readability_js 0.0.1

data/lib/readability_js/node/node_modules/parse5/dist/common/html.js

@@ -0,0 +1,523 @@
+/** All valid namespaces in HTML. */
+export var NS;
+(function (NS) {
+    NS["HTML"] = "http://www.w3.org/1999/xhtml";
+    NS["MATHML"] = "http://www.w3.org/1998/Math/MathML";
+    NS["SVG"] = "http://www.w3.org/2000/svg";
+    NS["XLINK"] = "http://www.w3.org/1999/xlink";
+    NS["XML"] = "http://www.w3.org/XML/1998/namespace";
+    NS["XMLNS"] = "http://www.w3.org/2000/xmlns/";
+})(NS || (NS = {}));
+export var ATTRS;
+(function (ATTRS) {
+    ATTRS["TYPE"] = "type";
+    ATTRS["ACTION"] = "action";
+    ATTRS["ENCODING"] = "encoding";
+    ATTRS["PROMPT"] = "prompt";
+    ATTRS["NAME"] = "name";
+    ATTRS["COLOR"] = "color";
+    ATTRS["FACE"] = "face";
+    ATTRS["SIZE"] = "size";
+})(ATTRS || (ATTRS = {}));
+/**
+ * The mode of the document.
+ *
+ * @see {@link https://dom.spec.whatwg.org/#concept-document-limited-quirks}
+ */
+export var DOCUMENT_MODE;
+(function (DOCUMENT_MODE) {
+    DOCUMENT_MODE["NO_QUIRKS"] = "no-quirks";
+    DOCUMENT_MODE["QUIRKS"] = "quirks";
+    DOCUMENT_MODE["LIMITED_QUIRKS"] = "limited-quirks";
+})(DOCUMENT_MODE || (DOCUMENT_MODE = {}));
+export var TAG_NAMES;
+(function (TAG_NAMES) {
+    TAG_NAMES["A"] = "a";
+    TAG_NAMES["ADDRESS"] = "address";
+    TAG_NAMES["ANNOTATION_XML"] = "annotation-xml";
+    TAG_NAMES["APPLET"] = "applet";
+    TAG_NAMES["AREA"] = "area";
+    TAG_NAMES["ARTICLE"] = "article";
+    TAG_NAMES["ASIDE"] = "aside";
+    TAG_NAMES["B"] = "b";
+    TAG_NAMES["BASE"] = "base";
+    TAG_NAMES["BASEFONT"] = "basefont";
+    TAG_NAMES["BGSOUND"] = "bgsound";
+    TAG_NAMES["BIG"] = "big";
+    TAG_NAMES["BLOCKQUOTE"] = "blockquote";
+    TAG_NAMES["BODY"] = "body";
+    TAG_NAMES["BR"] = "br";
+    TAG_NAMES["BUTTON"] = "button";
+    TAG_NAMES["CAPTION"] = "caption";
+    TAG_NAMES["CENTER"] = "center";
+    TAG_NAMES["CODE"] = "code";
+    TAG_NAMES["COL"] = "col";
+    TAG_NAMES["COLGROUP"] = "colgroup";
+    TAG_NAMES["DD"] = "dd";
+    TAG_NAMES["DESC"] = "desc";
+    TAG_NAMES["DETAILS"] = "details";
+    TAG_NAMES["DIALOG"] = "dialog";
+    TAG_NAMES["DIR"] = "dir";
+    TAG_NAMES["DIV"] = "div";
+    TAG_NAMES["DL"] = "dl";
+    TAG_NAMES["DT"] = "dt";
+    TAG_NAMES["EM"] = "em";
+    TAG_NAMES["EMBED"] = "embed";
+    TAG_NAMES["FIELDSET"] = "fieldset";
+    TAG_NAMES["FIGCAPTION"] = "figcaption";
+    TAG_NAMES["FIGURE"] = "figure";
+    TAG_NAMES["FONT"] = "font";
+    TAG_NAMES["FOOTER"] = "footer";
+    TAG_NAMES["FOREIGN_OBJECT"] = "foreignObject";
+    TAG_NAMES["FORM"] = "form";
+    TAG_NAMES["FRAME"] = "frame";
+    TAG_NAMES["FRAMESET"] = "frameset";
+    TAG_NAMES["H1"] = "h1";
+    TAG_NAMES["H2"] = "h2";
+    TAG_NAMES["H3"] = "h3";
+    TAG_NAMES["H4"] = "h4";
