diff options
Diffstat (limited to 'plugins/readability')
| -rw-r--r-- | plugins/readability/README.md | 2 | ||||
| -rwxr-xr-x | plugins/readability/readability | bin | 0 -> 17056 bytes | |||
| -rw-r--r-- | plugins/readability/readability.c | 29 | ||||
| -rw-r--r-- | plugins/readability/readability.h | 8 | ||||
| -rw-r--r-- | plugins/readability/readability.js | 2340 | ||||
| -rwxr-xr-x | plugins/readability/recompute_READABILITY_N.sh | 8 |
6 files changed, 2387 insertions, 0 deletions
diff --git a/plugins/readability/README.md b/plugins/readability/README.md new file mode 100644 index 0000000..2a19a8e --- /dev/null +++ b/plugins/readability/README.md @@ -0,0 +1,2 @@ +This code reimplements firefox readability mode. Code taken from <https://raw.githubusercontent.com/ushnisha/readability-reader-webextensions/master/content_scripts/tranquilize.js> + diff --git a/plugins/readability/readability b/plugins/readability/readability Binary files differnew file mode 100755 index 0000000..2c70985 --- /dev/null +++ b/plugins/readability/readability diff --git a/plugins/readability/readability.c b/plugins/readability/readability.c new file mode 100644 index 0000000..d7d1ed9 --- /dev/null +++ b/plugins/readability/readability.c @@ -0,0 +1,29 @@ +#include <stdio.h> +#include <string.h> +#include <stdlib.h> +#define READABILITY_N 84251 + 1 + +void read_readability_js(char* string){ + FILE *fp=fopen("/home/loki/Documents/core/software/fresh/C/rose-browser/rose-bud-personal/plugins/readability/readability.js", "r"); + if (!fp) { // fp is NULL, fopen failed + fprintf(stderr, "Failed to open file\n"); + string=NULL; + return; + } + int i=0; + int c; + while ((c = fgetc(fp)) != EOF){ + string[i++] = c; + } + string[i]='\0'; + fclose(fp); +} + +/* +int main(){ + char* readability_js = malloc(READABILITY_N+1); + read_readability_js(readability_js); + printf("%s", readability_js); + free(readability_js); +} +*/ diff --git a/plugins/readability/readability.h b/plugins/readability/readability.h new file mode 100644 index 0000000..5fe0628 --- /dev/null +++ b/plugins/readability/readability.h @@ -0,0 +1,8 @@ +#ifndef READABILITY +#define READABILITY + +#define READABILITY_N 84251 + 1 + +void read_readability_js(char* string); + +#endif diff --git a/plugins/readability/readability.js b/plugins/readability/readability.js new file mode 100644 index 0000000..21f3e3a --- /dev/null +++ b/plugins/readability/readability.js @@ -0,0 +1,2340 @@ +/*eslint-env es6:false*/ +/* + * Copyright (c) 2010 Arc90 Inc + * + * Licensed under the Apache License, Version 2.0 (the "License"); + * you may not use this file except in compliance with the License. + * You may obtain a copy of the License at + * + * http://www.apache.org/licenses/LICENSE-2.0 + * + * Unless required by applicable law or agreed to in writing, software + * distributed under the License is distributed on an "AS IS" BASIS, + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + * See the License for the specific language governing permissions and + * limitations under the License. + */ + +/* + * This code is heavily based on Arc90's readability.js (1.7.1) script + * available at: http://code.google.com/p/arc90labs-readability + */ + +/** + * Public constructor. + * @param {HTMLDocument} doc The document to parse. + * @param {Object} options The options object. + */ +function Readability(doc, options) { + // In some older versions, people passed a URI as the first argument. Cope: + 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 = []; + + // Configurable options + 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; + + // Start with all flags set + this._flags = this.FLAG_STRIP_UNLIKELYS | + this.FLAG_WEIGHT_CLASSES | + this.FLAG_CLEAN_CONDITIONALLY; + + + // Control whether log messages are sent to the console + 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 dump !