"""
Extracted from the UTF-8 to ASCII conversion program found at
https://metamap.nlm.nih.gov/ReplaceUTF8.shtml
"""
UNICODE_TO_ASCII = {
'\u221E': " ",
'\u02C2': " ",
'\u2218': " ",
'\u0020': " ",
'\u0021': "!",
'\u0023': "#",
'\u0024': "$",
'\u0025': "%",
'\u0026': "&",
'\u0028': "(",
'\u0029': ")",
'\u002a': "*",
'\u002b': "+",
'\u002c': ",",
'\u002d': "-",
'\u002e': ".",
'\u002f': "/",
'\u0030': "0",
'\u0031': "1",
'\u0032': "2",
'\u0033': "3",
'\u0034': "4",
'\u0035': "5",
'\u0036': "6",
'\u0037': "7",
'\u0038': "8",
'\u0039': "9",
'\u003a': ":",
'\u003b': ";",
'\u003c': "<",
'\u003d': "=",
'\u003e': ">",
'\u003f': "?",
'\u0040': "@",
'\u0041': "A",
'\u0042': "B",
'\u0043': "C",
'\u0044': "D",
'\u0045': "E",
'\u0046': "F",
'\u0047': "G",
'\u0048': "H",
'\u0049': "I",
'\u004a': "J",
'\u004b': "K",
'\u004c': "L",
'\u004d': "M",
'\u004e': "N",
'\u004f': "O",
'\u0050': "P",
'\u0051': "Q",
'\u0052': "R",
'\u0053': "S",
'\u0054': "T",
'\u0055': "U",
'\u0056': "V",
'\u0057': "W",
'\u0058': "X",
'\u0059': "Y",
'\u005a': "Z",
'\u005b': "[",
'\u005d': "]",
'\u005e': "^",
'\u005f': "_",
'\u0060': "`",
'\u0061': "a",
'\u0062': "b",
'\u0063': "c",
'\u0064': "d",
'\u0065': "e",
'\u0066': "f",
'\u0067': "g",
'\u0068': "h",
'\u0069': "i",
'\u006a': "j",
'\u006b': "k",
'\u006c': "l",
'\u006d': "m",
'\u006e': "n",
'\u006f': "o",
'\u0070': "p",
'\u0071': "q",
'\u0072': "r",
'\u0073': "s",
'\u0074': "t",
'\u0075': "u",
'\u0076': "v",
'\u0077': "w",
'\u0078': "x",
'\u0079': "y",
'\u007a': "z",
'\u007b': "{",
'\u007c': "|",
'\u007d': "}",
'\u007e': "~",
'\u00a0': " ",
'\u00a1': "!",
'\u00a2': "c", # Cent sign
'\u00a3': "L", # Pound sign
'\u00a4': " ", # Currency sign
'\u00a5': "Y", # Yen sign
'\u00a6': "|",
'\u00a7': "S", # Section Sign
'\u00a8': "",
'\u00a9': "(c)",
'\u00aa': "a",
'\u00ac': "!", # Not sign
'\u00ad': "-",
'\u00ae': "(r)",
'\u00af': "-", # Macron
'\u00b0': " degrees ", # Degree
'\u00b1': "+/-",
'\u00b2': "2",
'\u00b3': "3",
'\u00b4': "'",
'\u00b5': "u",
'\u00b6': " ", # Paragraph
'\u00b7': ".", # Middle dot
'\u00b8': " ", # Cedilla
'\u00b9': "1",
'\u00ba': "o",
'\u00bc': "1/4",
'\u00bd': "1/2",
'\u00be': "3/4",
'\u00bf': "?", # Inverted Question Mark
'\u00c0': "A",
'\u00c1': "A",
'\u00c2': "A",
'\u00c3': "A",
'\u00c4': "A",
'\u00c5': "A",
'\u00c6': "AE",
'\u00c7': "C",
'\u00c8': "E",
'\u00c9': "E",
'\u00ca': "E",
'\u00cb': "E",
'\u00cc': "I",
'\u00cd': "I",
'\u00ce': "I",
'\u00cf': "I",
'\u00d0': "D", # Cap Eth
'\u00d1': "N",
'\u00d2': "O",
'\u00d3': "O",
'\u00d4': "O",
'\u00d5': "O",
'\u00d6': "O",
'\u00d7': "*",
'\u00d8': "O",
'\u00d9': "U",
'\u00da': "U",
'\u00db': "U",
'\u00dc': "U",
'\u00dd': "Y",
'\u00de': "P", # Cap Thorn
'\u00df': "beta",
'\u00e0': "a",
'\u00e1': "a",
'\u00e2': "a",
'\u00e3': "a",
'\u00e4': "a",
'\u00e5': "a",
'\u00e6': "ae",
'\u00e7': "c",
'\u00e8': "e",
'\u00e9': "e",
'\u00ea': "e",
'\u00eb': "e",
'\u00ec': "i",
'\u00ed': "i",
'\u00ee': "i",
'\u00ef': "i",
'\u00f1': "n",
'\u00f2': "o",
'\u00f3': "o",
'\u00f4': "o",
'\u00f5': "o",
'\u00f6': "o",
'\u00f7': "/",
'\u00f8': "o",
'\u00f9': "u",
'\u00fa': "u",
'\u00fb': "u",
'\u00fc': "u",
'\u00fd': "y",
'\u00fe': "p", # Thorn