+    TAG_NAMES["H5"] = "h5";
+    TAG_NAMES["H6"] = "h6";
+    TAG_NAMES["HEAD"] = "head";
+    TAG_NAMES["HEADER"] = "header";
+    TAG_NAMES["HGROUP"] = "hgroup";
+    TAG_NAMES["HR"] = "hr";
+    TAG_NAMES["HTML"] = "html";
+    TAG_NAMES["I"] = "i";
+    TAG_NAMES["IMG"] = "img";
+    TAG_NAMES["IMAGE"] = "image";
+    TAG_NAMES["INPUT"] = "input";
+    TAG_NAMES["IFRAME"] = "iframe";
+    TAG_NAMES["KEYGEN"] = "keygen";
+    TAG_NAMES["LABEL"] = "label";
+    TAG_NAMES["LI"] = "li";
+    TAG_NAMES["LINK"] = "link";
+    TAG_NAMES["LISTING"] = "listing";
+    TAG_NAMES["MAIN"] = "main";
+    TAG_NAMES["MALIGNMARK"] = "malignmark";
+    TAG_NAMES["MARQUEE"] = "marquee";
+    TAG_NAMES["MATH"] = "math";
+    TAG_NAMES["MENU"] = "menu";
+    TAG_NAMES["META"] = "meta";
+    TAG_NAMES["MGLYPH"] = "mglyph";
+    TAG_NAMES["MI"] = "mi";
+    TAG_NAMES["MO"] = "mo";
+    TAG_NAMES["MN"] = "mn";
+    TAG_NAMES["MS"] = "ms";
+    TAG_NAMES["MTEXT"] = "mtext";
+    TAG_NAMES["NAV"] = "nav";
+    TAG_NAMES["NOBR"] = "nobr";
+    TAG_NAMES["NOFRAMES"] = "noframes";
+    TAG_NAMES["NOEMBED"] = "noembed";
+    TAG_NAMES["NOSCRIPT"] = "noscript";
+    TAG_NAMES["OBJECT"] = "object";
+    TAG_NAMES["OL"] = "ol";
+    TAG_NAMES["OPTGROUP"] = "optgroup";
+    TAG_NAMES["OPTION"] = "option";
+    TAG_NAMES["P"] = "p";
+    TAG_NAMES["PARAM"] = "param";
+    TAG_NAMES["PLAINTEXT"] = "plaintext";
+    TAG_NAMES["PRE"] = "pre";
+    TAG_NAMES["RB"] = "rb";
+    TAG_NAMES["RP"] = "rp";
+    TAG_NAMES["RT"] = "rt";
+    TAG_NAMES["RTC"] = "rtc";
+    TAG_NAMES["RUBY"] = "ruby";
+    TAG_NAMES["S"] = "s";
+    TAG_NAMES["SCRIPT"] = "script";
+    TAG_NAMES["SEARCH"] = "search";
+    TAG_NAMES["SECTION"] = "section";
+    TAG_NAMES["SELECT"] = "select";
+    TAG_NAMES["SOURCE"] = "source";
+    TAG_NAMES["SMALL"] = "small";
+    TAG_NAMES["SPAN"] = "span";
+    TAG_NAMES["STRIKE"] = "strike";
+    TAG_NAMES["STRONG"] = "strong";
+    TAG_NAMES["STYLE"] = "style";
+    TAG_NAMES["SUB"] = "sub";
+    TAG_NAMES["SUMMARY"] = "summary";
+    TAG_NAMES["SUP"] = "sup";
+    TAG_NAMES["TABLE"] = "table";
+    TAG_NAMES["TBODY"] = "tbody";
+    TAG_NAMES["TEMPLATE"] = "template";
+    TAG_NAMES["TEXTAREA"] = "textarea";
+    TAG_NAMES["TFOOT"] = "tfoot";
+    TAG_NAMES["TD"] = "td";
+    TAG_NAMES["TH"] = "th";
+    TAG_NAMES["THEAD"] = "thead";
+    TAG_NAMES["TITLE"] = "title";
+    TAG_NAMES["TR"] = "tr";
+    TAG_NAMES["TRACK"] = "track";
+    TAG_NAMES["TT"] = "tt";
+    TAG_NAMES["U"] = "u";
+    TAG_NAMES["UL"] = "ul";
+    TAG_NAMES["SVG"] = "svg";
+    TAG_NAMES["VAR"] = "var";
+    TAG_NAMES["WBR"] = "wbr";
+    TAG_NAMES["XMP"] = "xmp";
+})(TAG_NAMES || (TAG_NAMES = {}));
+/**
+ * Tag IDs are numeric IDs for known tag names.
+ *
+ * We use tag IDs to improve the performance of tag name comparisons.
