18. Python Package & Environment Management#

1. uv — Overview#

  • Fast modern Python package and project manager
  • Written in Rust → 10-100x faster than pip
  • Replaces: pip + venv + pyenv combined
  • Created by Astral (makers of Ruff linter)

2. uv — Correct Command Sequence — Exam Answer:#

# ✅ Correct sequence (exam answer May_FN Q359)
uv init                      # Step 1: initialize project
uv python install 3.11       # Step 2: install specific Python version
uv add pandas numpy          # Step 3: add dependencies

Wrong Sequences:#

❌ uv init python 3.11 → uv add pandas numpy
❌ uv create --version 3.11 → uv install pandas numpy
❌ uv new --python=3.11 → uv sync pandas numpy

3. uv — Complete Command Reference:#

# Project
uv init                      # create new project
uv init my-project           # create in new folder

# Python versions
uv python install 3.11       # install Python 3.11
uv python list               # list available versions
uv python pin 3.11           # pin version for project

# Dependencies
uv add pandas                # add package
uv add pandas==2.1.0         # specific version
uv add pandas numpy scipy    # multiple packages
uv remove pandas             # remove package
uv sync                      # install from lockfile

# Running
uv run script.py             # run in project environment
uv run jupyter notebook      # run Jupyter

# Virtual environment
uv venv                      # create venv
uv venv --python 3.11        # with specific Python

# pip compatibility
uv pip install pandas
uv pip install -r requirements.txt

4. pip — Traditional Package Manager:#

pip install pandas              # install
pip install pandas==2.1.0       # specific version
pip install -r requirements.txt # from requirements file
pip uninstall pandas            # remove
pip list                        # list installed packages
pip freeze > requirements.txt   # save current environment
pip show pandas                 # package details
pip install --upgrade pandas    # upgrade package
pip install --no-cache-dir pandas  # skip cache

5. venv — Built-in Virtual Environment:#

# Create
python -m venv .venv

# Activate
source .venv/bin/activate      # Mac/Linux
.venv\Scripts\activate         # Windows
.venv\Scripts\Activate.ps1     # Windows PowerShell

# Deactivate
deactivate

# Delete
rm -rf .venv                   # Mac/Linux
rmdir /s .venv                 # Windows

360a. conda — Data Science Package Manager:#

# Create environment
conda create -n myenv python=3.11

# Activate
conda activate myenv

# Deactivate
conda deactivate

# Install packages
conda install pandas numpy

# List environments
conda env list

# Export environment
conda env export > environment.yml

# Create from file
conda env create -f environment.yml

# Remove environment
conda env remove -n myenv

conda vs pip:#

Featurepipconda
Package sourcePyPIAnaconda + PyPI
Environment mgmtNeeds venvBuilt-in
Non-Python packages❌✅ (C libraries etc.)
Data science focused❌✅
SpeedMediumSlower

6. requirements.txt — Version Pinning:#

# ✅ Exact version (fully reproducible)
pandas==2.1.0
numpy==1.24.0
scikit-learn==1.3.0
matplotlib==3.7.0

# ✅ Allow patches (minor updates)
pandas>=2.0.0,<3.0.0

# ❌ Unpinned (unpredictable)
pandas
numpy

Generate requirements.txt:#

pip freeze > requirements.txt     # current environment
pip install -r requirements.txt   # install from file

7. python-dotenv — Load .env Files:#

pip install python-dotenv
from dotenv import load_dotenv
import os

load_dotenv()                          # loads .env file
api_key = os.getenv('API_KEY')        # read variable
db_url = os.getenv('DATABASE_URL')
# .env file (NEVER commit to git!)
API_KEY=your_key_here
DATABASE_URL=postgresql://user:pass@localhost/db
DEBUG=false
SECRET_KEY=supersecretkey
# .gitignore — always exclude
.env
.env.*
*.env

Tool Comparison — Complete:#

TaskTraditionaluvconda
Create venvpython -m venvuv venvconda create
Install packagepip installuv addconda install
Install Pythonpyenv installuv python installBuilt-in
From requirementspip install -ruv syncconda env create
SpeedMedium✅ FastestSlowest
Data science❌❌✅

Why Virtual Environments?#

Without venv:
Project A needs pandas==1.5.0
Project B needs pandas==2.1.0
→ Conflict! Can only have one globally

With venv:
Project A → its own environment → pandas==1.5.0
Project B → its own environment → pandas==2.1.0
→ No conflict ✅

Quick Reference#

uv sequence (exam answer):
  uv init → uv python install 3.11 → uv add pandas ✅

uv commands:
  uv add      → add dependency
  uv remove   → remove dependency
  uv sync     → install from lockfile
  uv run      → run in project environment

pip commands:
  pip install package==version
  pip freeze > requirements.txt
  pip install -r requirements.txt

venv:
  python -m venv .venv
  source .venv/bin/activate (Mac/Linux)
  .venv\Scripts\activate (Windows)

conda:
  conda create -n env python=3.11
  conda activate env
  conda install package

.env file:
  ✅ Store secrets here
  ✅ Load with python-dotenv
  ✅ Always add to .gitignore
  ❌ Never commit to git

requirements.txt:
  pandas==2.1.0   ✅ pinned
  pandas          ❌ unpinned
Backward 17. Kumu 19. Visualization Libraries Forward