From f0007935915e73d54793a8ad1f3fe9b82a2a414a Mon Sep 17 00:00:00 2001
From: LawyZheng <lawyzheng1106@gmail.com>
Date: Thu, 25 Apr 2024 09:38:39 +0800
Subject: [PATCH] Upgrade context limit (#235)

---
 skyvern/webeye/browser_manager.py  |  7 ++++---
 skyvern/webeye/scraper/domUtils.js | 15 ++++-----------
 2 files changed, 8 insertions(+), 14 deletions(-)

diff --git a/skyvern/webeye/browser_manager.py b/skyvern/webeye/browser_manager.py
index f0fd697d..93922bad 100644
--- a/skyvern/webeye/browser_manager.py
+++ b/skyvern/webeye/browser_manager.py
@@ -1,5 +1,7 @@
 from __future__ import annotations
 
+import random
+
 import structlog
 from playwright.async_api import Browser, Playwright, async_playwright
 
@@ -52,9 +54,8 @@ class BrowserManager:
             self.pages[task.task_id] = self.pages[task.workflow_run_id]
             return self.pages[task.task_id]
 
-        # TODO: percentage to use new context tree, starting from 20%
-        # new_ctx = random.choices([False, True], weights=[0.8, 0.2], k=1)[0]
-        new_ctx = False
+        # TODO: percentage (50%) to use new context tree
+        new_ctx = random.choices([False, True], weights=[0.5, 0.5], k=1)[0]
         LOG.info("Creating browser state for task", task_id=task.task_id, new_ctx=new_ctx)
         browser_state = await self._create_browser_state(task.proxy_location, task.url, new_ctx)
 
diff --git a/skyvern/webeye/scraper/domUtils.js b/skyvern/webeye/scraper/domUtils.js
index 1c9df42a..314086e6 100644
--- a/skyvern/webeye/scraper/domUtils.js
+++ b/skyvern/webeye/scraper/domUtils.js
@@ -469,11 +469,6 @@ function getElementContext(element) {
     if (childContext.length > 0) {
       fullContext.push(childContext);
     }
-
-    const charLimit = 1000;
-    if (fullContext.join(";").length > charLimit) {
-      fullContext = new Array();
-    }
   }
   return fullContext.join(";");
 }
@@ -509,10 +504,9 @@ function getElementContent(element, skipped_element = null) {
     nodeContent = cleanupText(nodeTextContentList.join(";"));
   }
   let finalTextContent = cleanupText(textContent);
-
   // Currently we don't support too much context. Character limit is 1000 per element.
   // we don't think element context has to be that big
-  const charLimit = 1000;
+  const charLimit = 5000;
   if (finalTextContent.length > charLimit) {
     if (nodeContent.length <= charLimit) {
       finalTextContent = nodeContent;
@@ -842,13 +836,13 @@ function buildTreeFromBody(new_ctx = false) {
       ) {
         let grandParentElement = parentElement.parentElement;
         if (grandParentElement) {
-          let context = getElementContext(grandParentElement, element.context);
+          let context = getElementContext(grandParentElement);
           if (context.length > 0) {
             ctx.push(context);
           }
         }
       }
-      let context = getElementContext(parentElement, element.context);
+      let context = getElementContext(parentElement);
       if (context.length > 0) {
         ctx.push(context);
       }
@@ -941,8 +935,7 @@ function buildTreeFromBody(new_ctx = false) {
     ctxList = getContextByParent(element, ctxList);
     ctxList = getContextByTable(element, ctxList);
     const context = ctxList.join(";");
-    // const context = getContextByParent(element)
-    if (context && context.length <= 1000) {
+    if (context && context.length <= 5000) {
       element.context = context;
     }