+ */
+export var TAG_ID;
+(function (TAG_ID) {
+    TAG_ID[TAG_ID["UNKNOWN"] = 0] = "UNKNOWN";
+    TAG_ID[TAG_ID["A"] = 1] = "A";
+    TAG_ID[TAG_ID["ADDRESS"] = 2] = "ADDRESS";
+    TAG_ID[TAG_ID["ANNOTATION_XML"] = 3] = "ANNOTATION_XML";
+    TAG_ID[TAG_ID["APPLET"] = 4] = "APPLET";
+    TAG_ID[TAG_ID["AREA"] = 5] = "AREA";
+    TAG_ID[TAG_ID["ARTICLE"] = 6] = "ARTICLE";
+    TAG_ID[TAG_ID["ASIDE"] = 7] = "ASIDE";
+    TAG_ID[TAG_ID["B"] = 8] = "B";
+    TAG_ID[TAG_ID["BASE"] = 9] = "BASE";
+    TAG_ID[TAG_ID["BASEFONT"] = 10] = "BASEFONT";
+    TAG_ID[TAG_ID["BGSOUND"] = 11] = "BGSOUND";
+    TAG_ID[TAG_ID["BIG"] = 12] = "BIG";
+    TAG_ID[TAG_ID["BLOCKQUOTE"] = 13] = "BLOCKQUOTE";
+    TAG_ID[TAG_ID["BODY"] = 14] = "BODY";
+    TAG_ID[TAG_ID["BR"] = 15] = "BR";
+    TAG_ID[TAG_ID["BUTTON"] = 16] = "BUTTON";
+    TAG_ID[TAG_ID["CAPTION"] = 17] = "CAPTION";
+    TAG_ID[TAG_ID["CENTER"] = 18] = "CENTER";
+    TAG_ID[TAG_ID["CODE"] = 19] = "CODE";
+    TAG_ID[TAG_ID["COL"] = 20] = "COL";
+    TAG_ID[TAG_ID["COLGROUP"] = 21] = "COLGROUP";
+    TAG_ID[TAG_ID["DD"] = 22] = "DD";
+    TAG_ID[TAG_ID["DESC"] = 23] = "DESC";
+    TAG_ID[TAG_ID["DETAILS"] = 24] = "DETAILS";
+    TAG_ID[TAG_ID["DIALOG"] = 25] = "DIALOG";
+    TAG_ID[TAG_ID["DIR"] = 26] = "DIR";
+    TAG_ID[TAG_ID["DIV"] = 27] = "DIV";
+    TAG_ID[TAG_ID["DL"] = 28] = "DL";
+    TAG_ID[TAG_ID["DT"] = 29] = "DT";
+    TAG_ID[TAG_ID["EM"] = 30] = "EM";
+    TAG_ID[TAG_ID["EMBED"] = 31] = "EMBED";
+    TAG_ID[TAG_ID["FIELDSET"] = 32] = "FIELDSET";
+    TAG_ID[TAG_ID["FIGCAPTION"] = 33] = "FIGCAPTION";
+    TAG_ID[TAG_ID["FIGURE"] = 34] = "FIGURE";
+    TAG_ID[TAG_ID["FONT"] = 35] = "FONT";
+    TAG_ID[TAG_ID["FOOTER"] = 36] = "FOOTER";
+    TAG_ID[TAG_ID["FOREIGN_OBJECT"] = 37] = "FOREIGN_OBJECT";
+    TAG_ID[TAG_ID["FORM"] = 38] = "FORM";
+    TAG_ID[TAG_ID["FRAME"] = 39] = "FRAME";
+    TAG_ID[TAG_ID["FRAMESET"] = 40] = "FRAMESET";
+    TAG_ID[TAG_ID["H1"] = 41] = "H1";
+    TAG_ID[TAG_ID["H2"] = 42] = "H2";
+    TAG_ID[TAG_ID["H3"] = 43] = "H3";
+    TAG_ID[TAG_ID["H4"] = 44] = "H4";
+    TAG_ID[TAG_ID["H5"] = 45] = "H5";
+    TAG_ID[TAG_ID["H6"] = 46] = "H6";
+    TAG_ID[TAG_ID["HEAD"] = 47] = "HEAD";
+    TAG_ID[TAG_ID["HEADER"] = 48] = "HEADER";
+    TAG_ID[TAG_ID["HGROUP"] = 49] = "HGROUP";
+    TAG_ID[TAG_ID["HR"] = 50] = "HR";
+    TAG_ID[TAG_ID["HTML"] = 51] = "HTML";
+    TAG_ID[TAG_ID["I"] = 52] = "I";
+    TAG_ID[TAG_ID["IMG"] = 53] = "IMG";
+    TAG_ID[TAG_ID["IMAGE"] = 54] = "IMAGE";
+    TAG_ID[TAG_ID["INPUT"] = 55] = "INPUT";
+    TAG_ID[TAG_ID["IFRAME"] = 56] = "IFRAME";
+    TAG_ID[TAG_ID["KEYGEN"] = 57] = "KEYGEN";
+    TAG_ID[TAG_ID["LABEL"] = 58] = "LABEL";
+    TAG_ID[TAG_ID["LI"] = 59] = "LI";
+    TAG_ID[TAG_ID["LINK"] = 60] = "LINK";
