Add networkidle exclusion to bundled driver via ts-morph at build time

patchright-python ships a vanilla driver (stealth is Python-layer), so it does NOT
consume the driver-repo's ts-morph-patched driver. To get the networkidle
captcha/analytics/heartbeat exclusion into the python package, patch the bundled
coreBundle.js (_inflightRequestStarted/_inflightRequestFinished) with ts-morph AST
after the wheels are built, then fix each wheel's RECORD hash so it stays valid.

- patch_bundled_networkidle.mjs: ts-morph transform (finds the method, reads the
  bundler-renamed param, inserts after the _isFavicon early-return; idempotent;
  hard-fails if the anchor is gone)
- patch_wheels_networkidle.py: runs it over dist/*.whl and rewrites RECORD
- build_release.yml: Setup Node + npm install + run the patcher before upload

Author: bugbase-automation

.github/workflows/build_release.yml

@@ -52,6 +52,20 @@ jobs:
             PLAYWRIGHT_TARGET_WHEEL=$wheel uv build --wheel
           done
 
+      # patchright-python bundles a vanilla driver (its stealth is Python-layer), so
+      # the driver-level networkidle exclusion has to be added to the bundled
+      # coreBundle.js here at build time - via ts-morph AST (no string replace, no
+      # marker), then the wheel RECORD is fixed so the wheel stays install-valid.
+      - name: Setup Node
+        uses: actions/setup-node@v4
+        with:
+          node-version: '22'
+
+      - name: Patch networkidle into bundled driver (ts-morph)
+        run: |
+          npm install
+          python3 patch_wheels_networkidle.py playwright-python/dist
+
       - name: Upload wheel artifacts
         uses: actions/upload-artifact@v4
         with:
@@ -77,12 +91,12 @@ jobs:
           body: |
             Custom patchright build based on playwright-python ${{ github.event.inputs.playwright_version }}.
 
-            The networkidle captcha/analytics/heartbeat exclusion is applied at the
-            **driver level** (ts-morph `framesPatch.ts`) in the bugbasesecurity
-            patchright driver fork, which this package downloads instead of the
-            upstream driver. No post-extraction JS patching in this package.
-
-            Driver: https://github.com/bugbasesecurity/patchright/releases/tag/v${{ github.event.inputs.patchright_release }}
+            Stealth is upstream patchright's Python-layer patching (unchanged). This
+            fork additionally excludes captcha/analytics/heartbeat requests from
+            `networkidle` by patching the bundled driver's `coreBundle.js`
+            (`_inflightRequestStarted`/`_inflightRequestFinished`) via **ts-morph AST**
+            at build time - no string replace, no marker comment - and fixing the
+            wheel RECORD. Same exclusion list as the driver-repo `framesPatch.ts`.
 
             Install: `pip install patchright@https://github.com/bugbasesecurity/patchright-python/releases/download/v${{ github.event.inputs.patchright_release }}/patchright-${{ github.event.inputs.patchright_release }}-py3-none-manylinux1_x86_64.whl`
           files: dist/*.whl

package.json

@@ -0,0 +1,7 @@
+{
+  "name": "patchright-python-networkidle-patch",
+  "private": true,
+  "type": "module",
+  "description": "ts-morph tooling to add the networkidle exclusion to the bundled driver at build time",
+  "dependencies": { "ts-morph": "^24.0.0" }
+}

