feat: Added support for HTML entites

Both &<name>; and &<name> are supported
OXY2DEV · Aug 3, 2024 · 3b270c1 · 3b270c1
1 parent ab0e54e
commit 3b270c1
Show file tree

Hide file tree

Showing 4 changed files with 363 additions and 17 deletions.
diff --git a/lua/markview.lua b/lua/markview.lua
@@ -1165,14 +1165,22 @@ markview.configuration = {
 		}
 	},
 	html = {
-		default = {
-			conceal = false
+		tags = {
+			enable = true,
+
+			default = {
+				conceal = false
+			},
+
+			config = {
+				p = { conceal = true, hl = "Title" },
+				u = { conceal = true, hl = "Underlined" },
+				i = { conceal = true, hl = "Italic" }
+			}
 		},
 
-		types = {
-			p = { conceal = true, hl = "Title" },
-			u = { conceal = true, hl = "Underlined" },
-			i = { conceal = true, hl = "Italic" }
+		entites = {
+			enable = true
 		}
 	},
 

diff --git a/lua/markview/entites.lua b/lua/markview/entites.lua
@@ -0,0 +1,256 @@
+local entites = {};
+
+entites.lookup = {
+    Aacute = "Á",
+    aacute = "á",
+    Acirc = "Â",
+    acirc = "â",
+    acute = "´",
+    AElig = "Æ",
+    aelig = "æ",
+    Agrave = "À",
+    agrave = "à",
+    alefsym = "ℵ",
+    Alpha = "Α",
+    alpha = "α",
+    amp = "&",
+    ["and"] = "∧",
+    ang = "∠",
+    Aring = "Å",
+    aring = "å",
+    asymp = "≈",
+    Atilde = "Ã",
+    atilde = "ã",
+    Auml = "Ä",
+    auml = "ä",
+    bdquo = "„",
+    Beta = "Β",
+    beta = "β",
+    brvbar = "¦",
+    bull = "•",
+    cap = "∩",
+    Ccedil = "Ç",
+    ccedil = "ç",
+    cedil = "¸",
+    cent = "¢",
+    Chi = "Χ",
+    chi = "χ",
+    circ = "ˆ",
+    clubs = "♣",
+    congc = "≅",
+    copy = "©",
+    crarr = "↵",
+    cup = "∪",
+    curren = "¤",
+    Dagger = "‡",
+    dagger = "†",
+    dArr = "⇓",
+    darr = "↓",
+    deg = "°",
+    Delta = "Δ",
+    delta = "δ",
+    diams = "♦",
+    divide = "÷",
+    Eacute = "É",
+    eacute = "é",
+    Ecirc = "Ê",
+    ecirc = "ê",
+    Egrave = "È",
+    egrave = "è",
+    empty = "∅",
+    Epsilon = "Ε",
+    epsilon = "ε",
+    equiv = "≡",
+    Eta = "Η",
+    eta = "η",
+    ETH = "Ð",
+    eth = "ð",
+    Euml = "Ë",
+    euml = "ë",
+    euro = "€",
+    exists = "∃",
+    fnof = "ƒ",
+    forall = "∀",
+    frac12 = "½",
+    frac14 = "¼",
+    frac34 = "¾",
+    frasl = "⁄",
+    Gamma = "Γ",
+    gamma = "γ",
+    ge = "≥",
+    gt = ">",
+    harr = "↔",
+    hArr = "⇔",
+    hearts = "♥",
+    hellip = "…",
+    Iacute = "Í",
+    iacute = "í",
+    Icirc = "Î",
+    icirc = "î",
+    iexcl = "¡",
+    Igrave = "Ì",
+    igrave = "ì",
+    image = "ℑ",
+    infin = "∞",
+    int = "∫",
+    Iota = "Ι",
+    iota = "ι",
+    iquest = "¿",
+    isin = "∈",
+    Iuml = "Ï",
+    iuml = "ï",
+    Kappa = "Κ",
+    kappa = "κ",
+    Lambda = "Λ",
+    lambda = "λ",
+    lang = "⟨",
+    laquo = "«",
+    lArr = "⇐",
+    larr = "←",
+    lceil = "⌈",
+    ldquo = "“",
+    le = "≤",
+    lfloor = "⌊",
+    lowast = "∗",
+    loz = "◊",
+    -- lrm = "‎",
+    lsaquo = "‹",
+    lsquo = "‘",
+    lt = "<",
+    macr = "¯",
+    mdash = "—",
+    micro = "µ",
+    middot = "·",
+    minus = "−",
+    Mu = "Μ",
+    mu = "μ",
+    nabla = "∇",
+    nbsp = " ",
+    ndash = "–",
+    ne = "≠",
+    ni = "∋",
+    ["not"] = "¬",
