Modify search to work with article contents

Co-authored-by: aider (gemini/gemini-2.5-pro) <aider@aider.chat>

apiserver/reindex.py

@@ -0,0 +1,67 @@
+import logging
+logging.basicConfig(
+        format='%(asctime)s - %(name)s - %(levelname)s - %(message)s',
+        level=logging.INFO)
+
+import database
+from sqlalchemy import select
+import search
+import sys
+
+import time
+import json
+import requests
+from bs4 import BeautifulSoup
+
+database.init()
+search.init()
+
+BATCH_SIZE = 1000
+
+def put_stories(stories):
+    return search.meili_api(requests.post, 'indexes/qotnews/documents', stories)
+
+def get_update(update_id):
+    return search.meili_api(requests.get, 'tasks/{}'.format(update_id))
+
+if __name__ == '__main__':
+    num_stories = database.count_stories()
+
+    print('Reindex {} stories?'.format(num_stories))
+    print('Press ENTER to continue, ctrl-c to cancel')
+    input()
+
+    story_list = database.get_story_list()
+
+    count = 1
+    while len(story_list):
+        stories = []
+
+        for _ in range(BATCH_SIZE):
+            try:
+                sid = story_list.pop()
+            except IndexError:
+                break
+
+            story = database.get_story(sid)
+            print('Indexing {}/{} id: {} title: {}'.format(count, num_stories, sid[0], story.title))
+            story_obj = json.loads(story.full_json)
+            story_obj.pop('comments', False)
+            if 'text' in story_obj and story_obj['text']:
+                soup = BeautifulSoup(story_obj['text'], 'html.parser')
+                story_obj['text'] = soup.get_text()
+            stories.append(story_obj)
+            count += 1
+
+        res = put_stories(stories)
+        update_id = res['taskUid']
+
+        print('Waiting for processing', end='')
+        while get_update(update_id)['status'] != 'succeeded':
+            time.sleep(0.5)
+            print('.', end='', flush=True)
+
+        print()
+
+    print('Done.')
+

apiserver/scripts/reindex.py

@@ -11,6 +11,7 @@ import sys
 import time
 import json
 import requests
+from bs4 import BeautifulSoup
 
 database.init()
 search.init()
@@ -45,6 +46,9 @@ if __name__ == '__main__':
             story = database.get_story(sid)
             print('Indexing {}/{} id: {} title: {}'.format(count, num_stories, sid[0], story.title))
             story_obj = json.loads(story.meta_json)
+            if 'text' in story_obj and story_obj['text']:
+                soup = BeautifulSoup(story_obj['text'], 'html.parser')
+                story_obj['text'] = soup.get_text()
             stories.append(story_obj)
             count += 1
 

apiserver/search.py

@@ -10,7 +10,8 @@ SEARCH_ENABLED = bool(settings.MEILI_URL)
 
 def meili_api(method, route, json=None, params=None, parse_json=True):
     try:
-        r = method(settings.MEILI_URL + route, json=json, params=params, timeout=4)
+        headers = {'Authorization': 'Bearer ' + settings.MEILI_API_KEY}
+        r = method(settings.MEILI_URL + route, json=json, params=params, headers=headers, timeout=4)
         if r.status_code > 299:
             raise Exception('Bad response code ' + str(r.status_code))
         if parse_json:
@@ -24,42 +25,40 @@ def meili_api(method, route, json=None, params=None, parse_json=True):
         logging.error('Problem with MeiliSearch api route: %s: %s', route, str(e))
         return False
 
