1661 lines
68 KiB
JavaScript
1661 lines
68 KiB
JavaScript
(() => {
|
|
var __create = Object.create;
|
|
var __getProtoOf = Object.getPrototypeOf;
|
|
var __defProp = Object.defineProperty;
|
|
var __getOwnPropNames = Object.getOwnPropertyNames;
|
|
var __hasOwnProp = Object.prototype.hasOwnProperty;
|
|
var __toESM = (mod, isNodeMode, target) => {
|
|
target = mod != null ? __create(__getProtoOf(mod)) : {};
|
|
const to = isNodeMode || !mod || !mod.__esModule ? __defProp(target, "default", { value: mod, enumerable: true }) : target;
|
|
for (let key of __getOwnPropNames(mod))
|
|
if (!__hasOwnProp.call(to, key))
|
|
__defProp(to, key, {
|
|
get: () => mod[key],
|
|
enumerable: true
|
|
});
|
|
return to;
|
|
};
|
|
var __commonJS = (cb, mod) => () => (mod || cb((mod = { exports: {} }).exports, mod), mod.exports);
|
|
|
|
// ../node_modules/.pnpm/@mozilla+readability@0.6.0/node_modules/@mozilla/readability/Readability.js
|
|
var require_Readability = __commonJS((exports, module) => {
|
|
function Readability(doc, options) {
|
|
if (options && options.documentElement) {
|
|
doc = options;
|
|
options = arguments[2];
|
|
} else if (!doc || !doc.documentElement) {
|
|
throw new Error("First argument to Readability constructor should be a document object.");
|
|
}
|
|
options = options || {};
|
|
this._doc = doc;
|
|
this._docJSDOMParser = this._doc.firstChild.__JSDOMParser__;
|
|
this._articleTitle = null;
|
|
this._articleByline = null;
|
|
this._articleDir = null;
|
|
this._articleSiteName = null;
|
|
this._attempts = [];
|
|
this._metadata = {};
|
|
this._debug = !!options.debug;
|
|
this._maxElemsToParse = options.maxElemsToParse || this.DEFAULT_MAX_ELEMS_TO_PARSE;
|
|
this._nbTopCandidates = options.nbTopCandidates || this.DEFAULT_N_TOP_CANDIDATES;
|
|
this._charThreshold = options.charThreshold || this.DEFAULT_CHAR_THRESHOLD;
|
|
this._classesToPreserve = this.CLASSES_TO_PRESERVE.concat(options.classesToPreserve || []);
|
|
this._keepClasses = !!options.keepClasses;
|
|
this._serializer = options.serializer || function(el) {
|
|
return el.innerHTML;
|
|
};
|
|
this._disableJSONLD = !!options.disableJSONLD;
|
|
this._allowedVideoRegex = options.allowedVideoRegex || this.REGEXPS.videos;
|
|
this._linkDensityModifier = options.linkDensityModifier || 0;
|
|
this._flags = this.FLAG_STRIP_UNLIKELYS | this.FLAG_WEIGHT_CLASSES | this.FLAG_CLEAN_CONDITIONALLY;
|
|
if (this._debug) {
|
|
let logNode = function(node) {
|
|
if (node.nodeType == node.TEXT_NODE) {
|
|
return `${node.nodeName} ("${node.textContent}")`;
|
|
}
|
|
let attrPairs = Array.from(node.attributes || [], function(attr) {
|
|
return `${attr.name}="${attr.value}"`;
|
|
}).join(" ");
|
|
return `<${node.localName} ${attrPairs}>`;
|
|
};
|
|
this.log = function() {
|
|
if (typeof console !== "undefined") {
|
|
let args = Array.from(arguments, (arg) => {
|
|
if (arg && arg.nodeType == this.ELEMENT_NODE) {
|
|
return logNode(arg);
|
|
}
|
|
return arg;
|
|
});
|
|
args.unshift("Reader: (Readability)");
|
|
console.log(...args);
|
|
} else if (typeof dump !== "undefined") {
|
|
var msg = Array.prototype.map.call(arguments, function(x) {
|
|
return x && x.nodeName ? logNode(x) : x;
|
|
}).join(" ");
|
|
dump("Reader: (Readability) " + msg + `
|
|
`);
|
|
}
|
|
};
|
|
} else {
|
|
this.log = function() {};
|
|
}
|
|
}
|
|
Readability.prototype = {
|
|
FLAG_STRIP_UNLIKELYS: 1,
|
|
FLAG_WEIGHT_CLASSES: 2,
|
|
FLAG_CLEAN_CONDITIONALLY: 4,
|
|
ELEMENT_NODE: 1,
|
|
TEXT_NODE: 3,
|
|
DEFAULT_MAX_ELEMS_TO_PARSE: 0,
|
|
DEFAULT_N_TOP_CANDIDATES: 5,
|
|
DEFAULT_TAGS_TO_SCORE: "section,h2,h3,h4,h5,h6,p,td,pre".toUpperCase().split(","),
|
|
DEFAULT_CHAR_THRESHOLD: 500,
|
|
REGEXPS: {
|
|
unlikelyCandidates: /-ad-|ai2html|banner|breadcrumbs|combx|comment|community|cover-wrap|disqus|extra|footer|gdpr|header|legends|menu|related|remark|replies|rss|shoutbox|sidebar|skyscraper|social|sponsor|supplemental|ad-break|agegate|pagination|pager|popup|yom-remote/i,
|
|
okMaybeItsACandidate: /and|article|body|column|content|main|shadow/i,
|
|
positive: /article|body|content|entry|hentry|h-entry|main|page|pagination|post|text|blog|story/i,
|
|
negative: /-ad-|hidden|^hid$| hid$| hid |^hid |banner|combx|comment|com-|contact|footer|gdpr|masthead|media|meta|outbrain|promo|related|scroll|share|shoutbox|sidebar|skyscraper|sponsor|shopping|tags|widget/i,
|
|
extraneous: /print|archive|comment|discuss|e[\-]?mail|share|reply|all|login|sign|single|utility/i,
|
|
byline: /byline|author|dateline|writtenby|p-author/i,
|
|
replaceFonts: /<(\/?)font[^>]*>/gi,
|
|
normalize: /\s{2,}/g,
|
|
videos: /\/\/(www\.)?((dailymotion|youtube|youtube-nocookie|player\.vimeo|v\.qq)\.com|(archive|upload\.wikimedia)\.org|player\.twitch\.tv)/i,
|
|
shareElements: /(\b|_)(share|sharedaddy)(\b|_)/i,
|
|
nextLink: /(next|weiter|continue|>([^\|]|$)|»([^\|]|$))/i,
|
|
prevLink: /(prev|earl|old|new|<|«)/i,
|
|
tokenize: /\W+/g,
|
|
whitespace: /^\s*$/,
|
|
hasContent: /\S$/,
|
|
hashUrl: /^#.+/,
|
|
srcsetUrl: /(\S+)(\s+[\d.]+[xw])?(\s*(?:,|$))/g,
|
|
b64DataUrl: /^data:\s*([^\s;,]+)\s*;\s*base64\s*,/i,
|
|
commas: /\u002C|\u060C|\uFE50|\uFE10|\uFE11|\u2E41|\u2E34|\u2E32|\uFF0C/g,
|
|
jsonLdArticleTypes: /^Article|AdvertiserContentArticle|NewsArticle|AnalysisNewsArticle|AskPublicNewsArticle|BackgroundNewsArticle|OpinionNewsArticle|ReportageNewsArticle|ReviewNewsArticle|Report|SatiricalArticle|ScholarlyArticle|MedicalScholarlyArticle|SocialMediaPosting|BlogPosting|LiveBlogPosting|DiscussionForumPosting|TechArticle|APIReference$/,
|
|
adWords: /^(ad(vertising|vertisement)?|pub(licité)?|werb(ung)?|广告|Реклама|Anuncio)$/iu,
|
|
loadingWords: /^((loading|正在加载|Загрузка|chargement|cargando)(…|\.\.\.)?)$/iu
|
|
},
|
|
UNLIKELY_ROLES: [
|
|
"menu",
|
|
"menubar",
|
|
"complementary",
|
|
"navigation",
|
|
"alert",
|
|
"alertdialog",
|
|
"dialog"
|
|
],
|
|
DIV_TO_P_ELEMS: new Set([
|
|
"BLOCKQUOTE",
|
|
"DL",
|
|
"DIV",
|
|
"IMG",
|
|
"OL",
|
|
"P",
|
|
"PRE",
|
|
"TABLE",
|
|
"UL"
|
|
]),
|
|
ALTER_TO_DIV_EXCEPTIONS: ["DIV", "ARTICLE", "SECTION", "P", "OL", "UL"],
|
|
PRESENTATIONAL_ATTRIBUTES: [
|
|
"align",
|
|
"background",
|
|
"bgcolor",
|
|
"border",
|
|
"cellpadding",
|
|
"cellspacing",
|
|
"frame",
|
|
"hspace",
|
|
"rules",
|
|
"style",
|
|
"valign",
|
|
"vspace"
|
|
],
|
|
DEPRECATED_SIZE_ATTRIBUTE_ELEMS: ["TABLE", "TH", "TD", "HR", "PRE"],
|
|
PHRASING_ELEMS: [
|
|
"ABBR",
|
|
"AUDIO",
|
|
"B",
|
|
"BDO",
|
|
"BR",
|
|
"BUTTON",
|
|
"CITE",
|
|
"CODE",
|
|
"DATA",
|
|
"DATALIST",
|
|
"DFN",
|
|
"EM",
|
|
"EMBED",
|
|
"I",
|
|
"IMG",
|
|
"INPUT",
|
|
"KBD",
|
|
"LABEL",
|
|
"MARK",
|
|
"MATH",
|
|
"METER",
|
|
"NOSCRIPT",
|
|
"OBJECT",
|
|
"OUTPUT",
|
|
"PROGRESS",
|
|
"Q",
|
|
"RUBY",
|
|
"SAMP",
|
|
"SCRIPT",
|
|
"SELECT",
|
|
"SMALL",
|
|
"SPAN",
|
|
"STRONG",
|
|
"SUB",
|
|
"SUP",
|
|
"TEXTAREA",
|
|
"TIME",
|
|
"VAR",
|
|
"WBR"
|
|
],
|
|
CLASSES_TO_PRESERVE: ["page"],
|
|
HTML_ESCAPE_MAP: {
|
|
lt: "<",
|
|
gt: ">",
|
|
amp: "&",
|
|
quot: '"',
|
|
apos: "'"
|
|
},
|
|
_postProcessContent(articleContent) {
|
|
this._fixRelativeUris(articleContent);
|
|
this._simplifyNestedElements(articleContent);
|
|
if (!this._keepClasses) {
|
|
this._cleanClasses(articleContent);
|
|
}
|
|
},
|
|
_removeNodes(nodeList, filterFn) {
|
|
if (this._docJSDOMParser && nodeList._isLiveNodeList) {
|
|
throw new Error("Do not pass live node lists to _removeNodes");
