Python code coverage for Lib/htmlentitydefs.py

#	count	content
1	1	"""HTML character entity references."""
2	n/a
3	n/a	# maps the HTML entity name to the Unicode codepoint
4	1	name2codepoint = {
5	1	'AElig': 0x00c6, # latin capital letter AE = latin capital ligature AE, U+00C6 ISOlat1
6	1	'Aacute': 0x00c1, # latin capital letter A with acute, U+00C1 ISOlat1
7	1	'Acirc': 0x00c2, # latin capital letter A with circumflex, U+00C2 ISOlat1
8	1	'Agrave': 0x00c0, # latin capital letter A with grave = latin capital letter A grave, U+00C0 ISOlat1
9	1	'Alpha': 0x0391, # greek capital letter alpha, U+0391
10	1	'Aring': 0x00c5, # latin capital letter A with ring above = latin capital letter A ring, U+00C5 ISOlat1
11	1	'Atilde': 0x00c3, # latin capital letter A with tilde, U+00C3 ISOlat1
12	1	'Auml': 0x00c4, # latin capital letter A with diaeresis, U+00C4 ISOlat1
13	1	'Beta': 0x0392, # greek capital letter beta, U+0392
14	1	'Ccedil': 0x00c7, # latin capital letter C with cedilla, U+00C7 ISOlat1
15	1	'Chi': 0x03a7, # greek capital letter chi, U+03A7
16	1	'Dagger': 0x2021, # double dagger, U+2021 ISOpub
17	1	'Delta': 0x0394, # greek capital letter delta, U+0394 ISOgrk3
18	1	'ETH': 0x00d0, # latin capital letter ETH, U+00D0 ISOlat1
19	1	'Eacute': 0x00c9, # latin capital letter E with acute, U+00C9 ISOlat1
20	1	'Ecirc': 0x00ca, # latin capital letter E with circumflex, U+00CA ISOlat1
21	1	'Egrave': 0x00c8, # latin capital letter E with grave, U+00C8 ISOlat1
22	1	'Epsilon': 0x0395, # greek capital letter epsilon, U+0395
23	1	'Eta': 0x0397, # greek capital letter eta, U+0397
24	1	'Euml': 0x00cb, # latin capital letter E with diaeresis, U+00CB ISOlat1
25	1	'Gamma': 0x0393, # greek capital letter gamma, U+0393 ISOgrk3
26	1	'Iacute': 0x00cd, # latin capital letter I with acute, U+00CD ISOlat1
27	1	'Icirc': 0x00ce, # latin capital letter I with circumflex, U+00CE ISOlat1
28	1	'Igrave': 0x00cc, # latin capital letter I with grave, U+00CC ISOlat1
29	1	'Iota': 0x0399, # greek capital letter iota, U+0399
30	1	'Iuml': 0x00cf, # latin capital letter I with diaeresis, U+00CF ISOlat1
31	1	'Kappa': 0x039a, # greek capital letter kappa, U+039A
32	1	'Lambda': 0x039b, # greek capital letter lambda, U+039B ISOgrk3
33	1	'Mu': 0x039c, # greek capital letter mu, U+039C
34	1	'Ntilde': 0x00d1, # latin capital letter N with tilde, U+00D1 ISOlat1
35	1	'Nu': 0x039d, # greek capital letter nu, U+039D
36	1	'OElig': 0x0152, # latin capital ligature OE, U+0152 ISOlat2
37	1	'Oacute': 0x00d3, # latin capital letter O with acute, U+00D3 ISOlat1
38	1	'Ocirc': 0x00d4, # latin capital letter O with circumflex, U+00D4 ISOlat1
39	1	'Ograve': 0x00d2, # latin capital letter O with grave, U+00D2 ISOlat1
40	1	'Omega': 0x03a9, # greek capital letter omega, U+03A9 ISOgrk3
41	1	'Omicron': 0x039f, # greek capital letter omicron, U+039F
42	1	'Oslash': 0x00d8, # latin capital letter O with stroke = latin capital letter O slash, U+00D8 ISOlat1
