astro-accelerator 0.0.94 → 0.0.96

package/package.json

@@ -1,5 +1,5 @@
 {
-  "version": "0.0.94",
+  "version": "0.0.96",
   "author": "Steve Fenton",
   "name": "astro-accelerator",
   "description": "A super-lightweight, accessible, SEO-friendly starter project for Astro",
@@ -28,6 +28,8 @@
     "astro": "^2.5.7",
     "astro-accelerator-utils": "^0.2.28",
     "hast-util-from-selector": "^2.0.1",
+    "html-to-text": "^9.0.5",
+    "keyword-extractor": "^0.0.25",
     "remark-directive": "^2.0.1",
     "sharp": "^0.32.1"
   },

package/public/css/main.css

@@ -916,8 +916,8 @@ pre.astro-code code {
     float: right;
     overflow: hidden;
     display: block;
-    width: 1.2em;
-    height: 1.5em;
+    width: 1.3em;
+    height: 1.6em;
     opacity: 0.5;
 }
 

package/public/js/search.js

@@ -57,47 +57,37 @@ function search(s) {
         item.score = 0;
         item.matchedHeadings = [];
 
-        // Imagine the user searched for "Kitchen Sink"
-        // The scores are arranged below from highest to lowest relevance
+        // The user searches for "Kitchen Sink"
 
-        // If the title contains a whole word match
-        queryTerms.forEach(t => {
-            if (containsWord(item.safeTitle, t)) {
-                item.score = item.score + 120;
-            }
-        });
+        // Part 1 - Phrase Matches, i.e. "Kitchen Sink"
 
-        // If the title contains "Kitchen Sink"
+        // Title
         if (contains(item.safeTitle, currentQuery)) {
             item.score = item.score + 60;
         }
 
-        // If a heading contains "Kitchen Sink"
+        // Headings
         item.headings.forEach(c => {
-            queryTerms.forEach(t => {
-                if (containsWord(c.safeText, t)) {
-                    item.score = item.score + 40;
-                }
-            });
-
             if (contains(c.safeText, currentQuery)) {
                 item.score = item.score + 20;
                 item.matchedHeadings.push(c);
             }
         });
 
-        // If the title contains "Kitchen Sink"
+        // Description
         if (contains(item.description, currentQuery)) {
             item.score = item.score + 20;
         }
+
+        // Part 2 - Term Matches, i.e. "Kitchen" or "Sink"
         
         queryTerms.forEach(term => {
-            // If the title contains "Kitchen" or "Sink"
+            // Title
             if (contains(item.safeTitle, term)) {
                 item.score = item.score + 40;
             }
 
-            // If a heading contains "Kitchen" or "Sink"
+            // Headings
             item.headings.forEach(c => {
                 if (contains(c.safeText, term)) {
                     item.score = item.score + 15;
@@ -107,17 +97,22 @@ function search(s) {
                 }
             });
 
-            // If the description contains "Kitchen" or "Sink"
+            // Description
             if (contains(item.description, term)) {
-                item.score = item.score + 10;
+                item.score = item.score + 15;
             }
 
-            // If a tag contains "Kitchen" or "Sink"
+            // Tags
             item.tags.forEach(t => {
                 if (contains(t, term)) {
-                    item.score = item.score + 5;
+                    item.score = item.score + 15;
                 }
             });
+
+            // Keywords
+            if (contains(item.keywords, term)) {
+                item.score = item.score + 15;
+            }
         })
 
         if (item.score > 0) {

package/src/env.d.ts

@@ -0,0 +1 @@
+/// <reference types="astro/client" />

package/src/pages/search.json.ts

@@ -3,6 +3,9 @@
 import { Accelerator, PostFiltering } from 'astro-accelerator-utils';
 import type { MarkdownInstance } from 'astro';
 import { SITE } from '@config';
+import { htmlToText } from 'html-to-text';
+import keywordExtractor from 'keyword-extractor';
+
 
 const getData = async () => {
     //@ts-ignore
@@ -26,13 +29,33 @@ const getData = async () => {
 
         const headings = await page.getHeadings();
         const title = await accelerator.markdown.getTextFrom(page.frontmatter.title ?? '');
-              
+        const content = page.compiledContent ? page.compiledContent() : '';
+        let counted: { word: string, count: number }[] = [];
+
+        if (content) {
+            const text = htmlToText(content, { wordwrap: false });
+
+            const words = keywordExtractor.extract(text, {
+                language: 'english',
+                return_changed_case: true
+            });
+
+            function unique (value: string, index: number, array: string[]) {
+                return array.indexOf(value) === index;
+            }
+            const uniques = words.filter(unique);
+            counted = uniques.map((w) => {
+                return { word: w, count: words.filter(wd => wd === w).length };
+            }).filter(e => e.word.replace(/[^a-z]+/g, '').length > 1 && e.count > 1);
+        }
+
         items.push({
             title: title,
             headings: headings.map(h => {
                 return {text: h.text, slug: h.slug }
             }),
             description: page.frontmatter.description ?? '',
+            keywords: counted.map(c => c.word).join(' '),
             tags: page.frontmatter.tags ?? [],
             url: SITE.url + accelerator.urlFormatter.formatAddress(url),
             date: page.frontmatter.pubDate ?? ''