From 5d174d2f194c6145e294e7e9de7fd518e977deac Mon Sep 17 00:00:00 2001
From: Jan Andrle <andrle.jan@centrum.cz>
Date: Mon, 8 Jul 2024 10:33:33 +0200
Subject: [PATCH] :zap: Now cli is able to update the last drop articles

---
 cli.mjs           | 26 +++++++++++++++++++++++---
 package-lock.json |  4 ++--
 package.json      |  2 +-
 3 files changed, 26 insertions(+), 6 deletions(-)

diff --git a/cli.mjs b/cli.mjs
index 9db102e..7db3bc2 100755
--- a/cli.mjs
+++ b/cli.mjs
@@ -96,28 +96,48 @@ async function sitemap(){
 	const json= s.test("-f", path) ? s.cat(path).xargs(JSON.parse) : { drops: [], articles: [] };
 	await syncDrops(json);
 	const [ { drop: drop_last } ]= json.drops;
-	const [ article_last= { drop: "" } ]= json.articles;
-	if(drop_last === article_last.drop) return { json, changed: [] };
 	
 	const res= await fetch(url_drops+drop_last);
 	if(res.status !== 200) return { json, changed: [] };
 	
+	const drop_articles= dropArticles(drop_last, json);
 	const dom= new JSDOM(await res.text());
 	const diff= [];
 	for(const article of dom.window.document.querySelectorAll("article")){
+		const loc= article.querySelector("a")?.href;
+		if(!loc){
+			echo("Article without link:", article.textContent);
+			continue;
+		}
+		if(drop_articles.has(loc)) continue;
 		diff.push({
 			title: article.querySelector("h2").textContent.trim(),
 			perex: article.querySelector("[class^=ArticleTile_perex]").textContent.trim(),
 			author: article.querySelector("[class^=ArticleTile_author]").textContent.trim(),
-			loc: article.querySelector("a").href,
+			loc,
 			drop: drop_last,
 		});
 	}
+	if(!diff.length) return { json, changed: [] };
 
 	json.articles.unshift(...diff);
 	s.echo(JSON.stringify(json, null, "\t")).to(path);
 	return { json, changed: [ path ] };
 }
+/**
+ * Assumes that articles are sorted from newest to oldest
+ * @param {Drop.drop} drop
+ * @param {Sitemap} json
+ * @returns {Set<Article.loc>}
+ * */
+function dropArticles(drop, { articles }){
+	const out= new Set();
+	for(const article of articles){
+		if(article.drop !== drop) break;
+		out.add(article.loc);
+	}
+	return out;
+}
 /** @param {Sitemap} json */
 async function syncDrops(json){
 	const [ { drop: drop_last } ]= json.drops;
diff --git a/package-lock.json b/package-lock.json
index 4705ecb..ae4d39b 100644
--- a/package-lock.json
+++ b/package-lock.json
@@ -1,12 +1,12 @@
 {
 	"name": "pagenotfound-cli",
-	"version": "1.1.0",
+	"version": "1.2.0",
 	"lockfileVersion": 3,
 	"requires": true,
 	"packages": {
 		"": {
 			"name": "pagenotfound-cli",
-			"version": "1.1.0",
+			"version": "1.2.0",
 			"license": "MIT",
 			"dependencies": {
 				"jsdom": "~24.1"
diff --git a/package.json b/package.json
index 4cffb0e..dfd1470 100644
--- a/package.json
+++ b/package.json
@@ -1,6 +1,6 @@
 {
 	"name": "pagenotfound-cli",
-	"version": "1.1.1",
+	"version": "1.2.0",
 	"description": "Utility primary for generating RSS feed for Pagenotfound",
 	"bin": "cli.js",
 	"scripts": {