|
|
}
|
|
for (var i = nodeList.length - 1;i >= 0; i--) {
|
|
var node = nodeList[i];
|
|
var parentNode = node.parentNode;
|
|
if (parentNode) {
|
|
if (!filterFn || filterFn.call(this, node, i, nodeList)) {
|
|
parentNode.removeChild(node);
|
|
}
|
|
}
|
|
}
|
|
},
|
|
_replaceNodeTags(nodeList, newTagName) {
|
|
if (this._docJSDOMParser && nodeList._isLiveNodeList) {
|
|
throw new Error("Do not pass live node lists to _replaceNodeTags");
|
|
}
|
|
for (const node of nodeList) {
|
|
this._setNodeTag(node, newTagName);
|
|
}
|
|
},
|
|
_forEachNode(nodeList, fn) {
|
|
Array.prototype.forEach.call(nodeList, fn, this);
|
|
},
|
|
_findNode(nodeList, fn) {
|
|
return Array.prototype.find.call(nodeList, fn, this);
|
|
},
|
|
_someNode(nodeList, fn) {
|
|
return Array.prototype.some.call(nodeList, fn, this);
|
|
},
|
|
_everyNode(nodeList, fn) {
|
|
return Array.prototype.every.call(nodeList, fn, this);
|
|
},
|
|
_getAllNodesWithTag(node, tagNames) {
|
|
if (node.querySelectorAll) {
|
|
return node.querySelectorAll(tagNames.join(","));
|
|
}
|
|
return [].concat.apply([], tagNames.map(function(tag) {
|
|
var collection = node.getElementsByTagName(tag);
|
|
return Array.isArray(collection) ? collection : Array.from(collection);
|
|
}));
|
|
},
|
|
_cleanClasses(node) {
|
|
var classesToPreserve = this._classesToPreserve;
|
|
var className = (node.getAttribute("class") || "").split(/\s+/).filter((cls) => classesToPreserve.includes(cls)).join(" ");
|
|
if (className) {
|
|
node.setAttribute("class", className);
|
|
} else {
|
|
node.removeAttribute("class");
|
|
}
|
|
for (node = node.firstElementChild;node; node = node.nextElementSibling) {
|
|
this._cleanClasses(node);
|
|
}
|
|
},
|
|
_isUrl(str) {
|
|
try {
|
|
new URL(str);
|
|
return true;
|
|
} catch {
|
|
return false;
|
|
}
|
|
},
|
|
_fixRelativeUris(articleContent) {
|
|
var baseURI = this._doc.baseURI;
|
|
var documentURI = this._doc.documentURI;
|
|
function toAbsoluteURI(uri) {
|
|
if (baseURI == documentURI && uri.charAt(0) == "#") {
|
|
return uri;
|
|
}
|
|
try {
|
|
return new URL(uri, baseURI).href;
|
|
} catch (ex) {}
|
|
return uri;
|
|
}
|
|
var links = this._getAllNodesWithTag(articleContent, ["a"]);
|
|
this._forEachNode(links, function(link) {
|
|
var href = link.getAttribute("href");
|
|
if (href) {
|
|
if (href.indexOf("javascript:") === 0) {
|
|
if (link.childNodes.length === 1 && link.childNodes[0].nodeType === this.TEXT_NODE) {
|
|
var text = this._doc.createTextNode(link.textContent);
|
|
link.parentNode.replaceChild(text, link);
|
|
} else {
|
|
var container = this._doc.createElement("span");
|
|
while (link.firstChild) {
|
|
container.appendChild(link.firstChild);
|
|
}
|
|
link.parentNode.replaceChild(container, link);
|
|
}
|
|
} else {
|
|
link.setAttribute("href", toAbsoluteURI(href));
|
|
}
|
|
}
|
|
});
|
|
var medias = this._getAllNodesWithTag(articleContent, [
|
|
"img",
|
|
"picture",
|
|
"figure",
|
|
"video",
|
|
"audio",
|
|
"source"
|
|
]);
|
|
this._forEachNode(medias, function(media) {
|
|
var src = media.getAttribute("src");
|
|
var poster = media.getAttribute("poster");
|
|
var srcset = media.getAttribute("srcset");
|
|
if (src) {
|
|
media.setAttribute("src", toAbsoluteURI(src));
|
|
}
|
|
if (poster) {
|
|
media.setAttribute("poster", toAbsoluteURI(poster));
|
|
}
|
|
if (srcset) {
|
|
var newSrcset = srcset.replace(this.REGEXPS.srcsetUrl, function(_, p1, p2, p3) {
|
|
return toAbsoluteURI(p1) + (p2 || "") + p3;
|
|
});
|
|
media.setAttribute("srcset", newSrcset);
|
|
}
|
|
});
|
|
},
|
|
_simplifyNestedElements(articleContent) {
|
|
var node = articleContent;
|
|
while (node) {
|
|
if (node.parentNode && ["DIV", "SECTION"].includes(node.tagName) && !(node.id && node.id.startsWith("readability"))) {
|
|
if (this._isElementWithoutContent(node)) {
|
|
node = this._removeAndGetNext(node);
|
|
continue;
|
|
} else if (this._hasSingleTagInsideElement(node, "DIV") || this._hasSingleTagInsideElement(node, "SECTION")) {
|
|
var child = node.children[0];
|
|
for (var i = 0;i < node.attributes.length; i++) {
|
|
child.setAttributeNode(node.attributes[i].cloneNode());
|
|
}
|
|
node.parentNode.replaceChild(child, node);
|
|
node = child;
|
|
continue;
|
|
}
|
|
}
|
|
node = this._getNextNode(node);
|
|
}
|
|
},
|
|
_getArticleTitle() {
|
|
var doc = this._doc;
|
|
var curTitle = "";
|
|
var origTitle = "";
|
|
try {
|
|
curTitle = origTitle = doc.title.trim();
|
|
if (typeof curTitle !== "string") {
|
|
curTitle = origTitle = this._getInnerText(doc.getElementsByTagName("title")[0]);
|
|
}
|
|
} catch (e) {}
|
|
var titleHadHierarchicalSeparators = false;
|
|
function wordCount(str) {
|
|
return str.split(/\s+/).length;
|
|
}
|
|
if (/ [\|\-\\\/>»] /.test(curTitle)) {
|
|
titleHadHierarchicalSeparators = / [\\\/>»] /.test(curTitle);
|
|
let allSeparators = Array.from(origTitle.matchAll(/ [\|\-\\\/>»] /gi));
|
|
curTitle = origTitle.substring(0, allSeparators.pop().index);
|
|
if (wordCount(curTitle) < 3) {
|
|
curTitle = origTitle.replace(/^[^\|\-\\\/>»]*[\|\-\\\/>»]/gi, "");
|
|
}
|
|
} else if (curTitle.includes(": ")) {
|
|
var headings = this._getAllNodesWithTag(doc, ["h1", "h2"]);
|
|
var trimmedTitle = curTitle.trim();
|
|
var match = this._someNode(headings, function(heading) {
|
|
return heading.textContent.trim() === trimmedTitle;
|
|
});
|
|
if (!match) {
|
|
curTitle = origTitle.substring(origTitle.lastIndexOf(":") + 1);
|
|
if (wordCount(curTitle) < 3) {
|
|
curTitle = origTitle.substring(origTitle.indexOf(":") + 1);
|
|
} else if (wordCount(origTitle.substr(0, origTitle.indexOf(":"))) > 5) {
|
|
curTitle = origTitle;
|
|
}
|
|
}
|
|
} else if (curTitle.length > 150 || curTitle.length < 15) {
|
|
var hOnes = doc.getElementsByTagName("h1");
|
|
if (hOnes.length === 1) {
|
|
curTitle = this._getInnerText(hOnes[0]);
|
|
}
|
|
}
|
|
curTitle = curTitle.trim().replace(this.REGEXPS.normalize, " ");
|
|
var curTitleWordCount = wordCount(curTitle);
|
|
if (curTitleWordCount <= 4 && (!titleHadHierarchicalSeparators || curTitleWordCount != wordCount(origTitle.replace(/[\|\-\\\/>»]+/g, "")) - 1)) {
|
|
curTitle = origTitle;
|
|
}
|
|
return curTitle;
|
|
},
|
|
_prepDocument() {
|
|
var doc = this._doc;
|
|
this._removeNodes(this._getAllNodesWithTag(doc, ["style"]));
|
|
if (doc.body) {
|
|
this._replaceBrs(doc.body);
|
|
}
|
|
this._replaceNodeTags(this._getAllNodesWithTag(doc, ["font"]), "SPAN");
|
|
},
|
|
_nextNode(node) {
|
|
var next = node;
|
|
while (next && next.nodeType != this.ELEMENT_NODE && this.REGEXPS.whitespace.test(next.textContent)) {
|
|
next = next.nextSibling;
|
|
}
|
|
return next;
|
|
},
|
|
_replaceBrs(elem) {
|
|
this._forEachNode(this._getAllNodesWithTag(elem, ["br"]), function(br) {
|
|
var next = br.nextSibling;
|
|
var replaced = false;
|
|
while ((next = this._nextNode(next)) && next.tagName == "BR") {
|
|
replaced = true;
|
|
var brSibling = next.nextSibling;
|
|
next.remove();
|
|
next = brSibling;
|
|
}
|
|
if (replaced) {
|
|
var p = this._doc.createElement("p");
|
|
br.parentNode.replaceChild(p, br);
|
|
next = p.nextSibling;
|
|
while (next) {
|
|
if (next.tagName == "BR") {
|
|
var nextElem = this._nextNode(next.nextSibling);
|
|
if (nextElem && nextElem.tagName == "BR") {
|
|
break;
|
|
}
|
|
}
|
|
if (!this._isPhrasingContent(next)) {
|
|
break;
|
|
}
|
|
var sibling = next.nextSibling;
|
|