'\u00ff': "y",
'\u0100': "A",
'\u0101': "a",
'\u0102': "A",
'\u0103': "a",
'\u0104': "A",
'\u0105': "a",
'\u0106': "C",
'\u0107': "c",
'\u0108': "C",
'\u0109': "c",
'\u010a': "C",
'\u010b': "c",
'\u010c': "C",
'\u010d': "c",
'\u010e': "D",
'\u010f': "d",
'\u0110': "D",
'\u0111': "d",
'\u0112': "E",
'\u0113': "e",
'\u0114': "E",
'\u0115': "e",
'\u0116': "E",
'\u0117': "e",
'\u0118': "E",
'\u0119': "e",
'\u011a': "E",
'\u011b': "e",
'\u011c': "G",
'\u011d': "g",
'\u011e': "G",
'\u011f': "g",
'\u0120': "G",
'\u0121': "g",
'\u0122': "G",
'\u0123': "g",
'\u0124': "H",
'\u0125': "h",
'\u0126': "H",
'\u0127': "h",
'\u0128': "I",
'\u0129': "i",
'\u012a': "I",
'\u012b': "i",
'\u012c': "I",
'\u012d': "i",
'\u012e': "I",
'\u012f': "i",
'\u0130': "I",
'\u0131': "i",
'\u0132': "IJ",
'\u0133': "ij",
'\u0134': "J",
'\u0135': "j",
'\u0136': "K",
'\u0137': "k",
'\u0139': "L",
'\u013a': "l",
'\u013b': "L",
'\u013c': "l",
'\u013d': "L",
'\u013e': "l",
'\u013f': "L",
'\u0140': "l",
'\u0141': "L",
'\u0142': "l",
'\u0143': "N",
'\u0144': "n",
'\u0145': "N",
'\u0146': "n",
'\u0147': "N",
'\u0148': "n",
'\u0149': "n",
'\u014c': "O",
'\u014d': "o",
'\u014e': "O",
'\u014f': "o",
'\u0150': "O",
'\u0151': "o",
'\u0152': "OE",
'\u0153': "oe",
'\u0154': "R",
'\u0155': "r",
'\u0156': "R",
'\u0157': "r",
'\u0158': "R",
'\u0159': "r",
'\u015a': "S",
'\u015b': "s",
'\u015c': "S",
'\u015d': "s",
'\u015e': "S",
'\u015f': "s",
'\u0160': "S",
'\u0161': "s",
'\u0162': "T",
'\u0163': "t",
'\u0164': "T",
'\u0165': "t",
'\u0166': "T",
'\u0167': "t",
'\u0168': "U",
'\u0169': "u",
'\u016a': "U",
'\u016b': "u",
'\u016c': "U",
'\u016d': "u",
'\u016e': "U",
'\u016f': "u",
'\u0170': "U",
'\u0171': "u",
'\u0172': "U",
'\u0173': "u",
'\u0174': "W",
'\u0175': "w",
'\u0176': "Y",
'\u0177': "y",
'\u0178': "Y",
'\u0179': "Z",
'\u017a': "z",
'\u017b': "Z",
'\u017c': "z",
'\u017d': "Z",
'\u017e': "z",
'\u017f': "s",
'\u0180': "b",
'\u0181': "B",
'\u0182': "B",
'\u0183': "b",
'\u0184': "6",
'\u0185': "6",
'\u0186': "O",
'\u0187': "C",
'\u0188': "c",
'\u0189': "D",
'\u018a': "D",
'\u018b': "D",
'\u018c': "d",
'\u018e': "E",
'\u0190': "E",
'\u0191': "F",
'\u0192': "f",
'\u0193': "G",
'\u0194': "Gamma",
'\u0195': "hv",
'\u0196': "Iota",
'\u0197': "I",
'\u0198': "K",
'\u0199': "k",
'\u019a': "l",
'\u019b': "lambda",
'\u019c': "M",
'\u019d': "N",
'\u019e': "n",
'\u019f': "O",
'\u01a0': "O",
'\u01a1': "o",
'\u01a2': "OI",
'\u01a3': "oi",
'\u01a4': "P",
'\u01a5': "p",
'\u01a6': "YR",
'\u01a7': "2",
'\u01a8': "2",
'\u01ab': "t",
'\u01ac': "T",
'\u01ad': "t",
'\u01ae': "T",
'\u01af': "U",
'\u01b0': "u",
'\u01b1': "Upsilon",
'\u01b2': "V",
'\u01b3': "Y",
'\u01b4': "y",
'\u01b5': "Z",
'\u01b6': "z",
'\u01bb': "2",
'\u01bc': "5",
'\u01bd': "5",
'\u01c0': "|",
'\u01c3': "!",
'\u01c4': "DZ",
'\u01c5': "Dz",
'\u01c6': "dz",
'\u01c7': "LJ",
'\u01c8': "Lj",
'\u01c9': "lj",
'\u01ca': "NJ",
'\u01cb': "Nj",
'\u01cc': "nj",
'\u01cd': "A",
'\u01ce': "a",
'\u01cf': "I",
'\u01d0': "i",
'\u01d1': "O",
'\u01d2': "o",
'\u01d3': "U",
'\u01d4': "u",
'\u01d5': "U",
'\u01d6': "u",
'\u01d7': "U",
'\u01d8': "u",
'\u01d9': "U",
'\u01da': "u",
'\u01db': "U",
'\u01dc': "u",
'\u01dd': "e",
