 ISOgrk3 -->
189 initialMap.put("\u03B2", "β"); // greek small letter beta, U+03B2 ISOgrk3 -->
190 initialMap.put("\u03B3", "γ"); // greek small letter gamma,U+03B3 ISOgrk3 -->
191 initialMap.put("\u03B4", "δ"); // greek small letter delta,U+03B4 ISOgrk3 -->
192 initialMap.put("\u03B5", "ε"); // greek small letter epsilon,U+03B5 ISOgrk3 -->
193 initialMap.put("\u03B6", "ζ"); // greek small letter zeta, U+03B6 ISOgrk3 -->
194 initialMap.put("\u03B7", "η"); // greek small letter eta, U+03B7 ISOgrk3 -->
195 initialMap.put("\u03B8", "θ"); // greek small letter theta,U+03B8 ISOgrk3 -->
196 initialMap.put("\u03B9", "ι"); // greek small letter iota, U+03B9 ISOgrk3 -->
197 initialMap.put("\u03BA", "κ"); // greek small letter kappa,U+03BA ISOgrk3 -->
198 initialMap.put("\u03BB", "λ"); // greek small letter lambda,U+03BB ISOgrk3 -->
199 initialMap.put("\u03BC", "μ"); // greek small letter mu, U+03BC ISOgrk3 -->
200 initialMap.put("\u03BD", "ν"); // greek small letter nu, U+03BD ISOgrk3 -->
201 initialMap.put("\u03BE", "ξ"); // greek small letter xi, U+03BE ISOgrk3 -->
202 initialMap.put("\u03BF", "ο"); // greek small letter omicron, U+03BF NEW -->
203 initialMap.put("\u03C0", "π"); // greek small letter pi, U+03C0 ISOgrk3 -->
204 initialMap.put("\u03C1", "ρ"); // greek small letter rho, U+03C1 ISOgrk3 -->
205 initialMap.put("\u03C2", "ς"); // greek small letter final sigma,U+03C2 ISOgrk3 -->
206 initialMap.put("\u03C3", "σ"); // greek small letter sigma,U+03C3 ISOgrk3 -->
207 initialMap.put("\u03C4", "τ"); // greek small letter tau, U+03C4 ISOgrk3 -->
208 initialMap.put("\u03C5", "υ"); // greek small letter upsilon,U+03C5 ISOgrk3 -->
209 initialMap.put("\u03C6", "φ"); // greek small letter phi, U+03C6 ISOgrk3 -->
210 initialMap.put("\u03C7", "χ"); // greek small letter chi, U+03C7 ISOgrk3 -->
211 initialMap.put("\u03C8", "ψ"); // greek small letter psi, U+03C8 ISOgrk3 -->
212 initialMap.put("\u03C9", "ω"); // greek small letter omega,U+03C9 ISOgrk3 -->
213 initialMap.put("\u03D1", "ϑ"); // greek small letter theta symbol,U+03D1 NEW -->
214 initialMap.put("\u03D2", "ϒ"); // greek upsilon with hook symbol,U+03D2 NEW -->
215 initialMap.put("\u03D6", "ϖ"); // greek pi symbol, U+03D6 ISOgrk3 -->
216 // <!-- General Punctuation -->
217 initialMap.put("\u2022", "•"); // bullet = black small circle,U+2022 ISOpub -->
218 // <!-- bullet is NOT the same as bullet operator, U+2219 -->
219 initialMap.put("\u2026", "…"); // horizontal ellipsis = three dot leader,U+2026 ISOpub -->
220 initialMap.put("\u2032", "′"); // prime = minutes = feet, U+2032 ISOtech -->
221 initialMap.put("\u2033", "″"); // double prime = seconds = inches,U+2033 ISOtech -->
222 initialMap.put("\u203E", "‾"); // overline = spacing overscore,U+203E NEW -->
223 initialMap.put("\u2044", "⁄"); // fraction slash, U+2044 NEW -->
224 // <!-- Letterlike Symbols -->
225 initialMap.put("\u2118", "℘"); // script capital P = power set= Weierstrass p, U+2118 ISOamso -->
226 initialMap.put("\u2111", "ℑ"); // blackletter capital I = imaginary part,U+2111 ISOamso -->
227 initialMap.put("\u211C", "ℜ"); // blackletter capital R = real part symbol,U+211C ISOamso -->