43	1	'Otilde': 0x00d5, # latin capital letter O with tilde, U+00D5 ISOlat1
44	1	'Ouml': 0x00d6, # latin capital letter O with diaeresis, U+00D6 ISOlat1
45	1	'Phi': 0x03a6, # greek capital letter phi, U+03A6 ISOgrk3
46	1	'Pi': 0x03a0, # greek capital letter pi, U+03A0 ISOgrk3
47	1	'Prime': 0x2033, # double prime = seconds = inches, U+2033 ISOtech
48	1	'Psi': 0x03a8, # greek capital letter psi, U+03A8 ISOgrk3
49	1	'Rho': 0x03a1, # greek capital letter rho, U+03A1
50	1	'Scaron': 0x0160, # latin capital letter S with caron, U+0160 ISOlat2
51	1	'Sigma': 0x03a3, # greek capital letter sigma, U+03A3 ISOgrk3
52	1	'THORN': 0x00de, # latin capital letter THORN, U+00DE ISOlat1
53	1	'Tau': 0x03a4, # greek capital letter tau, U+03A4
54	1	'Theta': 0x0398, # greek capital letter theta, U+0398 ISOgrk3
55	1	'Uacute': 0x00da, # latin capital letter U with acute, U+00DA ISOlat1
56	1	'Ucirc': 0x00db, # latin capital letter U with circumflex, U+00DB ISOlat1
57	1	'Ugrave': 0x00d9, # latin capital letter U with grave, U+00D9 ISOlat1
58	1	'Upsilon': 0x03a5, # greek capital letter upsilon, U+03A5 ISOgrk3
59	1	'Uuml': 0x00dc, # latin capital letter U with diaeresis, U+00DC ISOlat1
60	1	'Xi': 0x039e, # greek capital letter xi, U+039E ISOgrk3
61	1	'Yacute': 0x00dd, # latin capital letter Y with acute, U+00DD ISOlat1
62	1	'Yuml': 0x0178, # latin capital letter Y with diaeresis, U+0178 ISOlat2
63	1	'Zeta': 0x0396, # greek capital letter zeta, U+0396
64	1	'aacute': 0x00e1, # latin small letter a with acute, U+00E1 ISOlat1
65	1	'acirc': 0x00e2, # latin small letter a with circumflex, U+00E2 ISOlat1
66	1	'acute': 0x00b4, # acute accent = spacing acute, U+00B4 ISOdia
67	1	'aelig': 0x00e6, # latin small letter ae = latin small ligature ae, U+00E6 ISOlat1
68	1	'agrave': 0x00e0, # latin small letter a with grave = latin small letter a grave, U+00E0 ISOlat1
69	1	'alefsym': 0x2135, # alef symbol = first transfinite cardinal, U+2135 NEW
70	1	'alpha': 0x03b1, # greek small letter alpha, U+03B1 ISOgrk3
71	1	'amp': 0x0026, # ampersand, U+0026 ISOnum
72	1	'and': 0x2227, # logical and = wedge, U+2227 ISOtech
73	1	'ang': 0x2220, # angle, U+2220 ISOamso
74	1	'aring': 0x00e5, # latin small letter a with ring above = latin small letter a ring, U+00E5 ISOlat1
75	1	'asymp': 0x2248, # almost equal to = asymptotic to, U+2248 ISOamsr
76	1	'atilde': 0x00e3, # latin small letter a with tilde, U+00E3 ISOlat1
77	1	'auml': 0x00e4, # latin small letter a with diaeresis, U+00E4 ISOlat1
78	1	'bdquo': 0x201e, # double low-9 quotation mark, U+201E NEW
79	1	'beta': 0x03b2, # greek small letter beta, U+03B2 ISOgrk3
80	1	'brvbar': 0x00a6, # broken bar = broken vertical bar, U+00A6 ISOnum
81	1	'bull': 0x2022, # bullet = black small circle, U+2022 ISOpub
82	1	'cap': 0x2229, # intersection = cap, U+2229 ISOtech
83	1	'ccedil': 0x00e7, # latin small letter c with cedilla, U+00E7 ISOlat1
84	1	'cedil': 0x00b8, # cedilla = spacing cedilla, U+00B8 ISOdia
85	1	'cent': 0x00a2, # cent sign, U+00A2 ISOnum