== "undefined") { + var msg = Array.prototype.map.call(arguments, function(x) { + return (x && x.nodeName) ? logNode(x) : x; + }).join(" "); + dump("Reader: (Readability) " + msg + "\n"); + } else 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.apply(console, args); + } + }; + } else { + this.log = function () {}; + } +} + +Readability.prototype = { + FLAG_STRIP_UNLIKELYS: 0x1, + FLAG_WEIGHT_CLASSES: 0x2, + FLAG_CLEAN_CONDITIONALLY: 0x4, + + // https://developer.mozilla.org/en-US/docs/Web/API/Node/nodeType + ELEMENT_NODE: 1, + TEXT_NODE: 3, + + // Max number of nodes supported by this parser. Default: 0 (no limit) + DEFAULT_MAX_ELEMS_TO_PARSE: 0, + + // The number of top candidates to consider when analysing how + // tight the competition is among candidates. + DEFAULT_N_TOP_CANDIDATES: 5, + + // Element tags to score by default. + DEFAULT_TAGS_TO_SCORE: "section,h2,h3,h4,h5,h6,p,td,pre".toUpperCase().split(","), + + // The default number of chars an article must have in order to return a result + DEFAULT_CHAR_THRESHOLD: 500, + + // All of the regular expressions in use within readability. + // Defined up here so we don't instantiate them repeatedly in loops. + REGEXPS: { + // NOTE: These two regular expressions are duplicated in + // Readability-readerable.js. Please keep both copies in sync. + 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$| hid |^hid |banner|combx|comment|com-|contact|foot|footer|footnote|gdpr|masthead|media|meta|outbrain|promo|related|scroll|share|shoutbox|sidebar|skyscraper|sponsor|shopping|tags|tool|widget/i, + extraneous: /print|archive|comment|discuss|e[\-]?mail|share|reply|all|login|sign|single|utility/i, + byline: /byline|author|dateline|writtenby|p-author/i, + replaceFonts: /<(\/?)font[^>]*>/gi, + normalize: /\s{2,}/g, + videos: /\/\/(www\.)?((dailymotion|youtube|youtube-nocookie|player\.vimeo|v\.qq)\.com|(archive|upload\.wikimedia)\.org|player\.twitch\.tv)/i, + shareElements: /(\b|_)(share|sharedaddy)(\b|_)/i, + nextLink: /(next|weiter|continue|>([^\|]|$)|»([^\|]|$))/i, + prevLink: /(prev|earl|old|new|<|«)/i, + tokenize: /\W+/g, + whitespace: /^\s*$/, + hasContent: /\S$/, + hashUrl: /^#.+/, + srcsetUrl: /(\S+)(\s+[\d.]+[xw])?(\s*(?:,|$))/g, + b64DataUrl: /^data:\s*([^\s;,]+)\s*;\s*base64\s*,/i, + // See: https://schema.org/Article + jsonLdArticleTypes: /^Article|AdvertiserContentArticle|NewsArticle|AnalysisNewsArticle|AskPublicNewsArticle|BackgroundNewsArticle|OpinionNewsArticle|ReportageNewsArticle|ReviewNewsArticle|Report|SatiricalArticle|ScholarlyArticle|MedicalScholarlyArticle|SocialMediaPosting|BlogPosting|LiveBlogPosting|DiscussionForumPosting|TechArticle|APIReference$/ + }, + + 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"], + + PRESENTATIONAL_ATTRIBUTES: [ "align", "background", "bgcolor", "border", "cellpadding", "cellspacing", "frame", "hspace", "rules", "style", "valign", "vspace" ], + + DEPRECATED_SIZE_ATTRIBUTE_ELEMS: [ "TABLE", "TH", "TD", "HR", "PRE" ], + + // The commented out elements qualify as phrasing content but tend to be + // removed by readability when put into paragraphs, so we ignore them here. + PHRASING_ELEMS: [ + // "CANVAS", "IFRAME", "SVG", "VIDEO", + "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" + ], + + // These are the classes that readability sets itself. + CLASSES_TO_PRESERVE: [ "page" ], + + // These are the list of HTML entities that need to be escaped. + HTML_ESCAPE_MAP: { + "lt": "<", + "gt": ">", + "amp": "&", + "quot": '"', + "apos": "'", + }, + + /** + * Run any post-process modifications to article content as necessary. + * + * @param Element + * @return void + **/ + _postProcessContent: function(articleContent) { + // Readability cannot open relative uris so we convert them to absolute uris. + this._fixRelativeUris(articleContent); + + this._simplifyNestedElements(articleContent); + + if (!this._keepClasses) { + // Remove classes. + this._cleanClasses(articleContent); + } + }, + + /** + * Iterates over a NodeList, calls `filterFn` for each node and removes node + * if function returned `true`. + * + * If function is not passed, removes all the nodes in node list. + * + * @param NodeList nodeList