clarion

Clarion 🚀

A lightweight, cloud-deployed web API for serving machine learning classification models.

The Mission

This project serves as the practical application for the "Cloud Computing for Data Science" diploma curriculum. Its goal is to take a trained machine learning model and make it accessible to the world through a simple, robust API.

This project picks up where the AutoCleanSE project leaves off. We will use the clean data produced by AutoCleanSE to train a model, and then we will build the infrastructure to deploy and serve that model in the cloud.

Core Features (The Roadmap)

• API Framework: A simple web server using Flask or FastAPI to define prediction endpoints.
• Cloud Storage Integration: Ability to connect to S3-compatible object storage (like MinIO or AWS S3) to retrieve assets.
• Model Serving: Load a serialized ML model (e.g., a .pkl file) and use it to make live predictions.
• Cloud Deployment: The entire application will be packaged and deployed to a cloud provider (e.g., as an AWS Lambda function).
• Containerization: The application will be containerized using Docker/Podman for portability.

Current Status

We have successfully established programmatic access to a local, S3-compatible MinIO server using boto3. This proves our ability to interact with cloud storage, which is the foundational first step.

Local Setup for Cloud Simulation

This project uses MinIO to simulate AWS S3 locally. This allows for rapid, offline development.

Running the Local MinIO Server

# First, ensure the MinIO container exists. If not, create it.
# This command runs the server in the background.
podman run -d -p 9000:9000 -p 9001:9001 --name minio-server -v ~/minio-data:/data -e "MINIO_ROOT_USER=admin" -e "MINIO_ROOT_PASSWORD=password" minio/minio server /data --console-address ":9001"

# To start the server for a new work session:
podman start minio-server

# To stop the server:
podman stop minio-server

Usage

The current test script demonstrates how to connect to the MinIO server.

# cloud_test.py
import boto3

# --- Configuration ---
access_key = "admin"
secret_key = "password"
endpoint_url = "http://127.0.0.1:9000"
# ---

s3_client = boto3.client(
    's3',
    endpoint_url=endpoint_url,
    aws_access_key_id=access_key,
    aws_secret_access_key=secret_key
)

response = s3_client.list_buckets()
print("Successfully connected. Found buckets:")
for bucket in response['Buckets']:
    print(f"- {bucket['Name']}")

License

Distributed under the MIT License.