p.appendChild(next);
|
|
next = sibling;
|
|
}
|
|
while (p.lastChild && this._isWhitespace(p.lastChild)) {
|
|
p.lastChild.remove();
|
|
}
|
|
if (p.parentNode.tagName === "P") {
|
|
this._setNodeTag(p.parentNode, "DIV");
|
|
}
|
|
}
|
|
});
|
|
},
|
|
_setNodeTag(node, tag) {
|
|
this.log("_setNodeTag", node, tag);
|
|
if (this._docJSDOMParser) {
|
|
node.localName = tag.toLowerCase();
|
|
node.tagName = tag.toUpperCase();
|
|
return node;
|
|
}
|
|
var replacement = node.ownerDocument.createElement(tag);
|
|
while (node.firstChild) {
|
|
replacement.appendChild(node.firstChild);
|
|
}
|
|
node.parentNode.replaceChild(replacement, node);
|
|
if (node.readability) {
|
|
replacement.readability = node.readability;
|
|
}
|
|
for (var i = 0;i < node.attributes.length; i++) {
|
|
replacement.setAttributeNode(node.attributes[i].cloneNode());
|
|
}
|
|
return replacement;
|
|
},
|
|
_prepArticle(articleContent) {
|
|
this._cleanStyles(articleContent);
|
|
this._markDataTables(articleContent);
|
|
this._fixLazyImages(articleContent);
|
|
this._cleanConditionally(articleContent, "form");
|
|
this._cleanConditionally(articleContent, "fieldset");
|
|
this._clean(articleContent, "object");
|
|
this._clean(articleContent, "embed");
|
|
this._clean(articleContent, "footer");
|
|
this._clean(articleContent, "link");
|
|
this._clean(articleContent, "aside");
|
|
var shareElementThreshold = this.DEFAULT_CHAR_THRESHOLD;
|
|
this._forEachNode(articleContent.children, function(topCandidate) {
|
|
this._cleanMatchedNodes(topCandidate, function(node, matchString) {
|
|
return this.REGEXPS.shareElements.test(matchString) && node.textContent.length < shareElementThreshold;
|
|
});
|
|
});
|
|
this._clean(articleContent, "iframe");
|
|
this._clean(articleContent, "input");
|
|
this._clean(articleContent, "textarea");
|
|
this._clean(articleContent, "select");
|
|
this._clean(articleContent, "button");
|
|
this._cleanHeaders(articleContent);
|
|
this._cleanConditionally(articleContent, "table");
|
|
this._cleanConditionally(articleContent, "ul");
|
|
this._cleanConditionally(articleContent, "div");
|
|
this._replaceNodeTags(this._getAllNodesWithTag(articleContent, ["h1"]), "h2");
|
|
this._removeNodes(this._getAllNodesWithTag(articleContent, ["p"]), function(paragraph) {
|
|
var contentElementCount = this._getAllNodesWithTag(paragraph, [
|
|
"img",
|
|
"embed",
|
|
"object",
|
|
"iframe"
|
|
]).length;
|
|
return contentElementCount === 0 && !this._getInnerText(paragraph, false);
|
|
});
|
|
this._forEachNode(this._getAllNodesWithTag(articleContent, ["br"]), function(br) {
|
|
var next = this._nextNode(br.nextSibling);
|
|
if (next && next.tagName == "P") {
|
|
br.remove();
|
|
}
|
|
});
|
|
this._forEachNode(this._getAllNodesWithTag(articleContent, ["table"]), function(table) {
|
|
var tbody = this._hasSingleTagInsideElement(table, "TBODY") ? table.firstElementChild : table;
|
|
if (this._hasSingleTagInsideElement(tbody, "TR")) {
|
|
var row = tbody.firstElementChild;
|
|
if (this._hasSingleTagInsideElement(row, "TD")) {
|
|
var cell = row.firstElementChild;
|
|
cell = this._setNodeTag(cell, this._everyNode(cell.childNodes, this._isPhrasingContent) ? "P" : "DIV");
|
|
table.parentNode.replaceChild(cell, table);
|
|
}
|
|
}
|
|
});
|
|
},
|
|
_initializeNode(node) {
|
|
node.readability = { contentScore: 0 };
|
|
switch (node.tagName) {
|
|
case "DIV":
|
|
node.readability.contentScore += 5;
|
|
break;
|
|
case "PRE":
|
|
case "TD":
|
|
case "BLOCKQUOTE":
|
|
node.readability.contentScore += 3;
|
|
break;
|
|
case "ADDRESS":
|
|
case "OL":
|
|
case "UL":
|
|
case "DL":
|
|
case "DD":
|
|
case "DT":
|
|
case "LI":
|
|
case "FORM":
|
|
node.readability.contentScore -= 3;
|
|
break;
|
|
case "H1":
|
|
case "H2":
|
|
case "H3":
|
|
case "H4":
|
|
case "H5":
|
|
case "H6":
|
|
case "TH":
|
|
node.readability.contentScore -= 5;
|
|
break;
|
|
}
|
|
node.readability.contentScore += this._getClassWeight(node);
|
|
},
|
|
_removeAndGetNext(node) {
|
|
var nextNode = this._getNextNode(node, true);
|
|
node.remove();
|
|
return nextNode;
|
|
},
|
|
_getNextNode(node, ignoreSelfAndKids) {
|
|
if (!ignoreSelfAndKids && node.firstElementChild) {
|
|
return node.firstElementChild;
|
|
}
|
|
if (node.nextElementSibling) {
|
|
return node.nextElementSibling;
|
|
}
|
|
do {
|
|
node = node.parentNode;
|
|
} while (node && !node.nextElementSibling);
|
|
return node && node.nextElementSibling;
|
|
},
|
|
_textSimilarity(textA, textB) {
|
|
var tokensA = textA.toLowerCase().split(this.REGEXPS.tokenize).filter(Boolean);
|
|
var tokensB = textB.toLowerCase().split(this.REGEXPS.tokenize).filter(Boolean);
|
|
if (!tokensA.length || !tokensB.length) {
|
|
return 0;
|
|
}
|
|
var uniqTokensB = tokensB.filter((token) => !tokensA.includes(token));
|
|
var distanceB = uniqTokensB.join(" ").length / tokensB.join(" ").length;
|
|
return 1 - distanceB;
|
|
},
|
|
_isValidByline(node, matchString) {
|
|
var rel = node.getAttribute("rel");
|
|
var itemprop = node.getAttribute("itemprop");
|
|
var bylineLength = node.textContent.trim().length;
|
|
return (rel === "author" || itemprop && itemprop.includes("author") || this.REGEXPS.byline.test(matchString)) && !!bylineLength && bylineLength < 100;
|
|
},
|
|
_getNodeAncestors(node, maxDepth) {
|
|
maxDepth = maxDepth || 0;
|
|
var i = 0, ancestors = [];
|
|
while (node.parentNode) {
|
|
ancestors.push(node.parentNode);
|
|
if (maxDepth && ++i === maxDepth) {
|
|
break;
|
|
}
|
|
node = node.parentNode;
|
|
}
|
|
return ancestors;
|
|
},
|
|
_grabArticle(page) {
|
|
this.log("**** grabArticle ****");
|
|
var doc = this._doc;
|
|
var isPaging = page !== null;
|
|
page = page ? page : this._doc.body;
|
|
if (!page) {
|
|
this.log("No body found in document. Abort.");
|
|
return null;
|
|
}
|
|
var pageCacheHtml = page.innerHTML;
|
|
while (true) {
|
|
this.log("Starting grabArticle loop");
|
|
var stripUnlikelyCandidates = this._flagIsActive(this.FLAG_STRIP_UNLIKELYS);
|
|
var elementsToScore = [];
|
|
var node = this._doc.documentElement;
|
|
let shouldRemoveTitleHeader = true;
|
|
while (node) {
|
|
if (node.tagName === "HTML") {
|
|
this._articleLang = node.getAttribute("lang");
|
|
}
|
|
var matchString = node.className + " " + node.id;
|
|
if (!this._isProbablyVisible(node)) {
|
|
this.log("Removing hidden node - " + matchString);
|
|
node = this._removeAndGetNext(node);
|
|
continue;
|
|
}
|
|
if (node.getAttribute("aria-modal") == "true" && node.getAttribute("role") == "dialog") {
|
|
node = this._removeAndGetNext(node);
|
|
continue;
|
|
}
|
|
if (!this._articleByline && !this._metadata.byline && this._isValidByline(node, matchString)) {
|
|
var endOfSearchMarkerNode = this._getNextNode(node, true);
|
|
var next = this._getNextNode(node);
|
|
var itemPropNameNode = null;
|
|
while (next && next != endOfSearchMarkerNode) {
|
|
var itemprop = next.getAttribute("itemprop");
|
|
if (itemprop && itemprop.includes("name")) {
|
|
itemPropNameNode = next;
|
|
break;
|
|
} else {
|
|
next = this._getNextNode(next);
|
|
}
|
|
}
|
|
this._articleByline = (itemPropNameNode ?? node).textContent.trim();
|
|
node = this._removeAndGetNext(node);
|
|
continue;
|
|
}
|
|
if (shouldRemoveTitleHeader && this._headerDuplicatesTitle(node)) {
|
|
this.log("Removing header: ", node.textContent.trim(), this._articleTitle.trim());
|
|
shouldRemoveTitleHeader = false;
