Project Showcase

Overview

Full-stack iOS fitness assistant made in nodejs and Swift deployed with a demo SQL backend.

Three-Tier Architecture: Frontend (SwiftUI), application logic (Node.js), and database (PostgreSQL) layers.
RESTful API Design: Implemented 15+ API endpoints following REST principles with SWAGGER Documentation.
Protocol-Oriented Network Layer: Custom Swift NetworkManager implementing protocol-oriented design for testability and abstraction.
Semantic AI Engine: Workout recommendation system utilizing NLP techniques to analyze workout history and generate personalized plans.
Database Optimization: Normalized PostgreSQL schema with proper indexing strategies.

RepBook was developed as a database optimization project, demonstrating backend architecture, and database design.

RepBook follows a strict three-tier architecture with clear boundaries between system components:

Presentation Tier (iOS): Swift and Objective-C implementation using MVVM architecture for separation of UI and business logic
Application Tier (Node.js): Express-based API server with middleware
Data Tier (PostgreSQL): 3NF relational database

The RESTful API follows the OpenAPI specification with comprehensive documentation. Core endpoints include:

POST /signup - Register a new user with account and metrics data.
POST /login - Authenticate a user and generate an auth key.
GET /checkUsername/:username - Check if a username is available.
POST /exercises - Add exercises to a workout.
POST /updateUserInfo/:memberId - Update personal user information.
GET /userDataAndMetrics/:memberId - Retrieve user data and physical metrics.
POST /setGymMembership - Set or update gym membership details.
GET /workouts/:memberId - Retrieve all workouts for a specific member.
GET /membersMetrics/:memberId - Retrieve a member's physical metrics.
POST /createWorkout/:memberId - Create a new workout.
GET /exercises - Retrieve a list of all exercises.
GET /exercises/search?workoutId=:workoutId - Retrieve exercises for a specific workout.
GET /exercises/search?q= - Search exercises by name.
DELETE /workouts/:workoutId - Delete a specific workout.
PUT /workouts/:workoutId - Rename an existing workout.
POST /loggedWorkouts - Log a completed workout session.
GET /fetchSafeData/:memberId - Fetch filtered user data and workout details.

The recommendation system implements semantic analysis of workout data to generate personalized plans:

Data Processing: Historical workout data is processed to extract training patterns and preferences
Feature Engineering: Extraction of meaningful features from raw workout data
Pattern Recognition: Analysis of exercise combinations, intensity progression, and recovery patterns
Recommendation Generation: Contextual recommendation based on user goals, history, and physiological constraints