Build Strong CSV converter

[radupana]

Part of #73 — Converter Architecture

Overview

Implement a Strong CSV to openweight converter. Strong is the most popular strength training app and the most common source of workout data people want to migrate. This is the first app-specific converter and will validate the shared converter architecture.

Strong CSV Format

Known columns

Column	Type	Maps to	Notes
Date	datetime (format varies)	WorkoutLog.date	See date format notes below
Workout Name	string	WorkoutLog.name	Workout name/title
Duration	string (e.g. "1h 23m")	WorkoutLog.durationSeconds	Human-readable duration, needs parsing
Exercise Name	string	ExerciseLog.exercise.name	App-specific exercise name
Set Order	integer	Set ordering within exercise	1-based
Weight	number	SetLog.weight	In user's preferred unit
Reps	integer	SetLog.reps	Rep count
Distance	number	SetLog.distance	Distance value
Seconds	integer	SetLog.durationSeconds	For timed exercises
Notes	string	SetLog.notes	Per-set notes
Workout Notes	string	WorkoutLog.notes	Workout-level notes
RPE	number	SetLog.rpe	May be present in newer exports

Known quirks

• Date format varies by locale: US exports use "2024-01-15 10:30:00" but other locales may differ. Must handle multiple date formats.
• Duration is human-readable: "1h 23m 45s" format, needs regex parsing to extract seconds.
• No unit column: Weight unit depends on the user's app setting at export time. Converter needs --unit kg|lb flag to specify.
• No set type: Strong does not distinguish warmup/working/failure sets. All sets are treated equally. Conversion report should note this as data that cannot be inferred.
• No superset info: Strong does not export superset groupings.
• No RPE in most exports: RPE column may be absent in older exports. Handle gracefully.
• Exercise names include equipment: e.g., "Bench Press (Barbell)", "Lateral Raise (Dumbbell)". Can be parsed to extract equipment.
• Export may be paywalled: Strong Pro required for CSV export in some versions.
• Must export in English: Non-English exports may have translated column headers.

Data loss (Strong to openweight)

Strong's format is minimal, so conversion is mostly additive. However:

• No set types can be inferred (all become untyped)
• No superset groupings
• No RPE (in most exports)
• No tempo, RIR, or rest data

This should be documented in the conversion report.

Implementation

Strong-specific transformer (tools/converters/src/converters/strong.ts)

const STRONG_COLUMNS: ColumnMap = {
  "Date":           "date",
  "Workout Name":   "workoutName",
  "Duration":       "workoutDuration",
  "Exercise Name":  "exerciseName",
  "Set Order":      "setOrder",
  "Weight":         "weight",
  "Reps":           "reps",
  "Distance":       "distance",
  "Seconds":        "durationSeconds",
  "Notes":          "notes",
  "Workout Notes":  "workoutNotes",
  "RPE":            "rpe",
};

Key implementation details

1. Duration parsing: Parse "1h 23m 45s" format with regex. Handle variations ("1h 23m", "45m", "1h").
2. Equipment extraction: Optionally parse "Exercise (Equipment)" pattern to split into exercise.name + exercise.equipment.
3. Workout grouping: Group rows by Date + Workout Name to form individual WorkoutLog objects.
4. Exercise grouping: Within a workout, group by Exercise Name maintaining Set Order.
5. Unit handling: Accept --unit flag. Default to preserving as-is with a warning in report if no unit specified.

Test Plan

• Fixture-based tests with sample Strong CSV data
• Verify workout grouping by date + name
• Verify exercise grouping and set ordering
• Verify duration parsing for various formats
• Verify equipment extraction from exercise names
• Verify handling of missing optional columns (RPE, Distance, Seconds)
• Verify empty/null field handling
• Roundtrip: converted output passes validateWorkoutLog()
• Edge cases: single-exercise workouts, timed-only exercises, bodyweight exercises (weight = 0)

Depends On

• Converter architecture: resilient, AI-augmented format conversion pipeline #73 (converter architecture — core framework must exist first)

Part Of

• Converter architecture: resilient, AI-augmented format conversion pipeline #73 (converter architecture)

[radupana]

Implemented in the converter framework (PR from #73). Strong CSV parser at tools/converters/src/parsers/strong.ts with full test coverage. Accessible via CLI: openweight convert --weight-unit kg strong-export.csv