+    TAG_ID[TAG_ID["LISTING"] = 61] = "LISTING";
+    TAG_ID[TAG_ID["MAIN"] = 62] = "MAIN";
+    TAG_ID[TAG_ID["MALIGNMARK"] = 63] = "MALIGNMARK";
+    TAG_ID[TAG_ID["MARQUEE"] = 64] = "MARQUEE";
+    TAG_ID[TAG_ID["MATH"] = 65] = "MATH";
+    TAG_ID[TAG_ID["MENU"] = 66] = "MENU";
+    TAG_ID[TAG_ID["META"] = 67] = "META";
+    TAG_ID[TAG_ID["MGLYPH"] = 68] = "MGLYPH";
+    TAG_ID[TAG_ID["MI"] = 69] = "MI";
+    TAG_ID[TAG_ID["MO"] = 70] = "MO";
+    TAG_ID[TAG_ID["MN"] = 71] = "MN";
+    TAG_ID[TAG_ID["MS"] = 72] = "MS";
+    TAG_ID[TAG_ID["MTEXT"] = 73] = "MTEXT";
+    TAG_ID[TAG_ID["NAV"] = 74] = "NAV";
+    TAG_ID[TAG_ID["NOBR"] = 75] = "NOBR";
+    TAG_ID[TAG_ID["NOFRAMES"] = 76] = "NOFRAMES";
+    TAG_ID[TAG_ID["NOEMBED"] = 77] = "NOEMBED";
+    TAG_ID[TAG_ID["NOSCRIPT"] = 78] = "NOSCRIPT";
+    TAG_ID[TAG_ID["OBJECT"] = 79] = "OBJECT";
+    TAG_ID[TAG_ID["OL"] = 80] = "OL";
+    TAG_ID[TAG_ID["OPTGROUP"] = 81] = "OPTGROUP";
+    TAG_ID[TAG_ID["OPTION"] = 82] = "OPTION";
+    TAG_ID[TAG_ID["P"] = 83] = "P";
+    TAG_ID[TAG_ID["PARAM"] = 84] = "PARAM";
+    TAG_ID[TAG_ID["PLAINTEXT"] = 85] = "PLAINTEXT";
+    TAG_ID[TAG_ID["PRE"] = 86] = "PRE";
+    TAG_ID[TAG_ID["RB"] = 87] = "RB";
+    TAG_ID[TAG_ID["RP"] = 88] = "RP";
+    TAG_ID[TAG_ID["RT"] = 89] = "RT";
+    TAG_ID[TAG_ID["RTC"] = 90] = "RTC";
+    TAG_ID[TAG_ID["RUBY"] = 91] = "RUBY";
+    TAG_ID[TAG_ID["S"] = 92] = "S";
+    TAG_ID[TAG_ID["SCRIPT"] = 93] = "SCRIPT";
+    TAG_ID[TAG_ID["SEARCH"] = 94] = "SEARCH";
+    TAG_ID[TAG_ID["SECTION"] = 95] = "SECTION";
+    TAG_ID[TAG_ID["SELECT"] = 96] = "SELECT";
+    TAG_ID[TAG_ID["SOURCE"] = 97] = "SOURCE";
+    TAG_ID[TAG_ID["SMALL"] = 98] = "SMALL";
+    TAG_ID[TAG_ID["SPAN"] = 99] = "SPAN";
+    TAG_ID[TAG_ID["STRIKE"] = 100] = "STRIKE";
+    TAG_ID[TAG_ID["STRONG"] = 101] = "STRONG";
+    TAG_ID[TAG_ID["STYLE"] = 102] = "STYLE";
+    TAG_ID[TAG_ID["SUB"] = 103] = "SUB";
+    TAG_ID[TAG_ID["SUMMARY"] = 104] = "SUMMARY";
+    TAG_ID[TAG_ID["SUP"] = 105] = "SUP";
+    TAG_ID[TAG_ID["TABLE"] = 106] = "TABLE";
+    TAG_ID[TAG_ID["TBODY"] = 107] = "TBODY";
+    TAG_ID[TAG_ID["TEMPLATE"] = 108] = "TEMPLATE";
+    TAG_ID[TAG_ID["TEXTAREA"] = 109] = "TEXTAREA";
+    TAG_ID[TAG_ID["TFOOT"] = 110] = "TFOOT";
+    TAG_ID[TAG_ID["TD"] = 111] = "TD";
+    TAG_ID[TAG_ID["TH"] = 112] = "TH";
+    TAG_ID[TAG_ID["THEAD"] = 113] = "THEAD";
+    TAG_ID[TAG_ID["TITLE"] = 114] = "TITLE";
+    TAG_ID[TAG_ID["TR"] = 115] = "TR";
+    TAG_ID[TAG_ID["TRACK"] = 116] = "TRACK";
+    TAG_ID[TAG_ID["TT"] = 117] = "TT";
+    TAG_ID[TAG_ID["U"] = 118] = "U";
+    TAG_ID[TAG_ID["UL"] = 119] = "UL";
+    TAG_ID[TAG_ID["SVG"] = 120] = "SVG";
+    TAG_ID[TAG_ID["VAR"] = 121] = "VAR";
+    TAG_ID[TAG_ID["WBR"] = 122] = "WBR";
