12_PyPi_Pkg

Publishing a Python Package to PyPI using UV: Part-1

This guide documents the end-to-end journey of creating, troubleshooting, and successfully publishing a Python package (autoedapkd) to PyPI using uv.

---

Step 1: Initialize the Project

The Action: We started by initializing a new Python package.

uv init --package autoeda
cd autoeda

Project State:

• Before: Empty directory.
• After: uv scaffolded a standard project structure:

  autoeda/
  ├── src/
  │   └── autoeda/
  │       └── __init__.py
  ├── .gitignore
  ├── .python-version
  ├── pyproject.toml
  └── README.md
  

---

Step 2: Add Dependencies

The Action: We added the necessary libraries for data analysis and visualization.

uv add pandas numpy matplotlib seaborn

Project State:

• Before: pyproject.toml had an empty dependencies = [] block.
• After: A uv.lock file was generated, and the dependencies were added to pyproject.toml:

  dependencies = [
      "matplotlib>=3.11.0",
      "numpy>=2.5.1",
      "pandas>=3.0.3",
      "seaborn>=0.13.2",
  ]

---

Step 3: Build the Package

The Action: Once the code was ready, we generated the distribution archives.

uv build

Project State:

• Before: No distribution files existed.
• After: A new dist/ folder was created containing the .tar.gz and .whl files:

  dist/
  ├── autoeda-0.1.0-py3-none-any.whl
  └── autoeda-0.1.0.tar.gz
  

---

Step 4: Generating a PyPI API Token

Before publishing, an API token is required.

1. Log in to pypi.org
2. Go to Account settings → API tokens
3. Click Add API token, choose the scope, and generate it.

Important

PyPI shows the token only once — copy and store it securely.

---

Step 5: Initial Publish Attempt (Error)

The Action: We attempted to upload the package to PyPI.

uv publish --token <YOUR_TOKEN>

Project State:

• Result: The upload failed with a 403 Forbidden error: 403 The user 'faizymohd' isn't allowed to upload to project 'autoeda'.
• Why: The name autoeda was already taken by someone else on PyPI.

---

Step 6: Fix the Name Conflict

The Action: We had to rename the package to something unique (autoedapkd). We updated the name field in pyproject.toml, and then renamed the module directory to match.

# Rename the source directory to match the new package name
mv src/autoeda src/autoedapkd

Project State:

• Before: Module was located at src/autoeda/ and pyproject.toml used name = "autoeda".
• After: Module moved to src/autoedapkd/ and pyproject.toml updated:

  name = "autoedapkd"
  ...
  [project.scripts]
  autoeda = "autoedapkd:main"

---

Step 7: Clean and Rebuild

The Action: We removed the old dist files and built the new renamed package.

rm -rf dist
uv build

Project State:

• Before: dist/ contained the old autoeda files.
• After: dist/ was refreshed with the new autoedapkd distribution files:

  dist/
  ├── autoedapkd-0.1.0-py3-none-any.whl
  └── autoedapkd-0.1.0.tar.gz
  

---

Step 8: Successful Publication

The Action: We ran the publish command one final time.

uv publish --token <YOUR_TOKEN>

Project State:

• Before: Wheels sat unpublished in the dist folder.
• After: The package autoedapkd was successfully uploaded to PyPI and became available for anyone to install via pip install autoedapkd!

---

Publishing an Updated Version: Part 2

When you make changes to your code and want to release a new version of your package, follow these steps:

Step 1: Bump the Version Number

The Action: You must increase the version number in pyproject.toml. PyPI does not allow you to overwrite an existing version.

• Open pyproject.toml
• Change version = "0.1.0" to version = "0.1.1" (or your next version).

Project State:

• Before: version = "0.1.0"
• After: version = "0.1.1"

---

Step 2: Clean the Old Builds

The Action: Remove the old distribution files so you don't accidentally try to upload them again.

rm -rf dist

Project State:

• Before: dist/ contains version 0.1.0 files.
• After: dist/ directory is completely removed.

---

Step 3: Build the New Version

The Action: Generate the new distribution archives for the updated version.

uv build

Project State:

• Before: No dist/ directory exists.
• After: A new dist/ folder is created with the updated version files:

  dist/
  ├── autoedapkd-0.1.1-py3-none-any.whl
  └── autoedapkd-0.1.1.tar.gz
  

---

Step 4: Publish the Update

The Action: Upload the new version to PyPI using your token.

uv publish --token <YOUR_TOKEN>

Project State:

• Before: New wheels sit unpublished.
• After: Version 0.1.1 is successfully uploaded to PyPI. Users can now run pip install --upgrade autoedapkd to get the latest features!

---

Package Link

You can view and install this package directly from PyPI: https://pypi.org/project/autoedapkd/

---

Running and Using the Package: Part 3

If you want to run the package locally using the example runner script run_EDA.py , follow these steps to set up your environment:

Step 1: Initialize the environment (Optional/First time)

If you haven't initialized the project space yet, you can do so using:

uv init

Step 2: Create a Virtual Environment

Create a virtual environment to isolate your dependencies:

uv venv

Step 3: Activate the Virtual Environment

Activate the environment (Windows PowerShell/CMD):

• PowerShell:

  .venv\Scripts\activate

Step 4: Install Dependencies from requirements.txt

Install the package dependencies into your virtual environment:

uv pip install -r requirements.txt

(Or manually add autoedapkd directly using uv add autoedapkd)

Step 5: Navigate and Run the Script

Navigate into the 12_PyPi_Pkg folder (if you aren't already in it) and run the run_EDA.py script:

cd 12_PyPi_Pkg/run_autoEDA
python run_EDA.py