Imported from SM23001/Gym1 (AGENTS.md). Install upstream with npx skills add SM23001/Gym1. Copyright stays with the author.

Agent guide — Gym1

Command-line gym management in Python with PostgreSQL. Entry point: python cli.py.

Database (required)

PostgreSQL runs on 192.168.1.34:5432 (LAN). The app and tests need network access to that host.

1. Copy env template: cp .env.example .env
2. Use gymdb_test for pytest (tests TRUNCATE all tables on every case)
3. Never point pytest at a production database

Create the test database once on the server if needed:

CREATE DATABASE gymdb_test OWNER gymuser;

Setup

python -m venv .venv
source .venv/bin/activate   # Windows: .venv\Scripts\activate
pip install -r requirements.txt
cp .env.example .env

Or: make setup

Verify changes

Always run after code changes:

make test
# or: .venv/bin/python -m pytest -q

Quick connectivity check:

make check-db

Run the app

make run
# or: .venv/bin/python cli.py

Seed demo data

Load sample trainers, members, classes, enrollments, and attendance via the service layer:

make seed
# or: .venv/bin/python seed.py --reset

Use GYM_DB_NAME=gymdb in .env (not gymdb_test). The script refuses to seed the test database. Without --reset, seeding is skipped if trainers already exist.

Architecture (do not violate)

Dependency flow: CLI → service → repository → db. Never put SQL in service.py or business rules in repository.py.

Adding a feature

Follow this order and match existing patterns:

1. models.py — entity if needed
2. repository.py — SQL functions
3. service.py — validation and business rules
4. cli.py / ui.py — menu and user messages
5. tests/test_*_crud.py or tests/test_service.py — copy fixture style from test_trainer_crud.py

Team ownership

MCP (Cursor)

Project MCP config: .cursor/mcp.json — gym-postgres server for schema inspection and read queries during development.

Requirements: Node.js (npx) and a reachable PostgreSQL test database (gymdb_test).

1. Copy .env.example to .env (same GYM_DB_* vars as the app).
2. Restart Cursor or toggle the server in Settings → MCP.
3. The wrapper scripts/run-mcp-postgres.sh loads .env and always connects to gymdb_test, not production gymdb.

Use MCP for debugging and exploration only. Business rules live in service.py; prefer make test to verify behavior.

Constraints

• Never commit .env (secrets stay local)
• Tests use real PostgreSQL — no mocks for persistence
• Keep CLI thin: catch BusinessError and ValueError, delegate to service
• Minimize scope: match existing naming, types, and patterns in each layer