diff --git a/packages/app-clipper/content_scripts/JSDOMParser.js b/packages/app-clipper/content_scripts/JSDOMParser.js index 4fb766c107..c21d5de1e4 100644 --- a/packages/app-clipper/content_scripts/JSDOMParser.js +++ b/packages/app-clipper/content_scripts/JSDOMParser.js @@ -1,4 +1,4 @@ -// https://github.com/mozilla/readability/tree/814f0a3884350b6f1adfdebb79ca3599e9806605 +// v0.4.1 - https://github.com/mozilla/readability/commit/28843b6de84447dd6cef04058fda336938e628dc /*eslint-env es6:false*/ /* This Source Code Form is subject to the terms of the Mozilla Public @@ -27,1164 +27,1172 @@ * want these lists to be updated when nodes are removed or added to the * document, you must take care to manually update them yourself. */ -(function (global) { + (function (global) { // XML only defines these and the numeric ones: - + var entityTable = { - 'lt': '<', - 'gt': '>', - 'amp': '&', - 'quot': '"', - 'apos': '\'', + "lt": "<", + "gt": ">", + "amp": "&", + "quot": '"', + "apos": "'", }; - + var reverseEntityTable = { - '<': '<', - '>': '>', - '&': '&', - '"': '"', - '\'': ''', + "<": "<", + ">": ">", + "&": "&", + '"': """, + "'": "'", }; - + function encodeTextContentHTML(s) { - return s.replace(/[&<>]/g, function(x) { - return reverseEntityTable[x]; - }); + return s.replace(/[&<>]/g, function(x) { + return reverseEntityTable[x]; + }); } - + function encodeHTML(s) { - return s.replace(/[&<>'"]/g, function(x) { - return reverseEntityTable[x]; - }); + return s.replace(/[&<>'"]/g, function(x) { + return reverseEntityTable[x]; + }); } - + function decodeHTML(str) { - return str.replace(/&(quot|amp|apos|lt|gt);/g, function(match, tag) { - return entityTable[tag]; - }).replace(/&#(?:x([0-9a-z]{1,4})|([0-9]{1,4}));/gi, function(match, hex, numStr) { - var num = parseInt(hex || numStr, hex ? 16 : 10); // read num - return String.fromCharCode(num); - }); + return str.replace(/&(quot|amp|apos|lt|gt);/g, function(match, tag) { + return entityTable[tag]; + }).replace(/&#(?:x([0-9a-z]{1,4})|([0-9]{1,4}));/gi, function(match, hex, numStr) { + var num = parseInt(hex || numStr, hex ? 16 : 10); // read num + return String.fromCharCode(num); + }); } - + // When a style is set in JS, map it to the corresponding CSS attribute var styleMap = { - 'alignmentBaseline': 'alignment-baseline', - 'background': 'background', - 'backgroundAttachment': 'background-attachment', - 'backgroundClip': 'background-clip', - 'backgroundColor': 'background-color', - 'backgroundImage': 'background-image', - 'backgroundOrigin': 'background-origin', - 'backgroundPosition': 'background-position', - 'backgroundPositionX': 'background-position-x', - 'backgroundPositionY': 'background-position-y', - 'backgroundRepeat': 'background-repeat', - 'backgroundRepeatX': 'background-repeat-x', - 'backgroundRepeatY': 'background-repeat-y', - 'backgroundSize': 'background-size', - 'baselineShift': 'baseline-shift', - 'border': 'border', - 'borderBottom': 'border-bottom', - 'borderBottomColor': 'border-bottom-color', - 'borderBottomLeftRadius': 'border-bottom-left-radius', - 'borderBottomRightRadius': 'border-bottom-right-radius', - 'borderBottomStyle': 'border-bottom-style', - 'borderBottomWidth': 'border-bottom-width', - 'borderCollapse': 'border-collapse', - 'borderColor': 'border-color', - 'borderImage': 'border-image', - 'borderImageOutset': 'border-image-outset', - 'borderImageRepeat': 'border-image-repeat', - 'borderImageSlice': 'border-image-slice', - 'borderImageSource': 'border-image-source', - 'borderImageWidth': 'border-image-width', - 'borderLeft': 'border-left', - 'borderLeftColor': 'border-left-color', - 'borderLeftStyle': 'border-left-style', - 'borderLeftWidth': 'border-left-width', - 'borderRadius': 'border-radius', - 'borderRight': 'border-right', - 'borderRightColor': 'border-right-color', - 'borderRightStyle': 'border-right-style', - 'borderRightWidth': 'border-right-width', - 'borderSpacing': 'border-spacing', - 'borderStyle': 'border-style', - 'borderTop': 'border-top', - 'borderTopColor': 'border-top-color', - 'borderTopLeftRadius': 'border-top-left-radius', - 'borderTopRightRadius': 'border-top-right-radius', - 'borderTopStyle': 'border-top-style', - 'borderTopWidth': 'border-top-width', - 'borderWidth': 'border-width', - 'bottom': 'bottom', - 'boxShadow': 'box-shadow', - 'boxSizing': 'box-sizing', - 'captionSide': 'caption-side', - 'clear': 'clear', - 'clip': 'clip', - 'clipPath': 'clip-path', - 'clipRule': 'clip-rule', - 'color': 'color', - 'colorInterpolation': 'color-interpolation', - 'colorInterpolationFilters': 'color-interpolation-filters', - 'colorProfile': 'color-profile', - 'colorRendering': 'color-rendering', - 'content': 'content', - 'counterIncrement': 'counter-increment', - 'counterReset': 'counter-reset', - 'cursor': 'cursor', - 'direction': 'direction', - 'display': 'display', - 'dominantBaseline': 'dominant-baseline', - 'emptyCells': 'empty-cells', - 'enableBackground': 'enable-background', - 'fill': 'fill', - 'fillOpacity': 'fill-opacity', - 'fillRule': 'fill-rule', - 'filter': 'filter', - 'cssFloat': 'float', - 'floodColor': 'flood-color', - 'floodOpacity': 'flood-opacity', - 'font': 'font', - 'fontFamily': 'font-family', - 'fontSize': 'font-size', - 'fontStretch': 'font-stretch', - 'fontStyle': 'font-style', - 'fontVariant': 'font-variant', - 'fontWeight': 'font-weight', - 'glyphOrientationHorizontal': 'glyph-orientation-horizontal', - 'glyphOrientationVertical': 'glyph-orientation-vertical', - 'height': 'height', - 'imageRendering': 'image-rendering', - 'kerning': 'kerning', - 'left': 'left', - 'letterSpacing': 'letter-spacing', - 'lightingColor': 'lighting-color', - 'lineHeight': 'line-height', - 'listStyle': 'list-style', - 'listStyleImage': 'list-style-image', - 'listStylePosition': 'list-style-position', - 'listStyleType': 'list-style-type', - 'margin': 'margin', - 'marginBottom': 'margin-bottom', - 'marginLeft': 'margin-left', - 'marginRight': 'margin-right', - 'marginTop': 'margin-top', - 'marker': 'marker', - 'markerEnd': 'marker-end', - 'markerMid': 'marker-mid', - 'markerStart': 'marker-start', - 'mask': 'mask', - 'maxHeight': 'max-height', - 'maxWidth': 'max-width', - 'minHeight': 'min-height', - 'minWidth': 'min-width', - 'opacity': 'opacity', - 'orphans': 'orphans', - 'outline': 'outline', - 'outlineColor': 'outline-color', - 'outlineOffset': 'outline-offset', - 'outlineStyle': 'outline-style', - 'outlineWidth': 'outline-width', - 'overflow': 'overflow', - 'overflowX': 'overflow-x', - 'overflowY': 'overflow-y', - 'padding': 'padding', - 'paddingBottom': 'padding-bottom', - 'paddingLeft': 'padding-left', - 'paddingRight': 'padding-right', - 'paddingTop': 'padding-top', - 'page': 'page', - 'pageBreakAfter': 'page-break-after', - 'pageBreakBefore': 'page-break-before', - 'pageBreakInside': 'page-break-inside', - 'pointerEvents': 'pointer-events', - 'position': 'position', - 'quotes': 'quotes', - 'resize': 'resize', - 'right': 'right', - 'shapeRendering': 'shape-rendering', - 'size': 'size', - 'speak': 'speak', - 'src': 'src', - 'stopColor': 'stop-color', - 'stopOpacity': 'stop-opacity', - 'stroke': 'stroke', - 'strokeDasharray': 'stroke-dasharray', - 'strokeDashoffset': 'stroke-dashoffset', - 'strokeLinecap': 'stroke-linecap', - 'strokeLinejoin': 'stroke-linejoin', - 'strokeMiterlimit': 'stroke-miterlimit', - 'strokeOpacity': 'stroke-opacity', - 'strokeWidth': 'stroke-width', - 'tableLayout': 'table-layout', - 'textAlign': 'text-align', - 'textAnchor': 'text-anchor', - 'textDecoration': 'text-decoration', - 'textIndent': 'text-indent', - 'textLineThrough': 'text-line-through', - 'textLineThroughColor': 'text-line-through-color', - 'textLineThroughMode': 'text-line-through-mode', - 'textLineThroughStyle': 'text-line-through-style', - 'textLineThroughWidth': 'text-line-through-width', - 'textOverflow': 'text-overflow', - 'textOverline': 'text-overline', - 'textOverlineColor': 'text-overline-color', - 'textOverlineMode': 'text-overline-mode', - 'textOverlineStyle': 'text-overline-style', - 'textOverlineWidth': 'text-overline-width', - 'textRendering': 'text-rendering', - 'textShadow': 'text-shadow', - 'textTransform': 'text-transform', - 'textUnderline': 'text-underline', - 'textUnderlineColor': 'text-underline-color', - 'textUnderlineMode': 'text-underline-mode', - 'textUnderlineStyle': 'text-underline-style', - 'textUnderlineWidth': 'text-underline-width', - 'top': 'top', - 'unicodeBidi': 'unicode-bidi', - 'unicodeRange': 'unicode-range', - 'vectorEffect': 'vector-effect', - 'verticalAlign': 'vertical-align', - 'visibility': 'visibility', - 'whiteSpace': 'white-space', - 'widows': 'widows', - 'width': 'width', - 'wordBreak': 'word-break', - 'wordSpacing': 'word-spacing', - 'wordWrap': 'word-wrap', - 'writingMode': 'writing-mode', - 'zIndex': 'z-index', - 'zoom': 'zoom', + "alignmentBaseline": "alignment-baseline", + "background": "background", + "backgroundAttachment": "background-attachment", + "backgroundClip": "background-clip", + "backgroundColor": "background-color", + "backgroundImage": "background-image", + "backgroundOrigin": "background-origin", + "backgroundPosition": "background-position", + "backgroundPositionX": "background-position-x", + "backgroundPositionY": "background-position-y", + "backgroundRepeat": "background-repeat", + "backgroundRepeatX": "background-repeat-x", + "backgroundRepeatY": "background-repeat-y", + "backgroundSize": "background-size", + "baselineShift": "baseline-shift", + "border": "border", + "borderBottom": "border-bottom", + "borderBottomColor": "border-bottom-color", + "borderBottomLeftRadius": "border-bottom-left-radius", + "borderBottomRightRadius": "border-bottom-right-radius", + "borderBottomStyle": "border-bottom-style", + "borderBottomWidth": "border-bottom-width", + "borderCollapse": "border-collapse", + "borderColor": "border-color", + "borderImage": "border-image", + "borderImageOutset": "border-image-outset", + "borderImageRepeat": "border-image-repeat", + "borderImageSlice": "border-image-slice", + "borderImageSource": "border-image-source", + "borderImageWidth": "border-image-width", + "borderLeft": "border-left", + "borderLeftColor": "border-left-color", + "borderLeftStyle": "border-left-style", + "borderLeftWidth": "border-left-width", + "borderRadius": "border-radius", + "borderRight": "border-right", + "borderRightColor": "border-right-color", + "borderRightStyle": "border-right-style", + "borderRightWidth": "border-right-width", + "borderSpacing": "border-spacing", + "borderStyle": "border-style", + "borderTop": "border-top", + "borderTopColor": "border-top-color", + "borderTopLeftRadius": "border-top-left-radius", + "borderTopRightRadius": "border-top-right-radius", + "borderTopStyle": "border-top-style", + "borderTopWidth": "border-top-width", + "borderWidth": "border-width", + "bottom": "bottom", + "boxShadow": "box-shadow", + "boxSizing": "box-sizing", + "captionSide": "caption-side", + "clear": "clear", + "clip": "clip", + "clipPath": "clip-path", + "clipRule": "clip-rule", + "color": "color", + "colorInterpolation": "color-interpolation", + "colorInterpolationFilters": "color-interpolation-filters", + "colorProfile": "color-profile", + "colorRendering": "color-rendering", + "content": "content", + "counterIncrement": "counter-increment", + "counterReset": "counter-reset", + "cursor": "cursor", + "direction": "direction", + "display": "display", + "dominantBaseline": "dominant-baseline", + "emptyCells": "empty-cells", + "enableBackground": "enable-background", + "fill": "fill", + "fillOpacity": "fill-opacity", + "fillRule": "fill-rule", + "filter": "filter", + "cssFloat": "float", + "floodColor": "flood-color", + "floodOpacity": "flood-opacity", + "font": "font", + "fontFamily": "font-family", + "fontSize": "font-size", + "fontStretch": "font-stretch", + "fontStyle": "font-style", + "fontVariant": "font-variant", + "fontWeight": "font-weight", + "glyphOrientationHorizontal": "glyph-orientation-horizontal", + "glyphOrientationVertical": "glyph-orientation-vertical", + "height": "height", + "imageRendering": "image-rendering", + "kerning": "kerning", + "left": "left", + "letterSpacing": "letter-spacing", + "lightingColor": "lighting-color", + "lineHeight": "line-height", + "listStyle": "list-style", + "listStyleImage": "list-style-image", + "listStylePosition": "list-style-position", + "listStyleType": "list-style-type", + "margin": "margin", + "marginBottom": "margin-bottom", + "marginLeft": "margin-left", + "marginRight": "margin-right", + "marginTop": "margin-top", + "marker": "marker", + "markerEnd": "marker-end", + "markerMid": "marker-mid", + "markerStart": "marker-start", + "mask": "mask", + "maxHeight": "max-height", + "maxWidth": "max-width", + "minHeight": "min-height", + "minWidth": "min-width", + "opacity": "opacity", + "orphans": "orphans", + "outline": "outline", + "outlineColor": "outline-color", + "outlineOffset": "outline-offset", + "outlineStyle": "outline-style", + "outlineWidth": "outline-width", + "overflow": "overflow", + "overflowX": "overflow-x", + "overflowY": "overflow-y", + "padding": "padding", + "paddingBottom": "padding-bottom", + "paddingLeft": "padding-left", + "paddingRight": "padding-right", + "paddingTop": "padding-top", + "page": "page", + "pageBreakAfter": "page-break-after", + "pageBreakBefore": "page-break-before", + "pageBreakInside": "page-break-inside", + "pointerEvents": "pointer-events", + "position": "position", + "quotes": "quotes", + "resize": "resize", + "right": "right", + "shapeRendering": "shape-rendering", + "size": "size", + "speak": "speak", + "src": "src", + "stopColor": "stop-color", + "stopOpacity": "stop-opacity", + "stroke": "stroke", + "strokeDasharray": "stroke-dasharray", + "strokeDashoffset": "stroke-dashoffset", + "strokeLinecap": "stroke-linecap", + "strokeLinejoin": "stroke-linejoin", + "strokeMiterlimit": "stroke-miterlimit", + "strokeOpacity": "stroke-opacity", + "strokeWidth": "stroke-width", + "tableLayout": "table-layout", + "textAlign": "text-align", + "textAnchor": "text-anchor", + "textDecoration": "text-decoration", + "textIndent": "text-indent", + "textLineThrough": "text-line-through", + "textLineThroughColor": "text-line-through-color", + "textLineThroughMode": "text-line-through-mode", + "textLineThroughStyle": "text-line-through-style", + "textLineThroughWidth": "text-line-through-width", + "textOverflow": "text-overflow", + "textOverline": "text-overline", + "textOverlineColor": "text-overline-color", + "textOverlineMode": "text-overline-mode", + "textOverlineStyle": "text-overline-style", + "textOverlineWidth": "text-overline-width", + "textRendering": "text-rendering", + "textShadow": "text-shadow", + "textTransform": "text-transform", + "textUnderline": "text-underline", + "textUnderlineColor": "text-underline-color", + "textUnderlineMode": "text-underline-mode", + "textUnderlineStyle": "text-underline-style", + "textUnderlineWidth": "text-underline-width", + "top": "top", + "unicodeBidi": "unicode-bidi", + "unicodeRange": "unicode-range", + "vectorEffect": "vector-effect", + "verticalAlign": "vertical-align", + "visibility": "visibility", + "whiteSpace": "white-space", + "widows": "widows", + "width": "width", + "wordBreak": "word-break", + "wordSpacing": "word-spacing", + "wordWrap": "word-wrap", + "writingMode": "writing-mode", + "zIndex": "z-index", + "zoom": "zoom", }; - + // Elements