86	1	'chi': 0x03c7, # greek small letter chi, U+03C7 ISOgrk3
87	1	'circ': 0x02c6, # modifier letter circumflex accent, U+02C6 ISOpub
88	1	'clubs': 0x2663, # black club suit = shamrock, U+2663 ISOpub
89	1	'cong': 0x2245, # approximately equal to, U+2245 ISOtech
90	1	'copy': 0x00a9, # copyright sign, U+00A9 ISOnum
91	1	'crarr': 0x21b5, # downwards arrow with corner leftwards = carriage return, U+21B5 NEW
92	1	'cup': 0x222a, # union = cup, U+222A ISOtech
93	1	'curren': 0x00a4, # currency sign, U+00A4 ISOnum
94	1	'dArr': 0x21d3, # downwards double arrow, U+21D3 ISOamsa
95	1	'dagger': 0x2020, # dagger, U+2020 ISOpub
96	1	'darr': 0x2193, # downwards arrow, U+2193 ISOnum
97	1	'deg': 0x00b0, # degree sign, U+00B0 ISOnum
98	1	'delta': 0x03b4, # greek small letter delta, U+03B4 ISOgrk3
99	1	'diams': 0x2666, # black diamond suit, U+2666 ISOpub
100	1	'divide': 0x00f7, # division sign, U+00F7 ISOnum
101	1	'eacute': 0x00e9, # latin small letter e with acute, U+00E9 ISOlat1
102	1	'ecirc': 0x00ea, # latin small letter e with circumflex, U+00EA ISOlat1
103	1	'egrave': 0x00e8, # latin small letter e with grave, U+00E8 ISOlat1
104	1	'empty': 0x2205, # empty set = null set = diameter, U+2205 ISOamso
105	1	'emsp': 0x2003, # em space, U+2003 ISOpub
106	1	'ensp': 0x2002, # en space, U+2002 ISOpub
107	1	'epsilon': 0x03b5, # greek small letter epsilon, U+03B5 ISOgrk3
108	1	'equiv': 0x2261, # identical to, U+2261 ISOtech
109	1	'eta': 0x03b7, # greek small letter eta, U+03B7 ISOgrk3
110	1	'eth': 0x00f0, # latin small letter eth, U+00F0 ISOlat1
111	1	'euml': 0x00eb, # latin small letter e with diaeresis, U+00EB ISOlat1
112	1	'euro': 0x20ac, # euro sign, U+20AC NEW
113	1	'exist': 0x2203, # there exists, U+2203 ISOtech
114	1	'fnof': 0x0192, # latin small f with hook = function = florin, U+0192 ISOtech
115	1	'forall': 0x2200, # for all, U+2200 ISOtech
116	1	'frac12': 0x00bd, # vulgar fraction one half = fraction one half, U+00BD ISOnum
117	1	'frac14': 0x00bc, # vulgar fraction one quarter = fraction one quarter, U+00BC ISOnum
118	1	'frac34': 0x00be, # vulgar fraction three quarters = fraction three quarters, U+00BE ISOnum
119	1	'frasl': 0x2044, # fraction slash, U+2044 NEW
120	1	'gamma': 0x03b3, # greek small letter gamma, U+03B3 ISOgrk3
121	1	'ge': 0x2265, # greater-than or equal to, U+2265 ISOtech
122	1	'gt': 0x003e, # greater-than sign, U+003E ISOnum
123	1	'hArr': 0x21d4, # left right double arrow, U+21D4 ISOamsa
124	1	'harr': 0x2194, # left right arrow, U+2194 ISOamsa
125	1	'hearts': 0x2665, # black heart suit = valentine, U+2665 ISOpub
126	1	'hellip': 0x2026, # horizontal ellipsis = three dot leader, U+2026 ISOpub
127	1	'iacute': 0x00ed, # latin small letter i with acute, U+00ED ISOlat1
128	1	'icirc': 0x00ee, # latin small letter i with circumflex, U+00EE ISOlat1
129	1	'iexcl': 0x00a1, # inverted exclamation mark, U+00A1 ISOnum
130	1	'igrave': 0x00ec, # latin small letter i with grave, U+00EC ISOlat1
131	1	'image': 0x2111, # blackletter capital I = imaginary part, U+2111 ISOamso