'\u01de': "A",
'\u01df': "a",
'\u01e0': "A",
'\u01e1': "a",
'\u01e2': "AE",
'\u01e3': "ae",
'\u01e4': "G",
'\u01e5': "g",
'\u01e6': "G",
'\u01e7': "g",
'\u01e8': "K",
'\u01e9': "k",
'\u01ea': "O",
'\u01eb': "o",
'\u01ec': "O",
'\u01ed': "o",
'\u01f0': "j",
'\u01f1': "DZ",
'\u01f2': "Dz",
'\u01f3': "dz",
'\u01f4': "G",
'\u01f5': "g",
'\u01f8': "N",
'\u01f9': "n",
'\u01fa': "A",
'\u01fb': "a",
'\u01fc': "AE",
'\u01fd': "ae",
'\u01fe': "O",
'\u01ff': "o",
'\u0200': "A",
'\u0201': "a",
'\u0202': "A",
'\u0203': "a",
'\u0204': "E",
'\u0205': "e",
'\u0206': "E",
'\u0207': "e",
'\u0208': "I",
'\u0209': "i",
'\u020a': "I",
'\u020b': "i",
'\u020c': "O",
'\u020d': "o",
'\u020e': "O",
'\u020f': "o",
'\u0210': "R",
'\u0211': "r",
'\u0212': "R",
'\u0213': "r",
'\u0214': "U",
'\u0215': "u",
'\u0216': "U",
'\u0217': "u",
'\u0218': "S",
'\u0219': "s",
'\u021a': "T",
'\u021b': "t",
'\u021e': "H",
'\u021f': "h",
'\u0220': "N",
'\u0221': "d",
'\u0222': "OU",
'\u0223': "ou",
'\u0224': "Z",
'\u0225': "z",
'\u0226': "A",
'\u0227': "a",
'\u0228': "E",
'\u0229': "e",
'\u022a': "O",
'\u022b': "o",
'\u022c': "O",
'\u022d': "o",
'\u022e': "O",
'\u022f': "o",
'\u0230': "O",
'\u0231': "o",
'\u0232': "Y",
'\u0233': "y",
'\u0234': "l",
'\u0235': "n",
'\u0236': "t",
'\u0237': "j",
'\u0238': "db",
'\u0239': "qp",
'\u023a': "A",
'\u023b': "C",
'\u023c': "c",
'\u023d': "L",
'\u023e': "T",
'\u023f': "s",
'\u0240': "z",
'\u0243': "B",
'\u0244': "U",
'\u0245': "V",
'\u0246': "E",
'\u0247': "e",
'\u0248': "J",
'\u0249': "j",
'\u024a': "Q",
'\u024b': "q",
'\u024c': "R",
'\u024d': "r",
'\u024e': "Y",
'\u024f': "y",
'\u0250': "a",
'\u0251': "alpha",
'\u0252': "alpha",
'\u0253': "b",
'\u0254': "o",
'\u0255': "c",
'\u0256': "d",
'\u0257': "d",
'\u025b': "E",
'\u025f': "j",
'\u0260': "g",
'\u0261': "g",
'\u0262': "G",
'\u0263': "gamma",
'\u0265': "h",
'\u0266': "h",
'\u0268': "i",
'\u0269': "iota",
'\u026a': "i",
'\u026b': "l",
'\u026c': "l",
'\u026d': "l",
'\u026f': "m",
'\u0270': "m",
'\u0271': "m",
'\u0272': "n",
'\u0273': "n",
'\u0274': "N",
'\u0275': "o",
'\u0276': "OE",
'\u0277': "omega",
'\u0278': "phi",
'\u0279': "r",
'\u027a': "r",
'\u027b': "r",
'\u027c': "r",
'\u027d': "r",
'\u027e': "r",
'\u0280': "R",
'\u0282': "s",
'\u0284': "j",
'\u0287': "t",
'\u0288': "t",
'\u0289': "u",
'\u028a': "upsilon",
'\u028b': "v",
'\u028c': "v",
'\u028d': "w",
'\u028e': "y",
'\u028f': "Y",
'\u0290': "z",
'\u0291': "z",
'\u0297': "C",
'\u0299': "B",
'\u029a': "e",
'\u029b': "G",
'\u029c': "H",
'\u029d': "j",
'\u029e': "k",
'\u029f': "L",
'\u02a0': "q",
'\u02a3': "DZ",
'\u02a5': "dz",
'\u02a6': "ts",
'\u02a8': "tc",
'\u02aa': "ls",
'\u02ab': "lz",
'\u02ae': "h",
'\u02af': "h",
'\u02c6': "^",
'\u02dc': "~",
'\u037e': ";",
'\u0386': "Alpha",
'\u0388': "Epsilon",
'\u0389': "Eta",
'\u038a': "Iota",
'\u038c': "Omicron",
'\u038e': "Upsilon",
'\u038f': "Omega",
'\u0390': "iota",
'\u0391': "Alpha",
'\u0392': "Beta",
'\u0393': "Gamma",
'\u0394': "Delta",
'\u0395': "Epsilon",
'\u0396': "Zeta",
'\u0397': "Eta",
'\u0398': "Theta",
'\u0399': "Iota",
'\u039a': "Kappa",
'\u039b': "Lambda",
'\u039c': "Mu",
'\u039d': "Nu",
'\u039e': "Xi",
'\u039f': "Omicron",
