Writing Python

Python Development with UV

Modern Python development using uv for package management, PEP 723 for single-file scripts, and best-in-class tooling.

Quick Start

Single-File Scripts (Default)

By default, create self-contained scripts using PEP 723 format:

#!/usr/bin/env -S uv run --script
# /// script
# dependencies = [
#     "typer",
#     "rich",
# ]
# ///
"""
Script description and usage examples.

Usage:
    uv run python3 script.py --help
    uv run python3 script.py --option value
"""

import sys
# ... rest of script

Run with: uv run python3 script.py

Multi-File Projects

For larger projects requiring multiple files:

# Pin Python version
uv python pin 3.12

# Create virtual environment
uv venv --python 3.12

# Activate environment
source .venv/bin/activate

# Add dependencies
uv add package-name

# Run script
uv run python script.py

Development Tools

Testing, Linting, Type Checking

All tooling configuration is centralized in /tests/pyproject.toml.

Run from /tests directory:

cd tests

# Run tests
uv run pytest
uv run pytest -v  # verbose

# Type checking
uv run pyright
uv run pyright --stats

# Linting & formatting
uv run ruff check ../.opencode/skill
uv run ruff check --fix ../.opencode/skill
uv run ruff format ../.opencode/skill

Tool choices:

• Ruff - Linter/formatter (replaces Black, isort, Flake8)
• Pyright - Type checker (replaces MyPy)
• Pytest - Test runner

For detailed usage, see:

• references/ruff.md - Linting and formatting
• references/pyright.md - Static type checking
• references/pytest.md - Testing framework

Script Development Workflow

Start Small - Build Incrementally

1. Basic structure - Create script with --help flag
2. Test immediately - Run with uv run python3 script.py --help
3. Add --dry-run - Show what would happen without executing
4. Test again - Verify dry-run output
5. Add --verbose - Detailed output for debugging
6. Test again - Verify verbose mode
7. Continue incrementally - Add features one at a time, testing each

Shebang Format

#!/usr/bin/env uv run python3

PEP 723 Dependencies

# /// script
# dependencies = [
#     "typer",      # Modern CLI framework
#     "rich",       # Beautiful terminal output
#     "httpx",      # Modern HTTP client
# ]
# ///

Minimize dependencies - Try using stdlib first.

UV Commands Reference

Package Management

uv add <package>           # Add package to pyproject.toml
uv remove <package>        # Remove package
uv sync                    # Install/sync dependencies
uv lock                    # Create/update lock file

Python Version Management

uv python install <version>  # Install Python version
uv python list               # List installed versions
uv python pin <version>      # Set project Python version

Running Scripts

uv run python script.py      # Run with project environment
uvx <tool>                   # Run tool in isolated environment
uv tool install <package>    # Install global tool

Preferred Libraries

Core Utilities

• uv - Package manager (never use pip/python3 directly)
• typer - Modern CLI framework (built on click)
• rich - Beautiful terminal output
• python-dotenv - Environment variables (or Pydantic-Settings)

Development

• pytest - Testing framework
• ruff - Fast linting and formatting
• pyright - Static type checking

When Needed

• httpx - Modern HTTP client (replaces requests)
• Pydantic-Settings - Type-safe configuration with validation
• Polars - Fast DataFrame library (pandas alternative)
• DuckDB - Embedded analytical database
• Loguru - Simple, powerful logging

Test-Driven Development

TDD Cycle

1. Red - Write failing test for new functionality
2. Green - Write minimal code to pass test
3. Refactor - Improve code while keeping tests green

When to Use TDD

General approach: Code directly as you see fit.

Use TDD when: Facing issues or building complex components.

Development Sequence (When Using TDD)

1. Stubs - Define basic structure and interfaces
2. Pseudocode - Plan detailed logic within stubs
3. Data Layer - Implement data persistence and management
4. Business Logic - Implement core application rules
5. CLI/Frontend - Implement user interaction

Test Structure

See references/pytest.md for comprehensive testing guide.

