How m4nu7/fsdsregressionproject works
Project type and purpose
• This is an end-to-end machine-learning application with two main use cases:
a. Training a model from raw data (via a training pipeline in src/pipeline/training_pipeline.py and Jupyter notebooks)
b. Serving predictions through a simple web interface (Flask app in application.py with HTML templates).
• It also includes CI/CD (GitHub Actions), containerization (Dockerfile), and AWS Elastic Beanstalk deployment (.ebextensions).
File structure
• Top-level:
• application.py – Flask entrypoint for prediction service
• src/ – core Python code
– components/ – building blocks: data_ingestion, data_transformation, model_trainer
– pipeline/ – orchestration: training vs. prediction pipelines
– utils.py, logger.py, exception.py – shared helpers
• artifacts/ – stored data and serialized objects (raw.csv, train.csv, test.csv, model.pkl, preprocessor.pkl)
• notebooks/ – exploratory data analysis & model-training experiments
• templates/ – front-end HTML for input form and results
• Dockerfile + .ebextensions/ – container build and AWS deployment
• .github/workflows/main.yaml – CI pipeline definition
• requirements.txt, setup.py – dependencies & packaging
README
• Minimal instructions for environment setup
• Confirms Python 3.8, pip dependencies, conda environment