-def create_index():
-    json = dict(uid='qotnews', primaryKey='id')
-    return meili_api(requests.post, 'indexes', json=json)
-
-def update_rankings():
-    json = ['typo', 'words', 'proximity', 'date:desc', 'exactness']
-    return meili_api(requests.post, 'indexes/qotnews/settings/ranking-rules', json=json)
-
-def update_attributes():
-    json = ['title', 'url', 'author']
-    r = meili_api(requests.post, 'indexes/qotnews/settings/searchable-attributes', json=json)
-    json = ['id', 'ref', 'source', 'author', 'author_link', 'score', 'date', 'title', 'link', 'url', 'num_comments']
-    r = meili_api(requests.post, 'indexes/qotnews/settings/displayed-attributes', json=json)
-    return r
+def update_settings():
+    json = {
+        'rankingRules': ['words', 'typo', 'proximity', 'attribute', 'date:desc', 'exactness'],
+        'searchableAttributes': ['title', 'url', 'author', 'text'],
+        'displayedAttributes': ['id', 'ref', 'source', 'author', 'author_link', 'score', 'date', 'title', 'link', 'url', 'num_comments', 'text'],
+        'stopWords': ['a', 'an', 'the', 'and', 'or', 'but', 'if', 'in', 'on', 'at', 'by', 'for', 'with', 'to', 'from', 'of', 'is', 'it', 'that', 'this'],
+    }
+    return meili_api(requests.patch, 'indexes/qotnews/settings', json=json)
 
 def init():
     if not SEARCH_ENABLED:
         logging.info('Search is not enabled, skipping init.')
         return
-    print(create_index())
-    update_rankings()
-    update_attributes()
+    update_settings()
 
 def put_story(story):
     if not SEARCH_ENABLED: return
     return meili_api(requests.post, 'indexes/qotnews/documents', [story])
 
-def search(q):
+def search(q, in_article=False):
     if not SEARCH_ENABLED: return []
-    params = dict(q=q, limit=settings.FEED_LENGTH)
-    r = meili_api(requests.get, 'indexes/qotnews/search', params=params, parse_json=False)
+
+    json = dict(q=q, limit=settings.FEED_LENGTH)
+
+    if True:
+        json['attributesToSearchOn'] = ['text']
+        json['attributesToCrop'] = ['text']
+        json['attributesToRetrieve'] = ['id', 'ref', 'source', 'author', 'author_link', 'score', 'date', 'title', 'link', 'url', 'num_comments']
+        json['cropLength'] = 80
+
+    r = meili_api(requests.post, 'indexes/qotnews/search', json=json, parse_json=False)
     return r
     
 if __name__ == '__main__':
     init()
 
-    print(update_rankings())
-
     print(search('facebook'))

apiserver/server.py

@@ -131,8 +131,9 @@ def apistats():
 @flask_app.route('/api/search', strict_slashes=False)
 def apisearch():
     q = request.args.get('q', '')
+    in_article = request.args.get('article', False)
     if len(q) >= 3:
-        results = search.search(q)
+        results = search.search(q, in_article)
     else:
         results = '[]'
     res = Response(results)
@@ -262,7 +263,7 @@ def static_story(sid):
         show_comments=request.path.endswith('/c'),
     )
 
-http_server = WSGIServer(('', 33842), flask_app)
+http_server = WSGIServer(('0.0.0.0', 33842), flask_app)
 
 def feed_thread():
     global news_index, ref_list, current_item

searchserver/.gitignore

@@ -4,3 +4,4 @@
 
 meilisearch-linux-amd64
 data.ms/
+data.ms.old/

webclient/public/index.html

@@ -30,11 +30,7 @@
 		-->
 		<title>{{ title }}</title>
 
-		<script>document.documentElement.className = 'js-enabled';</script>
 		<style>
-			.js-enabled .static-content {
-				display: none;
-			}
 			html {
 				overflow-y: scroll;
 			}
@@ -45,20 +41,7 @@
 	</head>
 