Directory Structure

.opencode/skill/<skill>/
├── SKILL.md
└── scripts/
    ├── <script>.py
    └── tests/
        └── test_<script>.py

Helper Function Pattern

from pathlib import Path
import subprocess

SCRIPT_PATH = Path(__file__).parent.parent / "script.py"

def run_script(*args, env=None):
    """Execute script with uv run."""
    cmd = ["uv", "run", str(SCRIPT_PATH)] + list(args)
    result = subprocess.run(cmd, capture_output=True, text=True, env=env)
    return result.stdout, result.stderr, result.returncode

Test Class Organization

class TestVersion:
    """Test --version flag."""

    def test_version_flag(self):
        """--version should output version and exit with 0."""
        stdout, stderr, code = run_script("--version")
        assert code == 0
        assert "version" in stdout.lower()

Exit Code Standards

Type Checking

See references/pyright.md for comprehensive type checking guide.

Basic Type Hints

def greet(name: str) -> str:
    return f"Hello, {name}"

def find_user(id: int) -> str | None:
    return None

def process(items: list[str]) -> dict[str, int]:
    return {item: len(item) for item in items}

Docstrings

Use structured docstrings with Args, Returns, and Raises sections:

def calculate_total(items: list[dict], tax_rate: float = 0.0) -> float:
    """Calculate the total cost of items including tax.

    Args:
        items: List of item dictionaries with 'price' keys
        tax_rate: Tax rate as decimal (e.g., 0.08 for 8%)

    Returns:
        Total cost including tax

    Raises:
        ValueError: If items is empty or tax_rate is negative
    """
    if not items:
        raise ValueError("Items list cannot be empty")
    if tax_rate < 0:
        raise ValueError("Tax rate cannot be negative")
    
    subtotal = sum(item["price"] for item in items)
    return subtotal * (1 + tax_rate)

Best Practices

1. Use uv exclusively - Never run python3 or pip directly
2. Start with PEP 723 - Single-file scripts by default
3. Minimize dependencies - Try stdlib first
4. Test incrementally - Build and test feature by feature
5. Use type hints - Catch errors early with pyright
6. Format with ruff - Consistent code style
7. Follow exit codes - 0 for success, 1 for runtime errors, 2 for validation

Common Pitfalls

Mutable Default Arguments

Never use mutable objects (lists, dicts) as default argument values:

# BAD - The list persists across calls!
def add_item(item, items=[]):
    items.append(item)
    return items

add_item("a")  # ['a']
add_item("b")  # ['a', 'b'] - Unexpected!

# GOOD - Use None and create inside function
def add_item(item, items=None):
    if items is None:
        items = []
    items.append(item)
    return items

Bare Except Clauses

Never use bare except: - always catch specific exceptions:

# BAD - Catches everything including KeyboardInterrupt
try:
    do_something()
except:
    pass

# GOOD - Catch specific exceptions
try:
    do_something()
except (ValueError, TypeError) as e:
    print(f"Error: {e}", file=sys.stderr)
    sys.exit(1)

Comparing with None

Use is / is not for None comparisons:

# BAD
if value == None:
    ...

# GOOD
if value is None:
    ...

Security

Environment Variables

Store secrets in .env files, never in code:

# Load from .env file
from dotenv import load_dotenv
import os

load_dotenv()
api_key = os.getenv("API_KEY")

if not api_key:
    print("Error: API_KEY not set", file=sys.stderr)
    sys.exit(1)

Required Practices

• Never commit secrets - Add .env to .gitignore
• Never log secrets - Don't print API keys, passwords, or tokens
• Never hardcode - Use environment variables for all credentials
• Validate early - Check for required env vars at startup

.gitignore Entry

# Environment variables
.env
.env.local
.env.*.local

Bundled Resources

• references/ruff.md - Linting and formatting guide
• references/pyright.md - Type checking guide
• references/pytest.md - Testing framework guide