that can be self-closing var voidElems = { - 'area': true, - 'base': true, - 'br': true, - 'col': true, - 'command': true, - 'embed': true, - 'hr': true, - 'img': true, - 'input': true, - 'link': true, - 'meta': true, - 'param': true, - 'source': true, - 'wbr': true, + "area": true, + "base": true, + "br": true, + "col": true, + "command": true, + "embed": true, + "hr": true, + "img": true, + "input": true, + "link": true, + "meta": true, + "param": true, + "source": true, + "wbr": true }; - - var whitespace = [' ', '\t', '\n', '\r']; - - // See http://www.w3schools.com/dom/dom_nodetype.asp + + var whitespace = [" ", "\t", "\n", "\r"]; + + // See https://developer.mozilla.org/en-US/docs/Web/API/Node/nodeType var nodeTypes = { - ELEMENT_NODE: 1, - ATTRIBUTE_NODE: 2, - TEXT_NODE: 3, - CDATA_SECTION_NODE: 4, - ENTITY_REFERENCE_NODE: 5, - ENTITY_NODE: 6, - PROCESSING_INSTRUCTION_NODE: 7, - COMMENT_NODE: 8, - DOCUMENT_NODE: 9, - DOCUMENT_TYPE_NODE: 10, - DOCUMENT_FRAGMENT_NODE: 11, - NOTATION_NODE: 12, + ELEMENT_NODE: 1, + ATTRIBUTE_NODE: 2, + TEXT_NODE: 3, + CDATA_SECTION_NODE: 4, + ENTITY_REFERENCE_NODE: 5, + ENTITY_NODE: 6, + PROCESSING_INSTRUCTION_NODE: 7, + COMMENT_NODE: 8, + DOCUMENT_NODE: 9, + DOCUMENT_TYPE_NODE: 10, + DOCUMENT_FRAGMENT_NODE: 11, + NOTATION_NODE: 12 }; - + function getElementsByTagName(tag) { - tag = tag.toUpperCase(); - var elems = []; - var allTags = (tag === '*'); - function getElems(node) { - var length = node.children.length; - for (var i = 0; i < length; i++) { - var child = node.children[i]; - if (allTags || (child.tagName === tag)) - elems.push(child); - getElems(child); - } + tag = tag.toUpperCase(); + var elems = []; + var allTags = (tag === "*"); + function getElems(node) { + var length = node.children.length; + for (var i = 0; i < length; i++) { + var child = node.children[i]; + if (allTags || (child.tagName === tag)) + elems.push(child); + getElems(child); } - getElems(this); - return elems; + } + getElems(this); + elems._isLiveNodeList = true; + return elems; } - + var Node = function () {}; - + Node.prototype = { - attributes: null, - childNodes: null, - localName: null, - nodeName: null, - parentNode: null, - textContent: null, - nextSibling: null, - previousSibling: null, - - get firstChild() { - return this.childNodes[0] || null; - }, - - get firstElementChild() { - return this.children[0] || null; - }, - - get lastChild() { - return this.childNodes[this.childNodes.length - 1] || null; - }, - - get lastElementChild() { - return this.children[this.children.length - 1] || null; - }, - - appendChild: function (child) { - if (child.parentNode) { - child.parentNode.removeChild(child); - } - - var last = this.lastChild; - if (last) - last.nextSibling = child; - child.previousSibling = last; - - if (child.nodeType === Node.ELEMENT_NODE) { - child.previousElementSibling = this.children[this.children.length - 1] || null; - this.children.push(child); - child.previousElementSibling && (child.previousElementSibling.nextElementSibling = child); - } - this.childNodes.push(child); - child.parentNode = this; - }, - - removeChild: function (child) { - var childNodes = this.childNodes; - var childIndex = childNodes.indexOf(child); - if (childIndex === -1) { - throw 'removeChild: node not found'; + attributes: null, + childNodes: null, + localName: null, + nodeName: null, + parentNode: null, + textContent: null, + nextSibling: null, + previousSibling: null, + + get firstChild() { + return this.childNodes[0] || null; + }, + + get firstElementChild() { + return this.children[0] || null; + }, + + get lastChild() { + return this.childNodes[this.childNodes.length - 1] || null; + }, + + get lastElementChild() { + return this.children[this.children.length - 1] || null; + }, + + appendChild: function (child) { + if (child.parentNode) { + child.parentNode.removeChild(child); + } + + var last = this.lastChild; + if (last) + last.nextSibling = child; + child.previousSibling = last; + + if (child.nodeType === Node.ELEMENT_NODE) { + child.previousElementSibling = this.children[this.children.length - 1] || null; + this.children.push(child); + child.previousElementSibling && (child.previousElementSibling.nextElementSibling = child); + } + this.childNodes.push(child); + child.parentNode = this; + }, + + removeChild: function (child) { + var childNodes = this.childNodes; + var childIndex = childNodes.indexOf(child); + if (childIndex === -1) { + throw "removeChild: node not found"; + } else { + child.parentNode = null; + var prev = child.previousSibling; + var next = child.nextSibling; + if (prev) + prev.nextSibling = next; + if (next) + next.previousSibling = prev; + + if (child.nodeType === Node.ELEMENT_NODE) { + prev = child.previousElementSibling; + next = child.nextElementSibling; + if (prev) + prev.nextElementSibling = next; + if (next) + next.previousElementSibling = prev; + this.children.splice(this.children.indexOf(child), 1); + } + + child.previousSibling = child.nextSibling = null; + child.previousElementSibling = child.nextElementSibling = null; + + return childNodes.splice(childIndex, 1)[0]; + } + }, + + replaceChild: function (newNode, oldNode) { + var childNodes = this.childNodes; + var childIndex = childNodes.indexOf(oldNode); + if (childIndex === -1) { + throw "replaceChild: node not found"; + } else { + // This will take care of updating the new node if it was somewhere else before: + if (newNode.parentNode) + newNode.parentNode.removeChild(newNode); + + childNodes[childIndex] = newNode; + + // update the new node's sibling properties, and its new siblings' sibling properties + newNode.nextSibling = oldNode.nextSibling; + newNode.previousSibling = oldNode.previousSibling; + if (newNode.nextSibling) + newNode.nextSibling.previousSibling = newNode; + if (newNode.previousSibling) + newNode.previousSibling.nextSibling = newNode; + + newNode.parentNode = this; + + // Now deal with elements before we clear out those values for the old node, + // because it can help us take shortcuts here: + if (newNode.nodeType === Node.ELEMENT_NODE) { + if (oldNode.nodeType === Node.ELEMENT_NODE) { + // Both were elements, which makes this easier, we just swap things out: + newNode.previousElementSibling = oldNode.previousElementSibling; + newNode.nextElementSibling = oldNode.nextElementSibling; + if (newNode.previousElementSibling) + newNode.previousElementSibling.nextElementSibling = newNode; + if (newNode.nextElementSibling) + newNode.nextElementSibling.previousElementSibling = newNode; + this.children[this.children.indexOf(oldNode)] = newNode; } else { - child.parentNode = null; - var prev = child.previousSibling; - var next = child.nextSibling; - if (prev) - prev.nextSibling = next; - if (next) - next.previousSibling = prev; - - if (child.nodeType === Node.ELEMENT_NODE) { - prev = child.previousElementSibling; - next = child.nextElementSibling; - if (prev) - prev.nextElementSibling = next; - if (next) - next.previousElementSibling = prev; - this.children.splice(this.children.indexOf(child), 1); - } - - child.previousSibling = child.nextSibling = null; - child.previousElementSibling = child.nextElementSibling = null; - - return childNodes.splice(childIndex, 1)[0]; - } - }, - - replaceChild: function (newNode, oldNode) { - var childNodes = this.childNodes; - var childIndex = childNodes.indexOf(oldNode); - if (childIndex === -1) { - throw 'replaceChild: node not found'; - } else { - // This will take care of updating the new node if it was somewhere else before: - if (newNode.parentNode) - newNode.parentNode.removeChild(newNode); - - childNodes[childIndex] = newNode; - - // update the new node's sibling properties, and its new siblings' sibling properties - newNode.nextSibling = oldNode.nextSibling; - newNode.previousSibling = oldNode.previousSibling; - if (newNode.nextSibling) - newNode.nextSibling.previousSibling = newNode; - if (newNode.previousSibling) - newNode.previousSibling.nextSibling = newNode; - - newNode.parentNode = this; - - // Now deal with elements before we clear out those values for the old node, - // because it can help us take shortcuts here: - if (newNode.nodeType === Node.ELEMENT_NODE) { - if (oldNode.nodeType === Node.ELEMENT_NODE) { - // Both were elements, which makes this easier, we just swap things out: - newNode.previousElementSibling = oldNode.previousElementSibling; - newNode.nextElementSibling = oldNode.nextElementSibling; - if (newNode.previousElementSibling) - newNode.previousElementSibling.nextElementSibling = newNode; - if (newNode.nextElementSibling) - newNode.nextElementSibling.previousElementSibling = newNode; - this.children[this.children.indexOf(oldNode)] = newNode; - } else { - // Hard way: - newNode.previousElementSibling = (function() { - for (var i = childIndex - 1; i >= 0; i--) { - if (childNodes[i].nodeType === Node.ELEMENT_NODE) - return childNodes[i]; - } - return null; - })(); - if (newNode.previousElementSibling) { - newNode.nextElementSibling = newNode.previousElementSibling.nextElementSibling; - } else { - newNode.nextElementSibling = (function() { - for (var i = childIndex + 1; i < childNodes.length; i++) { - if (childNodes[i].nodeType === Node.ELEMENT_NODE) - return childNodes[i]; - } - return null; - })(); - } - if (newNode.previousElementSibling) - newNode.previousElementSibling.nextElementSibling = newNode; - if (newNode.nextElementSibling) - newNode.nextElementSibling.previousElementSibling = newNode; - - if (newNode.nextElementSibling) - this.children.splice(this.children.indexOf(newNode.nextElementSibling), 0, newNode); - else - this.children.push(newNode); - } - } else if (oldNode.nodeType === Node.ELEMENT_NODE) { - // new node is not an element node. - // if the old one was, update its element siblings: - if (oldNode.previousElementSibling) - oldNode.previousElementSibling.nextElementSibling = oldNode.nextElementSibling; - if (oldNode.nextElementSibling) - oldNode.nextElementSibling.previousElementSibling = oldNode.previousElementSibling; - this.children.splice(this.children.indexOf(oldNode), 1); - - // If the old node wasn't an element, neither the new nor the old node was an element, - // and the children array and its members shouldn't need any updating. - } - - - oldNode.parentNode = null; - oldNode.previousSibling = null; - oldNode.nextSibling = null; - if (oldNode.nodeType === Node.ELEMENT_NODE) { - oldNode.previousElementSibling = null; - oldNode.nextElementSibling = null; - } - return oldNode; - } - }, - - __JSDOMParser__: true, - }; - - for (var nodeType in nodeTypes) { - Node[nodeType] = Node.prototype[nodeType] = nodeTypes[nodeType]; - } - - var Attribute = function (name, value) { - this.name = name; - this._value = value; - }; - - Attribute.prototype = { - get value() { - return this._value; - }, - setValue: function(newValue) { - this._value = newValue; - }, - getEncodedValue: function() { - return encodeHTML(this._value); - }, - }; - - var Comment = function () { - this.childNodes = []; - }; - - Comment.prototype = { - __proto__: Node.prototype, - - nodeName: '#comment', - nodeType: Node.COMMENT_NODE, - }; - - var Text = function () { - this.childNodes = []; - }; - - Text.prototype = { - __proto__: Node.prototype, - - nodeName: '#text', - nodeType: Node.TEXT_NODE, - get textContent() { - if (typeof this._textContent === 'undefined') { - this._textContent = decodeHTML(this._innerHTML || ''); - } - return this._textContent; - }, - get innerHTML() { - if (typeof this._innerHTML === 'undefined') { - this._innerHTML = encodeTextContentHTML(this._textContent || ''); - } - return this._innerHTML; - }, - - set innerHTML(newHTML) { - this._innerHTML = newHTML; - delete this._textContent; - }, - set textContent(newText) { - this._textContent = newText; - delete this._innerHTML; - }, - }; - - var Document = function (url) { - this.documentURI = url; - this.styleSheets = []; - this.childNodes = []; - this.children = []; - }; - - Document.prototype = { - __proto__: Node.prototype, - - nodeName: '#document', - nodeType: Node.DOCUMENT_NODE, - title: '', - - getElementsByTagName: getElementsByTagName, - - getElementById: function (id) { - function getElem(node) { - var length = node.children.length; - if (node.id === id) - return node; - for (var i = 0; i < length; i++) { - var el = getElem(node.children[i]); - if (el) - return el; + // Hard way: + newNode.previousElementSibling = (function() { + for (var i = childIndex - 1; i >= 0; i--) { + if (childNodes[i].nodeType === Node.ELEMENT_NODE) + return childNodes[i]; } return null; + })(); + if (newNode.previousElementSibling) { + newNode.nextElementSibling = newNode.previousElementSibling.nextElementSibling; + } else { + newNode.nextElementSibling = (function() { + for (var i = childIndex + 1; i < childNodes.length; i++) { + if (childNodes[i].nodeType === Node.ELEMENT_NODE) + return childNodes[i]; + } + return null; + })(); + } + if (newNode.previousElementSibling) + newNode.previousElementSibling.nextElementSibling = newNode; + if (newNode.nextElementSibling) + newNode.nextElementSibling.previousElementSibling = newNode; + + if (newNode.nextElementSibling) + this.children.splice(this.children.indexOf(newNode.nextElementSibling), 0, newNode); + else + this.children.push(newNode); } - return getElem(this); - }, - - createElement: function (tag) { - var node = new Element(tag); - return node; - }, - - createTextNode: function (text) { - var node = new Text(); - node.textContent = text; - return node; - }, - - get baseURI() { - if (!this.hasOwnProperty('_baseURI')) { - this._baseURI = this.documentURI; - var baseElements = this.getElementsByTagName('base'); - var href = baseElements[0] && baseElements[0].getAttribute('href'); - if (href) { - try { - this._baseURI = (new URL(href, this._baseURI)).href; - } catch (ex) {/* Just fall back to documentURI */} - } - } - return this._baseURI; - }, - }; - - var Element = function (tag) { - // We use this to find the closing tag. - this._matchingTag = tag; - // We're explicitly a non-namespace aware parser, we just pretend it's all HTML. - var lastColonIndex = tag.lastIndexOf(':'); - if (lastColonIndex != -1) { - tag = tag.substring(lastColonIndex + 1); + } else if (oldNode.nodeType === Node.ELEMENT_NODE) { + // new node is not an element node. + // if the old one was, update its element siblings: + if (oldNode.previousElementSibling) + oldNode.previousElementSibling.nextElementSibling = oldNode.nextElementSibling; + if (oldNode.nextElementSibling) + oldNode.nextElementSibling.previousElementSibling = oldNode.previousElementSibling; + this.children.splice(this.children.indexOf(oldNode), 1); + + // If the old node wasn't an element, neither the new nor the old node was an element, + // and the children array and its members shouldn't need any updating. + } + + + oldNode.parentNode = null; + oldNode.previousSibling = null; + oldNode.nextSibling = null; + if (oldNode.nodeType === Node.ELEMENT_NODE) { + oldNode.previousElementSibling = null; + oldNode.nextElementSibling = null; + } + return oldNode; } - this.attributes = []; - this.childNodes = []; - this.children = []; - this.nextElementSibling = this.previousElementSibling = null; - this.localName = tag.toLowerCase(); - this.tagName = tag.toUpperCase(); - this.style = new Style(this); + }, + + __JSDOMParser__: true, }; - + + for (var nodeType in nodeTypes) { + Node[nodeType] = Node.prototype[nodeType] = nodeTypes[nodeType]; + } + + var Attribute = function (name, value) { + this.name = name; + this._value = value; + }; + + Attribute.prototype = { + get value() { + return this._value; + }, + setValue: function(newValue) { + this._value = newValue; + }, + getEncodedValue: function() { + return encodeHTML(this._value); + }, + }; + + var Comment = function () { + this.childNodes = []; + }; + + Comment.prototype = { + __proto__: Node.prototype, + + nodeName: "#comment", + nodeType: Node.COMMENT_NODE + }; + + var Text = function () { + this.childNodes = []; + }; + + Text.prototype = { + __proto__: Node.prototype, + + nodeName: "#text", + nodeType: Node.TEXT_NODE, + get textContent() { + if (typeof this._textContent === "undefined") { + this._textContent = decodeHTML(this._innerHTML || ""); + } + return this._textContent; + }, + get innerHTML() { + if (typeof this._innerHTML === "undefined") { + this._innerHTML = encodeTextContentHTML(this._textContent || ""); + } + return this._innerHTML; + }, + + set innerHTML(newHTML) { + this._innerHTML = newHTML; + delete this._textContent; + }, + set textContent(newText) { + this._textContent = newText; + delete this._innerHTML; + }, + }; + + var Document = function (url) { + this.documentURI = url; + this.styleSheets = []; + this.childNodes = []; + this.children = []; + }; + + Document.prototype = { + __proto__: Node.prototype, + + nodeName: "#document", + nodeType: Node.DOCUMENT_NODE, + title: "", + + getElementsByTagName: getElementsByTagName, + + getElementById: function (id) { + function getElem(node) { + var length = node.children.length; + if (node.id === id) + return node; + for (var i = 0; i < length; i++) { + var el = getElem(node.children[i]); + if (el) + return el; + } + return null; + } + return getElem(this); + }, + + createElement: function (tag) { + var node = new Element(tag); + return node; + }, + + createTextNode: function (text) { + var node = new Text(); + node.textContent = text; + return node; + }, + + get baseURI() { + if (!this.hasOwnProperty("_baseURI")) { + this._baseURI = this.documentURI; + var baseElements = this.getElementsByTagName("base"); + var href = baseElements[0] && baseElements[0].getAttribute("href"); + if (href) { + try { + this._baseURI = (new URL(href, this._baseURI)).href; + } catch (ex) {/* Just fall back to documentURI */} + } + } + return this._baseURI; + }, + }; + + var Element = function (tag) { + // We use this to find the closing tag. + this._matchingTag = tag; + // We're explicitly a non-namespace aware parser, we just pretend it's all HTML. + var lastColonIndex = tag.lastIndexOf(":"); + if (lastColonIndex != -1) { + tag = tag.substring(lastColonIndex + 1); + } + this.attributes = []; + this.childNodes = []; + this.children = []; + this.nextElementSibling = this.previousElementSibling = null; + this.localName = tag.toLowerCase(); + this.tagName = tag.toUpperCase(); + this.style = new Style(this); + }; + Element.prototype = { - __proto__: Node.prototype, - - nodeType: Node.ELEMENT_NODE, - - getElementsByTagName: getElementsByTagName, - - get className() { - return this.getAttribute('class') || ''; - }, - - set className(str) { - this.setAttribute('class', str); - }, - - get id() { - return this.getAttribute('id') || ''; - }, - - set id(str) { - this.setAttribute('id', str); - }, - - get href() { - return this.getAttribute('href') || ''; - }, - - set href(str) { - this.setAttribute('href', str); - }, - - get src() { - return this.getAttribute('src') || ''; - }, - - set src(str) { - this.setAttribute('src', str); - }, - - get srcset() { - return this.getAttribute('srcset') || ''; - }, - - set srcset(str) { - this.setAttribute('srcset', str); - }, - - get nodeName() { - return this.tagName; - }, - - get innerHTML() { - function getHTML(node) { - var i = 0; - for (i = 0; i < node.childNodes.length; i++) { - var child = node.childNodes[i]; - if (child.localName) { - arr.push('<' + child.localName); - - // serialize attribute list - for (var j = 0; j < child.attributes.length; j++) { - var attr = child.attributes[j]; - // the attribute value will be HTML escaped. - var val = attr.getEncodedValue(); - var quote = (val.indexOf('"') === -1 ? '"' : '\''); - arr.push(' ' + attr.name + '=' + quote + val + quote); - } - - if (child.localName in voidElems && !child.childNodes.length) { - // if this is a self-closing element, end it here - arr.push('/>'); - } else { - // otherwise, add its children - arr.push('>'); - getHTML(child); - arr.push(''); - } - } else { - // This is a text node, so asking for innerHTML won't recurse. - arr.push(child.innerHTML); - } - } + __proto__: Node.prototype, + + nodeType: Node.ELEMENT_NODE, + + getElementsByTagName: getElementsByTagName, + + get className() { + return this.getAttribute("class") || ""; + }, + + set className(str) { + this.setAttribute("class", str); + }, + + get id() { + return this.getAttribute("id") || ""; + }, + + set id(str) { + this.setAttribute("id", str); + }, + + get href() { + return this.getAttribute("href") || ""; + }, + + set href(str) { + this.setAttribute("href", str); + }, + + get src() { + return this.getAttribute("src") || ""; + }, + + set src(str) { + this.setAttribute("src", str); + }, + + get srcset() { + return this.getAttribute("srcset") || ""; + }, + + set srcset(str) { + this.setAttribute("srcset", str); + }, + + get nodeName() { + return this.tagName; + }, + + get innerHTML() { + function getHTML(node) { + var i = 0; + for (i = 0; i < node.childNodes.length; i++) { + var child = node.childNodes[i]; + if (child.localName) { + arr.push("<" + child.localName); + + // serialize attribute list + for (var j = 0; j < child.attributes.length; j++) { + var attr = child.attributes[j]; + // the attribute value will be HTML escaped. + var val = attr.getEncodedValue(); + var quote = (val.indexOf('"') === -1 ? '"' : "'"); + arr.push(" " + attr.name + "=" + quote + val + quote); + } + + if (child.localName in voidElems && !child.childNodes.length) { + // if this is a self-closing element, end it here + arr.push("/>"); + } else { + // otherwise, add its children + arr.push(">"); + getHTML(child); + arr.push(""); + } + } else { + // This is a text node, so asking for innerHTML won't recurse. + arr.push(child.innerHTML); } - - // Using Array.join() avoids the overhead from lazy string concatenation. - // See http://blog.cdleary.com/2012/01/string-representation-in-spidermonkey/#ropes - var arr = []; - getHTML(this); - return arr.join(''); - }, - - set innerHTML(html) { - var parser = new JSDOMParser(); - var node = parser.parse(html); - var i; - for (i = this.childNodes.length; --i >= 0;) { - this.childNodes[i].parentNode = null; + } + } + + // Using Array.join() avoids the overhead from lazy string concatenation. + var arr = []; + getHTML(this); + return arr.join(""); + }, + + set innerHTML(html) { + var parser = new JSDOMParser(); + var node = parser.parse(html); + var i; + for (i = this.childNodes.length; --i >= 0;) { + this.childNodes[i].parentNode = null; + } + this.childNodes = node.childNodes; + this.children = node.children; + for (i = this.childNodes.length; --i >= 0;) { + this.childNodes[i].parentNode = this; + } + }, + + set textContent(text) { + // clear parentNodes for existing children + for (var i = this.childNodes.length; --i >= 0;) { + this.childNodes[i].parentNode = null; + } + + var node = new Text(); + this.childNodes = [ node ]; + this.children = []; + node.textContent = text; + node.parentNode = this; + }, + + get textContent() { + function getText(node) { + var nodes = node.childNodes; + for (var i = 0; i < nodes.length; i++) { + var child = nodes[i]; + if (child.nodeType === 3) { + text.push(child.textContent); + } else { + getText(child); } - this.childNodes = node.childNodes; - this.children = node.children; - for (i = this.childNodes.length; --i >= 0;) { - this.childNodes[i].parentNode = this; - } - }, - - set textContent(text) { - // clear parentNodes for existing children - for (var i = this.childNodes.length; --i >= 0;) { - this.childNodes[i].parentNode = null; - } - - var node = new Text(); - this.childNodes = [ node ]; - this.children = []; - node.textContent = text; - node.parentNode = this; - }, - - get textContent() { - function getText(node) { - var nodes = node.childNodes; - for (var i = 0; i < nodes.length; i++) { - var child = nodes[i]; - if (child.nodeType === 3) { - text.push(child.textContent); - } else { - getText(child); - } - } - } - - // Using Array.join() avoids the overhead from lazy string concatenation. - // See http://blog.cdleary.com/2012/01/string-representation-in-spidermonkey/#ropes - var text = []; - getText(this); - return text.join(''); - }, - - getAttribute: function (name) { - for (var i = this.attributes.length; --i >= 0;) { - var attr = this.attributes[i]; - if (attr.name === name) { - return attr.value; - } - } - return undefined; - }, - - setAttribute: function (name, value) { - for (var i = this.attributes.length; --i >= 0;) { - var attr = this.attributes[i]; - if (attr.name === name) { - attr.setValue(value); - return; - } - } - this.attributes.push(new Attribute(name, value)); - }, - - removeAttribute: function (name) { - for (var i = this.attributes.length; --i >= 0;) { - var attr = this.attributes[i]; - if (attr.name === name) { - this.attributes.splice(i, 1); - break; - } - } - }, - - hasAttribute: function (name) { - return this.attributes.some(function (attr) { - return attr.name == name; - }); - }, + } + } + + // Using Array.join() avoids the overhead from lazy string concatenation. + // See http://blog.cdleary.com/2012/01/string-representation-in-spidermonkey/#ropes + var text = []; + getText(this); + return text.join(""); + }, + + getAttribute: function (name) { + for (var i = this.attributes.length; --i >= 0;) { + var attr = this.attributes[i]; + if (attr.name === name) { + return attr.value; + } + } + return undefined; + }, + + setAttribute: function (name, value) { + for (var i = this.attributes.length; --i >= 0;) { + var attr = this.attributes[i]; + if (attr.name === name) { + attr.setValue(value); + return; + } + } + this.attributes.push(new Attribute(name, value)); + }, + + removeAttribute: function (name) { + for (var i = this.attributes.length; --i >= 0;) { + var attr = this.attributes[i]; + if (attr.name === name) { + this.attributes.splice(i, 1); + break; + } + } + }, + + hasAttribute: function (name) { + return this.attributes.some(function (attr) { + return attr.name == name; + }); + }, }; - + var Style = function (node) { - this.node = node; + this.node = node; }; - + // getStyle() and setStyle() use the style attribute string directly. This // won't be very efficient if there are a lot of style manipulations, but // it's the easiest way to make sure the style attribute string and the JS // style property stay in sync. Readability.js doesn't do many style // manipulations, so this should be okay. Style.prototype = { - getStyle: function (styleName) { - var attr = this.node.getAttribute('style'); - if (!attr) - return undefined; - - var styles = attr.split(';'); - for (var i = 0; i < styles.length; i++) { - var style = styles[i].split(':'); - var name = style[0].trim(); - if (name === styleName) - return style[1].trim(); - } - - return undefined; - }, - - setStyle: function (styleName, styleValue) { - var value = this.node.getAttribute('style') || ''; - var index = 0; - do { - var next = value.indexOf(';', index) + 1; - var length = next - index - 1; - var style = (length > 0 ? value.substr(index, length) : value.substr(index)); - if (style.substr(0, style.indexOf(':')).trim() === styleName) { - value = value.substr(0, index).trim() + (next ? ' ' + value.substr(next).trim() : ''); - break; - } - index = next; - } while (index); - - value += ' ' + styleName + ': ' + styleValue + ';'; - this.node.setAttribute('style', value.trim()); - }, + getStyle: function (styleName) { + var attr = this.node.getAttribute("style"); + if (!attr) + return undefined; + + var styles = attr.split(";"); + for (var i = 0; i < styles.length; i++) { + var style = styles[i].split(":"); + var name = style[0].trim(); + if (name === styleName) + return style[1].trim(); + } + + return undefined; + }, + + setStyle: function (styleName, styleValue) { + var value = this.node.getAttribute("style") || ""; + var index = 0; + do { + var next = value.indexOf(";", index) + 1; + var length = next - index - 1; + var style = (length > 0 ? value.substr(index, length) : value.substr(index)); + if (style.substr(0, style.indexOf(":")).trim() === styleName) { + value = value.substr(0, index).trim() + (next ? " " + value.substr(next).trim() : ""); + break; + } + index = next; + } while (index); + + value += " " + styleName + ": " + styleValue + ";"; + this.node.setAttribute("style", value.trim()); + } }; - + // For each item in styleMap, define a getter and setter on the style // property. for (var jsName in styleMap) { - (function (cssName) { - Style.prototype.