 	<body>
-		<script>
-			(function() {
-				try {
-					var theme = localStorage.getItem('theme');
-					if (theme === 'dark') {
-						document.body.style.backgroundColor = '#1a1a1a';
-					} else if (theme === 'black' || theme === 'red') {
-						document.body.style.backgroundColor = '#000';
-					}
-				} catch (e) {}
-			})();
-		</script>
 		<div id="root">
-			<div class="static-content">
 			<div class="container menu">
 				<p>
 					<a href="/">QotNews</a>
@@ -151,7 +134,6 @@
 				</div>
 			{% endif %}
 		</div>
-		</div>
 		<!--
 			This HTML file is a template.
 			If you open it directly in the browser, you will see an empty page.

webclient/src/App.js

@@ -19,7 +19,6 @@ function App() {
 	const [theme, setTheme] = useState(localStorage.getItem('theme') || '');
 	const cache = useRef({});
 	const [isFullScreen, setIsFullScreen] = useState(!!document.fullscreenElement);
-	const [waitingWorker, setWaitingWorker] = useState(null);
 
 	const updateCache = useCallback((key, value) => {
 		cache.current[key] = value;
@@ -45,14 +44,6 @@ function App() {
 		localStorage.setItem('theme', 'red');
 	};
 
-	useEffect(() => {
-		const onSWUpdate = e => {
-			setWaitingWorker(e.detail.waiting);
-		};
-		window.addEventListener('swUpdate', onSWUpdate);
-		return () => window.removeEventListener('swUpdate', onSWUpdate);
-	}, []);
-
 	useEffect(() => {
 		if (Object.keys(cache.current).length === 0) {
 			localForage.iterate((value, key) => {
@@ -99,19 +90,6 @@ function App() {
 
 	return (
 		<div className={theme}>
-			{waitingWorker &&
-				<div className='update-banner'>
-					Client version mismatch, please refresh:{' '}
-					<button onClick={() => {
-						waitingWorker.postMessage({ type: 'SKIP_WAITING' });
-						navigator.serviceWorker.addEventListener('controllerchange', () => {
-							window.location.reload();
-						});
-					}}>
-						Refresh
-					</button>
-				</div>
-			}
 			<Router>
 				<div className='container menu'>
 					<p>

webclient/src/Article.js

@@ -2,7 +2,7 @@ import React, { useState, useEffect } from 'react';
 import { useParams } from 'react-router-dom';
 import { Helmet } from 'react-helmet';
 import localForage from 'localforage';
-import { sourceLink, similarLink, infoLine, ToggleDot } from './utils.js';
+import { sourceLink, infoLine, ToggleDot } from './utils.js';
 import Latex from 'react-latex-next';
 import 'katex/dist/katex.min.css';
 
@@ -12,7 +12,6 @@ const DANGEROUS_TAGS = ['svg', 'math'];
 const latexDelimiters = [
     { left: '$$', right: '$$', display: true },
     { left: '\\[', right: '\\]', display: true },
-    { left: '$', right: '$', display: false },
     { left: '\\(', right: '\\)', display: false }
 ];
 
@@ -106,7 +105,7 @@ function Article({ cache }) {
 
 			if (v.nodeName === '#text') {
 				const text = v.data;
-				if (text.includes('\\[') || text.includes('\\(') || text.includes('$$') || /\$(?:[^$]*[^\s$])\$/.test(text)) {
+				if (text.includes('\\[') || text.includes('\\(') || text.includes('$$')) {
 					return <Latex key={key} delimiters={latexDelimiters}>{text}</Latex>;
 				}
 
@@ -138,8 +137,7 @@ function Article({ cache }) {
 			const textContent = v.textContent.trim();
 			const isMath = (textContent.startsWith('\\(') && textContent.endsWith('\\)')) ||
 				(textContent.startsWith('\\[') && textContent.endsWith('\\]')) ||
-				(textContent.startsWith('$$') && textContent.endsWith('$$')) ||
-				(textContent.startsWith('$') && textContent.endsWith('$') && textContent.indexOf('$') !== textContent.lastIndexOf('$') && !/\s/.test(textContent.charAt(textContent.length - 2)));
+				(textContent.startsWith('$$') && textContent.endsWith('$$'));
 
