Jupyter Notebooks

Download ready-to-use Jupyter notebooks from the Kanva dashboard.

Downloading Notebooks

Each trained model in Kanva can generate customized Jupyter notebooks:

1. Open your project in the Kanva dashboard
2. Navigate to the Use Model section
3. Select API tab
4. Click Download Notebook

The notebook is pre-configured with:

• Your project ID
• Example input features matching your model's schema
• Code for both synchronous and asynchronous predictions

Notebook Contents

Setup Cell

import requests
import time

# Configuration - replace with your actual values
API_URL = "https://kanva.human-driven.ai/api/v1"
API_KEY = "YOUR_API_KEY"
PROJECT_ID = "your-project-id"

Sync Prediction Cell

def predict_sync(input_data: dict, timeout: int = 60) -> dict:
    response = requests.post(
        f"{API_URL}/projects/{PROJECT_ID}/predict",
        headers={"X-API-Key": API_KEY, "Content-Type": "application/json"},
        json={"input": input_data},
        params={"sync": "true", "timeout": str(timeout)}
    )
    response.raise_for_status()
    return response.json()

# Example usage
result = predict_sync({
    "date": ["2024-01-15"],
    "category": ["A"],
    "value": [123.45]
})

print(f"Prediction: {result['data']['predictions']}")

Async Prediction Cell

def predict_async(input_data: dict, poll_interval: float = 2.0) -> dict:
    # Start job
    response = requests.post(
        f"{API_URL}/projects/{PROJECT_ID}/predict",
        headers={"X-API-Key": API_KEY, "Content-Type": "application/json"},
        json={"input": input_data}
    )
    response.raise_for_status()
    job_id = response.json()["data"]["jobId"]

    # Poll for completion
    while True:
        status = requests.get(
            f"{API_URL}/projects/{PROJECT_ID}/jobs/{job_id}",
            headers={"X-API-Key": API_KEY}
        ).json()

        if status["data"]["status"] == "completed":
            return status
        elif status["data"]["status"] == "failed":
            raise Exception(status["data"]["error"])

        time.sleep(poll_interval)

Running in JupyterLab

1. Install JupyterLab:

   pip install jupyterlab

2. Start the server:

   jupyter lab

3. Open the downloaded .ipynb file

4. Update the API_KEY variable with your actual key

5. Run cells with Shift+Enter

Running in Google Colab

1. Go to colab.research.google.com
2. File > Upload notebook
3. Select your downloaded .ipynb file
4. Update credentials and run

Security Note

warning

The downloaded notebook contains placeholder credentials. Before sharing or committing notebooks:

• Remove your actual API key
• Replace with "YOUR_API_KEY" placeholder