132	1	'infin': 0x221e, # infinity, U+221E ISOtech
133	1	'int': 0x222b, # integral, U+222B ISOtech
134	1	'iota': 0x03b9, # greek small letter iota, U+03B9 ISOgrk3
135	1	'iquest': 0x00bf, # inverted question mark = turned question mark, U+00BF ISOnum
136	1	'isin': 0x2208, # element of, U+2208 ISOtech
137	1	'iuml': 0x00ef, # latin small letter i with diaeresis, U+00EF ISOlat1
138	1	'kappa': 0x03ba, # greek small letter kappa, U+03BA ISOgrk3
139	1	'lArr': 0x21d0, # leftwards double arrow, U+21D0 ISOtech
140	1	'lambda': 0x03bb, # greek small letter lambda, U+03BB ISOgrk3
141	1	'lang': 0x2329, # left-pointing angle bracket = bra, U+2329 ISOtech
142	1	'laquo': 0x00ab, # left-pointing double angle quotation mark = left pointing guillemet, U+00AB ISOnum
143	1	'larr': 0x2190, # leftwards arrow, U+2190 ISOnum
144	1	'lceil': 0x2308, # left ceiling = apl upstile, U+2308 ISOamsc
145	1	'ldquo': 0x201c, # left double quotation mark, U+201C ISOnum
146	1	'le': 0x2264, # less-than or equal to, U+2264 ISOtech
147	1	'lfloor': 0x230a, # left floor = apl downstile, U+230A ISOamsc
148	1	'lowast': 0x2217, # asterisk operator, U+2217 ISOtech
149	1	'loz': 0x25ca, # lozenge, U+25CA ISOpub
150	1	'lrm': 0x200e, # left-to-right mark, U+200E NEW RFC 2070
151	1	'lsaquo': 0x2039, # single left-pointing angle quotation mark, U+2039 ISO proposed
152	1	'lsquo': 0x2018, # left single quotation mark, U+2018 ISOnum
153	1	'lt': 0x003c, # less-than sign, U+003C ISOnum
154	1	'macr': 0x00af, # macron = spacing macron = overline = APL overbar, U+00AF ISOdia
155	1	'mdash': 0x2014, # em dash, U+2014 ISOpub
156	1	'micro': 0x00b5, # micro sign, U+00B5 ISOnum
157	1	'middot': 0x00b7, # middle dot = Georgian comma = Greek middle dot, U+00B7 ISOnum
158	1	'minus': 0x2212, # minus sign, U+2212 ISOtech
159	1	'mu': 0x03bc, # greek small letter mu, U+03BC ISOgrk3
160	1	'nabla': 0x2207, # nabla = backward difference, U+2207 ISOtech
161	1	'nbsp': 0x00a0, # no-break space = non-breaking space, U+00A0 ISOnum
162	1	'ndash': 0x2013, # en dash, U+2013 ISOpub
163	1	'ne': 0x2260, # not equal to, U+2260 ISOtech
164	1	'ni': 0x220b, # contains as member, U+220B ISOtech
165	1	'not': 0x00ac, # not sign, U+00AC ISOnum
166	1	'notin': 0x2209, # not an element of, U+2209 ISOtech
167	1	'nsub': 0x2284, # not a subset of, U+2284 ISOamsn
168	1	'ntilde': 0x00f1, # latin small letter n with tilde, U+00F1 ISOlat1
169	1	'nu': 0x03bd, # greek small letter nu, U+03BD ISOgrk3
170	1	'oacute': 0x00f3, # latin small letter o with acute, U+00F3 ISOlat1
171	1	'ocirc': 0x00f4, # latin small letter o with circumflex, U+00F4 ISOlat1
172	1	'oelig': 0x0153, # latin small ligature oe, U+0153 ISOlat2
173	1	'ograve': 0x00f2, # latin small letter o with grave, U+00F2 ISOlat1
174	1	'oline': 0x203e, # overline = spacing overscore, U+203E NEW
175	1	'omega': 0x03c9, # greek small letter omega, U+03C9 ISOgrk3
176	1	'omicron': 0x03bf, # greek small letter omicron, U+03BF NEW