__defineGetter__(jsName, function () { - return this.getStyle(cssName); - }); - Style.prototype.__defineSetter__(jsName, function (value) { - this.setStyle(cssName, value); - }); - })(styleMap[jsName]); + (function (cssName) { + Style.prototype.__defineGetter__(jsName, function () { + return this.getStyle(cssName); + }); + Style.prototype.__defineSetter__(jsName, function (value) { + this.setStyle(cssName, value); + }); + })(styleMap[jsName]); } - + var JSDOMParser = function () { - this.currentChar = 0; - - // In makeElementNode() we build up many strings one char at a time. Using - // += for this results in lots of short-lived intermediate strings. It's - // better to build an array of single-char strings and then join() them - // together at the end. And reusing a single array (i.e. |this.strBuf|) - // over and over for this purpose uses less memory than using a new array - // for each string. - this.strBuf = []; - - // Similarly, we reuse this array to return the two arguments from - // makeElementNode(), which saves us from having to allocate a new array - // every time. - this.retPair = []; - - this.errorState = ''; + this.currentChar = 0; + + // In makeElementNode() we build up many strings one char at a time. Using + // += for this results in lots of short-lived intermediate strings. It's + // better to build an array of single-char strings and then join() them + // together at the end. And reusing a single array (i.e. |this.strBuf|) + // over and over for this purpose uses less memory than using a new array + // for each string. + this.strBuf = []; + + // Similarly, we reuse this array to return the two arguments from + // makeElementNode(), which saves us from having to allocate a new array + // every time. + this.retPair = []; + + this.errorState = ""; }; - + JSDOMParser.prototype = { - error: function(m) { - dump('JSDOMParser error: ' + m + '\n'); - this.errorState += m + '\n'; - }, - - /** - * Look at the next character without advancing the index. - */ - peekNext: function () { - return this.html[this.currentChar]; - }, - - /** - * Get the next character and advance the index. - */ - nextChar: function () { - return this.html[this.currentChar++]; - }, - - /** - * Called after a quote character is read. This finds the next quote - * character and returns the text string in between. - */ - readString: function (quote) { - var str; - var n = this.html.indexOf(quote, this.currentChar); - if (n === -1) { - this.currentChar = this.html.length; - str = null; - } else { - str = this.html.substring(this.currentChar, n); - this.currentChar = n + 1; - } - - return str; - }, - - /** - * Called when parsing a node. This finds the next name/value attribute - * pair and adds the result to the attributes list. - */ - readAttribute: function (node) { - var name = ''; - - var n = this.html.indexOf('=', this.currentChar); - if (n === -1) { - this.currentChar = this.html.length; - } else { - // Read until a '=' character is hit; this will be the attribute key - name = this.html.substring(this.currentChar, n); - this.currentChar = n + 1; - } - - if (!name) - return; - - // After a '=', we should see a '"' for the attribute value - var c = this.nextChar(); - if (c !== '"' && c !== '\'') { - this.error('Error reading attribute ' + name + ', expecting \'"\''); - return; - } - - // Read the attribute value (and consume the matching quote) - var value = this.readString(c); - - node.attributes.push(new Attribute(name, decodeHTML(value))); - - return; - }, - - /** - * Parses and returns an Element node. This is called after a '<' has been - * read. - * - * @returns an array; the first index of the array is the parsed node; - * the second index is a boolean indicating whether this is a void - * Element - */ - makeElementNode: function (retPair) { - var c = this.nextChar(); - - // Read the Element tag name - var strBuf = this.strBuf; - strBuf.length = 0; - while (whitespace.indexOf(c) == -1 && c !== '>' && c !== '/') { - if (c === undefined) - return false; - strBuf.push(c); - c = this.nextChar(); - } - var tag = strBuf.join(''); - - if (!tag) - return false; - - var node = new Element(tag); - - // Read Element attributes - while (c !== '/' && c !== '>') { - if (c === undefined) - return false; - while (whitespace.indexOf(this.html[this.currentChar++]) != -1) { - // Advance cursor to first non-whitespace char. - } - this.currentChar--; - c = this.nextChar(); - if (c !== '/' && c !== '>') { - --this.currentChar; - this.readAttribute(node); - } - } - - // If this is a self-closing tag, read '/>' - var closed = false; - if (c === '/') { - closed = true; - c = this.nextChar(); - if (c !== '>') { - this.error('expected \'>\' to close ' + tag); - return false; - } - } - - retPair[0] = node; - retPair[1] = closed; - return true; - }, - - /** - * If the current input matches this string, advance the input index; - * otherwise, do nothing. - * - * @returns whether input matched string - */ - match: function (str) { - var strlen = str.length; - if (this.html.substr(this.currentChar, strlen).toLowerCase() === str.toLowerCase()) { - this.currentChar += strlen; - return true; - } + error: function(m) { + if (typeof dump !== "undefined") { + dump("JSDOMParser error: " + m + "\n"); + } else if (typeof console !== "undefined") { + console.log("JSDOMParser error: " + m + "\n"); + } + this.errorState += m + "\n"; + }, + + /** + * Look at the next character without advancing the index. + */ + peekNext: function () { + return this.html[this.currentChar]; + }, + + /** + * Get the next character and advance the index. + */ + nextChar: function () { + return this.html[this.currentChar++]; + }, + + /** + * Called after a quote character is read. This finds the next quote + * character and returns the text string in between. + */ + readString: function (quote) { + var str; + var n = this.html.indexOf(quote, this.currentChar); + if (n === -1) { + this.currentChar = this.html.length; + str = null; + } else { + str = this.html.substring(this.currentChar, n); + this.currentChar = n + 1; + } + + return str; + }, + + /** + * Called when parsing a node. This finds the next name/value attribute + * pair and adds the result to the attributes list. + */ + readAttribute: function (node) { + var name = ""; + + var n = this.html.indexOf("=", this.currentChar); + if (n === -1) { + this.currentChar = this.html.length; + } else { + // Read until a '=' character is hit; this will be the attribute key + name = this.html.substring(this.currentChar, n); + this.currentChar = n + 1; + } + + if (!name) + return; + + // After a '=', we should see a '"' for the attribute value + var c = this.nextChar(); + if (c !== '"' && c !== "'") { + this.error("Error reading attribute " + name + ", expecting '\"'"); + return; + } + + // Read the attribute value (and consume the matching quote) + var value = this.readString(c); + + node.attributes.push(new Attribute(name, decodeHTML(value))); + + return; + }, + + /** + * Parses and returns an Element node. This is called after a '<' has been + * read. + * + * @returns an array; the first index of the array is the parsed node; + * the second index is a boolean indicating whether this is a void + * Element + */ + makeElementNode: function (retPair) { + var c = this.nextChar(); + + // Read the Element tag name + var strBuf = this.strBuf; + strBuf.length = 0; + while (whitespace.indexOf(c) == -1 && c !== ">" && c !== "/") { + if (c === undefined) return false; - }, - - /** - * Searches the input until a string is found and discards all input up to - * and including the matched string. - */ - discardTo: function (str) { - var index = this.html.indexOf(str, this.currentChar) + str.length; - if (index === -1) - this.currentChar = this.html.length; - this.currentChar = index; - }, - - /** - * Reads child nodes for the given node. - */ - readChildren: function (node) { - var child; - while ((child = this.readNode())) { - // Don't keep Comment nodes - if (child.nodeType !== 8) { - node.appendChild(child); - } - } - }, - - discardNextComment: function() { - if (this.match('--')) { - this.discardTo('-->'); - } else { - var c = this.nextChar(); - while (c !== '>') { - if (c === undefined) - return null; - if (c === '"' || c === '\'') - this.readString(c); - c = this.nextChar(); - } - } - return new Comment(); - }, - - - /** - * Reads the next child node from the input. If we're reading a closing - * tag, or if we've reached the end of input, return null. - * - * @returns the node - */ - readNode: function () { - var c = this.nextChar(); - + strBuf.push(c); + c = this.nextChar(); + } + var tag = strBuf.join(""); + + if (!tag) + return false; + + var node = new Element(tag); + + // Read Element attributes + while (c !== "/" && c !== ">") { + if (c === undefined) + return false; + while (whitespace.indexOf(this.html[this.currentChar++]) != -1) { + // Advance cursor to first non-whitespace char. + } + this.currentChar--; + c = this.nextChar(); + if (c !== "/" && c !== ">") { + --this.currentChar; + this.readAttribute(node); + } + } + + // If this is a self-closing tag, read '/>' + var closed = false; + if (c === "/") { + closed = true; + c = this.nextChar(); + if (c !== ">") { + this.error("expected '>' to close " + tag); + return false; + } + } + + retPair[0] = node; + retPair[1] = closed; + return true; + }, + + /** + * If the current input matches this string, advance the input index; + * otherwise, do nothing. + * + * @returns whether input matched string + */ + match: function (str) { + var strlen = str.length; + if (this.html.substr(this.currentChar, strlen).toLowerCase() === str.toLowerCase()) { + this.currentChar += strlen; + return true; + } + return false; + }, + + /** + * Searches the input until a string is found and discards all input up to + * and including the matched string. + */ + discardTo: function (str) { + var index = this.html.indexOf(str, this.currentChar) + str.length; + if (index === -1) + this.currentChar = this.html.length; + this.currentChar = index; + }, + + /** + * Reads child nodes for the given node. + */ + readChildren: function (node) { + var child; + while ((child = this.readNode())) { + // Don't keep Comment nodes + if (child.nodeType !== 8) { + node.appendChild(child); + } + } + }, + + discardNextComment: function() { + if (this.match("--")) { + this.discardTo("-->"); + } else { + var c = this.nextChar(); + while (c !== ">") { if (c === undefined) - return null; - - // Read any text as Text node - var textNode; - if (c !== '<') { - --this.currentChar; - textNode = new Text(); - var n = this.html.indexOf('<', this.currentChar); - if (n === -1) { - textNode.innerHTML = this.html.substring(this.currentChar, this.html.length); - this.currentChar = this.html.length; - } else { - textNode.innerHTML = this.html.substring(this.currentChar, n); - this.currentChar = n; - } - return textNode; + return null; + if (c === '"' || c === "'") + this.readString(c); + c = this.nextChar(); + } + } + return new Comment(); + }, + + + /** + * Reads the next child node from the input. If we're reading a closing + * tag, or if we've reached the end of input, return null. + * + * @returns the node + */ + readNode: function () { + var c = this.nextChar(); + + if (c === undefined) + return null; + + // Read any text as Text node + var textNode; + if (c !== "<") { + --this.currentChar; + textNode = new Text(); + var n = this.html.indexOf("<", this.currentChar); + if (n === -1) { + textNode.innerHTML = this.html.substring(this.currentChar, this.html.length); + this.currentChar = this.html.length; + } else { + textNode.innerHTML = this.html.substring(this.currentChar, n); + this.currentChar = n; + } + return textNode; + } + + if (this.match("![CDATA[")) { + var endChar = this.html.indexOf("]]>", this.currentChar); + if (endChar === -1) { + this.error("unclosed CDATA section"); + return null; + } + textNode = new Text(); + textNode.textContent = this.html.substring(this.currentChar, endChar); + this.currentChar = endChar + ("]]>").length; + return textNode; + } + + c = this.peekNext(); + + // Read Comment node. Normally, Comment nodes know their inner + // textContent, but we don't really care about Comment nodes (we throw + // them away in readChildren()). So just returning an empty Comment node + // here is sufficient. + if (c === "!" || c === "?") { + // We're still before the ! or ? that is starting this comment: + this.currentChar++; + return this.discardNextComment(); + } + + // If we're reading a closing tag, return null. This means we've reached + // the end of this set of child nodes. + if (c === "/") { + --this.currentChar; + return null; + } + + // Otherwise, we're looking at an Element node + var result = this.makeElementNode(this.retPair); + if (!result) + return null; + + var node = this.retPair[0]; + var closed = this.retPair[1]; + var localName = node.localName; + + // If this isn't a void Element, read its child nodes + if (!closed) { + this.readChildren(node); + var closingTag = ""; + if (!this.match(closingTag)) { + this.error("expected '" + closingTag + "' and got " + this.html.substr(this.currentChar, closingTag.length)); + return null; + } + } + + // Only use the first title, because SVG might have other + // title elements which we don't care about (medium.com + // does this, at least). + if (localName === "title" && !this.doc.title) { + this.doc.title = node.textContent.trim(); + } else if (localName === "head") { + this.doc.head = node; + } else if (localName === "body") { + this.doc.body = node; + } else if (localName === "html") { + this.doc.documentElement = node; + } + + return node; + }, + + /** + * Parses an HTML string and returns a JS implementation of the Document. + */ + parse: function (html, url) { + this.html = html; + var doc = this.doc = new Document(url); + this.readChildren(doc); + + // If this is an HTML document, remove root-level children except for the + // node + if (doc.documentElement) { + for (var i = doc.childNodes.length; --i >= 0;) { + var child = doc.childNodes[i]; + if (child !