Setup & Tooling

Five minutes of setup will save you fifty hours of pain. Let's build your Python workspace the right way — once.

▶ Watch this reel

What you'll learn

Python versions & install
venv & uv/pip
pyproject.toml & requirements
IDE setup
Jupyter notebooks

Remember this

Pin one Python version per project; uv or pyenv switches between them
Every project gets its own venv — never install into system Python
Declare dependencies in pyproject.toml + a lockfile; use Jupyter only for experiments

Python versions

Many versions coexist; pin one per project (requires-python = ">=3.13").
Install via python.org, Homebrew, or uv python install 3.13.
python --version + which python are your sanity checks.

Virtual environments (venv)

Isolated per-project package folders — mandatory, not optional.
python -m venv .venv → source .venv/bin/activate → pip install ...
Windows: .venv\Scripts\activate.
uv (astral.sh): drop-in pip replacement, 10-100x faster; also manages Python versions.

Dependency declaration

File	Role
`pyproject.toml`	Metadata + dependencies + tool config (the standard)
`uv.lock` / `requirements.txt`	Exact pinned versions for reproducibility

IDE

VS Code + Python + Pylance + Ruff extensions.
Ruff = formatter + linter in one (replaces Black/isort/flake8).
Debugger: one launch.json, then F5, breakpoints, stepping.

Jupyter

Cells run individually (Shift+Enter); output stays inline.
Ideal for data exploration and LLM prompt experiments.
Pitfalls: out-of-order execution hides broken state; poor git diffs. Keep shippable code in .py files.

Code: Day-one setup — from zero to a working project

# 1. Install a Python version (once)
#    uv python install 3.13

# 2. Create a project + virtual environment
mkdir myproject && cd myproject
uv venv .venv
source .venv/bin/activate        # Windows: .venv\Scripts\activate

# 3. Declare dependencies in pyproject.toml
#    [project]
#    name = "myproject"
#    requires-python = ">=3.13"
#    dependencies = ["httpx"]

# 4. Install + lock (reproducible!)
uv pip install httpx
uv lock

# 5. Verify which Python you're using
python --version
which python                     # should point inside .venv/