How to build a workout log with ChatGPT

Set up a workout log for me in databasedmcp, in one go.

1. Create a database called workouts, described as "Workout log: exercises, sessions and sets, logged by ChatGPT from what I tell it after training."
2. Create three tables:
   - exercises: id, name (required and unique, like "Back squat"), muscle_group (required).
   - workouts: id, done_on (date, defaults to today), notes (how it felt).
   - sets: id, workout_id (references workouts, required), exercise_id (references exercises, required), set_no (the order within the workout, required), reps (required), weight_kg (defaults to 0 for bodyweight), rpe (effort from 1 to 10, only if I mention it).
3. Write the database a guide that says what each table is for, to reuse an existing exercise name before adding a new one, that sets are one row each in order, and the query for "the last time I did an exercise", plus the estimated one-rep max formula: weight_kg * (1 + reps / 30.0).

From then on: when I tell you about a session, add a workout for the day with how it felt, then log every set with its number, reps, weight and effort if I gave one, adding any exercise that's new. Show me the tables when you're finished.

CREATE TABLE exercises (
  id INTEGER PRIMARY KEY,
  name TEXT NOT NULL UNIQUE,
  muscle_group TEXT NOT NULL
);

CREATE TABLE workouts (
  id INTEGER PRIMARY KEY,
  done_on TEXT NOT NULL DEFAULT (date('now')),
  notes TEXT
);

CREATE TABLE sets (
  id INTEGER PRIMARY KEY,
  workout_id INTEGER NOT NULL REFERENCES workouts(id),
  exercise_id INTEGER NOT NULL REFERENCES exercises(id),
  set_no INTEGER NOT NULL,
  reps INTEGER NOT NULL,
  weight_kg REAL NOT NULL DEFAULT 0,
  rpe REAL CHECK (rpe BETWEEN 1 AND 10)
);

# workouts: exercises, workouts, sets

- **exercises**: one row per exercise, names unique ("Back squat", not "squats"). Reuse an existing name before adding one.
- **workouts**: one row per session, with the date and how it felt.
- **sets**: one row per set, in order. `weight_kg` is the load (0 for bodyweight), `rpe` effort from 1 to 10 if I mention it.

Last time I did an exercise: `SELECT w.done_on, s.set_no, s.reps, s.weight_kg FROM sets s JOIN workouts w ON w.id = s.workout_id JOIN exercises e ON e.id = s.exercise_id WHERE e.name = ? ORDER BY w.done_on DESC, s.set_no LIMIT 10`

Estimated one-rep max (Epley): weight_kg * (1 + reps / 30.0).

INSERT INTO exercises (name, muscle_group) VALUES (?, 'legs'), (?, 'chest'), (?, 'back'), (?, 'back')
-- params: ["Back squat","Bench press","Deadlift","Pull-up"]

INSERT INTO workouts (notes) VALUES (?) RETURNING id
-- params: ["Leg day. Felt strong."]

INSERT INTO sets (workout_id, exercise_id, set_no, reps, weight_kg, rpe) VALUES (?, ?, 1, 5, 100, NULL), (?, ?, 2, 5, 100, NULL), (?, ?, 3, 5, 100, NULL), (?, ?, 4, 3, 110, 9)
-- params: [1,1,1,1,1,1,1,1]

INSERT INTO sets (workout_id, exercise_id, set_no, reps) VALUES (?, ?, 1, 8), (?, ?, 2, 8), (?, ?, 3, 8)
-- params: [1,4,1,4,1,4]

SELECT w.done_on, s.reps, s.weight_kg, round(s.weight_kg * (1 + s.reps / 30.0), 1) AS est_1rm FROM sets s JOIN workouts w ON w.id = s.workout_id JOIN exercises e ON e.id = s.exercise_id WHERE e.name = ? ORDER BY est_1rm DESC LIMIT 3
-- params: ["Back squat"]

SELECT strftime('%Y-%W', w.done_on) AS week, e.muscle_group, count(*) AS sets, round(sum(s.reps * s.weight_kg)) AS volume_kg FROM sets s JOIN workouts w ON w.id = s.workout_id JOIN exercises e ON e.id = s.exercise_id WHERE w.done_on >= date('now', '-56 days') GROUP BY week, e.muscle_group ORDER BY week, e.muscle_group

SELECT e.name, max(w.done_on) AS last_done FROM exercises e LEFT JOIN sets s ON s.exercise_id = e.id LEFT JOIN workouts w ON w.id = s.workout_id GROUP BY e.id HAVING last_done IS NULL OR last_done < date('now', '-7 days') ORDER BY last_done