+    TAG_ID[TAG_ID["XMP"] = 123] = "XMP";
+})(TAG_ID || (TAG_ID = {}));
+const TAG_NAME_TO_ID = new Map([
+    [TAG_NAMES.A, TAG_ID.A],
+    [TAG_NAMES.ADDRESS, TAG_ID.ADDRESS],
+    [TAG_NAMES.ANNOTATION_XML, TAG_ID.ANNOTATION_XML],
+    [TAG_NAMES.APPLET, TAG_ID.APPLET],
+    [TAG_NAMES.AREA, TAG_ID.AREA],
+    [TAG_NAMES.ARTICLE, TAG_ID.ARTICLE],
+    [TAG_NAMES.ASIDE, TAG_ID.ASIDE],
+    [TAG_NAMES.B, TAG_ID.B],
+    [TAG_NAMES.BASE, TAG_ID.BASE],
+    [TAG_NAMES.BASEFONT, TAG_ID.BASEFONT],
+    [TAG_NAMES.BGSOUND, TAG_ID.BGSOUND],
+    [TAG_NAMES.BIG, TAG_ID.BIG],
+    [TAG_NAMES.BLOCKQUOTE, TAG_ID.BLOCKQUOTE],
+    [TAG_NAMES.BODY, TAG_ID.BODY],
+    [TAG_NAMES.BR, TAG_ID.BR],
+    [TAG_NAMES.BUTTON, TAG_ID.BUTTON],
+    [TAG_NAMES.CAPTION, TAG_ID.CAPTION],
+    [TAG_NAMES.CENTER, TAG_ID.CENTER],
+    [TAG_NAMES.CODE, TAG_ID.CODE],
+    [TAG_NAMES.COL, TAG_ID.COL],
+    [TAG_NAMES.COLGROUP, TAG_ID.COLGROUP],
+    [TAG_NAMES.DD, TAG_ID.DD],
+    [TAG_NAMES.DESC, TAG_ID.DESC],
+    [TAG_NAMES.DETAILS, TAG_ID.DETAILS],
+    [TAG_NAMES.DIALOG, TAG_ID.DIALOG],
+    [TAG_NAMES.DIR, TAG_ID.DIR],
+    [TAG_NAMES.DIV, TAG_ID.DIV],
+    [TAG_NAMES.DL, TAG_ID.DL],
+    [TAG_NAMES.DT, TAG_ID.DT],
+    [TAG_NAMES.EM, TAG_ID.EM],
+    [TAG_NAMES.EMBED, TAG_ID.EMBED],
+    [TAG_NAMES.FIELDSET, TAG_ID.FIELDSET],
+    [TAG_NAMES.FIGCAPTION, TAG_ID.FIGCAPTION],
+    [TAG_NAMES.FIGURE, TAG_ID.FIGURE],
+    [TAG_NAMES.FONT, TAG_ID.FONT],
+    [TAG_NAMES.FOOTER, TAG_ID.FOOTER],
+    [TAG_NAMES.FOREIGN_OBJECT, TAG_ID.FOREIGN_OBJECT],
+    [TAG_NAMES.FORM, TAG_ID.FORM],
+    [TAG_NAMES.FRAME, TAG_ID.FRAME],
+    [TAG_NAMES.FRAMESET, TAG_ID.FRAMESET],
+    [TAG_NAMES.H1, TAG_ID.H1],
+    [TAG_NAMES.H2, TAG_ID.H2],
+    [TAG_NAMES.H3, TAG_ID.H3],
+    [TAG_NAMES.H4, TAG_ID.H4],
+    [TAG_NAMES.H5, TAG_ID.H5],
+    [TAG_NAMES.H6, TAG_ID.H6],
+    [TAG_NAMES.HEAD, TAG_ID.HEAD],
+    [TAG_NAMES.HEADER, TAG_ID.HEADER],
+    [TAG_NAMES.HGROUP, TAG_ID.HGROUP],
+    [TAG_NAMES.HR, TAG_ID.HR],
+    [TAG_NAMES.HTML, TAG_ID.HTML],
+    [TAG_NAMES.I, TAG_ID.I],
+    [TAG_NAMES.IMG, TAG_ID.IMG],
+    [TAG_NAMES.IMAGE, TAG_ID.IMAGE],
+    [TAG_NAMES.INPUT, TAG_ID.INPUT],
+    [TAG_NAMES.IFRAME, TAG_ID.IFRAME],
+    [TAG_NAMES.KEYGEN, TAG_ID.KEYGEN],
+    [TAG_NAMES.LABEL, TAG_ID.LABEL],
+    [TAG_NAMES.LI, TAG_ID.LI],
+    [TAG_NAMES.LINK, TAG_ID.LINK],
+    [TAG_NAMES.LISTING, TAG_ID.LISTING],
+    [TAG_NAMES.MAIN, TAG_ID.MAIN],
+    [TAG_NAMES.MALIGNMARK, TAG_ID.MALIGNMARK],
+    [TAG_NAMES.MARQUEE, TAG_ID.MARQUEE],
+    [TAG_NAMES.MATH, TAG_ID.MATH],
+    [TAG_NAMES.MENU, TAG_ID.MENU],
+    [TAG_NAMES.META, TAG_ID.META],
+    [TAG_NAMES.MGLYPH, TAG_ID.MGLYPH],
+    [TAG_NAMES.MI, TAG_ID.MI],
+    [TAG_NAMES.MO, TAG_ID.MO],
+    [TAG_NAMES.MN, TAG_ID.MN],