228 initialMap.put("\u2122", "™"); // trade mark sign, U+2122 ISOnum -->
229 initialMap.put("\u2135", "ℵ"); // alef symbol = first transfinite cardinal,U+2135 NEW -->
230 // <!-- alef symbol is NOT the same as hebrew letter alef,U+05D0 although the
231 // same glyph could be used to depict both characters -->
232 // <!-- Arrows -->
233 initialMap.put("\u2190", "←"); // leftwards arrow, U+2190 ISOnum -->
234 initialMap.put("\u2191", "↑"); // upwards arrow, U+2191 ISOnum-->
235 initialMap.put("\u2192", "→"); // rightwards arrow, U+2192 ISOnum -->
236 initialMap.put("\u2193", "↓"); // downwards arrow, U+2193 ISOnum -->
237 initialMap.put("\u2194", "↔"); // left right arrow, U+2194 ISOamsa -->
238 initialMap.put("\u21B5", "↵"); // downwards arrow with corner leftwards= carriage return, U+21B5 NEW -->
239 initialMap.put("\u21D0", "⇐"); // leftwards double arrow, U+21D0 ISOtech -->
240 // <!-- ISO 10646 does not say that lArr is the same as the 'is implied by'
241 // arrow but also does not have any other character for that function.
242 // So ? lArr canbe used for 'is implied by' as ISOtech suggests -->
243 initialMap.put("\u21D1", "⇑"); // upwards double arrow, U+21D1 ISOamsa -->
244 initialMap.put("\u21D2", "⇒"); // rightwards double arrow,U+21D2 ISOtech -->
245 // <!-- ISO 10646 does not say this is the 'implies' character but does not
246 // have another character with this function so ?rArr can be used for
247 // 'implies' as ISOtech suggests -->
248 initialMap.put("\u21D3", "⇓"); // downwards double arrow, U+21D3 ISOamsa -->
249 initialMap.put("\u21D4", "⇔"); // left right double arrow,U+21D4 ISOamsa -->
250 // <!-- Mathematical Operators -->
251 initialMap.put("\u2200", "∀"); // for all, U+2200 ISOtech -->
252 initialMap.put("\u2202", "∂"); // partial differential, U+2202 ISOtech -->
253 initialMap.put("\u2203", "∃"); // there exists, U+2203 ISOtech -->
254 initialMap.put("\u2205", "∅"); // empty set = null set = diameter,U+2205 ISOamso -->
255 initialMap.put("\u2207", "∇"); // nabla = backward difference,U+2207 ISOtech -->
256 initialMap.put("\u2208", "∈"); // element of, U+2208 ISOtech -->
257 initialMap.put("\u2209", "∉"); // not an element of, U+2209 ISOtech -->
258 initialMap.put("\u220B", "∋"); // contains as member, U+220B ISOtech -->
259 // <!-- should there be a more memorable name than 'ni'? -->
260 initialMap.put("\u220F", "∏"); // n-ary product = product sign,U+220F ISOamsb -->
261 // <!-- prod is NOT the same character as U+03A0 'greek capital letter pi'
262 // though the same glyph might be used for both -->
263 initialMap.put("\u2211", "∑"); // n-ary summation, U+2211 ISOamsb -->
264 // <!-- sum is NOT the same character as U+03A3 'greek capital letter sigma'
265 // though the same glyph might be used for both -->
266 initialMap.put("\u2212", "−"); // minus sign, U+2212 ISOtech -->
267 initialMap.put("\u2217", "∗"); // asterisk operator, U+2217 ISOtech -->
268 initialMap.put("\u221A", "√"); // square root = radical sign,U+221A ISOtech -->
269 initialMap.put("\u221D", "∝"); // proportional to, U+221D ISOtech -->
270 initialMap.put("\u221E", "∞"); // infinity, U+221E ISOtech -->
