Python Dev Setup for University Coursework

Deactivate when you switch projects

deactivate

# Prompt returns to normal. The .venv folder
# stays on disk, ready for next time.

The deactivate command works on every OS once the venv is active. The prompt loses the (.venv) prefix and Python falls back to the system interpreter. Never commit the .venv folder to git; the .gitignore step at the bottom handles that.

What a requirements.txt looks like

# requirements.txt
pandas==2.2.0
numpy==1.26.4
matplotlib==3.8.2
scikit-learn==1.4.0
pytest==8.0.0
ruff==0.3.4

Pin the exact version with ==. If your professor accepts >= ranges, fine, but the strict pin is what prevents the 11 p.m. surprise where pandas updates on the TA's machine and your DataFrame.append call breaks.

requirements.txt vs pyproject.toml

requirements.txt is the right default for a single assignment. One file, one job: list packages. The pyproject.toml format wraps build metadata, entry points, and tool config; it shines for shippable libraries, not for hw3. Use pyproject.toml only if your course names it. The CS50P, DATA 100, and CSE 163 graders all read requirements.txt.

Drop this into .vscode/settings.json

{
  "python.defaultInterpreterPath": ".venv/bin/python",
  "python.terminal.activateEnvironment": true,
  "[python]": {
    "editor.defaultFormatter": "charliermarsh.ruff",
    "editor.formatOnSave": true,
    "editor.codeActionsOnSave": {
      "source.fixAll.ruff": "explicit",
      "source.organizeImports.ruff": "explicit"
    }
  },
  "python.analysis.typeCheckingMode": "basic",
  "python.testing.pytestEnabled": true,
  "python.testing.pytestArgs": ["tests"],
  "files.exclude": {
    "**/__pycache__": true,
    "**/.pytest_cache": true
  }
}

On Windows, swap .venv/bin/python for .venv\Scripts\python.exe. Commit this file to the repo so anyone cloning the project gets the same editor behavior; the .gitignore snippet at the bottom keeps everything except settings.json out.

A runnable tests/test_example.py

# tests/test_example.py
import pytest
from src.analysis import normalize_grade

def test_normalize_grade_handles_perfect_score():
    assert normalize_grade(100) == 1.0

def test_normalize_grade_handles_zero():
    assert normalize_grade(0) == 0.0

def test_normalize_grade_rejects_negative():
    with pytest.raises(ValueError):
        normalize_grade(-5)

@pytest.mark.parametrize(
    "raw,expected",
    [
        (75, 0.75),
        (50, 0.50),
        (88, 0.88),
    ],
)
def test_normalize_grade_scales_correctly(raw, expected):
    assert normalize_grade(raw) == expected

Run the tests three ways

# Run every test in the project
pytest

# Verbose output, one line per test
pytest -v

# Run a single file
pytest tests/test_example.py

# Run a single test by name
pytest tests/test_example.py::test_normalize_grade_handles_zero

# Stop after the first failure
pytest -x

# Show the locals at the failure point
pytest -l

conftest.py for shared fixtures

# tests/conftest.py
import pytest
import pandas as pd

@pytest.fixture
def sample_grades():
    return pd.DataFrame({
        "student_id": [1, 2, 3, 4],
        "raw_score": [88, 72, 95, 60],
    })

# In tests/test_analysis.py
def test_class_mean(sample_grades):
    assert sample_grades["raw_score"].mean() == 78.75

A fixture declared in conftest.py is auto-discovered by every test in the same folder. Use it for any object three or more tests want to share: a sample DataFrame, a temp file, a fake database session.

Minimal ruff.toml at the project root

# ruff.toml
line-length = 100
target-version = "py312"

[lint]
select = [
    "E",    # pycodestyle errors
    "F",    # pyflakes
    "I",    # isort
    "B",    # flake8-bugbear
    "UP",   # pyupgrade
    "SIM",  # flake8-simplify
]
ignore = [
    "E501",  # line-too-long handled by formatter
]

[format]
quote-style = "double"
indent-style = "space"

target-version = "py312" tells ruff which interpreter to lint against. Set it to "py310" or "py311" if your course pins an older interpreter. Skip the ruff.toml entirely and the defaults still work fine; the file is for the moment a professor says 120-character lines and you need to widen line-length.