+    [TAG_NAMES.MS, TAG_ID.MS],
+    [TAG_NAMES.MTEXT, TAG_ID.MTEXT],
+    [TAG_NAMES.NAV, TAG_ID.NAV],
+    [TAG_NAMES.NOBR, TAG_ID.NOBR],
+    [TAG_NAMES.NOFRAMES, TAG_ID.NOFRAMES],
+    [TAG_NAMES.NOEMBED, TAG_ID.NOEMBED],
+    [TAG_NAMES.NOSCRIPT, TAG_ID.NOSCRIPT],
+    [TAG_NAMES.OBJECT, TAG_ID.OBJECT],
+    [TAG_NAMES.OL, TAG_ID.OL],
+    [TAG_NAMES.OPTGROUP, TAG_ID.OPTGROUP],
+    [TAG_NAMES.OPTION, TAG_ID.OPTION],
+    [TAG_NAMES.P, TAG_ID.P],
+    [TAG_NAMES.PARAM, TAG_ID.PARAM],
+    [TAG_NAMES.PLAINTEXT, TAG_ID.PLAINTEXT],
+    [TAG_NAMES.PRE, TAG_ID.PRE],
+    [TAG_NAMES.RB, TAG_ID.RB],
+    [TAG_NAMES.RP, TAG_ID.RP],
+    [TAG_NAMES.RT, TAG_ID.RT],
+    [TAG_NAMES.RTC, TAG_ID.RTC],
+    [TAG_NAMES.RUBY, TAG_ID.RUBY],
+    [TAG_NAMES.S, TAG_ID.S],
+    [TAG_NAMES.SCRIPT, TAG_ID.SCRIPT],
+    [TAG_NAMES.SEARCH, TAG_ID.SEARCH],
+    [TAG_NAMES.SECTION, TAG_ID.SECTION],
+    [TAG_NAMES.SELECT, TAG_ID.SELECT],
+    [TAG_NAMES.SOURCE, TAG_ID.SOURCE],
+    [TAG_NAMES.SMALL, TAG_ID.SMALL],
+    [TAG_NAMES.SPAN, TAG_ID.SPAN],
+    [TAG_NAMES.STRIKE, TAG_ID.STRIKE],
+    [TAG_NAMES.STRONG, TAG_ID.STRONG],
+    [TAG_NAMES.STYLE, TAG_ID.STYLE],
+    [TAG_NAMES.SUB, TAG_ID.SUB],
+    [TAG_NAMES.SUMMARY, TAG_ID.SUMMARY],
+    [TAG_NAMES.SUP, TAG_ID.SUP],
+    [TAG_NAMES.TABLE, TAG_ID.TABLE],
+    [TAG_NAMES.TBODY, TAG_ID.TBODY],
+    [TAG_NAMES.TEMPLATE, TAG_ID.TEMPLATE],
+    [TAG_NAMES.TEXTAREA, TAG_ID.TEXTAREA],
+    [TAG_NAMES.TFOOT, TAG_ID.TFOOT],
+    [TAG_NAMES.TD, TAG_ID.TD],
+    [TAG_NAMES.TH, TAG_ID.TH],
+    [TAG_NAMES.THEAD, TAG_ID.THEAD],
+    [TAG_NAMES.TITLE, TAG_ID.TITLE],
+    [TAG_NAMES.TR, TAG_ID.TR],
+    [TAG_NAMES.TRACK, TAG_ID.TRACK],
+    [TAG_NAMES.TT, TAG_ID.TT],
+    [TAG_NAMES.U, TAG_ID.U],
+    [TAG_NAMES.UL, TAG_ID.UL],
+    [TAG_NAMES.SVG, TAG_ID.SVG],
+    [TAG_NAMES.VAR, TAG_ID.VAR],
+    [TAG_NAMES.WBR, TAG_ID.WBR],
+    [TAG_NAMES.XMP, TAG_ID.XMP],
+]);
+export function getTagID(tagName) {
+    var _a;
+    return (_a = TAG_NAME_TO_ID.get(tagName)) !== null && _a !== void 0 ? _a : TAG_ID.UNKNOWN;
+}
+const $ = TAG_ID;
+export const SPECIAL_ELEMENTS = {
+    [NS.HTML]: new Set([
+        $.ADDRESS,
+        $.APPLET,
+        $.AREA,
+        $.ARTICLE,
+        $.ASIDE,
+        $.BASE,
+        $.BASEFONT,
+        $.BGSOUND,
+        $.BLOCKQUOTE,
+        $.BODY,
+        $.BR,
+        $.BUTTON,
+        $.CAPTION,
+        $.CENTER,
+        $.COL,
+        $.COLGROUP,
+        $.DD,
+        $.DETAILS,
+        $.DIR,
+        $.DIV,
+        $.DL,
+        $.DT,
+        $.EMBED,
+        $.FIELDSET,
+        $.FIGCAPTION,
+        $.FIGURE,
+        $.FOOTER,
+        $.FORM,
+        $.FRAME,
+        $.FRAMESET,
+        $.H1,
+        $.H2,
+        $.H3,
+        $.H4,
+        $.H5,