'\u03a0': "Pi",
'\u03a1': "Rho",
'\u03a3': "Sigma",
'\u03a4': "Tau",
'\u03a5': "Upsilon",
'\u03a6': "Phi",
'\u03a7': "Chi",
'\u03a8': "Psi",
'\u03a9': "Omega",
'\u03aa': "Iota",
'\u03ab': "Upsilon",
'\u03ac': "alpha",
'\u03ad': "epsilon",
'\u03ae': "eta",
'\u03af': "iota",
'\u03b0': "upsilon",
'\u03b1': "alpha",
'\u03b2': "beta",
'\u03b3': "gamma",
'\u03b4': "delta",
'\u03b5': "epsilon",
'\u03b6': "zeta",
'\u03b7': "eta",
'\u03b8': "theta",
'\u03b9': "iota",
'\u03ba': "kappa",
'\u03bb': "lambda",
'\u03bc': "mu",
'\u03bd': "nu",
'\u03be': "xi",
'\u03bf': "omicron",
'\u03c0': "pi",
'\u03c1': "rho",
'\u03c2': "sigma",
'\u03c3': "sigma",
'\u03c4': "tau",
'\u03c5': "upsilon",
'\u03c6': "phi",
'\u03c7': "chi",
'\u03c8': "psi",
'\u03c9': "omega",
'\u03ca': "iota",
'\u03cb': "upsilon",
'\u03cc': "omicron",
'\u03cd': "upsilon",
'\u03ce': "omega",
'\u03d0': "beta",
'\u03d1': "theta",
'\u03d2': "Upsilon",
'\u03d3': "Upsilon",
'\u03d4': "Upsilon",
'\u03d5': "phi",
'\u03d6': "pi",
'\u03f0': "kappa",
'\u03f1': "rho",
'\u03f2': "sigma",
'\u03f4': "Theta",
'\u03f5': "epsilon",
'\u03f9': "Sigma",
'\u03fc': "Rho",
'\u1e00': "A",
'\u1e01': "a",
'\u1e02': "B",
'\u1e03': "b",
'\u1e04': "B",
'\u1e05': "b",
'\u1e06': "B",
'\u1e07': "b",
'\u1e08': "C",
'\u1e09': "c",
'\u1e0a': "D",
'\u1e0b': "d",
'\u1e0c': "D",
'\u1e0d': "d",
'\u1e0e': "D",
'\u1e0f': "d",
'\u1e10': "D",
'\u1e11': "d",
'\u1e12': "D",
'\u1e13': "d",
'\u1e14': "E",
'\u1e15': "e",
'\u1e16': "E",
'\u1e17': "e",
'\u1e18': "E",
'\u1e19': "e",
'\u1e1a': "E",
'\u1e1b': "e",
'\u1e1c': "E",
'\u1e1d': "e",
'\u1e1e': "F",
'\u1e1f': "f",
'\u1e20': "G",
'\u1e21': "g",
'\u1e22': "H",
'\u1e23': "h",
'\u1e24': "H",
'\u1e25': "h",
'\u1e26': "H",
'\u1e27': "h",
'\u1e28': "H",
'\u1e29': "h",
'\u1e2a': "H",
'\u1e2b': "h",
'\u1e2c': "I",
'\u1e2d': "i",
'\u1e2e': "I",
'\u1e2f': "i",
'\u1e30': "K",
'\u1e31': "k",
'\u1e32': "K",
'\u1e33': "k",
'\u1e34': "K",
'\u1e35': "k",
'\u1e36': "L",
'\u1e37': "l",
'\u1e38': "L",
'\u1e39': "l",
'\u1e3a': "L",
'\u1e3b': "l",
'\u1e3c': "L",
'\u1e3d': "l",
'\u1e3e': "M",
'\u1e3f': "m",
'\u1e40': "M",
'\u1e41': "m",
'\u1e42': "M",
'\u1e43': "m",
'\u1e44': "N",
'\u1e45': "n",
'\u1e46': "N",
'\u1e47': "n",
'\u1e48': "N",
'\u1e49': "n",
'\u1e4a': "N",
'\u1e4b': "n",
'\u1e4c': "O",
'\u1e4d': "o",
'\u1e4e': "O",
'\u1e4f': "o",
'\u1e50': "O",
'\u1e51': "o",
'\u1e52': "O",
'\u1e53': "o",
'\u1e54': "P",
'\u1e55': "p",
'\u1e56': "P",
'\u1e57': "p",
'\u1e58': "R",
'\u1e59': "r",
'\u1e5a': "R",
'\u1e5b': "r",
'\u1e5c': "R",
'\u1e5d': "r",
'\u1e5e': "R",
'\u1e5f': "r",
'\u1e60': "S",
'\u1e61': "s",
'\u1e62': "S",
'\u1e63': "s",
'\u1e64': "S",
'\u1e65': "s",
'\u1e66': "S",
'\u1e67': "s",
'\u1e68': "S",
'\u1e69': "s",
'\u1e6a': "T",
'\u1e6b': "t",
'\u1e6c': "T",
'\u1e6d': "t",
'\u1e6e': "T",
'\u1e6f': "t",
'\u1e70': "T",
'\u1e71': "t",
'\u1e72': "U",
'\u1e73': "u",
'\u1e74': "U",
'\u1e75': "u",
'\u1e76': "U",
'\u1e77': "u",
'\u1e78': "U",
'\u1e79': "u",
'\u1e7a': "U",
'\u1e7b': "u",
'\u1e7c': "V",
'\u1e7d': "v",
'\u1e7e': "V",
'\u1e7f': "v",
'\u1e80': "W",
'\u1e81': "w",
'\u1e82': "W",
'\u1e83': "w",
'\u1e84': "W",
'\u1e85': "w",
'\u1e86': "W",
'\u1e87': "w",
