Convert simple inline LateX to unicode

latex_to_unicode.py

import re

__SYMBOLS_DICTIONARY__ = {
    # Greek Letters
    "Gamma": "Γ",
    "Delta": "Δ",
    "Lambda": "Λ",
    "Phi": "Φ",
    "Pi": "Π",
    "Psi": "Ψ",
    "Sigma": "Σ",
    "Theta": "Θ",
    "Upsilon": "Υ",
    "Xi": "Ξ",
    "Omega": "Ω",
    "alpha": "α",
    "beta": "β",
    "gamma": "γ",
    "delta": "δ",
    "epsilon": "ϵ",
    "zeta": "ζ",
    "eta": "η",
    "theta": "θ",
    "iota": "ι",
    "kappa": "κ",
    "lambda": "λ",
    "mu": "μ",
    "nu": "ν",
    "xi": "ξ",
    "pi": "π",
    "rho": "ρ",
    "sigma": "σ",
    "tau": "τ",
    "upsilon": "υ",
    "phi": "ϕ",
    "chi": "χ",
    "psi": "ψ",
    "omega": "ω",
    "digamma": "ϝ",
    "varepsilon": "ε",
    "varkappa": "ϰ",
    "varphi": "φ",
    "varpi": "ϖ",
    "varrho": "ϱ",
    "varsigma": "ς",
    "vartheta": "ϑ",

    # Other alphabetic characters
    "aleph": "ℵ",
    "beth": "ℶ",
    "daleth": "ℸ",
    "gimel": "ℷ",
    "complement": "∁",
    "ell": "ℓ",
    "eth": "ð",
    "hbar": "ℏ",
    "hslash": "ℏ",
    "mho": "℧",
    "partial": "∂",
    "wp": "℘",
    "circledS": "Ⓢ",
    "Bbbk": "k",
    "Finv": "Ⅎ",
    "Game": "⅁",
    "Im": "J",
    "Re": "R",

    # Other simple symbols
    "#": "#",
    "angle": "∠",
    "backprime": "‵",
    "bigstar": "★",
    "blacklozenge": "⧫",
    "blacksquare": "■",
    "blacktriangle": "▲",
    "blacktriangledown": "▼",
    "bot": "⊥",
    "clubsuit": "♣",
    "diagdown": "╲",
    "diagup": "╱",
    "diamondsuit": "♢",
    "emptyset": "∅",
    "exists": "∃",
    "flat": "♭",
    "forall": "∀",
    "heartsuit": "♡",
    "infty": "∞",
    "lozenge": "◊",
    "measuredangle": "∡",
    "nabla": "∇",
    "natural": "♮",
    "neg": "¬",
    "lnot": "¬",
    "nexists": "∄",
    "prime": "′",
    "S": "§",
    "sharp": "♯",
    "spadesuit": "♠",
    "sphericalangle": "∢",
    "square p": "□p",
    "surd": "√",
    "top": "⊤",
    "triangle": "△",
    "triangledown": "▽",
    "varnothing": "∅",

    # Accumulation operators: sum, integral, union, etc
    "int": "∫",
    "oint": "∮",
    "bigcap": "⋂",
    "bigcup": "⋃",
    "bigodot": "⨀",
    "bigoplus": "⨁",
    "bigotimes": "⨂",
    "bigsqcup": "⨆",
    "biguplus": "⨄",
    "bigvee": "⋁",
    "bigwedge": "⋀",
    "coprod": "∐",
    "prod": "ℿ",
    "sum": "⅀",
    "smallint": "∫",

    # Binary operators
    "amalg": "⨿",
    "ast": "∗",
    "barwedge": "⊼",
    "bigcirc": "◯",
    "bigtriangledown": "▽",
    "bigtriangleup": "△",
    "boxdot": "⊡",
    "boxminus": "⊟",
    "boxplus": "⊞",
    "boxtimes": "⊠",
    "bullet": "∙",
    "cap": "∩",
    "Cap": "⋒",
    "doublecap": "⋒",
    "cdot": "⋅",
    "centerdot": "⋅",
    "circ": "∘",
    "circledast": "⊛",
    "circledcirc": "⊚",
    "circleddash": "⊝",
    "cup": "∪",
    "Cup": "⋓",
    "doublecup": "⋓",
    "curlyvee": "⋎",
    "curlywedge": "⋏",
    "dagger": "†",
    "ddagger": "‡",
    "diamond": "⋄",
    "div": "÷",
    "divideontimes": "⋇",
    "dotplus": "∔",
    "doublebarwedge": "⩞",
    "gtrdot": "⋗",
    "intercal": "⊺",
    "leftthreetimes": "⋋",
    "lessdot": "⋖",
    "ltimes": "⋉",
    "mp": "∓",
    "odot": "⊙",
    "ominus": "⊖",
    "oplus": "⊕",
    "oslash": "⊘",
    "otimes": "⊗",
    "pm": "±",
    "rightthreetimes": "⋌",
    "rtimes": "⋊",
    "setminus": "∖",
    "smallsetminus": "∖",
    "sqcap": "⊓",
    "sqcup": "⊔",
    "star": "⋆",
    "times": "×",
    "triangleleft": "◃",
    "triangleright": "▹",
    "uplus": "⊎",
    "vee": "∨",
    "lor": "∨",
    "veebar": "⊻",
    "wedge": "∧",
    "land": "∧",
    "wr": "≀",

