adding more rules

Re #2039

changedetectionio/processors/restock_diff.py

@@ -61,4 +61,4 @@ class perform_site_check(difference_detection_processor):
         # Always record the new checksum
         update_obj["previous_md5"] = fetched_md5
 
-        return changed_detected, update_obj, self.fetcher.instock_data.encode('utf-8')
+        return changed_detected, update_obj, self.fetcher.instock_data.encode('utf-8').strip()

changedetectionio/res/stock-not-in-stock.js

@@ -48,12 +48,22 @@ function isItemInStock() {
         'tijdelijk uitverkocht',
         'unavailable tickets',
         'we do not currently have an estimate of when this product will be back in stock.',
+        'we don\'t know when or if this item will be back in stock.',
         'zur zeit nicht an lager',
         '品切れ',
         '已售完',
         '품절'
     ];
 
+    function getElementBaseText(element) {
+        // .textContent can include text from children which may give the wrong results
+        // scan only immediate TEXT_NODEs, which will be a child of the element
+        var text = "";
+        for (var i = 0; i < element.childNodes.length; ++i)
+            if (element.childNodes[i].nodeType === Node.TEXT_NODE)
+                text += element.childNodes[i].textContent;
+        return text.toLowerCase().trim();
+    }
 
     const negateOutOfStockRegexs = [
         '[0-9] in stock'
@@ -63,18 +73,22 @@ function isItemInStock() {
         negateOutOfStockRegexs_r.push(new RegExp(negateOutOfStockRegexs[0], 'g'));
     }
 
+    // The out-of-stock or in-stock-text is generally always above-the-fold
+    // and often below-the-fold is a list of related products that may or may not contain trigger text
+    // so it's good to filter to just the 'above the fold' elements
+    // and it should be atleast 100px from the top to ignore items in the toolbar, sometimes menu items like "Coming soon" exist
+    const elementsToScan = Array.from(document.getElementsByTagName('*')).filter(element => element.getBoundingClientRect().top + window.scrollY <= window.innerHeight && element.getBoundingClientRect().top + window.scrollY >= 100);
 
-  const elementsWithZeroChildren = Array.from(document.getElementsByTagName('*')).filter(element => element.children.length === 0);
+    var elementText = "";
 
     // REGEXS THAT REALLY MEAN IT'S IN STOCK
-  for (let i = elementsWithZeroChildren.length - 1; i >= 0; i--) {
-    const element = elementsWithZeroChildren[i];
-    if (element.offsetWidth > 0 || element.offsetHeight > 0 || element.getClientRects().length > 0) {
-      var elementText="";
+    for (let i = elementsToScan.length - 1; i >= 0; i--) {
+        const element = elementsToScan[i];
+        elementText = "";
         if (element.tagName.toLowerCase() === "input") {
             elementText = element.value.toLowerCase();
         } else {
-        elementText = element.textContent.toLowerCase();
+            elementText = getElementBaseText(element);
         }
 
         if (elementText.length) {
@@ -86,24 +100,23 @@ function isItemInStock() {
             }
         }
     }
-  }
 
     // OTHER STUFF THAT COULD BE THAT IT'S OUT OF STOCK
-  for (let i = elementsWithZeroChildren.length - 1; i >= 0; i--) {
-    const element = elementsWithZeroChildren[i];
+    for (let i = elementsToScan.length - 1; i >= 0; i--) {
+        const element = elementsToScan[i];
         if (element.offsetWidth > 0 || element.offsetHeight > 0 || element.getClientRects().length > 0) {
-      var elementText="";
+            elementText = "";
             if (element.tagName.toLowerCase() === "input") {
                 elementText = element.value.toLowerCase();
             } else {
-        elementText = element.textContent.toLowerCase();
+                elementText = getElementBaseText(element);
             }
 
             if (elementText.length) {
                 // and these mean its out of stock
                 for (const outOfStockText of outOfStockTexts) {
                     if (elementText.includes(outOfStockText)) {
-            return elementText; // item is out of stock
+                        return outOfStockText; // item is out of stock
                     }
                 }
             }
@@ -114,4 +127,5 @@ function isItemInStock() {
 }
 
 // returns the element text that makes it think it's out of stock
-return isItemInStock();
+return isItemInStock().trim()
+