seed-expense

Prompt file imported from Akshats-git/Agentic-Coding-Using-Claude-Code (.claude/commands/seed-expense.md). Fill in {{arguments}} before use. Copyright stays with the author.

Arguments received: {{arguments}}

Spendly lives in Expense-Tracker/ in this repo — all paths below are relative to that folder, and any script must run with its venv active (cd Expense-Tracker && source venv/bin/activate).

1. Parse {{arguments}} into three integers, in this order: user_id, count, months_back. If any of the three cannot be parsed, stop and print the expected usage: /seed-expense <user_id> <count> <months_back> — do not guess missing values.
2. Read Expense-Tracker/database/db.py for the expenses schema, the CATEGORIES list, and get_db(). Reuse get_db() rather than opening a new connection.
3. Verify the user_id exists in users. If it does not, stop and say so — do not create expenses for a user that does not exist.
4. Write and run a Python script that inserts count expenses for that user:
   • Spread the date values roughly evenly across the last months_back months (today going backwards).
   • Distribute category roughly proportionally across all eight categories in CATEGORIES rather than picking one category repeatedly.
   • Pick a realistic amount per category using sensible ₹ ranges (for example Food ₹50–800, Health ₹100–2000, Bills ₹200–3000) — do not use the same amount twice in a row.
   • Give each expense a short, realistic description.
5. Print a summary: how many expenses were inserted, for which user, and the date range they span.

Do not touch any file other than running this one-off script. Do not run any git commands.