|
|
node = this._removeAndGetNext(node);
|
|
continue;
|
|
}
|
|
if (stripUnlikelyCandidates) {
|
|
if (this.REGEXPS.unlikelyCandidates.test(matchString) && !this.REGEXPS.okMaybeItsACandidate.test(matchString) && !this._hasAncestorTag(node, "table") && !this._hasAncestorTag(node, "code") && node.tagName !== "BODY" && node.tagName !== "A") {
|
|
this.log("Removing unlikely candidate - " + matchString);
|
|
node = this._removeAndGetNext(node);
|
|
continue;
|
|
}
|
|
if (this.UNLIKELY_ROLES.includes(node.getAttribute("role"))) {
|
|
this.log("Removing content with role " + node.getAttribute("role") + " - " + matchString);
|
|
node = this._removeAndGetNext(node);
|
|
continue;
|
|
}
|
|
}
|
|
if ((node.tagName === "DIV" || node.tagName === "SECTION" || node.tagName === "HEADER" || node.tagName === "H1" || node.tagName === "H2" || node.tagName === "H3" || node.tagName === "H4" || node.tagName === "H5" || node.tagName === "H6") && this._isElementWithoutContent(node)) {
|
|
node = this._removeAndGetNext(node);
|
|
continue;
|
|
}
|
|
if (this.DEFAULT_TAGS_TO_SCORE.includes(node.tagName)) {
|
|
elementsToScore.push(node);
|
|
}
|
|
if (node.tagName === "DIV") {
|
|
var p = null;
|
|
var childNode = node.firstChild;
|
|
while (childNode) {
|
|
var nextSibling = childNode.nextSibling;
|
|
if (this._isPhrasingContent(childNode)) {
|
|
if (p !== null) {
|
|
p.appendChild(childNode);
|
|
} else if (!this._isWhitespace(childNode)) {
|
|
p = doc.createElement("p");
|
|
node.replaceChild(p, childNode);
|
|
p.appendChild(childNode);
|
|
}
|
|
} else if (p !== null) {
|
|
while (p.lastChild && this._isWhitespace(p.lastChild)) {
|
|
p.lastChild.remove();
|
|
}
|
|
p = null;
|
|
}
|
|
childNode = nextSibling;
|
|
}
|
|
if (this._hasSingleTagInsideElement(node, "P") && this._getLinkDensity(node) < 0.25) {
|
|
var newNode = node.children[0];
|
|
node.parentNode.replaceChild(newNode, node);
|
|
node = newNode;
|
|
elementsToScore.push(node);
|
|
} else if (!this._hasChildBlockElement(node)) {
|
|
node = this._setNodeTag(node, "P");
|
|
elementsToScore.push(node);
|
|
}
|
|
}
|
|
node = this._getNextNode(node);
|
|
}
|
|
var candidates = [];
|
|
this._forEachNode(elementsToScore, function(elementToScore) {
|
|
if (!elementToScore.parentNode || typeof elementToScore.parentNode.tagName === "undefined") {
|
|
return;
|
|
}
|
|
var innerText = this._getInnerText(elementToScore);
|
|
if (innerText.length < 25) {
|
|
return;
|
|
}
|
|
var ancestors2 = this._getNodeAncestors(elementToScore, 5);
|
|
if (ancestors2.length === 0) {
|
|
return;
|
|
}
|
|
var contentScore = 0;
|
|
contentScore += 1;
|
|
contentScore += innerText.split(this.REGEXPS.commas).length;
|
|
contentScore += Math.min(Math.floor(innerText.length / 100), 3);
|
|
this._forEachNode(ancestors2, function(ancestor, level) {
|
|
if (!ancestor.tagName || !ancestor.parentNode || typeof ancestor.parentNode.tagName === "undefined") {
|
|
return;
|
|
}
|
|
if (typeof ancestor.readability === "undefined") {
|
|
this._initializeNode(ancestor);
|
|
candidates.push(ancestor);
|
|
}
|
|
if (level === 0) {
|
|
var scoreDivider = 1;
|
|
} else if (level === 1) {
|
|
scoreDivider = 2;
|
|
} else {
|
|
scoreDivider = level * 3;
|
|
}
|
|
ancestor.readability.contentScore += contentScore / scoreDivider;
|
|
});
|
|
});
|
|
var topCandidates = [];
|
|
for (var c = 0, cl = candidates.length;c < cl; c += 1) {
|
|
var candidate = candidates[c];
|
|
var candidateScore = candidate.readability.contentScore * (1 - this._getLinkDensity(candidate));
|
|
candidate.readability.contentScore = candidateScore;
|
|
this.log("Candidate:", candidate, "with score " + candidateScore);
|
|
for (var t = 0;t < this._nbTopCandidates; t++) {
|
|
var aTopCandidate = topCandidates[t];
|
|
if (!aTopCandidate || candidateScore > aTopCandidate.readability.contentScore) {
|
|
topCandidates.splice(t, 0, candidate);
|
|
if (topCandidates.length > this._nbTopCandidates) {
|
|
topCandidates.pop();
|
|
}
|
|
break;
|
|
}
|
|
}
|
|
}
|
|
var topCandidate = topCandidates[0] || null;
|
|
var neededToCreateTopCandidate = false;
|
|
var parentOfTopCandidate;
|
|
if (topCandidate === null || topCandidate.tagName === "BODY") {
|
|
topCandidate = doc.createElement("DIV");
|
|
neededToCreateTopCandidate = true;
|
|
while (page.firstChild) {
|
|
this.log("Moving child out:", page.firstChild);
|
|
topCandidate.appendChild(page.firstChild);
|
|
}
|
|
page.appendChild(topCandidate);
|
|
this._initializeNode(topCandidate);
|
|
} else if (topCandidate) {
|
|
var alternativeCandidateAncestors = [];
|
|
for (var i = 1;i < topCandidates.length; i++) {
|
|
if (topCandidates[i].readability.contentScore / topCandidate.readability.contentScore >= 0.75) {
|
|
alternativeCandidateAncestors.push(this._getNodeAncestors(topCandidates[i]));
|
|
}
|
|
}
|
|
var MINIMUM_TOPCANDIDATES = 3;
|
|
if (alternativeCandidateAncestors.length >= MINIMUM_TOPCANDIDATES) {
|
|
parentOfTopCandidate = topCandidate.parentNode;
|
|
while (parentOfTopCandidate.tagName !== "BODY") {
|
|
var listsContainingThisAncestor = 0;
|
|
for (var ancestorIndex = 0;ancestorIndex < alternativeCandidateAncestors.length && listsContainingThisAncestor < MINIMUM_TOPCANDIDATES; ancestorIndex++) {
|
|
listsContainingThisAncestor += Number(alternativeCandidateAncestors[ancestorIndex].includes(parentOfTopCandidate));
|
|
}
|
|
if (listsContainingThisAncestor >= MINIMUM_TOPCANDIDATES) {
|
|
topCandidate = parentOfTopCandidate;
|
|
break;
|
|
}
|
|
parentOfTopCandidate = parentOfTopCandidate.parentNode;
|
|
}
|
|
}
|
|
if (!topCandidate.readability) {
|
|
this._initializeNode(topCandidate);
|
|
}
|
|
parentOfTopCandidate = topCandidate.parentNode;
|
|
var lastScore = topCandidate.readability.contentScore;
|
|
var scoreThreshold = lastScore / 3;
|
|
while (parentOfTopCandidate.tagName !== "BODY") {
|
|
if (!parentOfTopCandidate.readability) {
|
|
parentOfTopCandidate = parentOfTopCandidate.parentNode;
|
|
continue;
|
|
}
|
|
var parentScore = parentOfTopCandidate.readability.contentScore;
|
|
if (parentScore < scoreThreshold) {
|
|
break;
|
|
}
|
|
if (parentScore > lastScore) {
|
|
topCandidate = parentOfTopCandidate;
|
|
break;
|
|
}
|
|
lastScore = parentOfTopCandidate.readability.contentScore;
|
|
parentOfTopCandidate = parentOfTopCandidate.parentNode;
|
|
}
|
|
parentOfTopCandidate = topCandidate.parentNode;
|
|
while (parentOfTopCandidate.tagName != "BODY" && parentOfTopCandidate.children.length == 1) {
|
|
topCandidate = parentOfTopCandidate;
|
|
parentOfTopCandidate = topCandidate.parentNode;
|
|
}
|
|
if (!topCandidate.readability) {
|
|
this._initializeNode(topCandidate);
|
|
}
|
|
}
|
|
var articleContent = doc.createElement("DIV");
|
|
if (isPaging) {
|
|
articleContent.id = "readability-content";
|
|
}
|
|
var siblingScoreThreshold = Math.max(10, topCandidate.readability.contentScore * 0.2);
|
|
parentOfTopCandidate = topCandidate.parentNode;
|
|
var siblings = parentOfTopCandidate.children;
|
|
for (var s = 0, sl = siblings.length;s < sl; s++) {
|
|
var sibling = siblings[s];
|
|
var append = false;
|
|
this.log("Looking at sibling node:", sibling, sibling.readability ? "with score " + sibling.readability.contentScore : "");
|
|
this.log("Sibling has score", sibling.readability ? sibling.readability.contentScore : "Unknown");
|
|
if (sibling === topCandidate) {
|
|
append = true;
|
|
} else {
|
|
