template_repo

---

📁 Directory Structure Explained

Here is a breakdown of what each file and folder is responsible for:

├── README.md           # Project documentation
├── artifacts/            # Stores output files like trained models (.pkl) or preprocessors (.pkl)
├── notebooks/            # Contains Jupyter notebooks for experimentation (EDA, model prototyping)
├── requirements.txt      # Lists all Python libraries required for the project
├── setup.py              # Makes the 'src' directory installable as a local Python package
└── src/                  # The main source code for the project
    ├── __init__.py       # Makes 'src' a Python package
    ├── components/       # Contains individual modules for each step of the ML pipeline
    │   ├── __init__.py
    │   ├── data_ingestion.py     # Script to get and save the data
    │   ├── data_transformation.py  # Script for all data preprocessing and feature engineering
    │   └── model_trainer.py      # Script to train and save the model
    │
    ├── pipeline/         # Contains scripts that orchestrate the components
    │   ├── __init__.py
    │   ├── predict_pipeline.py   # Script to load the trained model and make new predictions
    │   └── train_pipeline.py     # Script to run the full training workflow (ingestion -> transformation -> training)
    │
    ├── exception.py      # Custom exception handling for detailed error messages
    ├── logger.py         # Configures a custom logger to save logs to a file
    └── utils.py          # Utility functions (e.g., save_object, load_object, evaluate_model)

Key Files

• setup.py: This file is crucial. It allows you to run pip install . in your terminal, which packages your entire src directory. This lets you import your code from anywhere using from src.components.data_ingestion import DataIngestion.
• src/logger.py: All logging.info("message") calls will be written to a log file in the logs/ directory (which this script creates).
• src/exception.py: Allows you to raise CustomException(e, sys) to get a perfectly formatted error message showing the exact file and line number where the error occurred.
• src/utils.py: A central place for helper functions. A common function here is save_object(), which is used in data_transformation.py to save the preprocessor and in model_trainer.py to save the trained model.
• artifacts/: This directory is not for code. It's where your train_pipeline.py will save its outputs (e.g., model.pkl, preprocessor.pkl).