The nodes to operate on + * @param Function filterFn the function to use as a filter + * @return void + */ + _removeNodes: function(nodeList, filterFn) { + // Avoid ever operating on live node lists. + 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); + } + } + } + }, + + /** + * Iterates over a NodeList, and calls _setNodeTag for each node. + * + * @param NodeList nodeList The nodes to operate on + * @param String newTagName the new tag name to use + * @return void + */ + _replaceNodeTags: function(nodeList, newTagName) { + // Avoid ever operating on live node lists. + 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); + } + }, + + /** + * Iterate over a NodeList, which doesn't natively fully implement the Array + * interface. + * + * For convenience, the current object context is applied to the provided + * iterate function. + * + * @param NodeList nodeList The NodeList. + * @param Function fn The iterate function. + * @return void + */ + _forEachNode: function(nodeList, fn) { + Array.prototype.forEach.call(nodeList, fn, this); + }, + + /** + * Iterate over a NodeList, and return the first node that passes + * the supplied test function + * + * For convenience, the current object context is applied to the provided + * test function. + * + * @param NodeList nodeList The NodeList. + * @param Function fn The test function. + * @return void + */ + _findNode: function(nodeList, fn) { + return Array.prototype.find.call(nodeList, fn, this); + }, + + /** + * Iterate over a NodeList, return true if any of the provided iterate + * function calls returns true, false otherwise. + * + * For convenience, the current object context is applied to the + * provided iterate function. + * + * @param NodeList nodeList The NodeList. + * @param Function fn The iterate function. + * @return Boolean + */ + _someNode: function(nodeList, fn) { + return Array.prototype.some.call(nodeList, fn, this); + }, + + /** + * Iterate over a NodeList, return true if all of the provided iterate + * function calls return true, false otherwise. + * + * For convenience, the current object context is applied to the + * provided iterate function. + * + * @param NodeList nodeList The NodeList. + * @param Function fn The iterate function. + * @return Boolean + */ + _everyNode: function(nodeList, fn) { + return Array.prototype.every.call(nodeList, fn, this); + }, + + /** + * Concat all nodelists passed as arguments. + * + * @return ...NodeList + * @return Array + */ + _concatNodeLists: function() { + var slice = Array.prototype.slice; + var args = slice.call(arguments); + var nodeLists = args.map(function(list) { + return slice.call(list); + }); + return Array.prototype.concat.apply([], nodeLists); + }, + + _getAllNodesWithTag: function(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); + })); + }, + + /** + * Removes the class="" attribute from every element in the given + * subtree, except those that match CLASSES_TO_PRESERVE and + * the classesToPreserve array from the options object. + * + * @param Element + * @return void + */ + _cleanClasses: function(node) { + var classesToPreserve = this._classesToPreserve; + var className = (node.getAttribute("class") || "") + .split(/\s+/) + .filter(function(cls) { + return classesToPreserve.indexOf(cls) != -1; + }) + .join(" "); + + if (className) { + node.setAttribute("class", className); + } else { + node.removeAttribute("class"); + } + + for (node = node.firstElementChild; node; node = node.nextElementSibling) { + this._cleanClasses(node); + } + }, + + /** + * Converts each <a> and <img> uri in the given element to an absolute URI, + * ignoring #ref URIs. + * + * @param Element + * @return void + */ + _fixRelativeUris: function(articleContent) { + var baseURI = this._doc.baseURI; + var documentURI = this._doc.documentURI; + function toAbsoluteURI(uri) { + // Leave hash links alone if the base URI matches the document URI: + if (baseURI == documentURI && uri.charAt(0) == "#") { + return uri; + } + + // Otherwise, resolve against base