 			const props = { key: key };
 			if (v.hasAttributes()) {
@@ -159,13 +157,6 @@ function Article({ cache }) {
 						const firstParen = mathContent.indexOf('\\(');
 						const lastParen = mathContent.lastIndexOf('\\)');
 						mathContent = mathContent.substring(0, firstParen) + '\\[' + mathContent.substring(firstParen + 2, lastParen) + '\\]' + mathContent.substring(lastParen + 2);
-					} else if (trimmed.startsWith('$') && !trimmed.startsWith('$$')) {
-						// Replace $ with $$
-						const firstDollar = mathContent.indexOf('$');
-						const lastDollar = mathContent.lastIndexOf('$');
-						if (firstDollar !== lastDollar) {
-							mathContent = mathContent.substring(0, firstDollar) + '$$' + mathContent.substring(firstDollar + 1, lastDollar) + '$$' + mathContent.substring(lastDollar + 1);
-						}
 					}
 				}
 				return <Tag {...props}><Latex delimiters={latexDelimiters}>{mathContent}</Latex></Tag>;
@@ -213,7 +204,7 @@ function Article({ cache }) {
 					<h1>{story.title} <button className='copy-button' onClick={copyLink}>{copyButtonText}</button></h1>
 
 					<div className='info'>
-						Source: {sourceLink(story)} | {similarLink(story)}
+						Source: {sourceLink(story)}
 					</div>
 
 					{infoLine(story)}

webclient/src/Feed.js

@@ -20,15 +20,6 @@ function Feed({ updateCache }) {
 	useEffect(() => {
 		const controller = new AbortController();
 
-		if ('serviceWorker' in navigator) {
-			navigator.serviceWorker.getRegistration().then(reg => {
-				if (reg) {
-					console.log('Checking for client update...');
-					reg.update();
-				}
-			});
-		}
-
 		fetch(filterSmallweb ? '/api?smallweb=true' : '/api', { signal: controller.signal })
 			.then(res => {
 				if (!res.ok) {
@@ -127,7 +118,7 @@ function Feed({ updateCache }) {
 
 			<div style={{marginBottom: '1rem'}}>
 				<input type="checkbox" id="filter-smallweb" className="checkbox" checked={filterSmallweb} onChange={handleFilterChange} />
-				<label htmlFor="filter-smallweb">Small websites</label>
+				<label htmlFor="filter-smallweb">Only Smallweb</label>
 			</div>
 
 			{error &&

webclient/src/Results.js

@@ -1,6 +1,7 @@
 import React, { useState, useEffect } from 'react';
-import { Link, useLocation } from 'react-router-dom';
+import { Link, useLocation, useHistory } from 'react-router-dom';
 import { Helmet } from 'react-helmet';
+import queryString from 'query-string';
 import { sourceLink, infoLine, logos } from './utils.js';
 import AbortController from 'abort-controller';
 
@@ -8,6 +9,19 @@ function Results() {
 	const [stories, setStories] = useState(false);
 	const [error, setError] = useState(false);
 	const location = useLocation();
+	const history = useHistory();
+
+	const handleFilterChange = e => {
+		const isChecked = e.target.checked;
+
+		const currentQuery = queryString.parse(location.search);
+		if (isChecked) {
+			currentQuery.article = 'true';
+		} else {
+			delete currentQuery.article;
+		}
+		history.push('/search?' + queryString.stringify(currentQuery));
+	};
 
 	useEffect(() => {
 		const controller = new AbortController();
@@ -32,11 +46,19 @@ function Results() {
 		};
 	}, [location.search]);
 
+	const searchInArticle = queryString.parse(location.search).article === 'true';
+
 	return (
 		<div className='container'>
 			<Helmet>
 				<title>Search Results | QotNews</title>
 			</Helmet>
+
+			<div style={{marginBottom: '1rem'}}>
+				<input type="checkbox" id="search-in-article" className="checkbox" checked={searchInArticle} onChange={handleFilterChange} />
+				<label htmlFor="search-in-article">Search in article</label>
+			</div>
+
 			{error && <p>Connection error?</p>}
 			{stories ?
 				<>
@@ -56,6 +78,10 @@ function Results() {
 									</div>
 
