add response modifier to prevent the removal of elements by paywall script

handlers/proxy.go

@@ -33,23 +33,23 @@ func NewProxySiteHandler(opts *ProxyOptions) fiber.Handler {
 			SetRequestModifications(
 				// rx.SpoofJA3fingerprint(ja3, "Googlebot"),
 				// rx.MasqueradeAsFacebookBot(),
-				rx.MasqueradeAsGoogleBot(),
+				//rx.MasqueradeAsGoogleBot(),
 				rx.DeleteOutgoingCookies(),
 				rx.ForwardRequestHeaders(),
 				//rx.SpoofReferrerFromGoogleSearch(),
 				rx.SpoofReferrerFromLinkedInPost(),
-				// rx.RequestWaybackMachine(),
+				//rx.RequestWaybackMachine(),
-				// rx.RequestArchiveIs(),
+				//rx.RequestArchiveIs(),
 			).
 			AddResponseModifications(
-				tx.InjectScriptBeforeDOMContentLoaded(`(() => {let d = document.createElement("div"); d.id = "adb-check"; document.body.append(d) })()`),
 				tx.ForwardResponseHeaders(),
 				tx.BypassCORS(),
 				tx.BypassContentSecurityPolicy(),
 				// tx.DeleteIncomingCookies(),
 				tx.RewriteHTMLResourceURLs(),
-				tx.PatchDynamicResourceURLs(),
 				tx.PatchTrackerScripts(),
+				tx.PatchDynamicResourceURLs(),
+				tx.BlockElementRemoval(".article-content"),
 			// tx.SetContentSecurityPolicy("default-src * 'unsafe-inline' 'unsafe-eval' data: blob:;"),
 			).
 			Execute()

proxychain/responsemodifers/block_element_removal.go

@@ -0,0 +1,42 @@
+package responsemodifers
+
+import (
+	_ "embed"
+	"strings"
+
+	"ladder/proxychain"
+	"ladder/proxychain/responsemodifers/rewriters"
+)
+
+//go:embed block_element_removal.js
+var blockElementRemoval string
+
+// BlockElementRemoval prevents paywall javascript from removing a
+// particular element by detecting the removal, then immediately reinserting it.
+// This is useful when a page will return a "fake" 404, after flashing the content briefly.
+// If the /outline/ API works, but the regular API doesn't, try this modifier.
+func BlockElementRemoval(cssSelector string) proxychain.ResponseModification {
+	return func(chain *proxychain.ProxyChain) error {
+		// don't add rewriter if it's not even html
+		ct := chain.Response.Header.Get("content-type")
+		if !strings.HasPrefix(ct, "text/html") {
+			return nil
+		}
+
+		params := map[string]string{
+			// ie: "div.article-content"
+			"{{CSS_SELECTOR}}": cssSelector,
+		}
+
+		rr := rewriters.NewScriptInjectorRewriterWithParams(
+			blockElementRemoval,
+			rewriters.BeforeDOMContentLoaded,
+			params,
+		)
+
+		htmlRewriter := rewriters.NewHTMLRewriter(chain.Response.Body, rr)
+		chain.Response.Body = htmlRewriter
+
+		return nil
+	}
+}

proxychain/responsemodifers/block_element_removal.js

@@ -0,0 +1,35 @@
+/**
+ * Monitors and restores specific DOM elements if they are removed.
+ *
+ * This self-invoking function creates a MutationObserver to watch for removal of elements matching
+ * "{{CSS_SELECTOR}}". If such an element is removed, it logs the event and attempts to restore the
+ * element after a 50ms delay. The restored element is reinserted at its original location or prepended
+ * to the document body if the original location is unavailable.
+ */
+(function() {
+  function handleMutation(mutationList) {
+    for (const mutation of mutationList) {
+      if (mutation.type === "childList") {
+        for (const node of Array.from(mutation.removedNodes)) {
+          if (node.outerHTML && node.querySelector("{{CSS_SELECTOR}}")) {
+            console.log(
+              "proxychain: prevented removal of element containing 'article-content'",
+            );
+            console.log(node.outerHTML);
+            setTimeout(() => {
+              let e = document.querySelector("{{CSS_SELECTOR}}");
+              if (e != null) {
+                e.replaceWith(node);
+              } else {
+                document.body.prepend(node);
+              }
+            }, 50);
+          }
+        }
+      }
+    }
+  }
+
+  const observer = new MutationObserver(handleMutation);
+  observer.observe(document, { childList: true, subtree: true });
+})();