URI: + try { + return new URL(uri, baseURI).href; + } catch (ex) { + // Something went wrong, just return the original: + } + return uri; + } + + var links = this._getAllNodesWithTag(articleContent, ["a"]); + this._forEachNode(links, function(link) { + var href = link.getAttribute("href"); + if (href) { + // Remove links with javascript: URIs, since + // they won't work after scripts have been removed from the page. + if (href.indexOf("javascript:") === 0) { + // if the link only contains simple text content, it can be converted to a text node + 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 { + // if the link has multiple children, they should all be preserved + 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: function(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.setAttribute(node.attributes[i].name, node.attributes[i].value); + } + node.parentNode.replaceChild(child, node); + node = child; + continue; + } + } + + node = this._getNextNode(node); + } + }, + + /** + * Get the article title as an H1. + * + * @return string + **/ + _getArticleTitle: function() { + var doc = this._doc; + var curTitle = ""; + var origTitle = ""; + + try { + curTitle = origTitle = doc.title.trim(); + + // If they had an element with id "title" in their HTML + if (typeof curTitle !== "string") + curTitle = origTitle = this._getInnerText(doc.getElementsByTagName("title")[0]); + } catch (e) {/* ignore exceptions setting the title. */} + + var titleHadHierarchicalSeparators = false; + function wordCount(str) { + return str.split(/\s+/).length; + } + + // If there's a separator in the title, first remove the final part + if ((/ [\|\-\\\/>»] /).test(curTitle)) { + titleHadHierarchicalSeparators = / [\\\/>»] /.test(curTitle); + curTitle = origTitle.replace(/(.*)[\|\-\\\/>»] .*/gi, "$1"); + + // If the resulting title is too short (3 words or fewer), remove + // the first part instead: + if (wordCount(curTitle) < 3) + curTitle = origTitle.replace(/[^\|\-\\\/>»]*[\|\-\\\/>»](.*)/gi, "$1"); + } else if (curTitle.indexOf(": ") !== -1) { + // Check if we have an heading containing this exact string, so we + // could assume it's the full title. + var headings = this._concatNodeLists( + doc.getElementsByTagName("h1"), + doc.getElementsByTagName("h2") + ); + var trimmedTitle = curTitle.trim(); + var match = this._someNode(headings, function(heading) { + return heading.textContent.trim() === trimmedTitle; + }); + + // If we don't, let's extract the title out of the original title string. + if (!match) { + curTitle = origTitle.substring(origTitle.lastIndexOf(":") + 1); + + // If the title is now too short, try the first colon instead: + if (wordCount(curTitle) < 3) { + curTitle = origTitle.substring(origTitle.indexOf(":") + 1); + // But if we have too many words before the colon there's something weird + // with the titles and the H tags so let's just use the original title instead + } 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, " "); + // If we now have 4 words or fewer as our title, and either no + // 'hierarchical' separators (\, /, > or ») were found in the original + // title or we decreased the number of words by more than 1 word, use + // the original title. + var curTitleWordCount = wordCount(curTitle); + if (curTitleWordCount <= 4 && + (!titleHadHierarchicalSeparators || + curTitleWordCount != wordCount(origTitle.replace(/[\|\-\\\/>»]+/g, "")) - 1)) { + curTitle = origTitle; + } + + return curTitle; + }, + + /** + * Prepare the HTML document for readability to scrape it. + * This includes things like stripping javascript, CSS, and handling terrible markup. + * + * @return void + **/ + _prepDocument: function() { + var doc = this._doc; + + // Remove all style tags in head + this._removeNodes(this._getAllNodesWithTag(doc, ["style"])); + + if (doc.body) { + this._replaceBrs(doc.body); + } + + this._replaceNodeTags(this._getAllNodesWithTag(doc, ["font"]), "SPAN"); + }, + + /** + * Finds the next node, starting from the given node, and ignoring + * whitespace in between. If the given node is an element, the same node is + * returned. + */ + _nextNode: function (node) { + var next = node; + while (next + && (next.nodeType != this.ELEMENT_NODE) + && this.REGEXPS.whitespace.test(next.textContent)) { + next = next.nextSibling; + } + return next; + }, + + /** + * Replaces 2 or more successive <br> elements with a single <p>. + * Whitespace between <br> elements are ignored. For example: + * <div>foo<br>bar<br> <br><br>abc</div> + * will become: + * <div>foo<br>bar<p>abc</p></div> + */ + _replaceBrs: function (elem) { + this._forEachNode(this._getAllNodesWithTag(elem, ["br"]), function(br) { + var next = br.nextSibling; + + // Whether 2 or more <br> elements have been found and replaced with a + // <p> block. + var replaced = false; + + // If we find a <br> chain, remove the <br>s until we hit another node + // or non-whitespace. This leaves behind the first <br> in the chain + // (which will be replaced with a <p> later). + while ((next = this._nextNode(next)) && (next.tagName == "BR")) { + replaced = true; + var brSibling = next.nextSibling; + next.parentNode.removeChild(next); + next = brSibling; + } + + // If we removed a <br> chain, replace the remaining <br> with a <p>. Add + // all sibling nodes as children of the <p> until we hit another <br> + // chain. + if (replaced) { + var p = this._doc.createElement("p"); + br.parentNode.replaceChild(p, br); + + next = p.nextSibling; + while (next) { + // If we've hit another <br><br>, we're done adding children to this <p>. + if (next.tagName == "BR") { + var nextElem = this._nextNode(next.nextSibling); + if (nextElem && nextElem.tagName == "BR") + break; + } + + if (!this._isPhrasingContent(next)) + break; + + // Otherwise, make this node a child of the new <p>. + var sibling = next.nextSibling; + p.appendChild(next); + next = sibling; + } + + while (p.lastChild && this._isWhitespace(p.lastChild)) { + p.removeChild(p.lastChild); + } + + if (p.parentNode.tagName === "P") + this._setNodeTag(p.parentNode, "DIV"); + } + }); + }, + + _setNodeTag: function (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++) { + try { + replacement.setAttribute(node.attributes[i].name, node.attributes[i].value); + } catch (ex) { + /* it's possible for setAttribute() to throw if the attribute name + * isn't a valid XML Name. Such attributes can however be parsed from + * source in HTML docs, see https://github.com/whatwg/html/issues/4275, + * so we can hit them here and then throw. We don't care about such + * attributes so we ignore them. + */ + } + } + return replacement; + }, + + /** + * Prepare the article node for display. Clean out any inline styles, + * iframes, forms, strip extraneous <p> tags, etc. + * + * @param Element + * @return void + **/ + _prepArticle: function(articleContent) { + this._cleanStyles(articleContent); + + // Check for data tables before we continue, to avoid removing items in + // those tables, which will often be isolated even though they're + // visually linked to other content-ful elements (text, images, etc.). + this._markDataTables(articleContent); + + this._fixLazyImages(articleContent); + + // Clean out junk from the article content + 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"); + + // Clean out elements with little content that have "share" in their id/class combinations from final top candidates, + // which means we don't remove the top candidates even they have "share". + + 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); + + // Do these last as the previous stuff may have removed junk + // that will affect these + this._cleanConditionally(articleContent, "table"); + this._cleanConditionally(articleContent, "ul"); + this._cleanConditionally(articleContent, "div"); + + // replace H1 with H2 as H1 should be only title that is displayed separately + this._replaceNodeTags(this._getAllNodesWithTag(articleContent, ["h1"]), "h2"); + + // Remove extra paragraphs + this._removeNodes(this._getAllNodesWithTag(articleContent, ["p"]), function (paragraph) { + var imgCount = paragraph.getElementsByTagName("img").length; + var embedCount = paragraph.getElementsByTagName("embed").length; + var objectCount = paragraph.getElementsByTagName("object").length; + // At this point, nasty iframes have been removed, only remain embedded video ones. + var iframeCount = paragraph.getElementsByTagName("iframe").length; + var totalCount = imgCount + embedCount + objectCount + iframeCount; + + return totalCount === 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.parentNode.removeChild(br); + }); + + // Remove single-cell tables + 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); + } + } + }); + }, + + /** + * Initialize a node with the readability object. Also checks the + * className/id for special names to add to its score. + * + * @param Element + * @return void + **/ + _initializeNode: function(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: function(node) { + var nextNode = this._getNextNode(node, true); + node.parentNode.removeChild(node); + return nextNode; + }, + + /** + * Traverse the DOM from node to node, starting at the node passed in. + * Pass true for the second parameter to indicate this node itself + * (and its kids) are going away, and we want the next node over. + * + * Calling this in a loop will traverse the DOM depth-first. + */ + _getNextNode: function(node, ignoreSelfAndKids) { + // First check for kids if those aren't being ignored + if (!ignoreSelfAndKids && node.firstElementChild) { + return node.firstElementChild; + } + // Then for siblings... + if (node.nextElementSibling) { + return node.nextElementSibling; + } + // And finally, move up the parent chain *and* find a sibling + // (because this is depth-first traversal, we will have already + // seen the parent nodes themselves). + do { + node = node.parentNode; + } while (node && !node.nextElementSibling); + return node && node.nextElementSibling; + }, + + // compares second text to first one + // 1 = same text, 0 = completely different text + // works the way that it splits both texts into words and then finds words that are unique in second text + // the result is given by the lower length of unique parts + _textSimilarity: function(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; + }, + + _checkByline: function(node, matchString) { + if (this._articleByline) { + return false; + } + + if (node.getAttribute !== undefined) { + var rel = node.getAttribute("rel"); + var itemprop = node.getAttribute("itemprop"); + } + + if ((rel === "author" || (itemprop && itemprop.indexOf("author") !== -1) || this.REGEXPS.byline.test(matchString)) && this._isValidByline(node.textContent)) { + this._articleByline = node.textContent.trim(); + return true; + } + + return false; + }, + + _getNodeAncestors: function(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 - Using a variety of metrics (content score, classname, element types), find the content that is + * most likely to be the stuff a user wants to read. Then return it wrapped up in a div. + * + * @param page a document to run upon. Needs to be a full document, complete with body. + * @return Element + **/ + _grabArticle: function (page) { + this.log("**** grabArticle ****"); + var doc = this._doc; + var isPaging = page !== null; + page = page ? page : this._doc.body; + + // We can't grab an article if we don't have a page! + 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); + + // First, node prepping. Trash nodes that look cruddy (like ones with the + // class name "comment", etc), and turn divs into P tags where they have been + // used inappropriately (as in, where they contain no other block level elements.) + 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; + } + + // Check to see if this node is a byline, and remove it if it is. + if (this._checkByline(node, matchString)) { + 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; + } + + // Remove unlikely candidates + 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(nod |