Merge branch 'main' into nsc/new-extract

apps/api/src/index.ts

@@ -80,6 +80,17 @@ function startServer(port = DEFAULT_PORT) {
       `For the Queue UI, open: http://${HOST}:${port}/admin/${process.env.BULL_AUTH_KEY}/queues`
     );
   });
+
+  const exitHandler = () => {
+    logger.info('SIGTERM signal received: closing HTTP server')
+    server.close(() => {
+      logger.info("Server closed.");
+      process.exit(0);
+    });
+  };
+
+  process.on('SIGTERM', exitHandler);
+  process.on('SIGINT', exitHandler);
   return server;
 }
 

apps/api/src/lib/crawl-redis.ts

@@ -166,10 +166,10 @@ export async function lockURLs(id: string, sc: StoredCrawl, urls: string[]): Pro
     return res;
 }
 
-export function crawlToCrawler(id: string, sc: StoredCrawl): WebCrawler {
+export function crawlToCrawler(id: string, sc: StoredCrawl, initialUrl?: string): WebCrawler {
     const crawler = new WebCrawler({
         jobId: id,
-        initialUrl: sc.originUrl!,
+        initialUrl: initialUrl ?? sc.originUrl!,
         includes: sc.crawlerOptions?.includes ?? [],
         excludes: sc.crawlerOptions?.excludes ?? [],
         maxCrawledLinks: sc.crawlerOptions?.maxCrawledLinks ?? 1000,

apps/api/src/scraper/WebScraper/crawler.ts

@@ -171,7 +171,7 @@ export class WebCrawler {
     let fullUrl = href;
     if (!href.startsWith("http")) {
       try {
-        fullUrl = new URL(href, this.baseUrl).toString();
+        fullUrl = new URL(href, url).toString();
       } catch (_) {
         return null;
       }

apps/api/src/services/queue-worker.ts

@@ -352,10 +352,10 @@ async function processJob(job: Job & { id: string }, token: string) {
 
       if (!job.data.sitemapped && job.data.crawlerOptions !== null) {
         if (!sc.cancelled) {
-          const crawler = crawlToCrawler(job.data.crawl_id, sc);
+          const crawler = crawlToCrawler(job.data.crawl_id, sc, doc.metadata?.url ?? doc.metadata?.sourceURL ?? undefined);
 
           const links = crawler.filterLinks(
-            crawler.extractLinksFromHTML(rawHtml ?? "", sc.originUrl as string),
+            crawler.extractLinksFromHTML(rawHtml ?? "", doc.metadata?.url ?? doc.metadata?.sourceURL ?? sc.originUrl as string),
             Infinity,
             sc.crawlerOptions?.maxDepth ?? 10
           );
@@ -504,19 +504,19 @@ async function processJob(job: Job & { id: string }, token: string) {
           job: job.id,
         },
       });
+
+      if (error instanceof CustomError) {
+        // Here we handle the error, then save the failed job
+        logger.error(error.message); // or any other error handling
+      }
+      logger.error(error);
+      if (error.stack) {
+        logger.error(error.stack);
+      }
     } else {
       logger.error(`🐂 Job timed out ${job.id}`);
     }
 
-    if (error instanceof CustomError) {
-      // Here we handle the error, then save the failed job
-      logger.error(error.message); // or any other error handling
-    }
-    logger.error(error);
-    if (error.stack) {
-      logger.error(error.stack);
-    }
-
     const data = {
       success: false,
       document: null,