177	1	'oplus': 0x2295, # circled plus = direct sum, U+2295 ISOamsb
178	1	'or': 0x2228, # logical or = vee, U+2228 ISOtech
179	1	'ordf': 0x00aa, # feminine ordinal indicator, U+00AA ISOnum
180	1	'ordm': 0x00ba, # masculine ordinal indicator, U+00BA ISOnum
181	1	'oslash': 0x00f8, # latin small letter o with stroke, = latin small letter o slash, U+00F8 ISOlat1
182	1	'otilde': 0x00f5, # latin small letter o with tilde, U+00F5 ISOlat1
183	1	'otimes': 0x2297, # circled times = vector product, U+2297 ISOamsb
184	1	'ouml': 0x00f6, # latin small letter o with diaeresis, U+00F6 ISOlat1
185	1	'para': 0x00b6, # pilcrow sign = paragraph sign, U+00B6 ISOnum
186	1	'part': 0x2202, # partial differential, U+2202 ISOtech
187	1	'permil': 0x2030, # per mille sign, U+2030 ISOtech
188	1	'perp': 0x22a5, # up tack = orthogonal to = perpendicular, U+22A5 ISOtech
189	1	'phi': 0x03c6, # greek small letter phi, U+03C6 ISOgrk3
190	1	'pi': 0x03c0, # greek small letter pi, U+03C0 ISOgrk3
191	1	'piv': 0x03d6, # greek pi symbol, U+03D6 ISOgrk3
192	1	'plusmn': 0x00b1, # plus-minus sign = plus-or-minus sign, U+00B1 ISOnum
193	1	'pound': 0x00a3, # pound sign, U+00A3 ISOnum
194	1	'prime': 0x2032, # prime = minutes = feet, U+2032 ISOtech
195	1	'prod': 0x220f, # n-ary product = product sign, U+220F ISOamsb
196	1	'prop': 0x221d, # proportional to, U+221D ISOtech
197	1	'psi': 0x03c8, # greek small letter psi, U+03C8 ISOgrk3
198	1	'quot': 0x0022, # quotation mark = APL quote, U+0022 ISOnum
199	1	'rArr': 0x21d2, # rightwards double arrow, U+21D2 ISOtech
200	1	'radic': 0x221a, # square root = radical sign, U+221A ISOtech
201	1	'rang': 0x232a, # right-pointing angle bracket = ket, U+232A ISOtech
202	1	'raquo': 0x00bb, # right-pointing double angle quotation mark = right pointing guillemet, U+00BB ISOnum
203	1	'rarr': 0x2192, # rightwards arrow, U+2192 ISOnum
204	1	'rceil': 0x2309, # right ceiling, U+2309 ISOamsc
205	1	'rdquo': 0x201d, # right double quotation mark, U+201D ISOnum
206	1	'real': 0x211c, # blackletter capital R = real part symbol, U+211C ISOamso
207	1	'reg': 0x00ae, # registered sign = registered trade mark sign, U+00AE ISOnum
208	1	'rfloor': 0x230b, # right floor, U+230B ISOamsc
209	1	'rho': 0x03c1, # greek small letter rho, U+03C1 ISOgrk3
210	1	'rlm': 0x200f, # right-to-left mark, U+200F NEW RFC 2070
211	1	'rsaquo': 0x203a, # single right-pointing angle quotation mark, U+203A ISO proposed
212	1	'rsquo': 0x2019, # right single quotation mark, U+2019 ISOnum
213	1	'sbquo': 0x201a, # single low-9 quotation mark, U+201A NEW
214	1	'scaron': 0x0161, # latin small letter s with caron, U+0161 ISOlat2
215	1	'sdot': 0x22c5, # dot operator, U+22C5 ISOamsb
216	1	'sect': 0x00a7, # section sign, U+00A7 ISOnum
217	1	'shy': 0x00ad, # soft hyphen = discretionary hyphen, U+00AD ISOnum
218	1	'sigma': 0x03c3, # greek small letter sigma, U+03C3 ISOgrk3
219	1	'sigmaf': 0x03c2, # greek small letter final sigma, U+03C2 ISOgrk3
220	1	'sim': 0x223c, # tilde operator = varies with = similar to, U+223C ISOtech
