How krixtalx/nimbus works
Main Components
• Application Layer (the “App” directory): This is where the main executable resides. It includes:
• The UI system: With dedicated files for different GUI windows (e.g., Inspector, Debug, Stats, Settings) and popups. It also manages user interactions.
• Entry point and application initialization (e.g., main.cpp, Application.cpp/h).
• Core Engine (the “Core” folder): This contains the lower-level functionality and infrastructure:
• Graphics Subsystem: Rendering, shader management, texture and framebuffer handling.
• Data Management: Structures for point clouds, meshes (meshlets/pointlets), and geometry types.
• Managers and Utilities: Input, memory, file management, resource management, and utility functions (e.g., for base64, JSON, file dialogs, and progress bars).
• Camera and scene management: Handling camera projection and navigation through the virtualized scene.
• Assets & Shaders (inside “App/Assets”): These include shader programs for both compute and rendering (e.g., ComputeShaders for composing images, depth and point cloud processing; RenderShaders for drawing points, triangles, grids, etc.).
• External Dependencies & Build Scripts: The vcpkg.json file and project scripts for Windows and Linux setup determine how dependencies are managed and the build process is organized.
Relationships and Interactions
• Trace user inputs passing into the Core engine where the camera, scene rendering, and data management take over.
• How the Core graphics subsystem interacts with the GPU via:
• Compute shaders: Offloading tasks like point cloud culling, LoD subsampling, hole-filling.
• Rendering shaders: Drawing the visual result (meshes, grids, overlays) to the screen.
• Interaction between on-demand disk I/O (via the file manager and utilities) and graphics resource updates. This shows how point cloud data is fetched and moved into GPU memory as needed.
Architectural Patterns and Design Principles
• The system leverages modular design, separating UI from the rendering core for maintainability.
• There is a clear division between front-end (interactive application presentation and UI) and back-end (data management and graphics rendering).
• Data flow is highly specialized toward streaming large point cloud data, indicating asynchronous and demand-driven data fetching from disk to CPU and GPU.