Week 10 Worklog

Week 10 Objectives:

  • Frontend:
    • Build the ChatScreen integrating AWS Bedrock (Claude 3.5 Haiku) with a sliding-window memory, and ProfileScreen with user account management.
    • Implement robust token refresh logic via Axios interceptor queues for seamless UX during session expiration.
    • Build HomeScreen as the central dashboard, resolve all outstanding bugs, and polish the overall UX across the app.
  • Backend:
    • Prepare for deployment: Docker Compose health checks, push image to Amazon ECR, define ECS Fargate task, enable AWS WAF.

Tasks to be carried out this week:

DayTaskStart DateCompletion DateReference Material
2- Backend endpoint refinements
  + Add @Valid annotations and custom constraint validators to all request DTOs
  + Standardize paginated responses: PageResponse<T> with page, size, totalPages, totalElements
  + Add GET /api/sessions/user/{userId}?startDate=&endDate= date-range filter for session history		03/16/202603/16/2026
3- Write unit tests (Spring Boot Test + JUnit 5 + Mockito)
  + UserWorkoutPlanServiceTest: clone method, activate logic, IDOR prevention
  + HealthCalculationServiceTest: BMI/BMR/TDEE formulas for various inputs
  + UserWorkoutSessionServiceTest: active session query, deactivate behavior		03/17/202603/17/2026
4- Build chatService (Frontend)
  + Direct AWS Bedrock Runtime API call (no Lambda proxy)
  + Model: anthropic.claude-3-5-haiku-20241022-v1:0
  + Vietnamese system prompt: fitness coach persona
  + Send last 12 conversation turns as context (sliding window memory)		03/18/202603/18/2026https://docs.aws.amazon.com/bedrock/
5- Build ChatScreen (Frontend)
  + Full-screen chat UI with message bubble list (user / bot)
  + Animated “typing” indicator (3 bouncing dots) while waiting for response
  + 4 quick-option chips: “Suggest exercises”, “Today’s menu”, “Calorie goal”, “Weight loss advice”
  + Vietnamese initial greeting from the fitness bot
  + Animated keyboard avoidance
  + notifyAlert error handling via global alert proxy		03/19/202603/19/2026
6- Build ProfileScreen (Frontend)
  + Display avatar (initial-letter fallback), name, email, username, birthdate, gender
  + Edit modal: update birthdate (YYYY-MM-DD) and gender via updateUserProfile API + dispatch(updateUserProfile)
  + Logout: signOut (Cognito revoke) + dispatch(logout) + clear secure storage
  + Delete account: deleteUserProfile + signOut — both gated by ConfirmModal		03/20/202603/20/2026
7- Build HomeScreen (Frontend) — 5 parallel data fetches on mount
  + Today’s meals: sum MealFood calories → daily calorie progress bar vs. 2500 kcal target
  + Latest HealthCalculation → show BMI
  + Latest BodyMetric → show current height/weight
  + Active workout plan name → quick link to PlanDetail
  + This-week session count → weekly progress vs. 4 sessions target		03/21/202603/21/2026
7- Backend deployment preparation
  + Refine Docker Compose with postgres health checks and Spring Actuator liveness/readiness probes
  + CI/CD: Push immutable backend image to Amazon ECR
  + Define ECS Fargate task definition mapped to ALB for serverless runtime
  + Protect public endpoints with AWS WAF		03/21/202603/21/2026
8- Comprehensive bug fix session across key Frontend screens
  + Fix WorkoutSessionScreen: edge case when all exercises completed before timer finishes
  + Fix DietScreen: ensure ensureDailyMeals not called on every render — move to useEffect with empty deps
  + Fix HealthDashboardScreen: loading state shown during calculateMetrics call		03/22/202603/22/2026

Week 10 Achievements:

  • Backend:
    • All request DTOs have @Valid constraints and validation was completed in Week 9.
    • API tested and ready: PageResponse<T> standardizes all paginated endpoints, date-range filtering works, unit tests all pass.
    • Production image pushed successfully to Amazon ECR, task definition created for ECS Fargate behind ALB.
    • AWS WAF configured and protecting public endpoints from common attacks.
    • Health checks on PostgreSQL and Spring Actuator ensure reliable container startup.
    • .env.example fully documents all 10+ required environment variables with descriptions.
  • Frontend:
    • chatService.sendChatToBedrock calls AWS Bedrock Runtime directly using credentials from .env.
    • Sliding window of 12 turns maintains conversation context economically.
    • Quick-option chips provide instant first interaction without typing.
    • Typing indicator creates natural chat feel; keyboard avoidance works on both iOS and Android.
    • Profile data loaded from Redux authSlice — no extra API call needed on screen mount.
    • Edit modal updates local state optimistically and confirms via API.
    • Logout and delete account flows both require confirmation — no accidental data loss.
    • 5 parallel Promise.all API calls complete in under 1.5s on localhost.
    • Daily calorie + weekly session progress indicators give users clear at-a-glance goals.
    • BMI and body metric snapshot visible on home without navigating away.
    • Workout session completion edge case resolved — app no longer freezes on final set.
    • ensureDailyMeals called only once on first mount — eliminated 4 redundant API calls per render.
    • Token refresh queue prevents logged-out flash when multiple concurrent requests get 401.
    • NotificationBox replaces native alerts — consistent in-app notification style.
    • All loading states and error messages standardized across the application.
    • Vietnamese labels complete and consistent throughout the UI.

AWS Knowledge Learned:

  • Backend:
    • Learned the full container artifact flow: optimized Docker builds, immutable image tagging, and publishing backend images to Amazon ECR.
    • Understood how ECS Fargate models deployment through task definitions, services, health checks, and desired task count.
    • Studied container networking on AWS, including private subnets, outbound access through NAT, and ALB target group integration.
    • Learned safe rollout patterns such as rolling updates and minimum healthy percent to reduce release risk.
    • Practiced separating runtime configuration from the image itself through task-level environment variables and secret injection.
    • Understood how GitHub Actions quality gates should protect AWS deployment stages by requiring lint, test, and build success first.
    • Connected release observability with rollback decisions so deployment safety is based on measurable runtime behavior.
    • In summary, connected AWS container platform knowledge directly to a realistic deployment path for the backend.
  • Frontend:
    • Learned frontend delivery architecture using S3 static hosting together with CloudFront for global caching and lower latency.
    • Studied HTTPS setup from end to end: ACM certificate issuance, validation workflow, and Route 53 DNS routing.
    • Understood the Origin Access Control model so S3 can remain private while CloudFront serves user traffic securely.
    • Learned cache-key and invalidation strategy to balance fresh UI delivery with strong CDN efficiency.
    • Studied AWS WAF as a baseline web-protection layer for common attacks and noisy traffic patterns.
    • Understood edge error handling and fallback strategies required to support SPA routing without broken deep links.
    • Practiced frontend release thinking with cache busting, controlled propagation, and rollback-friendly asset versioning.
    • In summary, focused on the AWS delivery layer that would make the frontend production-ready and globally accessible.

Next Week Plan:

  • Run full end-to-end integration testing of the entire application (all features, all screens) — 6 user journeys.
  • Write comprehensive project documentation (Backend README, Frontend guide, architecture overview).
  • Code cleanup: remove debug statements, fix linting, ensure CI/CD builds pass.
  • Response interceptor refinement + UX polish + final bug fixes (BMITrendChart empty state handling).
  • Project retrospective and preparation for handoff.