var contentBonus = 0;
|
|
if (sibling.className === topCandidate.className && topCandidate.className !== "") {
|
|
contentBonus += topCandidate.readability.contentScore * 0.2;
|
|
}
|
|
if (sibling.readability && sibling.readability.contentScore + contentBonus >= siblingScoreThreshold) {
|
|
append = true;
|
|
} else if (sibling.nodeName === "P") {
|
|
var linkDensity = this._getLinkDensity(sibling);
|
|
var nodeContent = this._getInnerText(sibling);
|
|
var nodeLength = nodeContent.length;
|
|
if (nodeLength > 80 && linkDensity < 0.25) {
|
|
append = true;
|
|
} else if (nodeLength < 80 && nodeLength > 0 && linkDensity === 0 && nodeContent.search(/\.( |$)/) !== -1) {
|
|
append = true;
|
|
}
|
|
}
|
|
}
|
|
if (append) {
|
|
this.log("Appending node:", sibling);
|
|
if (!this.ALTER_TO_DIV_EXCEPTIONS.includes(sibling.nodeName)) {
|
|
this.log("Altering sibling:", sibling, "to div.");
|
|
sibling = this._setNodeTag(sibling, "DIV");
|
|
}
|
|
articleContent.appendChild(sibling);
|
|
siblings = parentOfTopCandidate.children;
|
|
s -= 1;
|
|
sl -= 1;
|
|
}
|
|
}
|
|
if (this._debug) {
|
|
this.log("Article content pre-prep: " + articleContent.innerHTML);
|
|
}
|
|
this._prepArticle(articleContent);
|
|
if (this._debug) {
|
|
this.log("Article content post-prep: " + articleContent.innerHTML);
|
|
}
|
|
if (neededToCreateTopCandidate) {
|
|
topCandidate.id = "readability-page-1";
|
|
topCandidate.className = "page";
|
|
} else {
|
|
var div = doc.createElement("DIV");
|
|
div.id = "readability-page-1";
|
|
div.className = "page";
|
|
while (articleContent.firstChild) {
|
|
div.appendChild(articleContent.firstChild);
|
|
}
|
|
articleContent.appendChild(div);
|
|
}
|
|
if (this._debug) {
|
|
this.log("Article content after paging: " + articleContent.innerHTML);
|
|
}
|
|
var parseSuccessful = true;
|
|
var textLength = this._getInnerText(articleContent, true).length;
|
|
if (textLength < this._charThreshold) {
|
|
parseSuccessful = false;
|
|
page.innerHTML = pageCacheHtml;
|
|
this._attempts.push({
|
|
articleContent,
|
|
textLength
|
|
});
|
|
if (this._flagIsActive(this.FLAG_STRIP_UNLIKELYS)) {
|
|
this._removeFlag(this.FLAG_STRIP_UNLIKELYS);
|
|
} else if (this._flagIsActive(this.FLAG_WEIGHT_CLASSES)) {
|
|
this._removeFlag(this.FLAG_WEIGHT_CLASSES);
|
|
} else if (this._flagIsActive(this.FLAG_CLEAN_CONDITIONALLY)) {
|
|
this._removeFlag(this.FLAG_CLEAN_CONDITIONALLY);
|
|
} else {
|
|
this._attempts.sort(function(a, b) {
|
|
return b.textLength - a.textLength;
|
|
});
|
|
if (!this._attempts[0].textLength) {
|
|
return null;
|
|
}
|
|
articleContent = this._attempts[0].articleContent;
|
|
parseSuccessful = true;
|
|
}
|
|
}
|
|
if (parseSuccessful) {
|
|
var ancestors = [parentOfTopCandidate, topCandidate].concat(this._getNodeAncestors(parentOfTopCandidate));
|
|
this._someNode(ancestors, function(ancestor) {
|
|
if (!ancestor.tagName) {
|
|
return false;
|
|
}
|
|
var articleDir = ancestor.getAttribute("dir");
|
|
if (articleDir) {
|
|
this._articleDir = articleDir;
|
|
return true;
|
|
}
|
|
return false;
|
|
});
|
|
return articleContent;
|
|
}
|
|
}
|
|
},
|
|
_unescapeHtmlEntities(str) {
|
|
if (!str) {
|
|
return str;
|
|
}
|
|
var htmlEscapeMap = this.HTML_ESCAPE_MAP;
|
|
return str.replace(/&(quot|amp|apos|lt|gt);/g, function(_, tag) {
|
|
return htmlEscapeMap[tag];
|
|
}).replace(/&#(?:x([0-9a-f]+)|([0-9]+));/gi, function(_, hex, numStr) {
|
|
var num = parseInt(hex || numStr, hex ? 16 : 10);
|
|
if (num == 0 || num > 1114111 || num >= 55296 && num <= 57343) {
|
|
num = 65533;
|
|
}
|
|
return String.fromCodePoint(num);
|
|
});
|
|
},
|
|
_getJSONLD(doc) {
|
|
var scripts = this._getAllNodesWithTag(doc, ["script"]);
|
|
var metadata;
|
|
this._forEachNode(scripts, function(jsonLdElement) {
|
|
if (!metadata && jsonLdElement.getAttribute("type") === "application/ld+json") {
|
|
try {
|
|
var content = jsonLdElement.textContent.replace(/^\s*<!\[CDATA\[|\]\]>\s*$/g, "");
|
|
var parsed = JSON.parse(content);
|
|
if (Array.isArray(parsed)) {
|
|
parsed = parsed.find((it) => {
|
|
return it["@type"] && it["@type"].match(this.REGEXPS.jsonLdArticleTypes);
|
|
});
|
|
if (!parsed) {
|
|
return;
|
|
}
|
|
}
|
|
var schemaDotOrgRegex = /^https?\:\/\/schema\.org\/?$/;
|
|
var matches = typeof parsed["@context"] === "string" && parsed["@context"].match(schemaDotOrgRegex) || typeof parsed["@context"] === "object" && typeof parsed["@context"]["@vocab"] == "string" && parsed["@context"]["@vocab"].match(schemaDotOrgRegex);
|
|
if (!matches) {
|
|
return;
|
|
}
|
|
if (!parsed["@type"] && Array.isArray(parsed["@graph"])) {
|
|
parsed = parsed["@graph"].find((it) => {
|
|
return (it["@type"] || "").match(this.REGEXPS.jsonLdArticleTypes);
|
|
});
|
|
}
|
|
if (!parsed || !parsed["@type"] || !parsed["@type"].match(this.REGEXPS.jsonLdArticleTypes)) {
|
|
return;
|
|
}
|
|
metadata = {};
|
|
if (typeof parsed.name === "string" && typeof parsed.headline === "string" && parsed.name !== parsed.headline) {
|
|
var title = this._getArticleTitle();
|
|
var nameMatches = this._textSimilarity(parsed.name, title) > 0.75;
|
|
var headlineMatches = this._textSimilarity(parsed.headline, title) > 0.75;
|
|
if (headlineMatches && !nameMatches) {
|
|
metadata.title = parsed.headline;
|
|
} else {
|
|
metadata.title = parsed.name;
|
|
}
|
|
} else if (typeof parsed.name === "string") {
|
|
metadata.title = parsed.name.trim();
|
|
} else if (typeof parsed.headline === "string") {
|
|
metadata.title = parsed.headline.trim();
|
|
}
|
|
if (parsed.author) {
|
|
if (typeof parsed.author.name === "string") {
|
|
metadata.byline = parsed.author.name.trim();
|
|
} else if (Array.isArray(parsed.author) && parsed.author[0] && typeof parsed.author[0].name === "string") {
|
|
metadata.byline = parsed.author.filter(function(author) {
|
|
return author && typeof author.name === "string";
|
|
}).map(function(author) {
|
|
return author.name.trim();
|
|
}).join(", ");
|
|
}
|
|
}
|
|
if (typeof parsed.description === "string") {
|
|
metadata.excerpt = parsed.description.trim();
|
|
}
|
|
if (parsed.publisher && typeof parsed.publisher.name === "string") {
|
|
metadata.siteName = parsed.publisher.name.trim();
|
|
}
|
|
if (typeof parsed.datePublished === "string") {
|
|
metadata.datePublished = parsed.datePublished.trim();
|
|
}
|
|
} catch (err) {
|
|
this.log(err.message);
|
|
}
|
|
}
|
|
});
|
|
return metadata ? metadata : {};
|
|
},
|
|
_getArticleMetadata(jsonld) {
|
|
var metadata = {};
|
|
var values = {};
|
|
var metaElements = this._doc.getElementsByTagName("meta");
|
|
var propertyPattern = /\s*(article|dc|dcterm|og|twitter)\s*:\s*(author|creator|description|published_time|title|site_name)\s*/gi;
|
|
var namePattern = /^\s*(?:(dc|dcterm|og|twitter|parsely|weibo:(article|webpage))\s*[-\.:]\s*)?(author|creator|pub-date|description|title|site_name)\s*$/i;
|
|
this._forEachNode(metaElements, function(element) {
|
|
var elementName = element.getAttribute("name");
|
|
var elementProperty = element.getAttribute("property");
|
|
var content = element.getAttribute("content");
|
|
if (!content) {
|
|
return;
|
|
}
|
|
var matches = null;
|
|
var name = null;
|
|
if (elementProperty) {
|
|
matches = elementProperty.match(propertyPattern);
|
|
if (matches) {
|
|
name = matches[0].toLowerCase().replace(/\s/g, "");
|
|
values[name] = content.trim();
|
|
}
|
|
}
|
|