'\u1e88': "W",
'\u1e89': "w",
'\u1e8a': "X",
'\u1e8b': "x",
'\u1e8c': "X",
'\u1e8d': "x",
'\u1e8e': "Y",
'\u1e8f': "y",
'\u1e90': "Z",
'\u1e91': "z",
'\u1e92': "Z",
'\u1e93': "z",
'\u1e94': "Z",
'\u1e95': "z",
'\u1e96': "h",
'\u1e97': "t",
'\u1e98': "w",
'\u1e99': "y",
'\u1e9a': "a",
'\u1e9b': "s",
'\u1ea0': "A",
'\u1ea1': "a",
'\u1ea2': "A",
'\u1ea3': "a",
'\u1ea4': "A",
'\u1ea5': "a",
'\u1ea6': "A",
'\u1ea7': "a",
'\u1ea8': "A",
'\u1ea9': "a",
'\u1eaa': "A",
'\u1eab': "a",
'\u1eac': "A",
'\u1ead': "a",
'\u1eae': "A",
'\u1eaf': "a",
'\u1eb0': "A",
'\u1eb1': "a",
'\u1eb2': "A",
'\u1eb3': "a",
'\u1eb4': "A",
'\u1eb5': "a",
'\u1eb6': "A",
'\u1eb7': "a",
'\u1eb8': "E",
'\u1eb9': "e",
'\u1eba': "E",
'\u1ebb': "e",
'\u1ebc': "E",
'\u1ebd': "e",
'\u1ebe': "E",
'\u1ebf': "e",
'\u1ec0': "E",
'\u1ec1': "e",
'\u1ec2': "E",
'\u1ec3': "e",
'\u1ec4': "E",
'\u1ec5': "e",
'\u1ec6': "E",
'\u1ec7': "e",
'\u1ec8': "I",
'\u1ec9': "i",
'\u1eca': "I",
'\u1ecb': "i",
'\u1ecc': "O",
'\u1ecd': "o",
'\u1ece': "O",
'\u1ecf': "o",
'\u1ed0': "O",
'\u1ed1': "o",
'\u1ed2': "O",
'\u1ed3': "o",
'\u1ed4': "O",
'\u1ed5': "o",
'\u1ed6': "O",
'\u1ed7': "o",
'\u1ed8': "O",
'\u1ed9': "o",
'\u1eda': "O",
'\u1edb': "o",
'\u1edc': "O",
'\u1edd': "o",
'\u1ede': "O",
'\u1edf': "o",
'\u1ee0': "O",
'\u1ee1': "o",
'\u1ee2': "O",
'\u1ee3': "o",
'\u1ee4': "U",
'\u1ee5': "u",
'\u1ee6': "U",
'\u1ee7': "u",
'\u1ee8': "U",
'\u1ee9': "u",
'\u1eea': "U",
'\u1eeb': "u",
'\u1eec': "U",
'\u1eed': "u",
'\u1eee': "U",
'\u1eef': "u",
'\u1ef0': "U",
'\u1ef1': "u",
'\u1ef2': "Y",
'\u1ef3': "y",
'\u1ef4': "Y",
'\u1ef5': "y",
'\u1ef6': "Y",
'\u1ef7': "y",
'\u1ef8': "Y",
'\u1ef9': "y",
'\u2002': " ",
'\u2003': " ",
'\u2009': " ",
'\u2010': "-",
'\u2011': "-",
'\u2012': "-",
'\u2013': "-",
'\u2014': "-",
'\u2015': "--",
'\u2016': "||",
'\u2017': "_",
'\u2018': "'",
'\u2019': "'",
'\u201a': ",",
'\u201b': "'",
'\u2024': ".",
'\u2025': "..",
'\u2026': "...",
'\u202f': " ",
'\u2030': "0/00",
'\u2031': "0/000",
'\u2032': "'",
'\u2034': "'''",
'\u2035': "`",
'\u2037': "'''",
'\u2038': "^",
'\u2039': "<",
'\u203a': ">",
'\u203c': "!!",
'\u203d': "?",
'\u2044': "/",
'\u2045': "[",
'\u2046': "]",
'\u2047': "??",
'\u2048': "?!",
'\u2049': "!?",
'\u204e': "*",
'\u2052': "%",
'\u2053': "~",
'\u2056': "...",
'\u2057': "''''",
'\u2058': "....",
'\u2059': ".....",
'\u205a': "..",
'\u205f': " ",
'\u2070': "0", # superscript 0
'\u2071': "i",
'\u2072': "2",
'\u2073': "3",
'\u2074': "4",
'\u2075': "5",
'\u2076': "6",
'\u2077': "7",
'\u2078': "8",
'\u2079': "9",
'\u207A': "+",
'\u207B': "-",
'\u207C': "=",
'\u207D': "(",
'\u207E': ")",
'\u207F': "n",
'\u2080': "0", # subscript 0
'\u2081': "1", # subscript 1
'\u2082': "2", # subscript 2
'\u2083': "3", # subscript 3
'\u2084': "4", # subscript 4
'\u2085': "5", # subscript 5
'\u2086': "6", # subscript 6
'\u2087': "7", # subscript 7
'\u2088': "8", # subscript 8
'\u2089': "9", # subscript 9
'\u208A': "+",
'\u208B': "-",
'\u208C': "=",
'\u208D': "(",
'\u208E': ")",
'\u2090': "a",
'\u2091': "e",
'\u2092': "o",
'\u2093': "x",
'\u2094': "e", # subscript schwa
'\u2095': "h",
'\u2096': "k",
'\u2097': "l",
'\u2098': "m",
'\u2099': "n",
