🎯 Custom Fitness

Inject arbitrary Python source code as your fitness metric, bypassing the pre-packaged defaults.

---

Prerequisites

1. VEOX Server: Start the local VEOX server (requires Docker):

   docker run -d \
     --name veox-enclave-server \
     -p 8090:8090 \
     714044927654.dkr.ecr.us-east-2.amazonaws.com/doug/single_enclave/veox-enclave-server:latest
   

   See the Quick Start for detailed server setup, health checks, and Docker Compose instructions.

2. Python SDK: Install the veox package via PyPI:

   pip install veox
   

---

What It Does

Usually, the VEOX engine chooses metrics automatically (like AUC for binary, R² for regression). For domain-specific evaluation, you can write a raw Python function and pass it as a string to custom_fitness_code.

The engine compiles and executes your code securely within the VEOX Enclave sandbox, feeding it y_prob (or predictions) and y_true.

Note

Custom fitness compilation happens securely within the VEOX Enclave sandbox.

Pro Feature

This is a Pro feature requiring an active VEOX Pro license.

---

Full Example

import os
from sklearn.datasets import make_classification
import pandas as pd
from veox import VeoxEvolver

# This Python source is sent to the server and used to evaluate candidates
CUSTOM_F1_CODE = """
from sklearn.metrics import f1_score
import numpy as np

def compute_fitness(y_prob, y_true):
    \"\"\"Custom fitness: F1-score (macro) instead of AUC.\"\"\"
    y_pred = (np.asarray(y_prob) > 0.5).astype(int)
    return f1_score(y_true, y_pred, average='macro', zero_division=0.0)
"""

api_url = os.environ.get("VEOX_API_URL", "http://127.0.0.1:8090")

# 1. Generate dataset
X, y = make_classification(n_samples=300, n_features=10, n_informative=6, random_state=42)
df = pd.DataFrame(X, columns=[f"f{i}" for i in range(10)])
df["target"] = y

# 2. Connect
evolver = VeoxEvolver("binary", api_url=api_url)
evolver.health_check()

# 3. Evolve with custom fitness metric
print("Injecting custom F1-score fitness function...")
evolver.fit(
    data=df,
    target_column="target",
    max_generations=3,
    population_size=15,
    num_islands=1,
    timeout_per_eval=30,
    max_poll_time=120,
    custom_fitness_code=CUSTOM_F1_CODE,       # 💎 PRO FEATURE: Custom fitness
)

# 4. Results
print(f"Best F1:       {evolver.best_fitness_:.4f}")
print(f"Best Pipeline: {evolver.best_pipeline_}")

---

Signature Requirements

Your injected code string must define a function exactly named compute_fitness(y_prob, y_true) that returns a numeric float.

You may safely import numpy, pandas, scipy, and scikit-learn inside the snippet.

---

💎 VEOX Pro Activation

To unlock Custom Fitness and other Pro features, you must authenticate your local node with a VEOX License Token.

from veox import VeoxEvolver

evolver = VeoxEvolver("binary", api_url="http://127.0.0.1:8090")

# 1. Fetch your unique Hardware Fingerprint
fingerprint = evolver.get_system_fingerprint()
print(f"My VEOX Node Fingerprint: {fingerprint}")

# 2. Provide this fingerprint to your VEOX Admin to receive a JWT Token
jwt_token = "eyJ0b2tlbiI6ICJVRExK...EXPIRES"  # Replace with your token

# 3. Activate the Enclave
if evolver.activate_license(jwt_token):
    print("VIP Features Unlocked!")

Tips

• Security: Your custom fitness code runs inside the encrypted sandbox — it never touches the host filesystem.
• Imports: Only numpy, pandas, scipy, and scikit-learn are available inside the snippet.
• Debugging: If your function raises an exception, the candidate receives a penalty fitness of 0.0.