if (!matches && elementName && namePattern.test(elementName)) {
|
|
name = elementName;
|
|
if (content) {
|
|
name = name.toLowerCase().replace(/\s/g, "").replace(/\./g, ":");
|
|
values[name] = content.trim();
|
|
}
|
|
}
|
|
});
|
|
metadata.title = jsonld.title || values["dc:title"] || values["dcterm:title"] || values["og:title"] || values["weibo:article:title"] || values["weibo:webpage:title"] || values.title || values["twitter:title"] || values["parsely-title"];
|
|
if (!metadata.title) {
|
|
metadata.title = this._getArticleTitle();
|
|
}
|
|
const articleAuthor = typeof values["article:author"] === "string" && !this._isUrl(values["article:author"]) ? values["article:author"] : undefined;
|
|
metadata.byline = jsonld.byline || values["dc:creator"] || values["dcterm:creator"] || values.author || values["parsely-author"] || articleAuthor;
|
|
metadata.excerpt = jsonld.excerpt || values["dc:description"] || values["dcterm:description"] || values["og:description"] || values["weibo:article:description"] || values["weibo:webpage:description"] || values.description || values["twitter:description"];
|
|
metadata.siteName = jsonld.siteName || values["og:site_name"];
|
|
metadata.publishedTime = jsonld.datePublished || values["article:published_time"] || values["parsely-pub-date"] || null;
|
|
metadata.title = this._unescapeHtmlEntities(metadata.title);
|
|
metadata.byline = this._unescapeHtmlEntities(metadata.byline);
|
|
metadata.excerpt = this._unescapeHtmlEntities(metadata.excerpt);
|
|
metadata.siteName = this._unescapeHtmlEntities(metadata.siteName);
|
|
metadata.publishedTime = this._unescapeHtmlEntities(metadata.publishedTime);
|
|
return metadata;
|
|
},
|
|
_isSingleImage(node) {
|
|
while (node) {
|
|
if (node.tagName === "IMG") {
|
|
return true;
|
|
}
|
|
if (node.children.length !== 1 || node.textContent.trim() !== "") {
|
|
return false;
|
|
}
|
|
node = node.children[0];
|
|
}
|
|
return false;
|
|
},
|
|
_unwrapNoscriptImages(doc) {
|
|
var imgs = Array.from(doc.getElementsByTagName("img"));
|
|
this._forEachNode(imgs, function(img) {
|
|
for (var i = 0;i < img.attributes.length; i++) {
|
|
var attr = img.attributes[i];
|
|
switch (attr.name) {
|
|
case "src":
|
|
case "srcset":
|
|
case "data-src":
|
|
case "data-srcset":
|
|
return;
|
|
}
|
|
if (/\.(jpg|jpeg|png|webp)/i.test(attr.value)) {
|
|
return;
|
|
}
|
|
}
|
|
img.remove();
|
|
});
|
|
var noscripts = Array.from(doc.getElementsByTagName("noscript"));
|
|
this._forEachNode(noscripts, function(noscript) {
|
|
if (!this._isSingleImage(noscript)) {
|
|
return;
|
|
}
|
|
var tmp = doc.createElement("div");
|
|
tmp.innerHTML = noscript.innerHTML;
|
|
var prevElement = noscript.previousElementSibling;
|
|
if (prevElement && this._isSingleImage(prevElement)) {
|
|
var prevImg = prevElement;
|
|
if (prevImg.tagName !== "IMG") {
|
|
prevImg = prevElement.getElementsByTagName("img")[0];
|
|
}
|
|
var newImg = tmp.getElementsByTagName("img")[0];
|
|
for (var i = 0;i < prevImg.attributes.length; i++) {
|
|
var attr = prevImg.attributes[i];
|
|
if (attr.value === "") {
|
|
continue;
|
|
}
|
|
if (attr.name === "src" || attr.name === "srcset" || /\.(jpg|jpeg|png|webp)/i.test(attr.value)) {
|
|
if (newImg.getAttribute(attr.name) === attr.value) {
|
|
continue;
|
|
}
|
|
var attrName = attr.name;
|
|
if (newImg.hasAttribute(attrName)) {
|
|
attrName = "data-old-" + attrName;
|
|
}
|
|
newImg.setAttribute(attrName, attr.value);
|
|
}
|
|
}
|
|
noscript.parentNode.replaceChild(tmp.firstElementChild, prevElement);
|
|
}
|
|
});
|
|
},
|
|
_removeScripts(doc) {
|
|
this._removeNodes(this._getAllNodesWithTag(doc, ["script", "noscript"]));
|
|
},
|
|
_hasSingleTagInsideElement(element, tag) {
|
|
if (element.children.length != 1 || element.children[0].tagName !== tag) {
|
|
return false;
|
|
}
|
|
return !this._someNode(element.childNodes, function(node) {
|
|
return node.nodeType === this.TEXT_NODE && this.REGEXPS.hasContent.test(node.textContent);
|
|
});
|
|
},
|
|
_isElementWithoutContent(node) {
|
|
return node.nodeType === this.ELEMENT_NODE && !node.textContent.trim().length && (!node.children.length || node.children.length == node.getElementsByTagName("br").length + node.getElementsByTagName("hr").length);
|
|
},
|
|
_hasChildBlockElement(element) {
|
|
return this._someNode(element.childNodes, function(node) {
|
|
return this.DIV_TO_P_ELEMS.has(node.tagName) || this._hasChildBlockElement(node);
|
|
});
|
|
},
|
|
_isPhrasingContent(node) {
|
|
return node.nodeType === this.TEXT_NODE || this.PHRASING_ELEMS.includes(node.tagName) || (node.tagName === "A" || node.tagName === "DEL" || node.tagName === "INS") && this._everyNode(node.childNodes, this._isPhrasingContent);
|
|
},
|
|
_isWhitespace(node) {
|
|
return node.nodeType === this.TEXT_NODE && node.textContent.trim().length === 0 || node.nodeType === this.ELEMENT_NODE && node.tagName === "BR";
|
|
},
|
|
_getInnerText(e, normalizeSpaces) {
|
|
normalizeSpaces = typeof normalizeSpaces === "undefined" ? true : normalizeSpaces;
|
|
var textContent = e.textContent.trim();
|
|
if (normalizeSpaces) {
|
|
return textContent.replace(this.REGEXPS.normalize, " ");
|
|
}
|
|
return textContent;
|
|
},
|
|
_getCharCount(e, s) {
|
|
s = s || ",";
|
|
return this._getInnerText(e).split(s).length - 1;
|
|
},
|
|
_cleanStyles(e) {
|
|
if (!e || e.tagName.toLowerCase() === "svg") {
|
|
return;
|
|
}
|
|
for (var i = 0;i < this.PRESENTATIONAL_ATTRIBUTES.length; i++) {
|
|
e.removeAttribute(this.PRESENTATIONAL_ATTRIBUTES[i]);
|
|
}
|
|
if (this.DEPRECATED_SIZE_ATTRIBUTE_ELEMS.includes(e.tagName)) {
|
|
e.removeAttribute("width");
|
|
e.removeAttribute("height");
|
|
}
|
|
var cur = e.firstElementChild;
|
|
while (cur !== null) {
|
|
this._cleanStyles(cur);
|
|
cur = cur.nextElementSibling;
|
|
}
|
|
},
|
|
_getLinkDensity(element) {
|
|
var textLength = this._getInnerText(element).length;
|
|
if (textLength === 0) {
|
|
return 0;
|
|
}
|
|
var linkLength = 0;
|
|
this._forEachNode(element.getElementsByTagName("a"), function(linkNode) {
|
|
var href = linkNode.getAttribute("href");
|
|
var coefficient = href && this.REGEXPS.hashUrl.test(href) ? 0.3 : 1;
|
|
linkLength += this._getInnerText(linkNode).length * coefficient;
|
|
});
|
|
return linkLength / textLength;
|
|
},
|
|
_getClassWeight(e) {
|
|
if (!this._flagIsActive(this.FLAG_WEIGHT_CLASSES)) {
|
|
return 0;
|
|
}
|
|
var weight = 0;
|
|
if (typeof e.className === "string" && e.className !== "") {
|
|
if (this.REGEXPS.negative.test(e.className)) {
|
|
weight -= 25;
|
|
}
|
|
if (this.REGEXPS.positive.test(e.className)) {
|
|
weight += 25;
|
|
}
|
|
}
|
|
if (typeof e.id === "string" && e.id !== "") {
|
|
if (this.REGEXPS.negative.test(e.id)) {
|
|
weight -= 25;
|
|
}
|
|
if (this.REGEXPS.positive.test(e.id)) {
|
|
weight += 25;
|
|
}
|
|
}
|
|
return weight;
|
|
},
|
|
_clean(e, tag) {
|
|
var isEmbed = ["object", "embed", "iframe"].includes(tag);
|
|
this._removeNodes(this._getAllNodesWithTag(e, [tag]), function(element) {
|
|
if (isEmbed) {
|
|
for (var i = 0;i < element.attributes.length; i++) {
|
|
if (this._allowedVideoRegex.test(element.attributes[i].value)) {
|
|
return false;
|
|
}
|
|
}
|
|
if (element.tagName === "object" && this._allowedVideoRegex.test(element.innerHTML)) {
|
|
return false;
|
|
}
|
|
}
|
|
return true;
|
|
});
|
|
},
|
|
_hasAncestorTag(node, tagName, maxDepth, filterFn) {