== doc.documentElement) { + doc.removeChild(child); } - - if (this.match('![CDATA[')) { - var endChar = this.html.indexOf(']]>', this.currentChar); - if (endChar === -1) { - this.error('unclosed CDATA section'); - return null; - } - textNode = new Text(); - textNode.textContent = this.html.substring(this.currentChar, endChar); - this.currentChar = endChar + (']]>').length; - return textNode; - } - - c = this.peekNext(); - - // Read Comment node. Normally, Comment nodes know their inner - // textContent, but we don't really care about Comment nodes (we throw - // them away in readChildren()). So just returning an empty Comment node - // here is sufficient. - if (c === '!' || c === '?') { - // We're still before the ! or ? that is starting this comment: - this.currentChar++; - return this.discardNextComment(); - } - - // If we're reading a closing tag, return null. This means we've reached - // the end of this set of child nodes. - if (c === '/') { - --this.currentChar; - return null; - } - - // Otherwise, we're looking at an Element node - var result = this.makeElementNode(this.retPair); - if (!result) - return null; - - var node = this.retPair[0]; - var closed = this.retPair[1]; - var localName = node.localName; - - // If this isn't a void Element, read its child nodes - if (!closed) { - this.readChildren(node); - var closingTag = ''; - if (!this.match(closingTag)) { - this.error('expected \'' + closingTag + '\' and got ' + this.html.substr(this.currentChar, closingTag.length)); - return null; - } - } - - // Only use the first title, because SVG might have other - // title elements which we don't care about (medium.com - // does this, at least). - if (localName === 'title' && !this.doc.title) { - this.doc.title = node.textContent.trim(); - } else if (localName === 'head') { - this.doc.head = node; - } else if (localName === 'body') { - this.doc.body = node; - } else if (localName === 'html') { - this.doc.documentElement = node; - } - - return node; - }, - - /** - * Parses an HTML string and returns a JS implementation of the Document. - */ - parse: function (html, url) { - this.html = html; - var doc = this.doc = new Document(url); - this.readChildren(doc); - - // If this is an HTML document, remove root-level children except for the - // node - if (doc.documentElement) { - for (var i = doc.childNodes.length; --i >= 0;) { - var child = doc.childNodes[i]; - if (child !== doc.documentElement) { - doc.removeChild(child); - } - } - } - - return doc; - }, + } + } + + return doc; + } }; - + // Attach the standard DOM types to the global scope global.Node = Node; global.Comment = Comment; global.Document = Document; global.Element = Element; global.Text = Text; - + // Attach JSDOMParser to the global scope global.JSDOMParser = JSDOMParser; - -})(this); + + })(this); + + if (typeof module === "object") { + module.exports = this.JSDOMParser; + } \ No newline at end of file diff --git a/packages/app-clipper/content_scripts/README.md b/packages/app-clipper/content_scripts/README.md new file mode 100644 index 0000000000..b957c81399 --- /dev/null +++ b/packages/app-clipper/content_scripts/README.md @@ -0,0 +1,11 @@ +# Updating Readability + +Because of the way content scripts are loaded, we need to manually copy the whole Readability files here. That should be fine since they rarely change. + +The files to update are: + +- Readability.js +- Readability-readerable.js (for the function isProbablyReaderable) +- JSDOMParser.js + +When updating, **make sure you add the commit and version number at the top of each file**. diff --git a/packages/app-clipper/content_scripts/Readability-readerable.js b/packages/app-clipper/content_scripts/Readability-readerable.js index 0e4e8f9904..a4ac2ade34 100644 --- a/packages/app-clipper/content_scripts/Readability-readerable.js +++ b/packages/app-clipper/content_scripts/Readability-readerable.js @@ -1,7 +1,6 @@ -// https://github.com/mozilla/readability/tree/814f0a3884350b6f1adfdebb79ca3599e9806605 +// v0.4.1 - https://github.com/mozilla/readability/commit/28843b6de84447dd6cef04058fda336938e628dc /* eslint-env es6:false */ -/* globals exports */ /* * Copyright (c) 2010 Arc90 Inc * @@ -26,27 +25,38 @@ var REGEXPS = { // NOTE: These two regular expressions are duplicated in // Readability.js. Please keep both copies in sync. - unlikelyCandidates: /-ad-|ai2html|banner|breadcrumbs|combx|comment|community|cover-wrap|disqus|extra|foot|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|main|shadow/i, -}; - -function isNodeVisible(node) { - // Have to null-check node.style to deal with SVG and MathML nodes. - return (!node.style || node.style.display != 'none') && !node.hasAttribute('hidden'); -} - -/** - * Decides whether or not the document is reader-able without parsing the whole thing. - * - * @return boolean Whether or not we suspect Readability.parse() will suceeed at returning an article object. - */ -function isProbablyReaderable(doc, isVisible) { - if (!isVisible) { - isVisible = isNodeVisible; + 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) { + // Have to null-check node.style and node.className.indexOf to deal with SVG and MathML nodes. + return (!node.style || node.style.display != "none") + && !node.hasAttribute("hidden") + //check for "fallback-image" so that wikimedia math images are displayed + && (!node.hasAttribute("aria-hidden") || node.getAttribute("aria-hidden") != "true" || (node.className && node.className.indexOf && node.className.indexOf("fallback-image") !== -1)); + } + + /** + * Decides whether or not the document is reader-able without parsing the whole thing. + * @param {Object} options Configuration object. + * @param {number} [options.minContentLength=140] The minimum node content length used to decide if the document is readerable. + * @param {number} [options.minScore=20] The minumum cumulated 'score' used to determine if the document is readerable. + * @param {Function} [options.visibilityChecker=isNodeVisible] The function used to determine if a node is visible. + * @return {boolean} Whether or not we suspect Readability.parse() will suceeed at returning an article object. + */ + function isProbablyReaderable(doc, options = {}) { + // For backward compatibility reasons 'options' can either be a configuration object or the function used + // to determine if a node is visible. + if (typeof options == "function") { + options = { visibilityChecker: options }; } - - var nodes = doc.querySelectorAll('p, pre'); - + + var defaultOptions = { minScore: 20, minContentLength: 140, visibilityChecker: isNodeVisible }; + options = Object.assign(defaultOptions, options); + + var nodes = doc.querySelectorAll("p, pre"); + // Get
nodes which have
node(s) and append them into the `nodes` variable. // Some articles' DOM structures might look like //
@@ -54,46 +64,47 @@ function isProbablyReaderable(doc, isVisible) { //
// Sentences
//
- var brNodes = doc.querySelectorAll('div > br'); + 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 set = new Set(nodes); + [].forEach.call(brNodes, function (node) { + set.add(node.parentNode); + }); + nodes = Array.from(set); } - + var score = 0; // This is a little cheeky, we use the accumulator 'score' to decide what to return from // this callback: - return [].some.call(nodes, function(node) { - if (!isVisible(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 < 140) { - return false; - } - - score += Math.sqrt(textContentLength - 140); - - if (score > 20) { - return true; - } + 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 exports === 'object') { - exports.isProbablyReaderable = isProbablyReaderable; -} + } + + if (typeof module === "object") { + module.exports = isProbablyReaderable; + } \ No newline at end of file diff --git a/packages/app-clipper/content_scripts/Readability.js b/packages/app-clipper/content_scripts/Readability.js index db5ab6ce07..6edbca1701 100644 --- a/packages/app-clipper/content_scripts/Readability.js +++ b/packages/app-clipper/content_scripts/Readability.js @@ -1,4 +1,4 @@ -// https://github.com/mozilla/readability/tree/814f0a3884350b6f1adfdebb79ca3599e9806605 +// v0.4.1 - https://github.com/mozilla/readability/commit/28843b6de84447dd6cef04058fda336938e628dc /*eslint-env es6:false*/ /* @@ -27,1827 +27,2237 @@ * @param {HTMLDocument} doc The document to parse. * @param {Object} options The options object. */ -function Readability(doc, options) { + 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]; + doc = options; + options = arguments[2]; } else if (!doc || !doc.documentElement) { - throw new Error('First argument to Readability constructor should be a document object.'); + 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; - - var logEl; - + this.FLAG_WEIGHT_CLASSES | + this.FLAG_CLEAN_CONDITIONALLY; + + // Control whether log messages are sent to the console if (this._debug) { - logEl = function(e) { - var rv = e.nodeName + ' '; - if (e.nodeType == e.TEXT_NODE) { - return rv + '("' + e.textContent + '")'; + 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); } - var classDesc = e.className && ('.' + e.className.replace(/ /g, '.')); - var elDesc = ''; - if (e.id) - elDesc = '(#' + e.id + classDesc + ')'; - else if (classDesc) - elDesc = '(' + classDesc + ')'; - return rv + elDesc; - }; - this.log = function () { - if (typeof dump !== 'undefined') { - var msg = Array.prototype.map.call(arguments, function(x) { - return (x && x.nodeName) ? logEl(x) : x; - }).join(' '); - dump('Reader: (Readability) ' + msg + '\n'); - } else if (typeof console !== 'undefined') { - var args = ['Reader: (Readability) '].concat(arguments); - console.log.apply(console, args); - } - }; + return arg; + }); + args.unshift("Reader: (Readability)"); + console.log.apply(console, args); + } + }; } else { - this.log = function () {}; + this.log = function () {}; } -} - -Readability.prototype = { + } + + 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(','), - + 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|foot|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|main|shadow/i, - - positive: /article|body|content|entry|hentry|h-entry|main|page|pagination|post|text|blog|story/i, - negative: /hidden|^hid$| hid$| hid |^hid |banner|combx|comment|com-|contact|foot|footer|footnote|gdpr|masthead|media|meta|outbrain|promo|related|scroll|share|shoutbox|sidebar|skyscraper|sponsor|shopping|tags|tool|widget/i, - extraneous: /print|archive|comment|discuss|e[\-]?mail|share|reply|all|login|sign|single|utility/i, - byline: /byline|author|dateline|writtenby|p-author/i, - replaceFonts: /<(\/?)font[^>]*>/gi, - normalize: /\s{2,}/g, - videos: /\/\/(www\.)?((dailymotion|youtube|youtube-nocookie|player\.vimeo|v\.qq)\.com|(archive|upload\.wikimedia)\.org|player\.twitch\.tv)/i, - nextLink: /(next|weiter|continue|>([^\|]|$)|»([^\|]|$))/i, - prevLink: /(prev|earl|old|new|<|«)/i, - whitespace: /^\s*$/, - hasContent: /\S$/, + // NOTE: These two regular expressions are duplicated in + // Readability-readerable.js. Please keep both copies in sync. + unlikelyCandidates: /-ad-|ai2html|banner|breadcrumbs|combx|comment|community|cover-wrap|disqus|extra|footer|gdpr|header|legends|menu|related|remark|replies|rss|shoutbox|sidebar|skyscraper|social|sponsor|supplemental|ad-break|agegate|pagination|pager|popup|yom-remote/i, + okMaybeItsACandidate: /and|article|body|column|content|main|shadow/i, + + positive: /article|body|content|entry|hentry|h-entry|main|page|pagination|post|text|blog|story/i, + negative: /-ad-|hidden|^hid$| 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$/ }, - - DIV_TO_P_ELEMS: [ 'A', 'BLOCKQUOTE', 'DL', 'DIV', 'IMG', 'OL', 'P', 'PRE', 'TABLE', 'UL', 'SELECT' ], - - 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' ], - + + 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', + // "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' ], - + 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 - **/ + * 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); - + // 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 - */ + * 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) { - 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); - } - } + // 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 - */ + * 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) { - for (var i = nodeList.length - 1; i >= 0; i--) { - var node = nodeList[i]; - this._setNodeTag(node, 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 - */ + * 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); + Array.prototype.forEach.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 - */ + * 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); + 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 - */ + * 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); + return Array.prototype.every.call(nodeList, fn, this); }, - + /** - * Concat all nodelists passed as arguments. - * - * @return ...NodeList - * @return Array - */ + * 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); + 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); - })); + 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 - */ + * 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); - } + 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 and uri in the given element to an absolute URI, - * ignoring #ref URIs. - * - * @param Element - * @return void - */ + * Converts each and 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 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; } - - var links = this._getAllNodesWithTag(articleContent, ['a']); - this._forEachNode(links, function(link) { - var href = link.getAttribute('href'); - if (href) { - // Replace links with javascript: URIs with text content, since - // they won't work after scripts have been removed from the page. - if (href.indexOf('javascript:') === 0) { - var text = this._doc.createTextNode(link.textContent); - link.parentNode.replaceChild(text, link); - } else { - link.setAttribute('href', toAbsoluteURI(href)); - } - } - }); - - var imgs = this._getAllNodesWithTag(articleContent, ['img']); - this._forEachNode(imgs, function(img) { - var src = img.getAttribute('src'); - if (src) { - img.setAttribute('src', toAbsoluteURI(src)); - } - }); - }, - - /** - * Get the article title as an H1. - * - * @return void - **/ - _getArticleTitle: function() { - var doc = this._doc; - var curTitle = ''; - var origTitle = ''; - + + // Otherwise, resolve against base URI: 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; + return new URL(uri, baseURI).href; + } catch (ex) { + // Something went wrong, just return the original: } - - // 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(doc.getElementsByTagName('style')); - - if (doc.body) { - this._replaceBrs(doc.body); - } - - this._replaceNodeTags(doc.getElementsByTagName('font'), 'SPAN'); - }, - - /** - * Finds the next element, starting from the given node, and ignoring - * whitespace in between. If the given node is an element, the same node is - * returned. - */ - _nextElement: 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
elements with a single

