DataNova Python SDK (datanova-sdk)

The official Python client library for the DataNova Smart Analytics Platform REST API. Supports both synchronous (requests) and asynchronous (httpx) execution modes with automatic retries, context manager resource cleanup, and safe error handling.

🚀 Quick Installation

# Standard local installation:
pip install ./datanova-sdk

# Install with Async support (httpx):
pip install ./datanova-sdk[async]

# Install in editable mode for local SDK development:
pip install -e ./datanova-sdk[dev]

⚡ Quickstart Example (Synchronous)

import logging
from datanova_sdk import DataNovaClient, DataNovaAPIError

# Optional: Enable logging for SDK debugging
logging.basicConfig(level=logging.INFO)

# Use context manager for automatic session cleanup & resource management
with DataNovaClient(
    api_key="dn_live_your_secret_key_here",
    base_url="https://data-nova-drab.vercel.app",
    max_retries=3
) as client:
    try:
        # 1. System Health Check
        health = client.health()
        print("System Status:", health.get("status"))

        # 2. Upload Dataset with MIME Type auto-detection
        upload_res = client.upload_dataset("sample_data.csv")
        dataset_id = upload_res.get("dataset_id")
        
        if dataset_id:
            print("Uploaded Dataset ID:", dataset_id)

            # 3. Run Automated Exploratory Data Analysis (EDA)
            eda_results = client.run_eda(dataset_id=dataset_id)
            print("EDA Summary:", eda_results)

            # 4. Train Classification ML Model
            ml_results = client.train_classification(dataset_id=dataset_id, target_column="Purchased")
            print("ML Model Accuracy:", ml_results.get("accuracy"))
        else:
            print("Upload completed without dataset_id key:", upload_res)

    except DataNovaAPIError as err:
        print(f"API Error ({err.status_code}): {err}")

⚡ Async Quickstart Example (FastAPI / Asyncio)

import asyncio
from datanova_sdk import AsyncDataNovaClient

async def main():
    async with AsyncDataNovaClient(
        api_key="dn_live_your_secret_key_here",
        base_url="https://data-nova-drab.vercel.app"
    ) as client:
        health = await client.health()
        print("Async Status:", health.get("status"))

        datasets = await client.list_datasets()
        print("Datasets:", datasets)

if __name__ == "__main__":
    asyncio.run(main())

🛡️ Production Features

• API Key Validation: Ensures client initialization requires a valid non-empty API key.
• Safe JSON & Response Parsing: Soft fallback for non-JSON or HTML error responses (raw_response).
• MIME Type Auto-Detection: Guesses content types dynamically (.csv, .xlsx, .xls, etc.) using mimetypes.
• Automatic Retry Mechanism: Built-in exponential backoff retries for status codes 429, 502, 503, and 504.
• Session Resource Management: Implements close() and Python context manager protocol (__enter__ / __exit__).
• Dynamic Version User-Agent: Headers include DataNova-Python-SDK/1.0.0 or DataNova-Python-SDK-Async/1.0.0.
• Integrated Logging: Debug logs and trace logging via standard Python logging.getLogger("datanova").

🛠️ Supported REST API Endpoints (All 17 APIs)