 									{infoLine(x)}
+
+									{!!x?._formatted &&
+										<p>{x._formatted.text.replace(/\n/g, ' ')}</p>
+									}
 								</div>
 							)
 						:

webclient/src/Search.js

@@ -15,7 +15,9 @@ function Search() {
 		const newSearch = event.target.value;
 		setSearch(newSearch);
 		if (newSearch.length >= 3) {
-			const searchQuery = queryString.stringify({ 'q': newSearch });
+			const currentQuery = queryString.parse(location.search);
+			currentQuery.q = newSearch;
+			const searchQuery = queryString.stringify(currentQuery);
 			history.replace('/search?' + searchQuery);
 		} else {
 			history.replace('/');
@@ -24,7 +26,9 @@ function Search() {
 
 	const searchAgain = (event) => {
 		event.preventDefault();
-		const searchString = queryString.stringify({ 'q': event.target[0].value });
+		const currentQuery = queryString.parse(location.search);
+		currentQuery.q = event.target[0].value;
+		const searchString = queryString.stringify(currentQuery);
 		history.push('/search?' + searchString);
 		inputRef.current.blur();
 	}

webclient/src/Style-black.css

@@ -69,20 +69,9 @@
 }
 
 .black .checkbox:checked + label::after {
-	border-color: #eee;
+	border-color: #ddd;
 }
 
 .black .copy-button {
 	color: #828282;
 }
-
-.black .update-banner {
-	background-color: #333;
-	color: #ddd;
-}
-
-.black .update-banner button {
-	background-color: #444444;
-	border-color: #bbb;
-	color: #ddd;
-}

webclient/src/Style-dark.css

@@ -65,20 +65,9 @@
 }
 
 .dark .checkbox:checked + label::after {
-	border-color: #eee;
+	border-color: #ddd;
 }
 
 .dark .copy-button {
 	color: #828282;
 }
-
-.dark .update-banner {
-	background-color: #333;
-	color: #ddd;
-}
-
-.dark .update-banner button {
-	background-color: #444444;
-	border-color: #bbb;
-	color: #ddd;
-}

webclient/src/Style-light.css

@@ -43,21 +43,6 @@ input {
 	border-radius: 4px;
 }
 
-.update-banner {
-	background-color: #ddd;
-	padding: 0.75rem;
-	text-align: center;
-}
-
-.update-banner button {
-	margin-left: 1rem;
-	padding: 0.25rem 0.75rem;
-	border: 1px solid #828282;
-	border-radius: 4px;
-	background-color: transparent;
-	cursor: pointer;
-}
-
 .fullscreen {
 	margin: 0.25rem;
 	padding: 0.25rem;
@@ -370,38 +355,7 @@ button.comment {
 	top: 0.2em;
 	width: 0.3rem;
 	height: 0.6rem;
-	border-style: solid;
-	border-color: #000;
+	border: solid #000;
 	border-width: 0 2px 2px 0;
 	transform: rotate(45deg);
 }
-
-.tooltip .tooltiptext {
-	visibility: hidden;
-	width: 140px;
-	background-color: #555;
-	color: #fff;
-	text-align: center;
-	border-radius: 6px;
-	padding: 5px 0;
-	position: absolute;
-	z-index: 1;
-	bottom: 110%;
-	left: 50%;
-	margin-left: -70px;
-	opacity: 0;
-	transition: opacity 0.2s;
-	font-size: 0.9rem;
-	line-height: 1.3;
-}
-
-.forwardDot .tooltiptext {
-	left: auto;
-	right: 0;
-	margin-left: 0;
-}
-
-.tooltip.show-tooltip .tooltiptext {
-	visibility: visible;
-	opacity: 1;
-}

webclient/src/Style-red.css

@@ -87,20 +87,9 @@
 }
 
 .red .checkbox:checked + label::after {
-	border-color: #dd0000;
+	border-color: #aa0000;
 }
 
