guidelinescraper 1.0.11 → 1.0.13

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "guidelinescraper",
-  "version": "1.0.11",
+  "version": "1.0.13",
   "type": "module",
   "description": "Scrape a Frontify brand portal and save every page as PDF and clean HTML",
   "bin": {

package/purge-html.mjs

@@ -92,12 +92,39 @@ export function purge(html) {
     }
   }
 
-  // Remove stray button labels left behind from copy/download widgets
-  const strayLabels = /^(Copy|Download)\s+(markdown|code|link|text)$/i;
+  // Remove Frontify UI noise from text nodes
+  const NOISE_PATTERNS = [
+    /^(Copy|Download)\s+(markdown|code|link|text)$/i,
+    /Lorem ipsum dolor sit amet[\s\S]{0,300}commodo consequat\./,
+    /Your changes could not be saved\.\s*Please reload the page and try it again\./,
+    /^Describe this color palette here$/i,
+  ];
+  const COLOR_FORMATS = "HEX|RGB|CMYK|LESS|HSL|HSB|RAL|ORA|PMS|PMS-C|PMS-U|PMS-CP|PMS-PQ|PMS-TCX|CMYK-C|CMYK-U|CMYK-N|NCS|HKS|3M|LAB|PANTONE";
+  const COLOR_EMPTY_RE = new RegExp(`^(${COLOR_FORMATS})$`);
+  const COLOR_VALUE_RE = new RegExp(`^(${COLOR_FORMATS})(.+)$`);
+
   const walk = (node) => {
     for (const child of [...node.childNodes]) {
-      if (child.nodeType === 3 && strayLabels.test(child.textContent.trim())) {
-        child.remove();
+      if (child.nodeType === 3) {
+        let text = child.textContent;
+        for (const pat of NOISE_PATTERNS) {
+          text = text.replace(pat, "");
+        }
+        // Color palette: remove format-only lines (no value)
+        if (COLOR_EMPTY_RE.test(text.trim())) {
+          child.remove();
+          continue;
+        }
+        // Color palette: insert ": " between format label and value
+        const m = text.trim().match(COLOR_VALUE_RE);
+        if (m) {
+          text = m[1] + ": " + m[2].trim();
+        }
+        if (text.trim()) {
+          child.textContent = text;
+        } else {
+          child.remove();
+        }
       } else if (child.nodeType === 1) {
         walk(child);
       }