// Since VuePress needs to extract the header from the markdown source // file and display it in the sidebar or title (#238), this file simply // removes some unnecessary elements to make header displays well at // sidebar or title. // // But header's parsing in the markdown content is done by the markdown // loader based on markdown-it. markdown-it parser will will always keep // HTML in headers, so in VuePress, after being parsed by the markdiwn // loader, the raw HTML in headers will finally be parsed by Vue-loader. // so that we can write HTML/Vue in the header. One exception is the HTML // wrapped by (markdown token: '`') tag. const { compose } = require('./shared') const parseEmojis = str => { const emojiData = require('markdown-it-emoji/lib/data/full.json') return String(str).replace(/:(.+?):/g, (placeholder, key) => emojiData[key] || placeholder) } const unescapeHtml = html => String(html) .replace(/"/g, '"') .replace(/'/g, '\'') .replace(/:/g, ':') .replace(/</g, '<') .replace(/>/g, '>') const removeMarkdownTokens = str => String(str) .replace(/\[(.*)\]\(.*\)/, '$1') // []() .replace(/(`|\*{1,3}|_)(.*?[^\\])\1/g, '$2') // `{t}` | *{t}* | **{t}** | ***{t}*** | _{t}_ .replace(/(\\)(\*|_|`|\!)/g, '$2') // remove escape char '\' const trim = str => str.trim() // This method remove the raw HTML but reserve the HTML wrapped by ``. // e.g. // Input: " b", Output: "b" // Input: "`` b", Output: "`` b" exports.removeNonCodeWrappedHTML = (str) => { return String(str).replace(/(^|[^><`])<.*>([^><`]|$)/g, '$1$2') } // Unescape html, parse emojis and remove some md tokens. exports.parseHeaders = compose( unescapeHtml, parseEmojis, removeMarkdownTokens, trim ) // Also clean the html that isn't wrapped by code. // Because we want to support using VUE components in headers. // e.g. https://vuepress.vuejs.org/guide/using-vue.html#badge exports.deeplyParseHeaders = compose( exports.removeNonCodeWrappedHTML, exports.parseHeaders, )