'\u209A': "p",
'\u209B': "s",
'\u209C': "t",
'\u2100': "a/c",
'\u2101': "a/s",
'\u2102': "C",
'\u2103': "C",
'\u2105': "c/o",
'\u2106': "c/u",
'\u2107': "E",
'\u2109': "F",
'\u210a': "g",
'\u210b': "H",
'\u210c': "H",
'\u210d': "H",
'\u210e': "h",
'\u210f': "h",
'\u2110': "I",
'\u2111': "I",
'\u2112': "L",
'\u2113': "l",
'\u2115': "N",
'\u2116': "No",
'\u2117': "(p)",
'\u2118': "P",
'\u2119': "P",
'\u211a': "Q",
'\u211b': "R",
'\u211c': "R",
'\u211d': "R",
'\u211e': "Px",
'\u2120': "SM",
'\u2121': "TEL",
'\u2122': "TM",
'\u2124': "Z",
'\u2126': "Omega",
'\u2128': "Z",
'\u2129': "iota",
'\u212a': "K",
'\u212b': "A",
'\u212c': "B",
'\u212d': "C",
'\u212e': "e",
'\u212f': "e",
'\u2130': "E",
'\u2131': "F",
'\u2132': "F",
'\u2133': "M",
'\u2134': "o",
'\u2139': "i",
'\u213a': "Q",
'\u213b': "FAX",
'\u213c': "pi",
'\u213d': "gamma",
'\u213e': "Gamma",
'\u213f': "PI",
'\u2141': "G",
'\u2142': "L",
'\u2144': "Y",
'\u2145': "D",
'\u2146': "d",
'\u2147': "e",
'\u2148': "i",
'\u2149': "j",
'\u214b': "&",
'\u2153': "1/3",
'\u2154': "2/3",
'\u2155': "1/5",
'\u2156': "2/5",
'\u2157': "3/5",
'\u2158': "4/5",
'\u2159': "1/6",
'\u215a': "5/6",
'\u215b': "1/8",
'\u215c': "3/8",
'\u215d': "5/8",
'\u215e': "7/8",
'\u215f': "1/",
'\u2160': "I",
'\u2161': "II",
'\u2162': "III",
'\u2163': "IV",
'\u2164': "V",
'\u2165': "VI",
'\u2166': "VII",
'\u2167': "VIII",
'\u2168': "IX",
'\u2169': "X",
'\u216a': "XI",
'\u216b': "XII",
'\u216c': "L",
'\u216d': "C",
'\u216e': "D",
'\u216f': "M",
'\u2170': "i",
'\u2171': "ii",
'\u2172': "iii",
'\u2173': "iv",
'\u2174': "v",
'\u2175': "vi",
'\u2176': "vii",
'\u2177': "viii",
'\u2178': "ix",
'\u2179': "x",
'\u217a': "xi",
'\u217b': "xii",
'\u217c': "l",
'\u217d': "c",
'\u217e': "d",
'\u217f': "m",
'\u2180': "CD",
'\u2190': "<-",
'\u2192': "->",
'\u2194': "<->",
'\u219a': "<-",
'\u219b': "->",
'\u21ae': "<->",
'\u21cd': "<=",
'\u21ce': "<=>",
'\u21cf': "=>",
'\u21d0': "<=",
'\u21d2': "=>",
'\u21d4': "<=>",
'\u2303': "^",
'\u2329': "<",
'\u232a': ">",
'\u239b': "(",
'\u239c': "(",
'\u239d': "(",
'\u239e': ")",
'\u239f': ")",
'\u23a0': ")",
'\u23a1': "[",
'\u23a2': "[",
'\u23a3': "[",
'\u23a4': "]",
'\u23a5': "]",
'\u23a6': "]",
'\u23a7': "{",
'\u23a8': "{",
'\u23a9': "{",
'\u23ab': "}",
'\u23ac': "}",
'\u23ad': "}",
'\u2460': "1",
'\u2461': "2",
'\u2462': "3",
'\u2463': "4",
'\u2464': "5",
'\u2465': "6",
'\u2466': "7",
'\u2467': "8",
'\u2468': "9",
'\u2469': "10",
'\u246a': "11",
'\u246b': "12",
'\u246c': "13",
'\u246d': "14",
'\u246e': "15",
'\u246f': "16",
'\u2470': "17",
'\u2471': "18",
'\u2472': "19",
'\u2473': "20",
'\u2474': "(1)",
'\u2475': "(2)",
'\u2476': "(3)",
'\u2477': "(4)",
'\u2478': "(5)",
'\u2479': "(6)",
'\u247a': "(7)",
'\u247b': "(8)",
'\u247c': "(9)",
'\u247d': "(10)",
'\u247e': "(11)",
'\u247f': "(12)",
'\u2480': "(13)",
'\u2481': "(14)",
'\u2482': "(15)",
'\u2483': "(16)",
'\u2484': "(17)",
'\u2485': "(18)",
'\u2486': "(19)",
'\u2487': "(20)",
'\u2488': "1.",
'\u2489': "2.",
'\u248a': "3.",
'\u248b': "4.",
'\u248c': "5.",
'\u248d': "6.",
'\u248e': "7.",
'\u248f': "8.",
'\u2490': "9.",
'\u2491': "10.",
'\u2492': "11.",
'\u2493': "12.",
'\u2494': "13.",
'\u2495': "14.",
