support cross domain css sheet parse (#2535)

skyvern/webeye/scraper/domUtils.js

@@ -1432,7 +1432,7 @@ async function buildElementTree(
 ) {
   // Generate hover styles map at the start
   if (hoverStylesMap === undefined) {
-    hoverStylesMap = getHoverStylesMap();
+    hoverStylesMap = await getHoverStylesMap();
   }
 
   var elements = [];
@@ -2163,13 +2163,11 @@ function scrollToElementTop(element) {
  * https://stackoverflow.com/questions/7013559/is-there-a-way-to-get-element-hover-style-while-the-element-not-in-hover-state
  * https://stackoverflow.com/questions/17226676/how-to-simulate-a-mouseover-in-pure-javascript-that-activates-the-css-hover
  */
-function getHoverStylesMap() {
+async function getHoverStylesMap() {
   const hoverMap = new Map();
-  const sheets = document.styleSheets;
+  const sheets = [...document.styleSheets];
 
-  try {
+  const parseCssSheet = (sheet) => {
-    for (const sheet of sheets) {
-      try {
     const rules = sheet.cssRules || sheet.rules;
     for (const rule of rules) {
       if (rule.type === 1 && rule.selectorText) {
@@ -2224,11 +2222,55 @@ function getHoverStylesMap() {
         }
       }
     }
+  };
+
+  try {
+    await Promise.all(
+      sheets.map(async (sheet) => {
+        try {
+          parseCssSheet(sheet);
         } catch (e) {
           _jsConsoleWarn("Could not access stylesheet:", e);
-        continue;
+
+          if ((e.name !== "SecurityError" && e.code !== 18) || !sheet.href) {
+            return;
+          }
+
+          let newLink = null;
+          try {
+            _jsConsoleLog("recreating the link element: ", sheet.href);
+            const oldLink = document.querySelector(
+              `link[href="${sheet.href}"]`,
+            );
+            newLink = document.createElement("link");
+            newLink.rel = "stylesheet";
+            newLink.href = oldLink.href + "?v=" + Date.now(); // to void cache
+            newLink.crossOrigin = "anonymous";
+            // until the new link loaded, removing the old one
+            document.head.append(newLink);
+
+            // wait for a while until the sheet is fully loaded
+            await asyncSleepFor(1500);
+            const newSheets = [...document.styleSheets];
+            const refreshedSheet = newSheets.find(
+              (s) => s.href === newLink.href,
+            );
+            if (!refreshedSheet) {
+              newLink.remove();
+              return;
+            }
+            _jsConsoleLog("parsing recreated the link element: ", newLink.href);
+            parseCssSheet(refreshedSheet);
+            oldLink.remove();
+          } catch (e) {
+            _jsConsoleWarn("Error recreating the link element:", e);
+            if (newLink) {
+              newLink.remove();
             }
           }
+        }
+      }),
+    );
   } catch (e) {
     _jsConsoleError("Error processing stylesheets:", e);
   }
@@ -2468,11 +2510,11 @@ if (window.globalObserverForDOMIncrement === undefined) {
   });
 }
 
-function startGlobalIncrementalObserver(element = null) {
+async function startGlobalIncrementalObserver(element = null) {
   window.globalListnerFlag = true;
   window.globalDomDepthMap = new Map();
   window.globalOneTimeIncrementElements = [];
-  window.globalHoverStylesMap = getHoverStylesMap();
+  window.globalHoverStylesMap = await getHoverStylesMap();
   window.globalParsedElementCounter = new SafeCounter();
   window.globalObserverForDOMIncrement.takeRecords(); // cleanup the older data
   window.globalObserverForDOMIncrement.observe(document.body, {

skyvern/webeye/scraper/scraper.py

@@ -725,7 +725,7 @@ class IncrementalScrapePage:
         return self.element_tree_trimmed
 
     async def start_listen_dom_increment(self, element: ElementHandle | None = None) -> None:
-        js_script = "(element) => startGlobalIncrementalObserver(element)"
+        js_script = "async (element) => await startGlobalIncrementalObserver(element)"
         await SkyvernFrame.evaluate(frame=self.skyvern_frame.get_frame(), expression=js_script, arg=element)
 
     async def stop_listen_dom_increment(self) -> None: