From 07b4c1c0ed1cbd67ad0459a582e7f8263041256c Mon Sep 17 00:00:00 2001
From: Guilume <49784245+TheCutestCat@users.noreply.github.com>
Date: Sun, 5 Jan 2025 17:04:34 +0800
Subject: [PATCH] fix: not working long page screenshot (#403)

---
 crawl4ai/async_crawler_strategy.py | 26 +++++++++++++++++++++-----
 1 file changed, 21 insertions(+), 5 deletions(-)

diff --git a/crawl4ai/async_crawler_strategy.py b/crawl4ai/async_crawler_strategy.py
index 32bd14b8..4723a836 100644
--- a/crawl4ai/async_crawler_strategy.py
+++ b/crawl4ai/async_crawler_strategy.py
@@ -1639,11 +1639,9 @@ class AsyncPlaywrightCrawlerStrategy(AsyncCrawlerStrategy):
         Returns:
             str: The base64-encoded screenshot data
         """
-        dimensions = await self.get_page_dimensions(page)
-        page_height = dimensions['height']        
-        if page_height < kwargs.get(
-            "screenshot_height_threshold", SCREENSHOT_HEIGHT_TRESHOLD
-        ):
+        need_scroll = await self.page_need_scroll(page)
+        
+        if not need_scroll:
             # Page is short enough, just take a screenshot
             return await self.take_screenshot_naive(page)
         else:
@@ -2158,4 +2156,22 @@ class AsyncPlaywrightCrawlerStrategy(AsyncCrawlerStrategy):
                 const {scrollWidth, scrollHeight} = document.documentElement;
                 return {width: scrollWidth, height: scrollHeight};
             }
+        """)
+    
+    async def page_need_scroll(self, page: Page):
+        """
+        Determine whether the page need to scroll
+        
+        Args:
+            page: Playwright page object
+            
+        Returns:
+            page should scroll or not
+        """
+        return await page.evaluate("""
+            () => {
+                const scrollHeight = document.documentElement.scrollHeight;
+                const viewportHeight = window.innerHeight;
+                return scrollHeight > viewportHeight;
+            }
         """)
\ No newline at end of file