|
|
maxDepth = maxDepth || 3;
|
|
tagName = tagName.toUpperCase();
|
|
var depth = 0;
|
|
while (node.parentNode) {
|
|
if (maxDepth > 0 && depth > maxDepth) {
|
|
return false;
|
|
}
|
|
if (node.parentNode.tagName === tagName && (!filterFn || filterFn(node.parentNode))) {
|
|
return true;
|
|
}
|
|
node = node.parentNode;
|
|
depth++;
|
|
}
|
|
return false;
|
|
},
|
|
_getRowAndColumnCount(table) {
|
|
var rows = 0;
|
|
var columns = 0;
|
|
var trs = table.getElementsByTagName("tr");
|
|
for (var i = 0;i < trs.length; i++) {
|
|
var rowspan = trs[i].getAttribute("rowspan") || 0;
|
|
if (rowspan) {
|
|
rowspan = parseInt(rowspan, 10);
|
|
}
|
|
rows += rowspan || 1;
|
|
var columnsInThisRow = 0;
|
|
var cells = trs[i].getElementsByTagName("td");
|
|
for (var j = 0;j < cells.length; j++) {
|
|
var colspan = cells[j].getAttribute("colspan") || 0;
|
|
if (colspan) {
|
|
colspan = parseInt(colspan, 10);
|
|
}
|
|
columnsInThisRow += colspan || 1;
|
|
}
|
|
columns = Math.max(columns, columnsInThisRow);
|
|
}
|
|
return { rows, columns };
|
|
},
|
|
_markDataTables(root) {
|
|
var tables = root.getElementsByTagName("table");
|
|
for (var i = 0;i < tables.length; i++) {
|
|
var table = tables[i];
|
|
var role = table.getAttribute("role");
|
|
if (role == "presentation") {
|
|
table._readabilityDataTable = false;
|
|
continue;
|
|
}
|
|
var datatable = table.getAttribute("datatable");
|
|
if (datatable == "0") {
|
|
table._readabilityDataTable = false;
|
|
continue;
|
|
}
|
|
var summary = table.getAttribute("summary");
|
|
if (summary) {
|
|
table._readabilityDataTable = true;
|
|
continue;
|
|
}
|
|
var caption = table.getElementsByTagName("caption")[0];
|
|
if (caption && caption.childNodes.length) {
|
|
table._readabilityDataTable = true;
|
|
continue;
|
|
}
|
|
var dataTableDescendants = ["col", "colgroup", "tfoot", "thead", "th"];
|
|
var descendantExists = function(tag) {
|
|
return !!table.getElementsByTagName(tag)[0];
|
|
};
|
|
if (dataTableDescendants.some(descendantExists)) {
|
|
this.log("Data table because found data-y descendant");
|
|
table._readabilityDataTable = true;
|
|
continue;
|
|
}
|
|
if (table.getElementsByTagName("table")[0]) {
|
|
table._readabilityDataTable = false;
|
|
continue;
|
|
}
|
|
var sizeInfo = this._getRowAndColumnCount(table);
|
|
if (sizeInfo.columns == 1 || sizeInfo.rows == 1) {
|
|
table._readabilityDataTable = false;
|
|
continue;
|
|
}
|
|
if (sizeInfo.rows >= 10 || sizeInfo.columns > 4) {
|
|
table._readabilityDataTable = true;
|
|
continue;
|
|
}
|
|
table._readabilityDataTable = sizeInfo.rows * sizeInfo.columns > 10;
|
|
}
|
|
},
|
|
_fixLazyImages(root) {
|
|
this._forEachNode(this._getAllNodesWithTag(root, ["img", "picture", "figure"]), function(elem) {
|
|
if (elem.src && this.REGEXPS.b64DataUrl.test(elem.src)) {
|
|
var parts = this.REGEXPS.b64DataUrl.exec(elem.src);
|
|
if (parts[1] === "image/svg+xml") {
|
|
return;
|
|
}
|
|
var srcCouldBeRemoved = false;
|
|
for (var i = 0;i < elem.attributes.length; i++) {
|
|
var attr = elem.attributes[i];
|
|
if (attr.name === "src") {
|
|
continue;
|
|
}
|
|
if (/\.(jpg|jpeg|png|webp)/i.test(attr.value)) {
|
|
srcCouldBeRemoved = true;
|
|
break;
|
|
}
|
|
}
|
|
if (srcCouldBeRemoved) {
|
|
var b64starts = parts[0].length;
|
|
var b64length = elem.src.length - b64starts;
|
|
if (b64length < 133) {
|
|
elem.removeAttribute("src");
|
|
}
|
|
}
|
|
}
|
|
if ((elem.src || elem.srcset && elem.srcset != "null") && !elem.className.toLowerCase().includes("lazy")) {
|
|
return;
|
|
}
|
|
for (var j = 0;j < elem.attributes.length; j++) {
|
|
attr = elem.attributes[j];
|
|
if (attr.name === "src" || attr.name === "srcset" || attr.name === "alt") {
|
|
continue;
|
|
}
|
|
var copyTo = null;
|
|
if (/\.(jpg|jpeg|png|webp)\s+\d/.test(attr.value)) {
|
|
copyTo = "srcset";
|
|
} else if (/^\s*\S+\.(jpg|jpeg|png|webp)\S*\s*$/.test(attr.value)) {
|
|
copyTo = "src";
|
|
}
|
|
if (copyTo) {
|
|
if (elem.tagName === "IMG" || elem.tagName === "PICTURE") {
|
|
elem.setAttribute(copyTo, attr.value);
|
|
} else if (elem.tagName === "FIGURE" && !this._getAllNodesWithTag(elem, ["img", "picture"]).length) {
|
|
var img = this._doc.createElement("img");
|
|
img.setAttribute(copyTo, attr.value);
|
|
elem.appendChild(img);
|
|
}
|
|
}
|
|
}
|
|
});
|
|
},
|
|
_getTextDensity(e, tags) {
|
|
var textLength = this._getInnerText(e, true).length;
|
|
if (textLength === 0) {
|
|
return 0;
|
|
}
|
|
var childrenLength = 0;
|
|
var children = this._getAllNodesWithTag(e, tags);
|
|
this._forEachNode(children, (child) => childrenLength += this._getInnerText(child, true).length);
|
|
return childrenLength / textLength;
|
|
},
|
|
_cleanConditionally(e, tag) {
|
|
if (!this._flagIsActive(this.FLAG_CLEAN_CONDITIONALLY)) {
|
|
return;
|
|
}
|
|
this._removeNodes(this._getAllNodesWithTag(e, [tag]), function(node) {
|
|
var isDataTable = function(t) {
|
|
return t._readabilityDataTable;
|
|
};
|
|
var isList = tag === "ul" || tag === "ol";
|
|
if (!isList) {
|
|
var listLength = 0;
|
|
var listNodes = this._getAllNodesWithTag(node, ["ul", "ol"]);
|
|
this._forEachNode(listNodes, (list) => listLength += this._getInnerText(list).length);
|
|
isList = listLength / this._getInnerText(node).length > 0.9;
|
|
}
|
|
if (tag === "table" && isDataTable(node)) {
|
|
return false;
|
|
}
|
|
if (this._hasAncestorTag(node, "table", -1, isDataTable)) {
|
|
return false;
|
|
}
|
|
if (this._hasAncestorTag(node, "code")) {
|
|
return false;
|
|
}
|
|
if ([...node.getElementsByTagName("table")].some((tbl) => tbl._readabilityDataTable)) {
|
|
return false;
|
|
}
|
|
var weight = this._getClassWeight(node);
|
|
this.log("Cleaning Conditionally", node);
|
|
var contentScore = 0;
|
|
if (weight + contentScore < 0) {
|
|
return true;
|
|
}
|
|
if (this._getCharCount(node, ",") < 10) {
|
|
var p = node.getElementsByTagName("p").length;
|
|
var img = node.getElementsByTagName("img").length;
|
|
var li = node.getElementsByTagName("li").length - 100;
|
|
var input = node.getElementsByTagName("input").length;
|
|
var headingDensity = this._getTextDensity(node, [
|
|
"h1",
|
|
"h2",
|
|
"h3",
|
|
"h4",
|
|
"h5",
|
|
"h6"
|
|
]);
|
|
var embedCount = 0;
|
|
var embeds = this._getAllNodesWithTag(node, [
|
|
"object",
|
|
"embed",
|
|
"iframe"
|
|
]);
|
|
for (var i = 0;i < embeds.length; i++) {
|
|
for (var j = 0;j < embeds[i].attributes.length; j++) {
|
|
if (this._allowedVideoRegex.test(embeds[i].attributes[j].value)) {
|
|
return false;
|
|
}
|
|
}
|
|
if (embeds[i].tagName === "object" && this._allowedVideoRegex.test(embeds[i].innerHTML)) {
|
|
return false;
|
|
}
|
|
embedCount++;
|
|
}
|
|
var innerText = this._getInnerText(node);
|
|
if (this.REGEXPS.adWords.test(innerText) || this.REGEXPS.loadingWords.test(innerText)) {
|
|
return true;
|
|
}
|
|
var contentLength = innerText.length;
|
|
var linkDensity = this._getLinkDensity(node);
|
|
var textishTags = ["SPAN", "LI", "TD"].concat(Array.from(this.DIV_TO_P_ELEMS));
|
|
var textDensity = this._getTextDensity(node, textishTags);
|
|
var isFigureChild = this._hasAncestorTag(node, "figure");
|
|
const shouldRemoveNode = () => {
|
|
const errs = [];
|
|
if (!isFigureChild && img > 1 && p / img < 0.5) {
|
|
errs.push(`Bad p to img ratio (img=${img}, p=${p})`);
|
|
}
|
|
if (!isList && li > p) {
|
|
errs.push(`Too many li's outside of a list. (li=${li} > p=${p})`);