. - * Whitespace between
elements are ignored. For example: - *

foo
bar


abc
- * will become: - *
foo
bar

abc

- */ - _replaceBrs: function (elem) { - this._forEachNode(this._getAllNodesWithTag(elem, ['br']), function(br) { - var next = br.nextSibling; - - // Whether 2 or more
elements have been found and replaced with a - //

block. - var replaced = false; - - // If we find a
chain, remove the
s until we hit another element - // or non-whitespace. This leaves behind the first
in the chain - // (which will be replaced with a

later). - while ((next = this._nextElement(next)) && (next.tagName == 'BR')) { - replaced = true; - var brSibling = next.nextSibling; - next.parentNode.removeChild(next); - next = brSibling; - } - - // If we removed a
chain, replace the remaining
with a

. Add - // all sibling nodes as children of the

until we hit another
- // chain. - if (replaced) { - var p = this._doc.createElement('p'); - br.parentNode.replaceChild(p, br); - - next = p.nextSibling; - while (next) { - // If we've hit another

, we're done adding children to this

. - if (next.tagName == 'BR') { - var nextElem = this._nextElement(next.nextSibling); - if (nextElem && nextElem.tagName == 'BR') - break; - } - - if (!this._isPhrasingContent(next)) - break; - - // Otherwise, make this node a child of the new

. - 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 (node.__JSDOMParser__) { - 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

