Exercise 1 — Blueprint the Data Foundation

You're building a personal expense tracker CLI from scratch. This first slice is the data foundation: the Expense record and the JSON store that saves and loads it. Before you write code, blueprint the data shape, its constraints, and its edge cases — then build to that blueprint.

What you'll build

• expense_tracker/models.py — the Expense dataclass and its to_dict / from_dict serialization.
• expense_tracker/store.py — load_expenses / save_expenses over a JSON file, plus StoreError.

Where the TODOs are

• expense_tracker/models.py — define the Expense fields and implement to_dict / from_dict.
• expense_tracker/store.py — implement load_expenses / save_expenses. Watch the two edge cases called out in the TODO: a missing file returns []; a corrupt / non-array file raises StoreError. Handle both, distinctly.

Provided for you

pyproject.toml, the package skeleton, the .cursor/rules/python-cli.mdc rule file, and the stage tests (tests/test_models.py, tests/test_store.py).

Run / verify

python3.12 -m venv .venv
.venv/bin/pip install -e ".[dev]"
.venv/bin/pytest tests/test_models.py tests/test_store.py

The exercise is complete when both test files pass.