The Neuroblyx ecosystem brings together our models, hardware accelerators, and Cloud APIs.

# ==========================================================
# 🧠 NEUROBLYX NCC PYTHON CLIENT (1-Click Google Colab Code)
# ==========================================================
import urllib.request, json, numpy as np

class NCC:
    def __init__(self, api_key="YOUR_API_KEY_HERE", endpoint="https://api.neuroblyx.in"):
        self.endpoint = endpoint.rstrip("/")
        self.api_key = api_key

    def _request(self, path, method="GET", data=None):
        url = f"{self.endpoint}{path}"
        req = urllib.request.Request(url, method=method)
        req.add_header("Content-Type", "application/json")
        req.add_header("User-Agent", "NCC-Python-SDK/2.0")
        if self.api_key:
            req.add_header("Authorization", f"Bearer {self.api_key}")
            req.add_header("X-API-KEY", self.api_key)
        body = json.dumps(data).encode("utf-8") if data is not None else None
        with urllib.request.urlopen(req, data=body, timeout=45) as resp:
            return json.loads(resp.read().decode("utf-8")), resp.status

    def status(self):
        res, code = self._request("/api/v1/status")
        return res if code == 200 else {"error": res}

    def vmm(self, vector, matrix, activation="linear"):
        payload = {"vector": np.array(vector, dtype=float).tolist(), "matrix": np.array(matrix, dtype=float).tolist(), "activation": activation}
        res, _ = self._request("/api/v1/vmm", method="POST", data=payload)
        return res

    def train(self, X_train, y_train, epochs=20, lr=0.03, hidden_dim=32, verbose=True):
        payload = {"X_train": np.array(X_train, dtype=float).tolist(), "y_train": np.array(y_train, dtype=int).tolist(), "epochs": int(epochs), "learning_rate": float(lr), "hidden_dim": int(hidden_dim)}
        if verbose:
            print(f"🔗 [NCC-CONNECT] Connected to {self.endpoint}")
            print(f"⚡ [NCC-DISPATCH] Dispatching {len(X_train)} samples across Analog Crossbar Tiles...")
        res, _ = self._request("/api/v1/train", method="POST", data=payload)
        if verbose and "final_train_accuracy" in res:
            acc = res["final_train_accuracy"] * 100
            loss = res["final_train_loss"]
            duration = res["hardware_telemetry"]["training_duration_ms"]
            energy = res["hardware_telemetry"]["energy_microjoules"]
            print(f"✅ [NCC-SUCCESS] Training Completed in {duration:.2f} ms")
            print(f"   • Final Accuracy: {acc:.2f}%")
            print(f"   • Final Loss: {loss:.4f}")
            print(f"   • Energy Dissipated: {energy} μJ (1,700x lower than GPU)")
        return res

# 1. Initialize NCC Client
ncc = NCC(api_key="YOUR_API_KEY_HERE", endpoint="https://api.neuroblyx.in")
print("🟢 Status:", ncc.status()["status"])

# 2. Train Machine Learning Model
np.random.seed(42)
X_train = np.random.randn(100, 4)
y_train = (X_train[:, 0] + X_train[:, 1] > 0).astype(int)
result = ncc.train(X_train=X_train, y_train=y_train, epochs=20, lr=0.03)

# 3. Real-Time VMM Inference
vmm_output = ncc.vmm(vector=[1.0, 0.5, -0.2, 0.8], matrix=[[0.25, 0.10], [0.40, -0.15], [-0.30, 0.50], [0.10, 0.20]])
print("
✨ Real-Time VMM Currents:", vmm_output["output_currents"])
print("⏱️ Hardware Physical Latency:", vmm_output["telemetry"]["latency_ms"], "ms")