271 initialMap.put("\u2220", "∠"); // angle, U+2220 ISOamso -->
272 initialMap.put("\u2227", "∧"); // logical and = wedge, U+2227 ISOtech -->
273 initialMap.put("\u2228", "∨"); // logical or = vee, U+2228 ISOtech -->
274 initialMap.put("\u2229", "∩"); // intersection = cap, U+2229 ISOtech -->
275 initialMap.put("\u222A", "∪"); // union = cup, U+222A ISOtech -->
276 initialMap.put("\u222B", "∫"); // integral, U+222B ISOtech -->
277 initialMap.put("\u2234", "∴"); // therefore, U+2234 ISOtech -->
278 initialMap.put("\u223C", "∼"); // tilde operator = varies with = similar to,U+223C ISOtech -->
279 // <!-- tilde operator is NOT the same character as the tilde, U+007E,although
280 // the same glyph might be used to represent both -->
281 initialMap.put("\u2245", "≅"); // approximately equal to, U+2245 ISOtech -->
282 initialMap.put("\u2248", "≈"); // almost equal to = asymptotic to,U+2248 ISOamsr -->
283 initialMap.put("\u2260", "≠"); // not equal to, U+2260 ISOtech -->
284 initialMap.put("\u2261", "≡"); // identical to, U+2261 ISOtech -->
285 initialMap.put("\u2264", "≤"); // less-than or equal to, U+2264 ISOtech -->
286 initialMap.put("\u2265", "≥"); // greater-than or equal to,U+2265 ISOtech -->
287 initialMap.put("\u2282", "⊂"); // subset of, U+2282 ISOtech -->
288 initialMap.put("\u2283", "⊃"); // superset of, U+2283 ISOtech -->
289 // <!-- note that nsup, 'not a superset of, U+2283' is not covered by the
290 // Symbol font encoding and is not included. Should it be, for symmetry?
291 // It is in ISOamsn -->,
292 initialMap.put("\u2284", "⊄"); // not a subset of, U+2284 ISOamsn -->
293 initialMap.put("\u2286", "⊆"); // subset of or equal to, U+2286 ISOtech -->
294 initialMap.put("\u2287", "⊇"); // superset of or equal to,U+2287 ISOtech -->
295 initialMap.put("\u2295", "⊕"); // circled plus = direct sum,U+2295 ISOamsb -->
296 initialMap.put("\u2297", "⊗"); // circled times = vector product,U+2297 ISOamsb -->
297 initialMap.put("\u22A5", "⊥"); // up tack = orthogonal to = perpendicular,U+22A5 ISOtech -->
298 initialMap.put("\u22C5", "⋅"); // dot operator, U+22C5 ISOamsb -->
299 // <!-- dot operator is NOT the same character as U+00B7 middle dot -->
300 // <!-- Miscellaneous Technical -->
301 initialMap.put("\u2308", "⌈"); // left ceiling = apl upstile,U+2308 ISOamsc -->
302 initialMap.put("\u2309", "⌉"); // right ceiling, U+2309 ISOamsc -->
303 initialMap.put("\u230A", "⌊"); // left floor = apl downstile,U+230A ISOamsc -->
304 initialMap.put("\u230B", "⌋"); // right floor, U+230B ISOamsc -->
305 initialMap.put("\u2329", "⟨"); // left-pointing angle bracket = bra,U+2329 ISOtech -->
306 // <!-- lang is NOT the same character as U+003C 'less than' or U+2039 'single left-pointing angle quotation
307 // mark' -->
308 initialMap.put("\u232A", "⟩"); // right-pointing angle bracket = ket,U+232A ISOtech -->
309 // <!-- rang is NOT the same character as U+003E 'greater than' or U+203A
310 // 'single right-pointing angle quotation mark' -->
311 // <!-- Geometric Shapes -->
312 initialMap.put("\u25CA", "◊"); // lozenge, U+25CA ISOpub -->
313 // <!-- Miscellaneous Symbols -->
