Fix Kaggle trigger: use kaggle CLI push instead of fake /run endpoint

The Kaggle REST API has no /run endpoint. The correct way to trigger
a new kernel run is kaggle kernels push, which creates a new kernel
version (which immediately starts executing on Kaggle's servers).

New approach:
- Reads notebook + kernel-metadata.json from the Space filesystem
- Writes temporary kaggle.json from KAGGLE_USERNAME/KAGGLE_KEY secrets
- Runs: python -m kaggle kernels push -p notebooks/
- Returns the CLI output so errors are visible in the UI

Also logs auto-trigger result to Space logs instead of silently failing.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

app.py

@@ -609,26 +609,56 @@ def _count_corrections() -> int:
 
 
 def _trigger_kaggle_training(lang: str = "bam") -> str:
-    """Fire the Kaggle kernel via REST API if credentials are configured."""
+    """
+    Push the master trainer notebook to Kaggle via the kaggle CLI,
+    which creates a new kernel version (i.e. triggers a run).
+
+    Requires KAGGLE_USERNAME + KAGGLE_KEY Space secrets.
+    The notebook is read from notebooks/ inside the Space filesystem.
+    """
     if not KAGGLE_USERNAME or not KAGGLE_KEY:
-        return "⚠️ KAGGLE_USERNAME / KAGGLE_KEY not set in Space secrets — auto-trigger disabled."
-    try:
-        import urllib.request, urllib.parse, base64
-        token = base64.b64encode(f"{KAGGLE_USERNAME}:{KAGGLE_KEY}".encode()).decode()
-        url   = f"https://www.kaggle.com/api/v1/kernels/{KAGGLE_KERNEL_SLUG}/run"
-        body  = json.dumps({"enableGpu": True}).encode()
-        req   = urllib.request.Request(
-            url, data=body, method="POST",
-            headers={
-                "Authorization": f"Basic {token}",
-                "Content-Type":  "application/json",
-            },
+        return "⚠️ KAGGLE_USERNAME / KAGGLE_KEY not set in Space secrets."
+
+    import subprocess, tempfile, shutil
+
+    notebooks_dir = ROOT / "notebooks"
+    meta_file     = notebooks_dir / "kernel-metadata.json"
+    nb_file       = notebooks_dir / "kaggle_master_trainer.ipynb"
+
+    if not nb_file.exists():
+        return "❌ notebooks/kaggle_master_trainer.ipynb not found in Space."
+    if not meta_file.exists():
+        return "❌ notebooks/kernel-metadata.json not found in Space."
+
+    # Write kaggle.json to a temp dir so the CLI authenticates
+    with tempfile.TemporaryDirectory() as tmpdir:
+        kaggle_cfg = Path(tmpdir) / ".kaggle"
+        kaggle_cfg.mkdir()
+        creds_file = kaggle_cfg / "kaggle.json"
+        creds_file.write_text(
+            json.dumps({"username": KAGGLE_USERNAME, "key": KAGGLE_KEY}),
+            encoding="utf-8",
         )
-        with urllib.request.urlopen(req, timeout=15) as r:
-            resp = json.loads(r.read())
-        return f"✅ Kaggle training triggered! Run ID: {resp.get('currentRunningVersion', 'started')}"
-    except Exception as e:
-        return f"❌ Kaggle trigger failed: {e}"
+        creds_file.chmod(0o600)
+
+        env = {
+            **__import__("os").environ,
+            "KAGGLE_CONFIG_DIR": str(kaggle_cfg),
+            "PYTHONUTF8": "1",
+            "PYTHONIOENCODING": "utf-8",
+        }
+
+        result = subprocess.run(
+            ["python", "-m", "kaggle", "kernels", "push", "-p", str(notebooks_dir)],
+            capture_output=True, text=True, timeout=60, env=env,
+        )
+
+    if result.returncode == 0:
+        output = (result.stdout or "").strip()
+        return f"✅ Kaggle training triggered!\n{output or 'Kernel version created — check Kaggle for progress.'}"
+    else:
+        err = (result.stderr or result.stdout or "unknown error").strip()
+        return f"❌ Kaggle push failed:\n{err}"
 
 
 def _maybe_auto_trigger() -> None:
@@ -637,7 +667,11 @@ def _maybe_auto_trigger() -> None:
         return
     count = _count_corrections()
     if count > 0 and count % AUTO_TRAIN_THRESHOLD == 0:
-        threading.Thread(target=_trigger_kaggle_training, daemon=True).start()
+        import logging
+        def _run():
+            msg = _trigger_kaggle_training()
+            logging.getLogger(__name__).info("Auto-trigger: %s", msg)
+        threading.Thread(target=_run, daemon=True).start()
 
 
 # ── Bulk upload handler ────────────────────────────────────────────────────────