Sync wip branches 3/23/2026

[nuwangeek]

This pull request introduces several improvements and fixes across the Docker Compose configuration, environment variables, frontend rendering, and documentation. The main highlights include standardizing the naming of the PostgreSQL service, improving Markdown rendering in the frontend, and adding initialization for the Langfuse database.

Docker Compose and Database Configuration Updates:

• Standardized the PostgreSQL service name from rag_search_db to rag-search-db across all Docker Compose files (docker-compose.yml, docker-compose-test.yml, docker-compose-ec2.yml). This includes updating all references in service dependencies, environment variables, and container names to ensure consistency and prevent connection issues. [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13]
• Added a database initialization script (init-langfuse.sql) for the Langfuse database, which ensures the langfuse-db is created if it does not exist. This script is now mounted into the PostgreSQL container for automatic execution on startup. [1] [2] [3]

Environment Variable Adjustments:

• Updated the example environment file (env.example) to reflect the new PostgreSQL service name and database, ensuring that the DATABASE_URL and POSTGRES_DB variables are accurate.

Frontend Improvements:

• Enhanced Markdown rendering in the TestModel page by customizing the rendering of ordered lists and links, ensuring better readability and that links open securely in a new tab. [1] [2] [3] [4]

Documentation:

• Added a new guide (kubernetes/CONTAINER_REGISTRY_SETUP.md) detailing how to build and push container images for deployment, with step-by-step instructions for each major component.

Build Configuration:

• Introduced a new Vite configuration file with a plugin for filtering out hidden menu items based on environment variables, improving frontend build flexibility.