Uniformisation de la vérification des liens

tools/add_link.js

@@ -6,6 +6,7 @@ const crypto = require("crypto");
 const path = require("path");
 const os = require("os");
 const YAML = require("yaml");
+const { buildUserAgent, checkUrl } = require("./lib/http");
 const { getArchiveUrl, saveToArchive } = require("./lib/archive");
 const { scrapePage } = require("./lib/puppeteer");
 
@@ -59,6 +60,14 @@ if (duplicateBundlePath) {
 
 // Check URL accessibility and Archive.org availability
 (async () => {
+    const userAgent = buildUserAgent();
+    const initialCheck = await checkUrl(url, { userAgent, timeoutMs: 8000 });
+    if (initialCheck.errorType || (typeof initialCheck.status === "number" && initialCheck.status >= 400)) {
+        console.warn(`⚠ Vérification HTTP avant scraping: ${initialCheck.errorType || initialCheck.status || "indéterminé"}`);
+    } else {
+        console.log(`🌐 Vérification HTTP avant scraping: ${initialCheck.status ?? "inconnue"}`);
+    }
+
     let archiveUrl = await getArchiveUrl(url);
 
     // If the URL is not archived, attempt to save it
@@ -106,13 +115,16 @@ if (duplicateBundlePath) {
 
     // Scrape the page and capture a screenshot
     console.log(`🔍 Scraping page and capturing screenshot...`);
-    const metadata = await scrapePage(url, tempScreenshotPath);
+    const metadata = await scrapePage(url, tempScreenshotPath, { userAgent });
 
     // If Puppeteer failed, do not proceed
     if (!metadata || !fs.existsSync(tempScreenshotPath)) {
         console.error(`❌ Scraping failed. No bundle will be created.`);
         process.exit(1);
     }
+    if (!metadata.httpStatus && typeof initialCheck.status === "number") {
+        metadata.httpStatus = initialCheck.status;
+    }
 
     // Create Hugo bundle only if scraping succeeded
     console.log(`📦 Creating Hugo bundle for ${url}...`);