sathish-create/fitcoach_ai_web

Purpose and entry points

This repository is a fitness/coaching application with a static browser frontend and a Python backend for pose-based exercise analysis. Primary entry points are backend/main.py for the service and frontend/index.html for the client. Deployment configuration is in render.yaml.

Backend analysis service

backend/main.py hosts the backend API/runtime and delegates exercise-form analysis into the engine. The packaged pose model backend/pose_landmarker.task is a material runtime asset, indicating local pose-landmark inference rather than a remote vision API. Python dependencies are defined in backend/requirements.txt.

Pose processing boundary

backend/utils/pose_utils.py is the boundary between raw pose-landmarker output and normalized/derived pose data consumed by analysis rules. backend/engine/analyzer.py coordinates analysis: it likely selects applicable exercise/body-region rules, evaluates landmarks, and returns coaching or validation results to the API layer.

Rule-based form evaluation

The engine is organized as modular rule groups under backend/engine/rules/. backend/engine/rules/base.py and backend/engine/rules/abs.py establish shared and abdominal-specific evaluation behavior; other groups encode body-region or exercise-form logic, including backend/engine/rules/arms.py, backend/engine/rules/back.py, backend/engine/rules/chest.py, backend/engine/rules/full_body.py, backend/engine/rules/legs.py, backend/engine/rules/shoulders.py, and backend/engine/rules/triceps.py. This keeps pose inference separate from domain-specific coaching criteria.

Frontend application surfaces

The frontend is a multi-page static application. Workout and coaching flows are represented by frontend/workout.html and frontend/coach.html; discovery and content pages include frontend/exercises.html, frontend/library.html, and frontend/gyms.html. Shared client behavior is concentrated in frontend/js/app.js, with gym-specific behavior in frontend/js/gyms.js and visual landing behavior in frontend/js/landing-effects.js.

Content, identity, and persistence

Exercise content is locally represented in frontend/exercises.json. Authentication and client-side Supabase integration are isolated in frontend/js/supabase-auth.js and frontend/js/supabase-config.js. The database contract is captured in supabase_schema.sql, making Supabase the primary external backend service for account and application persistence rather than the Python pose-analysis service.

Operational boundaries

The frontend is browser-hosted HTML/CSS/JavaScript, while the backend requires a Python runtime plus the local pose model. render.yaml is the deployment/infrastructure boundary. backend/error_log.txt and backend/verify_logic.py suggest local operational diagnostics and rule-validation support, but the production architecture centers on browser UI → Supabase for app data/auth and browser/UI → Python service for pose/form analysis.