+    notin = "∉",
+    nsub = "⊄",
+    Ntilde = "Ñ",
+    ntilde = "ñ",
+    Nu = "Ν",
+    nu = "ν",
+    Oacute = "Ó",
+    oacute = "ó",
+    Ocirc = "Ô",
+    ocirc = "ô",
+    OElig = "Œ",
+    oelig = "œ",
+    Ograve = "Ò",
+    ograve = "ò",
+    Omeg = "Ω",
+    omega = "ω",
+    Omicron = "Ο",
+    omicron = "ο",
+    oline = "‾",
+    ["or"] = "∨",
+    Oslash = "Ø",
+    oslash = "ø",
+    Otilde = "Õ",
+    otilde = "õ",
+    Ouml = "Ö",
+    ouml = "ö",
+    para = "¶",
+    part = "∂",
+    permil = "‰",
+    perp = "⊥",
+    Phi = "Φ",
+    phi = "φ",
+    Pi = "Π",
+    pi = "π",
+    piv = "ϖ",
+    plusmn = "±",
+    pound = "£",
+    Prime = "″",
+    prime = "′",
+    prod = "∏",
+    prop = "∝",
+    Psi = "Ψ",
+    psi = "ψ",
+    quot = "\"",
+    radic = "√",
+    rang = "⟩",
+    raquo = "»",
+    rArr = "⇒",
+    rarr = "→",
+    rceil = "⌉",
+    rdquo = "”",
+    real = "ℜ",
+    reg = "®",
+    rflo = "⌋",
+    Rho = "Ρ",
+    rho = "ρ",
+    -- rlm = "‏",
+    rsaquo = "›",
+    rsquo = "’",
+    sbquo = "‚",
+    Scaron = "Š",
+    scaron = "š",
+    sdot = "⋅",
+    sect = "§",
+    shy = "",
+    Sigma = "Σ",
+    sigma = "σ",
+    sigmaf = "ς",
+    sim = "∼",
+    spades = "♠",
+    sub = "⊂",
+    sube = "⊆",
+    sum = "∑",
+    sup = "⊃",
+    sup1 = "¹",
+    sup2 = "²",
+    sup3 = "³",
+    supe = "⊇",
+    szlig = "ß",
+    Tau = "Τ",
+    tau = "τ",
+    there4 = "∴",
+    Theta = "Θ",
+    theta = "θ",
+    thetasym = "ϑ",
+    thinsp = " ",
+    THORN = "Þ",
+    thorn = "þ",
+    tilde = "˜",
+    times = "×",
+    trade = "™",
+    Uacute = "Ú",
+    uacute = "ú",
+    uArr = "⇑",
+    uarr = "↑",
+    Ucirc = "Û",
+    ucirc = "û",
+    Ugrave = "Ù",
+    ugrave = "ù",
+    uml = "¨",
+    upsih = "ϒ",
+    Upsilon = "Υ",
+    upsilon = "υ",
+    Uuml = "Ü",
+    uuml = "ü",
+    weierp = "℘",
+    Xi = "Ξ",
+    xi = "ξ",
+    Yacute = "Ý",
+    yacute = "ý",
+    yen = "¥",
+    yuml = "ÿ",
+    Yuml = "Ÿ",
+    Zeta = "Ζ",
+    zeta = "ζ",
+    -- zwj = "‍",
+    -- zwnj = "‌"
+}
+
+entites.get = function (string)
+	return entites.lookup[string], vim.fn.strdisplaywidth(entites.lookup[string]);
+end
+
+return entites;
diff --git a/lua/markview/parser.lua b/lua/markview/parser.lua
@@ -430,6 +430,8 @@ parser.md_inline = function (buffer, TStree, from, to)
 			] @link)
 
 		((code_span) @code)
+
+		((entity_reference) @entity)
 	]]);
 
 	-- The last 2 _ represent the metadata & query
@@ -511,6 +513,19 @@ parser.md_inline = function (buffer, TStree, from, to)
 				row_start = row_start,
 				row_end = row_end,
 
+				col_start = col_start,
+				col_end = col_end
+			})
+		elseif capture_name == "entity" then
+			table.insert(parser.parsed_content, {
+				node = capture_node,
+				type = "html_entity",
+
+				text = capture_text,
+
+				row_start = row_start,
+				row_end = row_end,
+
 				col_start = col_start,
 				col_end = col_end
 			})
@@ -521,8 +536,6 @@ end
 parser.html = function (buffer, TStree, from, to)
 	local scanned_queies = vim.treesitter.query.parse("html", [[
 		((element) @elem)
-
-		;((entity_reference) @entity)
 	]]);
 
 	for capture_id, capture_node, _, _ in scanned_queies:iter_captures(TStree:root(), buffer, from, to) do