 .red .copy-button {
 	color: #690000;
 }
-
-.red .update-banner {
-	background-color: #300;
-	color: #d00;
-}
-
-.red .update-banner button {
-	background-color: #440000;
-	border-color: #b00;
-	color: #d00;
-}

webclient/src/index.js

@@ -3,15 +3,9 @@ import ReactDOM from 'react-dom';
 import App from './App';
 import * as serviceWorker from './serviceWorker';
 
-// version 4
-
 ReactDOM.render(<App />, document.getElementById('root'));
 
 // If you want your app to work offline and load faster, you can change
 // // unregister() to register() below. Note this comes with some pitfalls.
 // // Learn more about service workers: https://bit.ly/CRA-PWA
-serviceWorker.register({
-  onUpdate: registration => {
-    window.dispatchEvent(new CustomEvent('swUpdate', { detail: registration }));
-  }
-});
+serviceWorker.unregister();

webclient/src/utils.js

@@ -1,4 +1,4 @@
-import React, { useState } from 'react';
+import React from 'react';
 import moment from 'moment';
 import { Link } from 'react-router-dom';
 
@@ -9,13 +9,6 @@ export const sourceLink = (story) => {
 	return (<a className='source' href={url}>{host}</a>);
 };
 
-export const similarLink = (story) => {
-	const url = story.url || story.link;
-	const urlObj = new URL(url);
-	const host = urlObj.hostname.replace(/^www\./, '');
-	return (<Link to={'/search?q="'+host+'"'} className='similar'>similar</Link>);
-};
-
 export const infoLine = (story) =>
 	<div className='info'>
 		{story.score} points
@@ -39,8 +32,6 @@ export const ToggleDot = ({ id, article }) => (
 );
 
 export const BackwardDot = () => {
-	const [showTooltip, setShowTooltip] = useState(false);
-
 	const goBackward = () => {
 		localStorage.setItem('scrollLock', 'True');
 		window.history.back();
@@ -51,25 +42,15 @@ export const BackwardDot = () => {
 	if (!document.fullscreenElement) return null;
 
 	return (
-		<div
-			className={'dot backwardDot tooltip' + (showTooltip ? ' show-tooltip' : '')}
-			onClick={goBackward}
-			onMouseEnter={() => setShowTooltip(true)}
-			onMouseLeave={() => setShowTooltip(false)}
-			onTouchStart={() => setShowTooltip(true)}
-			onTouchEnd={() => setShowTooltip(false)}
-		>
+		<div className='dot backwardDot' onClick={goBackward}>
 			<div className='button'>
 				
 			</div>
-			<span className="tooltiptext">Browser Back</span>
 		</div>
 	);
 };
 
 export const ForwardDot = () => {
-	const [showTooltip, setShowTooltip] = useState(false);
-
 	const goForward = () => {
 		localStorage.setItem('scrollLock', 'True');
 		window.history.forward();
@@ -79,18 +60,10 @@ export const ForwardDot = () => {
 	if (!isMobile) return null;
 
 	return (
-		<div
-			className={'dot forwardDot tooltip' + (showTooltip ? ' show-tooltip' : '')}
-			onClick={goForward}
-			onMouseEnter={() => setShowTooltip(true)}
-			onMouseLeave={() => setShowTooltip(false)}
-			onTouchStart={() => setShowTooltip(true)}
-			onTouchEnd={() => setShowTooltip(false)}
-		>
+		<div className='dot forwardDot' onClick={goForward}>
 			<div className='button'>
 				
 			</div>
-			<span className="tooltiptext">Browser Forward</span>
 		</div>
 	);
 };