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, 'h1'); - 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 /share/.test(matchString) && node.textContent.length < shareElementThreshold; - }); - }); - - // If there is only one h2 and its text content substantially equals article title, - // they are probably using it as a header and not a subheader, - // so remove it since we already extract the title separately. - var h2 = articleContent.getElementsByTagName('h2'); - if (h2.length === 1) { - var lengthSimilarRate = (h2[0].textContent.length - this._articleTitle.length) / this._articleTitle.length; - if (Math.abs(lengthSimilarRate) < 0.5) { - var titlesMatch = false; - if (lengthSimilarRate > 0) { - titlesMatch = h2[0].textContent.includes(this._articleTitle); - } else { - titlesMatch = this._articleTitle.includes(h2[0].textContent); - } - if (titlesMatch) { - this._clean(articleContent, 'h2'); - } - } - } - - 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'); - - // Remove extra paragraphs - this._removeNodes(articleContent.getElementsByTagName('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._nextElement(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; - }, - - _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 ? true: false); - 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) { - 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; - - while (node) { - 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; - } - - // Remove unlikely candidates - if (stripUnlikelyCandidates) { - if (this.REGEXPS.unlikelyCandidates.test(matchString) && - !this.REGEXPS.okMaybeItsACandidate.test(matchString) && - !this._hasAncestorTag(node, 'table') && - node.tagName !== 'BODY' && - node.tagName !== 'A') { - this.log('Removing unlikely candidate - ' + matchString); - node = this._removeAndGetNext(node); - continue; - } - } - - // Remove DIV, SECTION, and HEADER nodes without any content(e.g. text, image, video, or iframe). - 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.indexOf(node.tagName) !== -1) { - elementsToScore.push(node); - } - - // Turn all divs that don't have children block level elements into p's - if (node.tagName === 'DIV') { - // Put phrasing content into paragraphs. - 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.removeChild(p.lastChild); - } - p = null; - } - childNode = nextSibling; - } - - // Sites like http://mobile.slate.com encloses each paragraph with a DIV - // element. DIVs with only a P element inside and no text content can be - // safely converted into plain P elements to avoid confusing the scoring - // algorithm with DIVs with are, in practice, paragraphs. - 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); - } - - /** - * Loop through all paragraphs, and assign a score to them based on how content-y they look. - * Then add their score to their parent node. - * - * A score is determined by things like number of commas, class names, etc. Maybe eventually link density. - **/ - var candidates = []; - this._forEachNode(elementsToScore, function(elementToScore) { - if (!elementToScore.parentNode || typeof(elementToScore.parentNode.tagName) === 'undefined') - return; - - // If this paragraph is less than 25 characters, don't even count it. - var innerText = this._getInnerText(elementToScore); - if (innerText.length < 25) - return; - - // Exclude nodes with no ancestor. - var ancestors = this._getNodeAncestors(elementToScore, 3); - if (ancestors.length === 0) - return; - - var contentScore = 0; - - // Add a point for the paragraph itself as a base. - contentScore += 1; - - // Add points for any commas within this paragraph. - contentScore += innerText.split(',').length; - - // For every 100 characters in this paragraph, add another point. Up to 3 points. - contentScore += Math.min(Math.floor(innerText.length / 100), 3); - - // Initialize and score ancestors. - this._forEachNode(ancestors, 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); - } - - // Node score divider: - // - parent: 1 (no division) - // - grandparent: 2 - // - great grandparent+: ancestor level * 3 - if (level === 0) - var scoreDivider = 1; - else if (level === 1) - scoreDivider = 2; - else - scoreDivider = level * 3; - ancestor.readability.contentScore += contentScore / scoreDivider; - }); - }); - - // After we've calculated scores, loop through all of the possible - // candidate nodes we found and find the one with the highest score. - var topCandidates = []; - for (var c = 0, cl = candidates.length; c < cl; c += 1) { - var candidate = candidates[c]; - - // Scale the final candidates score based on link density. Good content - // should have a relatively small link density (5% or less) and be mostly - // unaffected by this operation. - 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 we still have no top candidate, just use the body as a last resort. - // We also have to copy the body node so it is something we can modify. - if (topCandidate === null || topCandidate.tagName === 'BODY') { - // Move all of the page's children into topCandidate - topCandidate = doc.createElement('DIV'); - neededToCreateTopCandidate = true; - // Move everything (not just elements, also text nodes etc.) into the container - // so we even include text directly in the body: - var kids = page.childNodes; - while (kids.length) { - this.log('Moving child out:', kids[0]); - topCandidate.appendChild(kids[0]); - } - - page.appendChild(topCandidate); - - this._initializeNode(topCandidate); - } else if (topCandidate) { - // Find a better top candidate node if it contains (at least three) nodes which belong to `topCandidates` array - // and whose scores are quite closed with current `topCandidate` node. - 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); - } - - // Because of our bonus system, parents of candidates might have scores - // themselves. They get half of the node. There won't be nodes with higher - // scores than our topCandidate, but if we see the score going *up* in the first - // few steps up the tree, that's a decent sign that there might be more content - // lurking in other places that we want to unify in. The sibling stuff - // below does some of that - but only if we've looked high enough up the DOM - // tree. - parentOfTopCandidate = topCandidate.parentNode; - var lastScore = topCandidate.readability.contentScore; - // The scores shouldn't get too low. - 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) { - // Alright! We found a better parent to use. - topCandidate = parentOfTopCandidate; - break; - } - lastScore = parentOfTopCandidate.readability.contentScore; - parentOfTopCandidate = parentOfTopCandidate.parentNode; - } - - // If the top candidate is the only child, use parent instead. This will help sibling - // joining logic when adjacent content is actually located in parent's sibling node. - parentOfTopCandidate = topCandidate.parentNode; - while (parentOfTopCandidate.tagName != 'BODY' && parentOfTopCandidate.children.length == 1) { - topCandidate = parentOfTopCandidate; - parentOfTopCandidate = topCandidate.parentNode; - } - if (!topCandidate.readability) { - this._initializeNode(topCandidate); - } - } - - // Now that we have the top candidate, look through its siblings for content - // that might also be related. Things like preambles, content split by ads - // that we removed, etc. - var articleContent = doc.createElement('DIV'); - if (isPaging) - articleContent.id = 'readability-content'; - - var siblingScoreThreshold = Math.max(10, topCandidate.readability.contentScore * 0.2); - // Keep potential top candidate's parent node to try to get text direction of it later. - 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; - - // Give a bonus if sibling nodes and top candidates have the example same classname - 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.indexOf(sibling.nodeName) === -1) { - // We have a node that isn't a common block level element, like a form or td tag. - // Turn it into a div so it doesn't get filtered out later by accident. - this.log('Altering sibling:', sibling, 'to div.'); - - sibling = this._setNodeTag(sibling, 'DIV'); - } - - articleContent.appendChild(sibling); - // siblings is a reference to the children array, and - // sibling is removed from the array when we call appendChild(). - // As a result, we must revisit this index since the nodes - // have been shifted. - s -= 1; - sl -= 1; - } - } - - if (this._debug) - this.log('Article content pre-prep: ' + articleContent.innerHTML); - // So we have all of the content that we need. Now we clean it up for presentation. - this._prepArticle(articleContent); - if (this._debug) - this.log('Article content post-prep: ' + articleContent.innerHTML); - - if (neededToCreateTopCandidate) { - // We already created a fake div thing, and there wouldn't have been any siblings left - // for the previous loop, so there's no point trying to create a new div, and then - // move all the children over. Just assign IDs and class names here. No need to append - // because that already happened anyway. - topCandidate.id = 'readability-page-1'; - topCandidate.className = 'page'; + 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 { - var div = doc.createElement('DIV'); - div.id = 'readability-page-1'; - div.className = 'page'; - var children = articleContent.childNodes; - while (children.length) { - div.appendChild(children[0]); - } - articleContent.appendChild(div); - } - - if (this._debug) - this.log('Article content after paging: ' + articleContent.innerHTML); - - var parseSuccessful = true; - - // Now that we've gone through the full algorithm, check to see if - // we got any meaningful content. If we didn't, we may need to re-run - // grabArticle with different flags set. This gives us a higher likelihood of - // finding the content, and the sieve approach gives us a higher likelihood of - // finding the -right- content. - var textLength = this._getInnerText(articleContent, true).length; - if (textLength < this._charThreshold) { - parseSuccessful = false; - page.innerHTML = pageCacheHtml; - - if (this._flagIsActive(this.FLAG_STRIP_UNLIKELYS)) { - this._removeFlag(this.FLAG_STRIP_UNLIKELYS); - this._attempts.push({articleContent: articleContent, textLength: textLength}); - } else if (this._flagIsActive(this.FLAG_WEIGHT_CLASSES)) { - this._removeFlag(this.FLAG_WEIGHT_CLASSES); - this._attempts.push({articleContent: articleContent, textLength: textLength}); - } else if (this._flagIsActive(this.FLAG_CLEAN_CONDITIONALLY)) { - this._removeFlag(this.FLAG_CLEAN_CONDITIONALLY); - this._attempts.push({articleContent: articleContent, textLength: textLength}); - } else { - this._attempts.push({articleContent: articleContent, textLength: textLength}); - // No luck after removing flags, just return the longest text we found during the different loops - this._attempts.sort(function (a, b) { - return b.textLength - a.textLength; - }); - - // But first check if we actually have something - if (!this._attempts[0].textLength) { - return null; - } - - articleContent = this._attempts[0].articleContent; - parseSuccessful = true; - } - } - - if (parseSuccessful) { - // Find out text direction from ancestors of final top candidate. - 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; + // 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); + } + }); }, - - /** - * Check whether the input string could be a byline. - * This verifies that the input is a string, and that the length - * is less than 100 chars. - * - * @param possibleByline {string} - a string to check whether its a byline. - * @return Boolean - whether the input string is a byline. - */ - _isValidByline: function(byline) { - if (typeof byline == 'string' || byline instanceof String) { - byline = byline.trim(); - return (byline.length > 0) && (byline.length < 100); + + _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
elements with a single

. + * Whitespace between
elements are ignored. For example: + *

foo
bar


abc
+ * will become: + *
foo
bar

abc

+ */ + _replaceBrs: function (elem) { + this._forEachNode(this._getAllNodesWithTag(elem, ["br"]), function(br) { + var next = br.nextSibling; + + // Whether 2 or more
elements have been found and replaced with a + //

block. + var replaced = false; + + // If we find a
chain, remove the
s until we hit another node + // or non-whitespace. This leaves behind the first
in the chain + // (which will be replaced with a

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
chain, replace the remaining
with a

. Add + // all sibling nodes as children of the

until we hit another
+ // chain. + if (replaced) { + var p = this._doc.createElement("p"); + br.parentNode.replaceChild(p, br); + + next = p.nextSibling; + while (next) { + // If we've hit another