'\u2496': "15.",
'\u2497': "16.",
'\u2498': "17.",
'\u2499': "18.",
'\u249a': "19.",
'\u249b': "20.",
'\u249c': "(a)",
'\u249d': "(b)",
'\u249e': "(c)",
'\u249f': "(d)",
'\u24a0': "(e)",
'\u24a1': "(f)",
'\u24a2': "(g)",
'\u24a3': "(h)",
'\u24a4': "(i)",
'\u24a5': "(j)",
'\u24a6': "(k)",
'\u24a7': "(l)",
'\u24a8': "(m)",
'\u24a9': "(n)",
'\u24aa': "(o)",
'\u24ab': "(p)",
'\u24ac': "(q)",
'\u24ad': "(r)",
'\u24ae': "(s)",
'\u24af': "(t)",
'\u24b0': "(u)",
'\u24b1': "(v)",
'\u24b2': "(w)",
'\u24b3': "(x)",
'\u24b4': "(y)",
'\u24b5': "(z)",
'\u24b6': "A",
'\u24b7': "B",
'\u24b8': "C",
'\u24b9': "D",
'\u24ba': "E",
'\u24bb': "F",
'\u24bc': "G",
'\u24bd': "H",
'\u24be': "I",
'\u24bf': "J",
'\u24c0': "K",
'\u24c1': "L",
'\u24c2': "M",
'\u24c3': "N",
'\u24c4': "O",
'\u24c5': "P",
'\u24c6': "Q",
'\u24c7': "R",
'\u24c8': "S",
'\u24c9': "T",
'\u24ca': "U",
'\u24cb': "V",
'\u24cc': "W",
'\u24cd': "X",
'\u24ce': "Y",
'\u24cf': "Z",
'\u24d0': "a",
'\u24d1': "b",
'\u24d2': "c",
'\u24d3': "d",
'\u24d4': "e",
'\u24d5': "f",
'\u24d6': "g",
'\u24d7': "h",
'\u24d8': "i",
'\u24d9': "j",
'\u24da': "k",
'\u24db': "l",
'\u24dc': "m",
'\u24dd': "n",
'\u24de': "o",
'\u24df': "p",
'\u24e0': "q",
'\u24e1': "r",
'\u24e2': "s",
'\u24e3': "t",
'\u24e4': "u",
'\u24e5': "v",
'\u24e6': "w",
'\u24e7': "x",
'\u24e8': "y",
'\u24e9': "z",
'\u24ea': "0",
'\u24eb': "(11)",
'\u24ec': "(12)",
'\u24ed': "(13)",
'\u24ee': "(14)",
'\u24ef': "(15)",
'\u24f0': "(16)",
'\u24f1': "(17)",
'\u24f2': "(18)",
'\u24f3': "(19)",
'\u24f4': "(20)",
'\u24f5': "(1)",
'\u24f6': "(2)",
'\u24f7': "(3)",
'\u24f8': "(4)",
'\u24f9': "(5)",
'\u24fa': "(6)",
'\u24fb': "(7)",
'\u24fc': "(8)",
'\u24fd': "(9)",
'\u24fe': "(10)",
'\u24ff': "(0)",
'\u3000': " ",
'\u3001': ",",
'\u3007': "0",
'\u3008': "<",
'\u3009': ">",
'\u300a': "<<",
'\u300b': ">>",
'\u301b': "]",
'\u301c': "~",
'\u3021': "1",
'\u3022': "2",
'\u3023': "3",
'\u3024': "4",
'\u3025': "5",
'\u3026': "6",
'\u3027': "7",
'\u3028': "8",
'\u3029': "9",
'\ufb00': "ff",
'\ufb01': "fi",
'\ufb02': "fl",
'\ufb03': "ffi",
'\ufb04': "ffl",
'\ufb05': "st",
'\ufb06': "st",
'\ufb29': "+",
'\ufe50': ",",
'\ufe51': ",",
'\ufe52': ".",
'\ufe54': ";",
'\ufe55': ":",
'\ufe56': "?",
'\ufe57': "!",
'\ufe58': "-",
'\ufe59': "(",
'\ufe5a': ")",
'\ufe5b': "{",
'\ufe5c': "}",
'\ufe5d': "(",
'\ufe5e': ")",
'\ufe5f': "#",
'\ufe60': "&",
'\ufe61': "*",
'\ufe62': "+",
'\ufe63': "-",
'\ufe64': "<",
'\ufe65': ">",
'\ufe66': "=",
'\ufe69': "$",
'\ufe6a': "%",
'\ufe6b': "@",
'\uff01': "!",
'\uff03': "#",
'\uff04': "$",
'\uff05': "%",
'\uff06': "&",
'\uff07': "'",
'\uff08': "(",
'\uff09': ")",
'\uff0a': "*",
'\uff0b': "+",
'\uff0c': ",",
'\uff0d': "-",
'\uff0e': ".",
'\uff0f': "/",
'\uff10': "0",
'\uff11': "1",
'\uff12': "2",
'\uff13': "3",
'\uff14': "4",
'\uff15': "5",
'\uff16': "6",
'\uff17': "7",
'\uff18': "8",
'\uff19': "9",
'\uff1a': ":",
'\uff1b': ";",
'\uff1c': "<",
'\uff1d': "=",
'\uff1e': ">",
'\uff1f': "?",
'\uff20': "@",
'\uff21': "A",
'\uff22': "B",
'\uff23': "C",
'\uff24': "D",
'\uff25': "E",
'\uff26': "F",
'\uff27': "G",
'\uff28': "H",
'\uff29': "I",
'\uff2a': "J",
'\uff2b': "K",
'\uff2c': "L",
'\uff2d': "M",