|
|
}
|
|
if (input > Math.floor(p / 3)) {
|
|
errs.push(`Too many inputs per p. (input=${input}, p=${p})`);
|
|
}
|
|
if (!isList && !isFigureChild && headingDensity < 0.9 && contentLength < 25 && (img === 0 || img > 2) && linkDensity > 0) {
|
|
errs.push(`Suspiciously short. (headingDensity=${headingDensity}, img=${img}, linkDensity=${linkDensity})`);
|
|
}
|
|
if (!isList && weight < 25 && linkDensity > 0.2 + this._linkDensityModifier) {
|
|
errs.push(`Low weight and a little linky. (linkDensity=${linkDensity})`);
|
|
}
|
|
if (weight >= 25 && linkDensity > 0.5 + this._linkDensityModifier) {
|
|
errs.push(`High weight and mostly links. (linkDensity=${linkDensity})`);
|
|
}
|
|
if (embedCount === 1 && contentLength < 75 || embedCount > 1) {
|
|
errs.push(`Suspicious embed. (embedCount=${embedCount}, contentLength=${contentLength})`);
|
|
}
|
|
if (img === 0 && textDensity === 0) {
|
|
errs.push(`No useful content. (img=${img}, textDensity=${textDensity})`);
|
|
}
|
|
if (errs.length) {
|
|
this.log("Checks failed", errs);
|
|
return true;
|
|
}
|
|
return false;
|
|
};
|
|
var haveToRemove = shouldRemoveNode();
|
|
if (isList && haveToRemove) {
|
|
for (var x = 0;x < node.children.length; x++) {
|
|
let child = node.children[x];
|
|
if (child.children.length > 1) {
|
|
return haveToRemove;
|
|
}
|
|
}
|
|
let li_count = node.getElementsByTagName("li").length;
|
|
if (img == li_count) {
|
|
return false;
|
|
}
|
|
}
|
|
return haveToRemove;
|
|
}
|
|
return false;
|
|
});
|
|
},
|
|
_cleanMatchedNodes(e, filter) {
|
|
var endOfSearchMarkerNode = this._getNextNode(e, true);
|
|
var next = this._getNextNode(e);
|
|
while (next && next != endOfSearchMarkerNode) {
|
|
if (filter.call(this, next, next.className + " " + next.id)) {
|
|
next = this._removeAndGetNext(next);
|
|
} else {
|
|
next = this._getNextNode(next);
|
|
}
|
|
}
|
|
},
|
|
_cleanHeaders(e) {
|
|
let headingNodes = this._getAllNodesWithTag(e, ["h1", "h2"]);
|
|
this._removeNodes(headingNodes, function(node) {
|
|
let shouldRemove = this._getClassWeight(node) < 0;
|
|
if (shouldRemove) {
|
|
this.log("Removing header with low class weight:", node);
|
|
}
|
|
return shouldRemove;
|
|
});
|
|
},
|
|
_headerDuplicatesTitle(node) {
|
|
if (node.tagName != "H1" && node.tagName != "H2") {
|
|
return false;
|
|
}
|
|
var heading = this._getInnerText(node, false);
|
|
this.log("Evaluating similarity of header:", heading, this._articleTitle);
|
|
return this._textSimilarity(this._articleTitle, heading) > 0.75;
|
|
},
|
|
_flagIsActive(flag) {
|
|
return (this._flags & flag) > 0;
|
|
},
|
|
_removeFlag(flag) {
|
|
this._flags = this._flags & ~flag;
|
|
},
|
|
_isProbablyVisible(node) {
|
|
return (!node.style || node.style.display != "none") && (!node.style || node.style.visibility != "hidden") && !node.hasAttribute("hidden") && (!node.hasAttribute("aria-hidden") || node.getAttribute("aria-hidden") != "true" || node.className && node.className.includes && node.className.includes("fallback-image"));
|
|
},
|
|
parse() {
|
|
if (this._maxElemsToParse > 0) {
|
|
var numTags = this._doc.getElementsByTagName("*").length;
|
|
if (numTags > this._maxElemsToParse) {
|
|
throw new Error("Aborting parsing document; " + numTags + " elements found");
|
|
}
|
|
}
|
|
this._unwrapNoscriptImages(this._doc);
|
|
var jsonLd = this._disableJSONLD ? {} : this._getJSONLD(this._doc);
|
|
this._removeScripts(this._doc);
|
|
this._prepDocument();
|
|
var metadata = this._getArticleMetadata(jsonLd);
|
|
this._metadata = metadata;
|
|
this._articleTitle = metadata.title;
|
|
var articleContent = this._grabArticle();
|
|
if (!articleContent) {
|
|
return null;
|
|
}
|
|
this.log("Grabbed: " + articleContent.innerHTML);
|
|
this._postProcessContent(articleContent);
|
|
if (!metadata.excerpt) {
|
|
var paragraphs = articleContent.getElementsByTagName("p");
|
|
if (paragraphs.length) {
|
|
metadata.excerpt = paragraphs[0].textContent.trim();
|
|
}
|
|
}
|
|
var textContent = articleContent.textContent;
|
|
return {
|
|
title: this._articleTitle,
|
|
byline: metadata.byline || this._articleByline,
|
|
dir: this._articleDir,
|
|
lang: this._articleLang,
|
|
content: this._serializer(articleContent),
|
|
textContent,
|
|
length: textContent.length,
|
|
excerpt: metadata.excerpt,
|
|
siteName: metadata.siteName || this._articleSiteName,
|
|
publishedTime: metadata.publishedTime
|
|
};
|
|
}
|
|
};
|
|
if (typeof module === "object") {
|
|
module.exports = Readability;
|
|
}
|
|
});
|
|
|
|
// ../node_modules/.pnpm/@mozilla+readability@0.6.0/node_modules/@mozilla/readability/Readability-readerable.js
|
|
var require_Readability_readerable = __commonJS((exports, module) => {
|
|
var REGEXPS = {
|
|
unlikelyCandidates: /-ad-|ai2html|banner|breadcrumbs|combx|comment|community|cover-wrap|disqus|extra|footer|gdpr|header|legends|menu|related|remark|replies|rss|shoutbox|sidebar|skyscraper|social|sponsor|supplemental|ad-break|agegate|pagination|pager|popup|yom-remote/i,
|
|
okMaybeItsACandidate: /and|article|body|column|content|main|shadow/i
|
|
};
|
|
function isNodeVisible(node) {
|
|
return (!node.style || node.style.display != "none") && !node.hasAttribute("hidden") && (!node.hasAttribute("aria-hidden") || node.getAttribute("aria-hidden") != "true" || node.className && node.className.includes && node.className.includes("fallback-image"));
|
|
}
|
|
function isProbablyReaderable(doc, options = {}) {
|
|
if (typeof options == "function") {
|
|
options = { visibilityChecker: options };
|
|
}
|
|
var defaultOptions = {
|
|
minScore: 20,
|
|
minContentLength: 140,
|
|
visibilityChecker: isNodeVisible
|
|
};
|
|
options = Object.assign(defaultOptions, options);
|
|
var nodes = doc.querySelectorAll("p, pre, article");
|
|
var brNodes = doc.querySelectorAll("div > br");
|
|
if (brNodes.length) {
|
|
var set = new Set(nodes);
|
|
[].forEach.call(brNodes, function(node) {
|
|
set.add(node.parentNode);
|
|
});
|
|
nodes = Array.from(set);
|
|
}
|
|
var score = 0;
|
|
return [].some.call(nodes, function(node) {
|
|
if (!options.visibilityChecker(node)) {
|
|
return false;
|
|
}
|
|
var matchString = node.className + " " + node.id;
|
|
if (REGEXPS.unlikelyCandidates.test(matchString) && !REGEXPS.okMaybeItsACandidate.test(matchString)) {
|
|
return false;
|
|
}
|
|
if (node.matches("li p")) {
|
|
return false;
|
|
}
|
|
var textContentLength = node.textContent.trim().length;
|
|
if (textContentLength < options.minContentLength) {
|
|
return false;
|
|
}
|
|
score += Math.sqrt(textContentLength - options.minContentLength);
|
|
if (score > options.minScore) {
|
|
return true;
|
|
}
|
|
return false;
|
|
});
|
|
}
|
|
if (typeof module === "object") {
|
|
module.exports = isProbablyReaderable;
|
|
}
|
|
});
|
|
|
|
// ../node_modules/.pnpm/@mozilla+readability@0.6.0/node_modules/@mozilla/readability/index.js
|
|
var require_readability = __commonJS((exports, module) => {
|
|
var Readability = require_Readability();
|
|
var isProbablyReaderable = require_Readability_readerable();
|
|
module.exports = {
|
|
Readability,
|
|
isProbablyReaderable
|
|
};
|
|
});
|
|
|
|
// dist/_readability-entry-46963-1772299414460.js
|
|
var import_readability = __toESM(require_readability(), 1);
|
|
globalThis.__readability = { Readability: import_readability.Readability, isProbablyReaderable: import_readability.isProbablyReaderable };
|
|
})();
|