From dba1fb2dc818c38c8b6bed9be56ad5b700476087 Mon Sep 17 00:00:00 2001
From: rafaelsideguide <150964962+rafaelsideguide@users.noreply.github.com>
Date: Tue, 16 Jul 2024 18:22:56 -0300
Subject: [PATCH] Update removeUnwantedElements.ts

---
 .../WebScraper/utils/removeUnwantedElements.ts       | 12 ++++++++++--
 1 file changed, 10 insertions(+), 2 deletions(-)

diff --git a/apps/api/src/scraper/WebScraper/utils/removeUnwantedElements.ts b/apps/api/src/scraper/WebScraper/utils/removeUnwantedElements.ts
index 38e4c5a0..7962a4a0 100644
--- a/apps/api/src/scraper/WebScraper/utils/removeUnwantedElements.ts
+++ b/apps/api/src/scraper/WebScraper/utils/removeUnwantedElements.ts
@@ -8,7 +8,11 @@ export const removeUnwantedElements = (
 ) => {
   const soup = cheerio.load(html);
 
-  if (pageOptions.onlyIncludeTags) {
+  if (
+    pageOptions.onlyIncludeTags &&
+    pageOptions.onlyIncludeTags.length > 0 &&
+    pageOptions.onlyIncludeTags[0] !== ''
+  ) {
     if (typeof pageOptions.onlyIncludeTags === "string") {
       pageOptions.onlyIncludeTags = [pageOptions.onlyIncludeTags];
     }
@@ -26,7 +30,11 @@ export const removeUnwantedElements = (
 
   soup("script, style, iframe, noscript, meta, head").remove();
 
-  if (pageOptions.removeTags) {
+  if (
+    pageOptions.removeTags &&
+    pageOptions.removeTags.length > 0 &&
+    pageOptions.removeTags[0] !== ''
+  ) {
     if (typeof pageOptions.removeTags === "string") {
       pageOptions.removeTags = [pageOptions.removeTags];
     }