221	1	'spades': 0x2660, # black spade suit, U+2660 ISOpub
222	1	'sub': 0x2282, # subset of, U+2282 ISOtech
223	1	'sube': 0x2286, # subset of or equal to, U+2286 ISOtech
224	1	'sum': 0x2211, # n-ary sumation, U+2211 ISOamsb
225	1	'sup': 0x2283, # superset of, U+2283 ISOtech
226	1	'sup1': 0x00b9, # superscript one = superscript digit one, U+00B9 ISOnum
227	1	'sup2': 0x00b2, # superscript two = superscript digit two = squared, U+00B2 ISOnum
228	1	'sup3': 0x00b3, # superscript three = superscript digit three = cubed, U+00B3 ISOnum
229	1	'supe': 0x2287, # superset of or equal to, U+2287 ISOtech
230	1	'szlig': 0x00df, # latin small letter sharp s = ess-zed, U+00DF ISOlat1
231	1	'tau': 0x03c4, # greek small letter tau, U+03C4 ISOgrk3
232	1	'there4': 0x2234, # therefore, U+2234 ISOtech
233	1	'theta': 0x03b8, # greek small letter theta, U+03B8 ISOgrk3
234	1	'thetasym': 0x03d1, # greek small letter theta symbol, U+03D1 NEW
235	1	'thinsp': 0x2009, # thin space, U+2009 ISOpub
236	1	'thorn': 0x00fe, # latin small letter thorn with, U+00FE ISOlat1
237	1	'tilde': 0x02dc, # small tilde, U+02DC ISOdia
238	1	'times': 0x00d7, # multiplication sign, U+00D7 ISOnum
239	1	'trade': 0x2122, # trade mark sign, U+2122 ISOnum
240	1	'uArr': 0x21d1, # upwards double arrow, U+21D1 ISOamsa
241	1	'uacute': 0x00fa, # latin small letter u with acute, U+00FA ISOlat1
242	1	'uarr': 0x2191, # upwards arrow, U+2191 ISOnum
243	1	'ucirc': 0x00fb, # latin small letter u with circumflex, U+00FB ISOlat1
244	1	'ugrave': 0x00f9, # latin small letter u with grave, U+00F9 ISOlat1
245	1	'uml': 0x00a8, # diaeresis = spacing diaeresis, U+00A8 ISOdia
246	1	'upsih': 0x03d2, # greek upsilon with hook symbol, U+03D2 NEW
247	1	'upsilon': 0x03c5, # greek small letter upsilon, U+03C5 ISOgrk3
248	1	'uuml': 0x00fc, # latin small letter u with diaeresis, U+00FC ISOlat1
249	1	'weierp': 0x2118, # script capital P = power set = Weierstrass p, U+2118 ISOamso
250	1	'xi': 0x03be, # greek small letter xi, U+03BE ISOgrk3
251	1	'yacute': 0x00fd, # latin small letter y with acute, U+00FD ISOlat1
252	1	'yen': 0x00a5, # yen sign = yuan sign, U+00A5 ISOnum
253	1	'yuml': 0x00ff, # latin small letter y with diaeresis, U+00FF ISOlat1
254	1	'zeta': 0x03b6, # greek small letter zeta, U+03B6 ISOgrk3
255	1	'zwj': 0x200d, # zero width joiner, U+200D NEW RFC 2070
256	1	'zwnj': 0x200c, # zero width non-joiner, U+200C NEW RFC 2070
257	n/a	}
258	n/a
259	n/a	# maps the Unicode codepoint to the HTML entity name
260	1	codepoint2name = {}
261	n/a
262	n/a	# maps the HTML entity name to the character
263	n/a	# (or a character reference if the character is outside the Latin-1 range)
264	1	entitydefs = {}
265	n/a
266	253	for (name, codepoint) in name2codepoint.iteritems():
267	252	codepoint2name[codepoint] = name
268	252	if codepoint <= 0xff:
269	100	entitydefs[name] = chr(codepoint)
270	n/a	else:
271	152	entitydefs[name] = '&#%d;' % codepoint
272	n/a
273	1	del name, codepoint