+        $.H6,
+        $.HEAD,
+        $.HEADER,
+        $.HGROUP,
+        $.HR,
+        $.HTML,
+        $.IFRAME,
+        $.IMG,
+        $.INPUT,
+        $.LI,
+        $.LINK,
+        $.LISTING,
+        $.MAIN,
+        $.MARQUEE,
+        $.MENU,
+        $.META,
+        $.NAV,
+        $.NOEMBED,
+        $.NOFRAMES,
+        $.NOSCRIPT,
+        $.OBJECT,
+        $.OL,
+        $.P,
+        $.PARAM,
+        $.PLAINTEXT,
+        $.PRE,
+        $.SCRIPT,
+        $.SECTION,
+        $.SELECT,
+        $.SOURCE,
+        $.STYLE,
+        $.SUMMARY,
+        $.TABLE,
+        $.TBODY,
+        $.TD,
+        $.TEMPLATE,
+        $.TEXTAREA,
+        $.TFOOT,
+        $.TH,
+        $.THEAD,
+        $.TITLE,
+        $.TR,
+        $.TRACK,
+        $.UL,
+        $.WBR,
+        $.XMP,
+    ]),
+    [NS.MATHML]: new Set([$.MI, $.MO, $.MN, $.MS, $.MTEXT, $.ANNOTATION_XML]),
+    [NS.SVG]: new Set([$.TITLE, $.FOREIGN_OBJECT, $.DESC]),
+    [NS.XLINK]: new Set(),
+    [NS.XML]: new Set(),
+    [NS.XMLNS]: new Set(),
+};
+export const NUMBERED_HEADERS = new Set([$.H1, $.H2, $.H3, $.H4, $.H5, $.H6]);
+const UNESCAPED_TEXT = new Set([
+    TAG_NAMES.STYLE,
+    TAG_NAMES.SCRIPT,
+    TAG_NAMES.XMP,
+    TAG_NAMES.IFRAME,
+    TAG_NAMES.NOEMBED,
+    TAG_NAMES.NOFRAMES,
+    TAG_NAMES.PLAINTEXT,
+]);
+export function hasUnescapedText(tn, scriptingEnabled) {
+    return UNESCAPED_TEXT.has(tn) || (scriptingEnabled && tn === TAG_NAMES.NOSCRIPT);
+}

data/lib/readability_js/node/node_modules/parse5/dist/common/token.d.ts

@@ -0,0 +1,84 @@
+import type { TAG_ID } from './html.js';
+export declare enum TokenType {
+    CHARACTER = 0,
+    NULL_CHARACTER = 1,
+    WHITESPACE_CHARACTER = 2,
+    START_TAG = 3,
+    END_TAG = 4,
+    COMMENT = 5,
+    DOCTYPE = 6,
+    EOF = 7,
+    HIBERNATION = 8
+}
+export interface Location {
+    /** One-based line index of the first character. */
+    startLine: number;
+    /** One-based column index of the first character. */
+    startCol: number;
+    /** Zero-based first character index. */
+    startOffset: number;
+    /** One-based line index of the last character. */
+    endLine: number;
+    /** One-based column index of the last character. Points directly *after* the last character. */
+    endCol: number;
+    /** Zero-based last character index. Points directly *after* the last character. */
+    endOffset: number;
+}
+export interface LocationWithAttributes extends Location {
+    /** Start tag attributes' location info. */
+    attrs?: Record<string, Location>;
+}
+export interface ElementLocation extends LocationWithAttributes {
+    /** Element's start tag location info. */
+    startTag?: Location;
+    /**
+     * Element's end tag location info.