314 initialMap.put("\u2660", "♠"); // black spade suit, U+2660 ISOpub -->
315 // <!-- black here seems to mean filled as opposed to hollow -->
316 initialMap.put("\u2663", "♣"); // black club suit = shamrock,U+2663 ISOpub -->
317 initialMap.put("\u2665", "♥"); // black heart suit = valentine,U+2665 ISOpub -->
318 initialMap.put("\u2666", "♦"); // black diamond suit, U+2666 ISOpub -->
319
320 // <!-- Latin Extended-A -->
321 initialMap.put("\u0152", "Œ"); // -- latin capital ligature OE,U+0152 ISOlat2 -->
322 initialMap.put("\u0153", "œ"); // -- latin small ligature oe, U+0153 ISOlat2 -->
323 // <!-- ligature is a misnomer, this is a separate character in some languages -->
324 initialMap.put("\u0160", "Š"); // -- latin capital letter S with caron,U+0160 ISOlat2 -->
325 initialMap.put("\u0161", "š"); // -- latin small letter s with caron,U+0161 ISOlat2 -->
326 initialMap.put("\u0178", "Ÿ"); // -- latin capital letter Y with dieresis,U+0178 ISOlat2 -->
327 // <!-- Spacing Modifier Letters -->
328 initialMap.put("\u02C6", "ˆ"); // -- modifier letter circumflex accent,U+02C6 ISOpub -->
329 initialMap.put("\u02DC", "˜"); // small tilde, U+02DC ISOdia -->
330 // <!-- General Punctuation -->
331 initialMap.put("\u2002", " "); // en space, U+2002 ISOpub -->
332 initialMap.put("\u2003", " "); // em space, U+2003 ISOpub -->
333 initialMap.put("\u2009", " "); // thin space, U+2009 ISOpub -->
334 initialMap.put("\u200C", "‌"); // zero width non-joiner,U+200C NEW RFC 2070 -->
335 initialMap.put("\u200D", "‍"); // zero width joiner, U+200D NEW RFC 2070 -->
336 initialMap.put("\u200E", "‎"); // left-to-right mark, U+200E NEW RFC 2070 -->
337 initialMap.put("\u200F", "‏"); // right-to-left mark, U+200F NEW RFC 2070 -->
338 initialMap.put("\u2013", "–"); // en dash, U+2013 ISOpub -->
339 initialMap.put("\u2014", "—"); // em dash, U+2014 ISOpub -->
340 initialMap.put("\u2018", "‘"); // left single quotation mark,U+2018 ISOnum -->
341 initialMap.put("\u2019", "’"); // right single quotation mark,U+2019 ISOnum -->
342 initialMap.put("\u201A", "‚"); // single low-9 quotation mark, U+201A NEW -->
343 initialMap.put("\u201C", "“"); // left double quotation mark,U+201C ISOnum -->
344 initialMap.put("\u201D", "”"); // right double quotation mark,U+201D ISOnum -->
345 initialMap.put("\u201E", "„"); // double low-9 quotation mark, U+201E NEW -->
346 initialMap.put("\u2020", "†"); // dagger, U+2020 ISOpub -->
347 initialMap.put("\u2021", "‡"); // double dagger, U+2021 ISOpub -->
348 initialMap.put("\u2030", "‰"); // per mille sign, U+2030 ISOtech -->
349 initialMap.put("\u2039", "‹"); // single left-pointing angle quotation mark,U+2039 ISO proposed -->
350 // <!-- lsaquo is proposed but not yet ISO standardized -->
351 initialMap.put("\u203A", "›"); // single right-pointing angle quotation mark,U+203A ISO proposed -->
352 // <!-- rsaquo is proposed but not yet ISO standardized -->
353 initialMap.put("\u20AC", "€"); // -- euro sign, U+20AC NEW -->
354 HTML40_EXTENDED_ESCAPE = Collections.unmodifiableMap(initialMap);
355 }
356
357 /**
358 * Reverse of {@link #HTML40_EXTENDED_ESCAPE} for unescaping purposes.
