diff --git a/apps/api/src/scraper/WebScraper/crawler.ts b/apps/api/src/scraper/WebScraper/crawler.ts index 0248df2..1b371fd 100644 --- a/apps/api/src/scraper/WebScraper/crawler.ts +++ b/apps/api/src/scraper/WebScraper/crawler.ts @@ -25,7 +25,7 @@ export class WebCrawler { initialUrl, includes, excludes, - maxCrawledLinks, + maxCrawledLinks = 10000, limit = 10000, generateImgAltText = false, maxCrawledDepth = 10, @@ -136,7 +136,7 @@ export class WebCrawler { inProgress?: (progress: Progress) => void ): Promise { const queue = async.queue(async (task: string, callback) => { - if (this.crawledUrls.size >= this.maxCrawledLinks) { + if (this.crawledUrls.size >= Math.min(this.maxCrawledLinks, this.limit)) { if (callback && typeof callback === "function") { callback(); } @@ -147,14 +147,14 @@ export class WebCrawler { if (inProgress && newUrls.length > 0) { inProgress({ current: this.crawledUrls.size, - total: this.maxCrawledLinks, + total: Math.min(this.maxCrawledLinks, this.limit), status: "SCRAPING", currentDocumentUrl: newUrls[newUrls.length - 1], }); } else if (inProgress) { inProgress({ current: this.crawledUrls.size, - total: this.maxCrawledLinks, + total: Math.min(this.maxCrawledLinks, this.limit), status: "SCRAPING", currentDocumentUrl: task, });