+     * This property is undefined, if the element has no closing tag.
+     */
+    endTag?: Location;
+}
+interface TokenBase {
+    readonly type: TokenType;
+    location: Location | null;
+}
+export interface DoctypeToken extends TokenBase {
+    readonly type: TokenType.DOCTYPE;
+    name: string | null;
+    forceQuirks: boolean;
+    publicId: string | null;
+    systemId: string | null;
+}
+export interface Attribute {
+    /** The name of the attribute. */
+    name: string;
+    /** The namespace of the attribute. */
+    namespace?: string;
+    /** The namespace-related prefix of the attribute. */
+    prefix?: string;
+    /** The value of the attribute. */
+    value: string;
+}
+export interface TagToken extends TokenBase {
+    readonly type: TokenType.START_TAG | TokenType.END_TAG;
+    tagName: string;
+    /** Used to cache the ID of the tag name. */
+    tagID: TAG_ID;
+    selfClosing: boolean;
+    ackSelfClosing: boolean;
+    attrs: Attribute[];
+    location: LocationWithAttributes | null;
+}
+export declare function getTokenAttr(token: TagToken, attrName: string): string | null;
+export interface CommentToken extends TokenBase {
+    readonly type: TokenType.COMMENT;
+    data: string;
+}
+export interface EOFToken extends TokenBase {
+    readonly type: TokenType.EOF;
+}
+export interface CharacterToken extends TokenBase {
+    type: TokenType.CHARACTER | TokenType.NULL_CHARACTER | TokenType.WHITESPACE_CHARACTER;
+    chars: string;
+}
+export type Token = DoctypeToken | TagToken | CommentToken | EOFToken | CharacterToken;
+export {};

data/lib/readability_js/node/node_modules/parse5/dist/common/token.js

@@ -0,0 +1,20 @@
+export var TokenType;
+(function (TokenType) {
+    TokenType[TokenType["CHARACTER"] = 0] = "CHARACTER";
+    TokenType[TokenType["NULL_CHARACTER"] = 1] = "NULL_CHARACTER";
+    TokenType[TokenType["WHITESPACE_CHARACTER"] = 2] = "WHITESPACE_CHARACTER";
+    TokenType[TokenType["START_TAG"] = 3] = "START_TAG";
+    TokenType[TokenType["END_TAG"] = 4] = "END_TAG";
+    TokenType[TokenType["COMMENT"] = 5] = "COMMENT";
+    TokenType[TokenType["DOCTYPE"] = 6] = "DOCTYPE";
+    TokenType[TokenType["EOF"] = 7] = "EOF";
+    TokenType[TokenType["HIBERNATION"] = 8] = "HIBERNATION";
+})(TokenType || (TokenType = {}));
+export function getTokenAttr(token, attrName) {
+    for (let i = token.attrs.length - 1; i >= 0; i--) {
+        if (token.attrs[i].name === attrName) {
+            return token.attrs[i].value;
+        }
+    }
+    return null;
+}

data/lib/readability_js/node/node_modules/parse5/dist/common/unicode.d.ts

@@ -0,0 +1,42 @@
+export declare const REPLACEMENT_CHARACTER = "\uFFFD";
+export declare enum CODE_POINTS {
+    EOF = -1,
+    NULL = 0,
+    TABULATION = 9,
+    CARRIAGE_RETURN = 13,
+    LINE_FEED = 10,
+    FORM_FEED = 12,
+    SPACE = 32,
+    EXCLAMATION_MARK = 33,
+    QUOTATION_MARK = 34,
+    AMPERSAND = 38,
+    APOSTROPHE = 39,
+    HYPHEN_MINUS = 45,
+    SOLIDUS = 47,
+    DIGIT_0 = 48,
+    DIGIT_9 = 57,
+    SEMICOLON = 59,
+    LESS_THAN_SIGN = 60,
+    EQUALS_SIGN = 61,
+    GREATER_THAN_SIGN = 62,
+    QUESTION_MARK = 63,
+    LATIN_CAPITAL_A = 65,
+    LATIN_CAPITAL_Z = 90,
+    RIGHT_SQUARE_BRACKET = 93,
+    GRAVE_ACCENT = 96,
+    LATIN_SMALL_A = 97,
+    LATIN_SMALL_Z = 122
+}
+export declare const SEQUENCES: {
+    DASH_DASH: string;
+    CDATA_START: string;
+    DOCTYPE: string;
+    SCRIPT: string;
+    PUBLIC: string;
+    SYSTEM: string;
+};
+export declare function isSurrogate(cp: number): boolean;
+export declare function isSurrogatePair(cp: number): boolean;
+export declare function getSurrogatePairCodePoint(cp1: number, cp2: number): number;
+export declare function isControlCodePoint(cp: number): boolean;
+export declare function isUndefinedCodePoint(cp: number): boolean;