Merge pull request #6 from lelouch77/build-restructure

Build restructuring

README.md

@@ -10,23 +10,25 @@ npm i docusurus-lunr-search --save
 ```
 npm run swizzle docusurus-lunr-search SearchBar
 ```
-3. Copy the [`build-search-data.js`](./build-search-data.js) to the root folder of your project
-4. Then build your Docusurus project
+3. Add the docusurus-lunr-search plugin to your `docusaurus.config.js`
 ```
-npm run build
+module.exports = {
+  // ...
+  plugins: [
+    'docusurus-lunr-search'
+  ]
+}
 ```
-5. Create the search data using by running `build-search-data.js`
+4. Then build your Docusurus project
 ```
-node build-search-data.js
+npm run build
 ```
-6. You are done!.
-\
- Now you can build your project again with new `search-data.js`
+5. Serve your application
 ```
-npm run build
-//or 
-npm start
+npx http-server ./build
 ```
+
+Note: Docusaurus search information can only be generated from a production build. Local development is currently not supported.
 ## Sample
 <p align="center">
   <img width="460" height="300" src="https://raw.githubusercontent.com/lelouch77/docusurus-lunr-search/master/assets/search-offline.png">

package.json

@@ -1,6 +1,6 @@
 {
   "name": "docusurus-lunr-search",
-  "version": "0.0.8",
+  "version": "1.0.0",
   "description": "Offline search component for Docusaurus V2",
   "main": "src/index.js",
   "publishConfig": {
@@ -19,6 +19,10 @@
     {
       "name": "Pieter C",
       "url": "https://github.com/pcallewaert"
+    },
+    {
+      "name": "ZachJW34",
+      "url": "https://github.com/ZachJW34"
     }
   ],
   "keywords": [
@@ -32,7 +36,6 @@
     "autocomplete.js": "^0.37.0",
     "cheerio": "^1.0.0-rc.3",
     "classnames": "^2.2.6",
-    "fs-extra": "^8.1.0",
     "hogan.js": "^3.0.2",
     "lunr": "^2.3.8"
   },

src/index.js

@@ -5,24 +5,26 @@
  * LICENSE file in the root directory of this source tree.
  */
 
-const path = require('path');
+const fs = require("fs");
+const path = require("path");
+const cheerio = require("cheerio");
 
-module.exports = function() {
-  return {
-    name: 'docusurus-lunr-search',
+const sectionHeaderElements = ["h2", "h3"];
 
+module.exports = function () {
+  return {
+    name: "docusurus-lunr-search",
     getThemePath() {
-      return path.resolve(__dirname, './theme');
+      return path.resolve(__dirname, "./theme");
     },
-
     configureWebpack() {
       // Ensure that algolia docsearch css is its own chunk
       return {
         optimization: {
           splitChunks: {
             cacheGroups: {
               algolia: {
-                name: 'algolia',
+                name: "algolia",
                 test: /algolia\.css$/,
                 chunks: `all`,
                 enforce: true,
@@ -34,5 +36,97 @@ module.exports = function() {
         },
       };
     },
+    async postBuild({ routesPaths = [], outDir, baseUrl }) {
+      const files = routesPaths
+        .filter((route) => route !== baseUrl && route !== `${baseUrl}404.html`)
+        .map((route) => route.substr(baseUrl.length))
+        .map((route) => ({
+          path: path.join(outDir, route, "index.html"),
+          url: route,
+        }));
+      const searchData = buildSearchData(files);
+      fs.writeFileSync(
+        path.join(outDir, "search-data.json"),
+        JSON.stringify(searchData)
+      );
+    },
   };
 };
+
+// Build search data for a html
+function buildSearchData(files) {
+  const searchData = [];
+  files.forEach(({ path, url }) => {
+    const htmlFile = fs.readFileSync(path);
+    //   const dom = new JSDOM(htmlFile);
+    const $ = cheerio.load(htmlFile);
+
+    const article = $("article");
+    if (!article.length) {
+      return;
+    }
+    const markdown = article.find(".markdown");
+    if (!markdown.length) {
+      return;
+    }
+
+    const pageTitleElement = article.find("h1");
+    if (!pageTitleElement.length) {
+      return;
+    }
+    const pageTitle = article.find("h1").text();
+    const sectionHeaders = sectionHeaderElements.reduce((acc, selector) => {
+      acc = acc.concat(Array.from(markdown.find(selector)));
+      return acc;
+    }, []);
+
+    keywords = $("meta[name='keywords']").attr("content");
+    if (typeof keywords !== "undefined" && keywords) {
+      keywords = keywords.replace(",", " ");
+    }
+
+    searchData.push({
+      title: pageTitle,
+      type: 0,
+      sectionRef: "#",
+      url,
+      // If there is no sections then push the complete content under page title
+      content: sectionHeaders.length === 0 ? getContent(markdown) : "",
+      keywords: keywords,
+    });
+
+    sectionHeaders.forEach((sectionHeader) => {
+      sectionHeader = $(sectionHeader);
+      const title = sectionHeader.text().slice(1);
+      const sectionRef = sectionHeader.children().first().attr("id");
+      const content = getSectionContent(sectionHeader);
+      searchData.push({
+        title,
+        type: 1,
+        pageTitle,
+        url: `${url}#${sectionRef}`,
+        content,
+      });
+    });
+  });
+  return searchData;
+}
+
+function getContent(element) {
+  const text =
+    element.is("table") || element.find("table").length !== 0
+      ? element.html().replace(/<[^>]*>/g, " ")
+      : element.text();
+  return text.replace(/\s\s+/g, " ").replace(/(\r\n|\n|\r)/gm, " ");
+}
+
+function getSectionContent(section) {
+  let content = "";
+  section = section.next();
+  while (section.length) {
+    if (sectionHeaderElements.some((s) => section.is(s))) break;
+    content += getContent(section) + " ";
+    section = section.next();
+  }
+  return content;
+}