    # Relation / comparison
    "approx": "≈",
    "approxeq": "≊",
    "asymp": "≍",
    "backsim": "∽",
    "backsimeq": "⋍",
    "bumpeq": "≏",
    "Bumpeq": "≎",
    "circeq": "≗",
    "cong": "≅",
    "curlyeqprec": "⋞",
    "curlyeqsucc": "⋟",
    "doteq": "≐",
    "doteqdot": "≑",
    "Doteq": "≑",
    "eqcirc": "≖",
    "eqsim": "≂",
    "eqslantgtr": "⪖",
    "eqslantless": "⪕",
    "equiv": "≡",
    "fallingdotseq": "≒",
    "geq": "≥",
    "ge": "≥",
    "geqq": "≧",
    "geqslant": "⩾",
    "gg": "≫",
    "ggg": "⋙",
    "gggtr": "⋙",
    "gnapprox": "⪊",
    "gneq": "⪈",
    "gneqq": "≩",
    "gnsim": "⋧",
    "gtrapprox": "⪆",
    "gtreqless": "⋛",
    "gtreqqless": "⪌",
    "gtrless": "≷",
    "gtrsim": "≳",
    "gvertneqq": "",
    "leq": "≤",
    "le": "≤",
    "leqq": "≦",
    "leqslant": "⩽",
    "lessapprox": "⪅",
    "lesseqgtr": "⋚",
    "lesseqqgtr": "⪋",
    "lessgtr": "≶",
    "lesssim": "≲",
    "ll": "≪",
    "lll": "⋘",
    "llless": "⋘",
    "lnapprox": "⪉",
    "lneq": "⪇",
    "lneqq": "≨",
    "lnsim": "⋦",
    "lvertneqq": "",
    "ncong": "≆",
    "neq": "≠",
    "ne": "≠",
    "ngeq": "≱",
    "ngeqq": "",
    "ngeqslant": "",
    "ngtr": "≯",
    "nleq": "≰",
    "nleqq": "",
    "nleqslant": "",
    "nless": "≮",
    "nprec": "⊀",
    "npreceq": "⋠",
    "nsim": "≁",
    "nsucc": "⊁",
    "nsucceq": "⋡",
    "prec": "≺",
    "precapprox": "⪷",
    "preccurlyeq": "≼",
    "preceq": "⪯",
    "precnapprox": "⪹",
    "precneqq": "⪵",
    "precnsim": "⋨",
    "precsim": "≾",
    "risingdotseq": "≓",
    "sim": "∼",
    "simeq": "≃",
    "succ": "≻",
    "succapprox": "⪸",
    "succcurlyeq": "≽",
    "succeq": "⪰",
    "succnapprox": "⪺",
    "succneqq": "⪶",
    "succnsim": "⋩",
    "succsim": "≿",
    "thickapprox": "≈",
    "thicksim": "∼",
    "triangleq": "≜",

    # Arrows
    "circlearrowleft": "↺",
    "circlearrowright": "↻",
    "curvearrowleft": "↶",
    "curvearrowright": "↷",
    "downdownarrows": "⇊",
    "downharpoonleft": "⇃",
    "downharpoonright": "⇂",
    "hookleftarrow": "↩",
    "hookrightarrow": "↪",
    "leftarrow": "←",
    "gets": "←",
    "Leftarrow": "⇐",
    "leftarrowtail": "↢",
    "leftharpoondown": "↽",
    "leftharpoonup": "↼",
    "leftleftarrows": "⇇",
    "leftrightarrow": "↔",
    "Leftrightarrow": "⇔",
    "leftrightarrows": "⇆",
    "leftrightharpoons": "⇋",
    "leftrightsquigarrow": "↭",
    "Lleftarrow": "⇚",
    "longleftarrow": "⟵",
    "Longleftarrow": "⟸",
    "longleftrightarrow": "⟷",
    "Longleftrightarrow": "⟺",
    "longmapsto": "⟼",
    "longrightarrow": "⟶",
    "Longrightarrow": "⟹",
    "implies": "⟹",
    "looparrowleft": "↫",
    "looparrowright": "↬",
    "Lsh": "↰",
    "mapsto": "↦",
    "multimap": "⊸",
    "nLeftarrow": "⇍",
    "nLeftrightarrow": "⇎",
    "nRightarrow": "⇏",
    "nearrow": "↗",
    "nleftarrow": "↚",
    "nleftrightarrow": "↮",
    "nrightarrow": "↛",
    "nwarrow": "↖",
    "rightarrow": "→",
    "to": "→",
    "Rightarrow": "⇒",
    "rightarrowtail": "↣",
    "rightharpoondown": "⇁",
    "rightharpoonup": "⇀",
    "rightleftarrows": "⇄",
    "rightleftharpoons": "⇌",
    "rightrightarrows": "⇉",
    "rightsquigarrow": "⇝",
    "Rrightarrow": "⇛",
    "Rsh": "↱",
    "searrow": "↘",
    "swarrow": "↙",
    "twoheadleftarrow": "↞",
    "twoheadrightarrow": "↠",
    "upharpoonleft": "↿",
    "upharpoonright": "↾",
    "restriction": "↾",
    "upuparrows": "⇈",
    "uparrow": "↑",
    "Uparrow": "⇑",
    "downarrow": "↓",
    "Downarrow": "⇓",
    "updownarrow": "↕",
    "Updownarrow": "⇕",

