Feature/tutor feedback 1

[linnewen]

[accidently created branch from branch genai-service, so llm-related files are included even though not in main yet, merge genai-service branch first, then tutor-feedback-1?]

CI/CD pipeline:

• Split build and deploy into separate workflow files: build.yml handles image building and pushing to GHCR, deploy-vm.yml triggers automatically after build completes successfully via workflow_run
• Added Docker layer caching to all build steps using GitHub Actions cache (type=gha). Each service has its own cache scope to prevent cache invalidation across services.

closes #44
closes #40

RoadmapService.java:

• Fixed wrong @Value placeholders on userHost and userPort: they were copying the llm.service.* keys instead of user.service.*
• Replaced bare RuntimeException with proper HTTP status exceptions throughout:
  • 404 when a user is not found
  • 503 when user-service or LLM service is unreachable
  • 404 when a roadmap is not found

closes #45

Roadmap.java / Milestone.java / Task.java:

• Fixed StackOverflow from @Data on bidirectional relationships: @Data generates toString() and hashCode() that traverse the full object graph, causing infinite recursion. Replaced with explicit @Getter, @Setter, @ToString(exclude = ...), and @EqualsAndHashCode(exclude = ...)

closes #47

server/compose.yaml:

• Fixedpostgres:latest to postgres:17.5: avoids unexpected behaviour

closes #38

Code Comments & Documentation:

Added explanatory comments across all files to clarify intent:

• build.yml
• deploy-vm.yml
• deploy-k8s.yml
• provision.yml
• lint.yml
• playbook.yml
• compose.azure.yml
• GatewayController.java
• main.py