mlops-zoomcamp-2024

This repository is a fork of the MLOps Zoomcamp course by DataTalks.Club, designed to teach MLOps principles and provide weekly homework assignments for each lecture.

• Environment Setup
• Managing Dependencies
• Acknowledgements

---

Environment Setup

• os: Ubuntu(wsl)
• python 3.10 ^
• poetry

Managing Dependencies

Using Poetry for Python Package and Dependency Management

Basic Steps

0. Install Poetry if not yet installed.

1. Navigate to your project directory:

   cd <PROJECT_DIR>

2. Initialize your project:

   poetry init -n

   • This command interactively creates a pyproject.toml file in your project directory.

3. Add dependencies:

   • Specify them in the tool.poetry.dependencies section of pyproject.toml:

     [tool.poetry.dependencies]
     numpy = "^1.19"

   • Use a mapping of package names and version constraints.

4. Use the add command to add dependencies:

   poetry add numpy

   • It finds a suitable version constraint and installs the package along with sub-dependencies.

   NOTE: If you already have pyproject.toml, you can use the install command to read the file, resolve dependencies, and install them.

   poetry install

Updating Dependencies

• To update to the latest compatible version:

  poetry update package

• To upgrade to the latest available version:

  poetry add package@latest

Installing Multiple Packages

$ poetry add pandas numpy
# or
$ poetry add $(cat requirements.txt)

Remove a dependency

The remove command removes required packages to your `pyproject.toml`` and installs them.

poetry remove numpy

Jupyter Notebook Integration

To integrate Poetry with Jupyter Notebook and manage Python packages effectively, consider using poetry-kernel. Follow the steps below, and choose the method that suits your workflow.

Method 1: Select Another Kernel

When running Jupyter Notebook, you can choose another kernel from Python environments. Use a kernel from the Python environment managed by Poetry Select the kernel named "<kernel_name>"

Method 2: Use Poetry-Managed Kernel

Create and use a new kernel managed by Poetry:

# Create a new kernel named "<kernel_name>"
$ poetry run python -m ipykernel install --user --name=<kernel_name>

Method 3: Run Jupyter Notebook Server

Start Jupyter Notebook server and connect using the provided URL:

# Run Jupyter Notebook server
$ poetry run jupyter notebook

---

Acknowledgements

Special thanks to DataTalks.Club for providing the MLOps Zoomcamp course, which forms the foundation of this repository.

Additional thanks to ChatGPT for assisting in composing, improving, and developing the content of this repository.

---