    # Other relation symbols
    "backepsilon": "∍",
    "because": "∵",
    "between": "≬",
    "blacktriangleleft": "◀",
    "blacktriangleright": "▶",
    "bowtie": "⋈",
    "dashv": "⊣",
    "frown": "⌢",
    "in": "∈",
    "mid": "∣",
    "models": "⊨",
    "ni": "∋",
    "owns": "∋",
    "nmid": "∤",
    "notin": "∉",
    "nparallel": "∦",
    "nshortmid": "",
    "nshortparallel": "",
    "nsubseteq": "⊈",
    "nsubseteqq": "",
    "nsupseteq": "⊉",
    "nsupseteqq": "",
    "ntriangleleft": "⋪",
    "ntrianglelefteq": "⋬",
    "ntriangleright": "⋫",
    "ntrianglerighteq": "⋭",
    "nvdash": "⊬",
    "nVdash": "⊮",
    "nvDash": "⊭",
    "nVDash": "⊯",
    "parallel": "∥",
    "perp": "⊥",
    "pitchfork": "⋔",
    "propto": "∝",
    "shortmid": "∣",
    "shortparallel": "∥",
    "smallfrown": "⌢",
    "smallsmile": "⌣",
    "smile": "⌣",
    "sqsubset": "⊏",
    "sqsubseteq": "⊑",
    "sqsupset": "⊐",
    "sqsupseteq": "⊒",
    "subset": "⊂",
    "Subset": "⋐",
    "subseteq": "⊆",
    "subseteqq": "⫅",
    "subsetneq": "⊊",
    "subsetneqq": "⫋",
    "supset": "⊃",
    "Supset": "⋑",
    "supseteq": "⊇",
    "supseteqq": "⫆",
    "supsetneq": "⊋",
    "supsetneqq": "⫌",
    "therefore": "∴",
    "trianglelefteq": "⊴",
    "trianglerighteq": "⊵",
    "varpropto": "∝",
    "varsubsetneq": "",
    "varsubsetneqq": "",
    "varsupsetneq": "",
    "varsupsetneqq": "",
    "vartriangle": "△",
    "vartriangleleft": "⊲",
    "vartriangleright": "⊳",
    "vdash": "⊢",
    "Vdash": "⊩",
    "vDash": "⊨",
    "Vvdash": "⊪",
    "vert": "|",
    "Vert": "∥",

    # Special fonts
    r"mathbb{C}": "ℂ",
    r"mathbb{H}": "ℍ",
    r"mathbb{N}": "ℕ",
    r"mathbb{P}": "ℙ",
    r"mathbb{Q}": "ℚ",
    r"mathbb{R}": "ℝ",
    r"mathbb{Z}": "ℤ",
}


class LatexCommandNotFound(KeyError):
    def __init__(self, command):
        super().__init__(f'Command not found: "{command}". See all available commands with "get_commands()"')


def get_commands():
    return __SYMBOLS_DICTIONARY__.keys()


def get_converter(symbol="\\"):
    regex = re.compile(re.escape(symbol) + r"\w+(\{\w+\})*")
    n_skip = len(symbol)

    def convert(txt):
        repl = lambda key: symbol_replace(key.group(0)[n_skip:])
        return re.sub(regex, repl, txt)

    return convert


def symbol_replace(symbol):
    try:
        return __SYMBOLS_DICTIONARY__[symbol]
    except KeyError:
        raise LatexCommandNotFound(symbol)


# https://latex.wikia.org/wiki/List_of_LaTeX_symbols
def __parse_text_from_wikia__(text):
    print("{")
    for line in text.split("\n"):
        if (len(line) < 3):
            continue
        sym, command = line.split("  ")
        if (" or" in command):
            commands = command.split(" or")
            for command_item in commands:
                command = command_item.strip()
                print(f'    "{command[1:]}": "{sym}",')
        else:
            print(f'    "{command[1:]}": "{sym}",')
    print("}")


if __name__ == '__main__':
    # print(re.findall(r"\\\w{2,}", r"ajskbasjds jkdhfd fdjfkds \\asjkhs jhfdk \latex"))
    c = get_converter("\\")
    print(c(r"\forall x \in \mathbb{N} \| x = 5 * 3 \sigma"))
    c(r"\forall x \in \mathbb{N} \| x = 5 * 3 \sigma")