'\uff2e': "N",
'\uff2f': "O",
'\uff30': "P",
'\uff31': "Q",
'\uff32': "R",
'\uff33': "S",
'\uff34': "T",
'\uff35': "U",
'\uff36': "V",
'\uff37': "W",
'\uff38': "X",
'\uff39': "Y",
'\uff3a': "Z",
'\uff3b': "[",
'\uff3d': "]",
'\uff3e': "^",
'\uff3f': "_",
'\uff40': "`",
'\uff41': "a",
'\uff42': "b",
'\uff43': "c",
'\uff44': "d",
'\uff45': "e",
'\uff46': "f",
'\uff47': "g",
'\uff48': "h",
'\uff49': "i",
'\uff4a': "j",
'\uff4b': "k",
'\uff4c': "l",
'\uff4d': "m",
'\uff4e': "n",
'\uff4f': "o",
'\uff50': "p",
'\uff51': "q",
'\uff52': "r",
'\uff53': "s",
'\uff54': "t",
'\uff55': "u",
'\uff56': "v",
'\uff57': "w",
'\uff58': "x",
'\uff59': "y",
'\uff5a': "z",
'\uff5b': "{",
'\uff5c': "|",
'\uff5d': "}",
'\uff5e': "~",
'\uff5f': "(",
'\uff60': ")",
'\uff64': ",",
'\uffe9': "<-",
'\uffeb': "->",
'\u2000': " ",
'\u2001': " ",
'\u2004': " ",
'\u2005': " ",
'\u2006': " ",
'\u2007': " ",
'\u2008': " ",
'\u200A': " ",
'\u200B': " ",
'\u2215': "/",
'\u207b': "-",
'\u208b': "-",
'\u2212': "-",
'\u2264': "<=",
'\u2265': ">=",
'\uc2a0': " ",
'\ufeff': " ",
'\u200E': "",
'\u2020': "",
'\u2217': "*",
'\u0092': "",
'\u00f0': "d", # Decimal=240 (Latin Small Letter Eth)
'\u0138': "K", # Decimal=312 (Latin Small Letter KRA)
'\u0150': "O", # Decimal=330 (Latin Capital Letter O With Double Acute)
'\u018d': "d", # Decimal=397 (Latin Small Letter Turned Delta)
'\u018f': "e", # Decimal=399 (Latin Capital Letter Schwa)
'\u0189': "D", # Decimal=425 (Latin Capital Letter African D)
'\u0190': "E", # Decimal=426 (Latin Capital Letter Open E)
'\u01b7': "3", # Decimal=439 (Latin Capital Letter Ezh)
'\u01b8': "E", # Decimal=440 (Latin Capital Letter Ezh Reversed)
'\u01b9': "e", # Decimal=441 (Latin Small Letter Ezh Reversed)
'\u01ba': "e", # Decimal=442 (Latin Small Letter Ezh With Tail)
'\u01be': "t", # Decimal=446 (Latin Letter Inverted Glottal Stop With Stroke)
'\u01bf': "p", # Decimal=447 (Latin Letter Wynn)
'\u01c1': "||", # Decimal=449 (Latin Letter Lateral Click)
'\u01ee': "3", # Decimal=494 (Latin Capital Letter Ezh With Caron)
'\u01ef': "3", # Decimal=495 (Latin Small Letter Ezh With Caron)
'\u0323': "",
'\u0410': "",
'\u0411': "",
'\u0412': "",
'\u0413': "",
'\u0414': "",
'\u0415': "",
'\u0416': "",
'\u0417': "",
'\u0418': "",
'\u0419': "",
'\u041a': "",
'\u041b': "",
'\u041c': "",
'\u041d': "",
'\u041e': "",
'\u041f': "",
'\u0420': "",
'\u0421': "",
'\u0422': "",
'\u0423': "",
'\u0424': "",
'\u0425': "",
'\u0426': "",
'\u0427': "",
'\u0428': "",
'\u0429': "",
'\u042a': "",
'\u042b': "",
'\u042c': "",
'\u042d': "",
'\u042e': "",
'\u042f': "",
'\u0430': "",
'\u0431': "",
'\u0432': "",
'\u0433': "",
'\u0434': "",
'\u0435': "",
'\u0436': "",
'\u0437': "",
'\u0438': "",
'\u0439': "",
'\u043a': "",
'\u043b': "",
'\u043c': "",
'\u043d': "",
'\u043e': "",
'\u043f': "",
'\u0440': "",
'\u0441': "",
'\u0442': "",
'\u0443': "",
'\u0444': "",
'\u0445': "",
'\u0446': "",
'\u0447': "",
'\u0448': "",
'\u0449': "",
'\u044a': "",
'\u044b': "",
'\u044c': "",
'\u044d': "",
'\u044e': "",
'\u044f': "",
'\u0451': "",
'\u200b': "",
'\u200e': "",
'\u2022': "",
'\u2261': "",
'\u8166': "",
'\u982d': "",
'\ub1cc': "",
'\ub450': "",
'\uf030': "",
}