EQuBitC18/Tezrisat

AI Summary: This issue aims to clean up dead and inconsistent frontend code paths to reduce confusion and maintenance overhead. It specifically targets legacy artifacts like unused services with stale imports/endpoints and duplicate component implementations, such as `MicrocourseDataService.js` and the `ScrollProgressBar` components. The goal is to establish a single, clear implementation path for each concern by identifying and removing as much dead code as possible.

AI Summary: This issue proposes adding an LLM provider abstraction to Tezrisat to decouple its content generation from OpenAI. The primary goal is to enable clean support for local models via Ollama, making the application more flexible and provider-agnostic. This change will involve introducing a provider interface, implementing it for OpenAI and Ollama, and adding configuration-driven model selection.

AI Summary: This issue proposes adding an official Docker Compose setup to provide a reproducible and isolated local development environment. The goal is to offer a one-command stack for contributors who prefer containerized workflows, orchestrating both backend and frontend components with proper environment wiring and persistent volumes.

AI Summary: This feature request proposes adding `/healthz` and `/readyz` endpoints to the backend for improved operational monitoring and deployment automation in self-hosted environments. The `/healthz` endpoint would indicate if the process is running, while `/readyz` would confirm the readiness of all critical dependencies. These endpoints would return stable status codes and JSON payloads, making it easier to configure liveness/readiness probes.