359 */
360 public static final Map<CharSequence, CharSequence> HTML40_EXTENDED_UNESCAPE;
361
362 static {
363 HTML40_EXTENDED_UNESCAPE = Collections.unmodifiableMap(invert(HTML40_EXTENDED_ESCAPE));
364 }
365
366 /**
367 * A Map<CharSequence, CharSequence> to escape the basic XML and HTML
368 * character entities.
369 *
370 * Namely: {@code " & < >}
371 */
372 public static final Map<CharSequence, CharSequence> BASIC_ESCAPE;
373
374 static {
375 final Map<CharSequence, CharSequence> initialMap = new HashMap<>();
376 initialMap.put("\"", """); // " - double-quote
377 initialMap.put("&", "&"); // & - ampersand
378 initialMap.put("<", "<"); // < - less-than
379 initialMap.put(">", ">"); // > - greater-than
380 BASIC_ESCAPE = Collections.unmodifiableMap(initialMap);
381 }
382
383 /**
384 * Reverse of {@link #BASIC_ESCAPE} for unescaping purposes.
385 */
386 public static final Map<CharSequence, CharSequence> BASIC_UNESCAPE;
387
388 static {
389 BASIC_UNESCAPE = Collections.unmodifiableMap(invert(BASIC_ESCAPE));
390 }
391
392 /**
393 * A Map<CharSequence, CharSequence> to escape the apostrophe character to
394 * its XML character entity.
395 */
396 public static final Map<CharSequence, CharSequence> APOS_ESCAPE;
397
398 static {
399 final Map<CharSequence, CharSequence> initialMap = new HashMap<>();
400 initialMap.put("'", "'"); // XML apostrophe
401 APOS_ESCAPE = Collections.unmodifiableMap(initialMap);
402 }
403
404 /**
405 * Reverse of {@link #APOS_ESCAPE} for unescaping purposes.
406 */
407 public static final Map<CharSequence, CharSequence> APOS_UNESCAPE;
408
409 static {
410 APOS_UNESCAPE = Collections.unmodifiableMap(invert(APOS_ESCAPE));
411 }
412
413 /**
414 * A Map<CharSequence, CharSequence> to escape the Java
415 * control characters.
416 *
417 * Namely: {@code \b \n \t \f \r}
418 */
419 public static final Map<CharSequence, CharSequence> JAVA_CTRL_CHARS_ESCAPE;
420
421 static {
422 final Map<CharSequence, CharSequence> initialMap = new HashMap<>();
423 initialMap.put("\b", "\\b");
424 initialMap.put("\n", "\\n");
425 initialMap.put("\t", "\\t");
426 initialMap.put("\f", "\\f");
427 initialMap.put("\r", "\\r");
428 JAVA_CTRL_CHARS_ESCAPE = Collections.unmodifiableMap(initialMap);
429 }
430
431 /**
432 * Reverse of {@link #JAVA_CTRL_CHARS_ESCAPE} for unescaping purposes.
433 */
434 public static final Map<CharSequence, CharSequence> JAVA_CTRL_CHARS_UNESCAPE;
435
436 static {
437 JAVA_CTRL_CHARS_UNESCAPE = Collections.unmodifiableMap(invert(JAVA_CTRL_CHARS_ESCAPE));
438 }
439
440 /**
441 * Inverts an escape Map into an unescape Map.
442 *
443 * @param map Map<String, String> to be inverted
444 * @return Map<String, String> inverted array
445 */
446 public static Map<CharSequence, CharSequence> invert(final Map<CharSequence, CharSequence> map) {
447 return map.entrySet().stream().collect(Collectors.toMap(Entry::getValue, Entry::getKey));
448 }
449
450 }