, we're done adding children to this

. + 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

. + 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

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; }, - - /** - * Attempts to get excerpt and byline metadata for the article. - * - * @return Object with optional "excerpt" and "byline" properties - */ - _getArticleMetadata: function() { - var metadata = {}; - var values = {}; - var metaElements = this._doc.getElementsByTagName('meta'); - - // property is a space-separated list of values - var propertyPattern = /\s*(dc|dcterm|og|twitter)\s*:\s*(author|creator|description|title|site_name)\s*/gi; - - // name is a single value - var namePattern = /^\s*(?:(dc|dcterm|og|twitter|weibo:(article|webpage))\s*[\.:]\s*)?(author|creator|description|title|site_name)\s*$/i; - - // Find description tags. - 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) { - for (var i = matches.length - 1; i >= 0; i--) { - // Convert to lowercase, and remove any whitespace - // so we can match below. - name = matches[i].toLowerCase().replace(/\s/g, ''); - // multiple authors - values[name] = content.trim(); - } - } - } - if (!matches && elementName && namePattern.test(elementName)) { - name = elementName; - if (content) { - // Convert to lowercase, remove any whitespace, and convert dots - // to colons so we can match below. - name = name.toLowerCase().replace(/\s/g, '').replace(/\./g, ':'); - values[name] = content.trim(); - } - } - }); - - // get title - metadata.title = values['dc:title'] || - values['dcterm:title'] || - values['og:title'] || - values['weibo:article:title'] || - values['weibo:webpage:title'] || - values['title'] || - values['twitter:title']; - - if (!metadata.title) { - metadata.title = this._getArticleTitle(); - } - - // get author - metadata.byline = values['dc:creator'] || - values['dcterm:creator'] || - values['author']; - - // get description - metadata.excerpt = values['dc:description'] || - values['dcterm:description'] || - values['og:description'] || - values['weibo:article:description'] || - values['weibo:webpage:description'] || - values['description'] || - values['twitter:description']; - - // get site name - metadata.siteName = values['og:site_name']; - - return metadata; + + _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; }, - - /** - * Removes script tags from the document. - * - * @param Element - **/ - _removeScripts: function(doc) { - this._removeNodes(doc.getElementsByTagName('script'), function(scriptNode) { - scriptNode.nodeValue = ''; - scriptNode.removeAttribute('src'); - return true; - }); - this._removeNodes(doc.getElementsByTagName('noscript')); - }, - - /** - * Check if this node has only whitespace and a single element with given tag - * Returns false if the DIV node contains non-empty text nodes - * or if it contains no element with given tag or more than 1 element. - * - * @param Element - * @param string tag of child element - **/ - _hasSingleTagInsideElement: function(element, tag) { - // There should be exactly 1 element child with given tag - if (element.children.length != 1 || element.children[0].tagName !== tag) { - return false; - } - - // And there should be no text nodes with real content - return !this._someNode(element.childNodes, function(node) { - return node.nodeType === this.TEXT_NODE && - this.REGEXPS.hasContent.test(node.textContent); - }); - }, - - _isElementWithoutContent: function(node) { - return node.nodeType === this.ELEMENT_NODE && - node.textContent.trim().length == 0 && - (node.children.length == 0 || - node.children.length == node.getElementsByTagName('br').length + node.getElementsByTagName('hr').length); - }, - - /** - * Determine whether element has any children block level elements. - * - * @param Element - */ - _hasChildBlockElement: function (element) { - return this._someNode(element.childNodes, function(node) { - return this.DIV_TO_P_ELEMS.indexOf(node.tagName) !== -1 || - this._hasChildBlockElement(node); - }); - }, - + /*** - * Determine if a node qualifies as phrasing content. - * https://developer.mozilla.org/en-US/docs/Web/Guide/HTML/Content_categories#Phrasing_content - **/ - _isPhrasingContent: function(node) { - return node.nodeType === this.TEXT_NODE || this.PHRASING_ELEMS.indexOf(node.tagName) !== -1 || - ((node.tagName === 'A' || node.tagName === 'DEL' || node.tagName === 'INS') && - this._everyNode(node.childNodes, this._isPhrasingContent)); - }, - - _isWhitespace: function(node) { - return (node.nodeType === this.TEXT_NODE && node.textContent.trim().length === 0) || - (node.nodeType === this.ELEMENT_NODE && node.tagName === 'BR'); - }, - - /** - * Get the inner text of a node - cross browser compatibly. - * This also strips out any excess whitespace to be found. - * - * @param Element - * @param Boolean normalizeSpaces (default: true) - * @return string - **/ - _getInnerText: function(e, normalizeSpaces) { - normalizeSpaces = (typeof normalizeSpaces === 'undefined') ? true : normalizeSpaces; - var textContent = e.textContent.trim(); - - if (normalizeSpaces) { - return textContent.replace(this.REGEXPS.normalize, ' '); + * 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) { + 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(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; + } + } + + // Remove DIV, SECTION, and HEADER nodes without any content(e.g. text, image, video, or iframe). + 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.indexOf(node.tagName) !== -1) { + elementsToScore.push(node); + } + + // Turn all divs that don't have children block level elements into p's + if (node.tagName === "DIV") { + // Put phrasing content into paragraphs. + 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.removeChild(p.lastChild); + } + p = null; + } + childNode = nextSibling; + } + + // Sites like http://mobile.slate.com encloses each paragraph with a DIV + // element. DIVs with only a P element inside and no text content can be + // safely converted into plain P elements to avoid confusing the scoring + // algorithm with DIVs with are, in practice, paragraphs. + 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); } - return textContent; - }, - - /** - * Get the number of times a string s appears in the node e. - * - * @param Element - * @param string - what to split on. Default is "," - * @return number (integer) - **/ - _getCharCount: function(e, s) { - s = s || ','; - return this._getInnerText(e).split(s).length - 1; - }, - - /** - * Remove the style attribute on every e and under. - * TODO: Test if getElementsByTagName(*) is faster. - * - * @param Element - * @return void - **/ - _cleanStyles: function(e) { - if (!e || e.tagName.toLowerCase() === 'svg') + + /** + * Loop through all paragraphs, and assign a score to them based on how content-y they look. + * Then add their score to their parent node. + * + * A score is determined by things like number of commas, class names, etc. Maybe eventually link density. + **/ + var candidates = []; + this._forEachNode(elementsToScore, function(elementToScore) { + if (!elementToScore.parentNode || typeof(elementToScore.parentNode.tagName) === "undefined") return; - - // Remove `style` and deprecated presentational attributes - for (var i = 0; i < this.PRESENTATIONAL_ATTRIBUTES.length; i++) { - e.removeAttribute(this.PRESENTATIONAL_ATTRIBUTES[i]); - } - - if (this.DEPRECATED_SIZE_ATTRIBUTE_ELEMS.indexOf(e.tagName) !== -1) { - e.removeAttribute('width'); - e.removeAttribute('height'); - } - - var cur = e.firstElementChild; - while (cur !== null) { - this._cleanStyles(cur); - cur = cur.nextElementSibling; - } - }, - - /** - * Get the density of links as a percentage of the content - * This is the amount of text that is inside a link divided by the total text in the node. - * - * @param Element - * @return number (float) - **/ - _getLinkDensity: function(element) { - var textLength = this._getInnerText(element).length; - if (textLength === 0) - return 0; - - var linkLength = 0; - - // XXX implement _reduceNodeList? - this._forEachNode(element.getElementsByTagName('a'), function(linkNode) { - linkLength += this._getInnerText(linkNode).length; - }); - - return linkLength / textLength; - }, - - /** - * Get an elements class/id weight. Uses regular expressions to tell if this - * element looks good or bad. - * - * @param Element - * @return number (Integer) - **/ - _getClassWeight: function(e) { - if (!this._flagIsActive(this.FLAG_WEIGHT_CLASSES)) - return 0; - - var weight = 0; - - // Look for a special classname - 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; - } - - // Look for a special ID - 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 a node of all elements of type "tag". - * (Unless it's a youtube/vimeo video. People love movies.) - * - * @param Element - * @param string tag to clean - * @return void - **/ - _clean: function(e, tag) { - var isEmbed = ['object', 'embed', 'iframe'].indexOf(tag) !== -1; - - this._removeNodes(e.getElementsByTagName(tag), function(element) { - // Allow youtube and vimeo videos through as people usually want to see those. - if (isEmbed) { - // First, check the elements attributes to see if any of them contain youtube or vimeo - for (var i = 0; i < element.attributes.length; i++) { - if (this.REGEXPS.videos.test(element.attributes[i].value)) { - return false; - } - } - - // For embed with tag, check inner HTML as well. - if (element.tagName === 'object' && this.REGEXPS.videos.test(element.innerHTML)) { - return false; - } + + // If this paragraph is less than 25 characters, don't even count it. + var innerText = this._getInnerText(elementToScore); + if (innerText.length < 25) + return; + + // Exclude nodes with no ancestor. + var ancestors = this._getNodeAncestors(elementToScore, 5); + if (ancestors.length === 0) + return; + + var contentScore = 0; + + // Add a point for the paragraph itself as a base. + contentScore += 1; + + // Add points for any commas within this paragraph. + contentScore += innerText.split(",").length; + + // For every 100 characters in this paragraph, add another point. Up to 3 points. + contentScore += Math.min(Math.floor(innerText.length / 100), 3); + + // Initialize and score ancestors. + this._forEachNode(ancestors, 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); } - - return true; + + // Node score divider: + // - parent: 1 (no division) + // - grandparent: 2 + // - great grandparent+: ancestor level * 3 + if (level === 0) + var scoreDivider = 1; + else if (level === 1) + scoreDivider = 2; + else + scoreDivider = level * 3; + ancestor.readability.contentScore += contentScore / scoreDivider; + }); }); - }, - - /** - * Check if a given node has one of its ancestor tag name matching the - * provided one. - * @param HTMLElement node - * @param String tagName - * @param Number maxDepth - * @param Function filterFn a filter to invoke to determine whether this node 'counts' - * @return Boolean - */ - _hasAncestorTag: function(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++; + + // After we've calculated scores, loop through all of the possible + // candidate nodes we found and find the one with the highest score. + var topCandidates = []; + for (var c = 0, cl = candidates.length; c < cl; c += 1) { + var candidate = candidates[c]; + + // Scale the final candidates score based on link density. Good content + // should have a relatively small link density (5% or less) and be mostly + // unaffected by this operation. + 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 we still have no top candidate, just use the body as a last resort. + // We also have to copy the body node so it is something we can modify. + if (topCandidate === null || topCandidate.tagName === "BODY") { + // Move all of the page's children into topCandidate + topCandidate = doc.createElement("DIV"); + neededToCreateTopCandidate = true; + // Move everything (not just elements, also text nodes etc.) into the container + // so we even include text directly in the body: + while (page.firstChild) { + this.log("Moving child out:", page.firstChild); + topCandidate.appendChild(page.firstChild); + } + + page.appendChild(topCandidate); + + this._initializeNode(topCandidate); + } else if (topCandidate) { + // Find a better top candidate node if it contains (at least three) nodes which belong to `topCandidates` array + // and whose scores are quite closed with current `topCandidate` node. + 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); + } + + // Because of our bonus system, parents of candidates might have scores + // themselves. They get half of the node. There won't be nodes with higher + // scores than our topCandidate, but if we see the score going *up* in the first + // few steps up the tree, that's a decent sign that there might be more content + // lurking in other places that we want to unify in. The sibling stuff + // below does some of that - but only if we've looked high enough up the DOM + // tree. + parentOfTopCandidate = topCandidate.parentNode; + var lastScore = topCandidate.readability.contentScore; + // The scores shouldn't get too low. + 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) { + // Alright! We found a better parent to use. + topCandidate = parentOfTopCandidate; + break; + } + lastScore = parentOfTopCandidate.readability.contentScore; + parentOfTopCandidate = parentOfTopCandidate.parentNode; + } + + // If the top candidate is the only child, use parent instead. This will help sibling + // joining logic when adjacent content is actually located in parent's sibling node. + parentOfTopCandidate = topCandidate.parentNode; + while (parentOfTopCandidate.tagName != "BODY" && parentOfTopCandidate.children.length == 1) { + topCandidate = parentOfTopCandidate; + parentOfTopCandidate = topCandidate.parentNode; + } + if (!topCandidate.readability) { + this._initializeNode(topCandidate); + } + } + + // Now that we have the top candidate, look through its siblings for content + // that might also be related. Things like preambles, content split by ads + // that we removed, etc. + var articleContent = doc.createElement("DIV"); + if (isPaging) + articleContent.id = "readability-content"; + + var siblingScoreThreshold = Math.max(10, topCandidate.readability.contentScore * 0.2); + // Keep potential top candidate's parent node to try to get text direction of it later. + 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; + + // Give a bonus if sibling nodes and top candidates have the example same classname + 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.indexOf(sibling.nodeName) === -1) { + // We have a node that isn't a common block level element, like a form or td tag. + // Turn it into a div so it doesn't get filtered out later by accident. + this.log("Altering sibling:", sibling, "to div."); + + sibling = this._setNodeTag(sibling, "DIV"); + } + + articleContent.appendChild(sibling); + // Fetch children again to make it compatible + // with DOM parsers without live collection support. + siblings = parentOfTopCandidate.children; + // siblings is a reference to the children array, and + // sibling is removed from the array when we call appendChild(). + // As a result, we must revisit this index since the nodes + // have been shifted. + s -= 1; + sl -= 1; + } + } + + if (this._debug) + this.log("Article content pre-prep: " + articleContent.innerHTML); + // So we have all of the content that we need. Now we clean it up for presentation. + this._prepArticle(articleContent); + if (this._debug) + this.log("Article content post-prep: " + articleContent.innerHTML); + + if (neededToCreateTopCandidate) { + // We already created a fake div thing, and there wouldn't have been any siblings left + // for the previous loop, so there's no point trying to create a new div, and then + // move all the children over. Just assign IDs and class names here. No need to append + // because that already happened anyway. + 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; + + // Now that we've gone through the full algorithm, check to see if + // we got any meaningful content. If we didn't, we may need to re-run + // grabArticle with different flags set. This gives us a higher likelihood of + // finding the content, and the sieve approach gives us a higher likelihood of + // finding the -right- content. + var textLength = this._getInnerText(articleContent, true).length; + if (textLength < this._charThreshold) { + parseSuccessful = false; + page.innerHTML = pageCacheHtml; + + if (this._flagIsActive(this.FLAG_STRIP_UNLIKELYS)) { + this._removeFlag(this.FLAG_STRIP_UNLIKELYS); + this._attempts.push({articleContent: articleContent, textLength: textLength}); + } else if (this._flagIsActive(this.FLAG_WEIGHT_CLASSES)) { + this._removeFlag(this.FLAG_WEIGHT_CLASSES); + this._attempts.push({articleContent: articleContent, textLength: textLength}); + } else if (this._flagIsActive(this.FLAG_CLEAN_CONDITIONALLY)) { + this._removeFlag(this.FLAG_CLEAN_CONDITIONALLY); + this._attempts.push({articleContent: articleContent, textLength: textLength}); + } else { + this._attempts.push({articleContent: articleContent, textLength: textLength}); + // No luck after removing flags, just return the longest text we found during the different loops + this._attempts.sort(function (a, b) { + return b.textLength - a.textLength; + }); + + // But first check if we actually have something + if (!this._attempts[0].textLength) { + return null; + } + + articleContent = this._attempts[0].articleContent; + parseSuccessful = true; + } + } + + if (parseSuccessful) { + // Find out text direction from ancestors of final top candidate. + 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; + } + } + }, + + /** + * Check whether the input string could be a byline. + * This verifies that the input is a string, and that the length + * is less than 100 chars. + * + * @param possibleByline {string} - a string to check whether its a byline. + * @return Boolean - whether the input string is a byline. + */ + _isValidByline: function(byline) { + if (typeof byline == "string" || byline instanceof String) { + byline = byline.trim(); + return (byline.length > 0) && (byline.length < 100); + } + return false; + }, + + /** + * Converts some of the common HTML entities in string to their corresponding characters. + * + * @param str {string} - a string to unescape. + * @return string without HTML entity. + */ + _unescapeHtmlEntities: function(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-z]{1,4})|([0-9]{1,4}));/gi, function(_, hex, numStr) { + var num = parseInt(hex || numStr, hex ? 16 : 10); + return String.fromCharCode(num); + }); + }, + + /** + * Try to extract metadata from JSON-LD object. + * For now, only Schema.org objects of type Article or its subtypes are supported. + * @return Object with any metadata that could be extracted (possibly none) + */ + _getJSONLD: function (doc) { + var scripts = this._getAllNodesWithTag(doc, ["script"]); + + var jsonLdElement = this._findNode(scripts, function(el) { + return el.getAttribute("type") === "application/ld+json"; + }); + + if (jsonLdElement) { + try { + // Strip CDATA markers if present + var content = jsonLdElement.textContent.replace(/^\s*\s*$/g, ""); + var parsed = JSON.parse(content); + var metadata = {}; + if ( + !parsed["@context"] || + !parsed["@context"].match(/^https?\:\/\/schema\.org$/) + ) { + return metadata; + } + + if (!parsed["@type"] && Array.isArray(parsed["@graph"])) { + parsed = parsed["@graph"].find(function(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") { + 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(); + } + return metadata; + } catch (err) { + this.log(err.message); + } + } + return {}; + }, + + /** + * Attempts to get excerpt and byline metadata for the article. + * + * @param {Object} jsonld — object containing any metadata that + * could be extracted from JSON-LD object. + * + * @return Object with optional "excerpt" and "byline" properties + */ + _getArticleMetadata: function(jsonld) { + var metadata = {}; + var values = {}; + var metaElements = this._doc.getElementsByTagName("meta"); + + // property is a space-separated list of values + var propertyPattern = /\s*(dc|dcterm|og|twitter)\s*:\s*(author|creator|description|title|site_name)\s*/gi; + + // name is a single value + var namePattern = /^\s*(?:(dc|dcterm|og|twitter|weibo:(article|webpage))\s*[\.:]\s*)?(author|creator|description|title|site_name)\s*$/i; + + // Find description tags. + 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) { + // Convert to lowercase, and remove any whitespace + // so we can match below. + name = matches[0].toLowerCase().replace(/\s/g, ""); + // multiple authors + values[name] = content.trim(); + } + } + if (!matches && elementName && namePattern.test(elementName)) { + name = elementName; + if (content) { + // Convert to lowercase, remove any whitespace, and convert dots + // to colons so we can match below. + name = name.toLowerCase().replace(/\s/g, "").replace(/\./g, ":"); + values[name] = content.trim(); + } + } + }); + + // get title + 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"]; + + if (!metadata.title) { + metadata.title = this._getArticleTitle(); + } + + // get author + metadata.byline = jsonld.byline || + values["dc:creator"] || + values["dcterm:creator"] || + values["author"]; + + // get description + 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"]; + + // get site name + metadata.siteName = jsonld.siteName || + values["og:site_name"]; + + // in many sites the meta value is escaped with HTML entities, + // so here we need to unescape it + metadata.title = this._unescapeHtmlEntities(metadata.title); + metadata.byline = this._unescapeHtmlEntities(metadata.byline); + metadata.excerpt = this._unescapeHtmlEntities(metadata.excerpt); + metadata.siteName = this._unescapeHtmlEntities(metadata.siteName); + + return metadata; + }, + + /** + * Check if node is image, or if node contains exactly only one image + * whether as a direct child or as its descendants. + * + * @param Element + **/ + _isSingleImage: function(node) { + if (node.tagName === "IMG") { + return true; + } + + if (node.children.length !== 1 || node.textContent.trim() !== "") { return false; + } + + return this._isSingleImage(node.children[0]); }, - + /** - * Return an object indicating how many rows and columns this table has. - */ + * Find all