patch_bundled_networkidle.mjs

@@ -0,0 +1,52 @@
+// AST patch (ts-morph) of the BUNDLED patchright driver (coreBundle.js) to exclude
+// captcha/analytics/heartbeat requests from networkidle. patchright-python ships a
+// vanilla driver bundle + Python-layer stealth, so networkidle (a driver-level concept)
+// must be added here at build time. This mirrors driver_patches/framesPatch.ts but runs
+// against the already-bundled coreBundle.js: it finds the _inflightRequestStarted /
+// _inflightRequestFinished methods, reads the real (bundler-renamed) param, and inserts
+// the exclusion right after the existing `if (<param>._isFavicon) return;` early-return.
+// No marker comment; hard-fails if an anchor is missing so upstream drift is caught.
+import { Project, SyntaxKind } from "ts-morph";
+
+const EXCLUDED = [
+  "challenges.cloudflare.com", "google.com/recaptcha", "www.gstatic.com/recaptcha",
+  "hcaptcha.com", "api.funcaptcha.com", "client-api.arkoselabs.com",
+  "google-analytics.com", "googletagmanager.com", "analytics.google.com",
+  "hotjar.com", "fullstory.com", "logrocket.com", "mouseflow.com", "clarity.ms",
+  "browser-intake-datadoghq.com", "sentry.io", "newrelic.com", "nr-data.net",
+  "forter.com", "/heartbeat", "/keepalive", "/keep-alive", "/beacon",
+];
+
+const target = process.argv[2];
+if (!target) { console.error("usage: node patch_bundled_networkidle.mjs <coreBundle.js>"); process.exit(2); }
+
+const project = new Project({ compilerOptions: { allowJs: true } });
+const sf = project.addSourceFileAtPath(target);
+
+function patchMethod(name) {
+  const methods = sf.getDescendantsOfKind(SyntaxKind.MethodDeclaration)
+    .filter((m) => m.getName() === name);
+  let patched = 0;
+  for (const m of methods) {
+    const body = m.getBody();
+    if (!body || body.getKind() !== SyntaxKind.Block) continue;
+    const stmts = body.getStatements();
+    const favIdx = stmts.findIndex((s) => s.getText().includes("._isFavicon"));
+    if (favIdx === -1) continue; // not the frame-manager method
+    if (body.getText().includes("some((p) => _reqUrl.includes(p))") ||
+        body.getText().includes("challenges.cloudflare.com")) { patched++; continue; } // idempotent
+    const param = m.getParameters()[0].getName();
+    body.insertStatements(favIdx + 1, [
+      `const _reqUrl = ${param}.url();`,
+      `if (${JSON.stringify(EXCLUDED)}.some((p) => _reqUrl.includes(p))) return;`,
+    ]);
+    patched++;
+  }
+  if (patched === 0) throw new Error(`networkidle patch: no '${name}' method with an _isFavicon anchor found — driver structure changed`);
+  return patched;
+}
+
+const a = patchMethod("_inflightRequestStarted");
+const b = patchMethod("_inflightRequestFinished");
+sf.saveSync();
+console.log(`patched _inflightRequestStarted x${a}, _inflightRequestFinished x${b}`);

patch_wheels_networkidle.py

@@ -0,0 +1,69 @@
+"""Apply the networkidle ts-morph patch to coreBundle.js inside built patchright
+wheels, then fix the wheel RECORD so the wheel stays install-valid.
+
+patchright-python bundles a vanilla driver (its stealth is Python-layer), so the
+driver-level networkidle exclusion has to be added to the bundled driver at build
+time. We do it with ts-morph (patch_bundled_networkidle.mjs) rather than a string
+replace, and rewrite RECORD so pip/uv accept the modified wheel.
+
+Usage: python patch_wheels_networkidle.py <wheel-or-dir> [more...]
+"""
+import base64
+import hashlib
+import subprocess
+import sys
+import tempfile
+import zipfile
+from pathlib import Path
+
+CORE = "patchright/driver/package/lib/coreBundle.js"
+MJS = Path(__file__).with_name("patch_bundled_networkidle.mjs")
+
+
+def record_hash(data: bytes) -> str:
+    digest = hashlib.sha256(data).digest()
+    return "sha256=" + base64.urlsafe_b64encode(digest).decode().rstrip("=")
+
+
+def patch_wheel(wheel: Path) -> None:
+    with zipfile.ZipFile(wheel) as z:
+        names = z.namelist()
+        if CORE not in names:
+            raise SystemExit(f"{wheel.name}: {CORE} not found in wheel")
+        contents = {n: z.read(n) for n in names}
+
+    with tempfile.TemporaryDirectory() as td:
+        core_path = Path(td) / "coreBundle.js"
+        core_path.write_bytes(contents[CORE])
+        subprocess.run(["node", str(MJS), str(core_path)], check=True)
+        patched = core_path.read_bytes()
+    contents[CORE] = patched
+
+    record_name = next(n for n in contents if n.endswith(".dist-info/RECORD"))
+    new_line = f"{CORE},{record_hash(patched)},{len(patched)}"
+    rows = []
+    for line in contents[record_name].decode().splitlines():
+        rows.append(new_line if line.startswith(CORE + ",") else line)
+    contents[record_name] = ("\n".join(rows) + "\n").encode()
+
+    tmp = wheel.with_suffix(".whl.tmp")
+    with zipfile.ZipFile(tmp, "w", zipfile.ZIP_DEFLATED) as z:
+        for name, data in contents.items():
+            z.writestr(name, data)
+    tmp.replace(wheel)
+    print(f"patched networkidle into {wheel.name}")
+
+
+def main() -> None:
+    targets: list[Path] = []
+    for arg in sys.argv[1:]:
+        p = Path(arg)
+        targets.extend(sorted(p.glob("*.whl")) if p.is_dir() else [p])
+    if not targets:
+        raise SystemExit("no wheels given")
+    for wheel in targets:
+        patch_wheel(wheel)
+
+
+if __name__ == "__main__":
+    main()