feat: NVIDIA GPU integration

[runleveldev]

This pull request adds support for NVIDIA GPU passthrough to LXC containers managed by the application. It introduces new database fields and UI elements to track and request GPU support at both the node and container level, updates the API and OpenAPI schema, and provides user and admin documentation for setup and usage. The most significant changes are grouped below.

Database & Model Changes:

• Added a nvidiaAvailable boolean field to the Node model and table to indicate if a node has NVIDIA GPU support, with corresponding migration and UI controls for admins. [1] [2] [3] [4] [5] [6] [7]
• Added a nvidiaRequested boolean field to the Container model and table to track if GPU passthrough was requested for a container, with corresponding migration. [1] [2] [3]

API & Logic Enhancements:

• Updated the container creation API and logic to allow users to request NVIDIA GPU support. When requested, the system ensures assignment to an NVIDIA-capable node, injects required environment variables, and errors if no suitable node is available. [1] [2] [3] [4] [5] [6]
• The /new container form and API response now indicate if NVIDIA GPU support is available in the site. [1] [2]

User Interface Updates:

• Added a checkbox to the container creation form for requesting NVIDIA GPU passthrough, which is enabled or disabled based on node availability, with explanatory tooltips.
• Added a checkbox to the node form for marking a node as NVIDIA-capable.

Documentation:

• Added a detailed admin guide for configuring NVIDIA GPU passthrough on Proxmox nodes, including driver and toolkit installation and marking nodes as GPU-capable in the app.
• Updated user and developer documentation to describe the new fields, form options, and database schema changes related to NVIDIA GPU support. [1] [2] [3] [4]