MayaFlux 0.1.0
Digital-First Multimedia Processing Framework
Loading...
Searching...
No Matches
MayaFlux::Buffers::DescriptorBuffer Class Reference

Specialized buffer for shader parameter bindings from nodes. More...

#include <DescriptorBuffer.hpp>

+ Inheritance diagram for MayaFlux::Buffers::DescriptorBuffer:
+ Collaboration diagram for MayaFlux::Buffers::DescriptorBuffer:

Public Member Functions

 DescriptorBuffer (const ShaderProcessorConfig &config, size_t initial_size=4096)
 Create descriptor buffer with shader configuration.
 
 ~DescriptorBuffer () override=default
 
void initialize ()
 Initialize the buffer and its processors.
 
void bind_scalar (const std::string &name, const std::shared_ptr< Nodes::Node > &node, const std::string &descriptor_name, uint32_t set=0)
 Bind scalar node output to uniform/SSBO.
 
void bind_vector (const std::string &name, const std::shared_ptr< Nodes::Node > &node, const std::string &descriptor_name, uint32_t set=0)
 Bind vector node output to SSBO.
 
void bind_matrix (const std::string &name, const std::shared_ptr< Nodes::Node > &node, const std::string &descriptor_name, uint32_t set=0)
 Bind matrix node output to SSBO.
 
void bind_structured (const std::string &name, const std::shared_ptr< Nodes::Node > &node, const std::string &descriptor_name, uint32_t set=0)
 Bind structured node output to SSBO.
 
void unbind (const std::string &name)
 Remove a binding.
 
std::shared_ptr< DescriptorBindingsProcessorget_bindings_processor () const
 Get the bindings processor.
 
std::vector< std::string > get_binding_names () const
 Get all binding names.
 
- Public Member Functions inherited from MayaFlux::Buffers::VKBuffer
 VKBuffer (size_t size_bytes, Usage usage, Kakshya::DataModality modality)
 Construct an unregistered VKBuffer.
 
 VKBuffer ()=default
 
 ~VKBuffer () override
 Virtual destructor.
 
void clear () override
 Clear buffer contents.
 
std::vector< Kakshya::DataVariantget_data ()
 Read buffer contents as Kakshya DataVariant.
 
void set_data (const std::vector< Kakshya::DataVariant > &data)
 Write data into the buffer.
 
void resize (size_t new_size, bool preserve_data=false)
 Resize buffer and recreate GPU resources if needed.
 
size_t get_size () const
 Get current logical size in bytes.
 
void process_default () override
 Run the buffer's default processor (if set and enabled)
 
void set_default_processor (std::shared_ptr< Buffers::BufferProcessor > processor) override
 Set the buffer's default processor.
 
std::shared_ptr< Buffers::BufferProcessorget_default_processor () const override
 Get the currently attached default processor.
 
std::shared_ptr< Buffers::BufferProcessingChainget_processing_chain () override
 Access the buffer's processing chain.
 
void set_processing_chain (std::shared_ptr< Buffers::BufferProcessingChain > chain, bool force=false) override
 Replace the buffer's processing chain.
 
bool has_data_for_cycle () const override
 Checks if the buffer has data for the current processing cycle.
 
bool needs_removal () const override
 Checks if the buffer should be removed from processing chains.
 
void mark_for_processing (bool has_data) override
 Marks the buffer's data availability for the current processing cycle.
 
void mark_for_removal () override
 Marks the buffer for removal from processing chains.
 
void enforce_default_processing (bool should_process) override
 Controls whether the buffer should use default processing.
 
bool needs_default_processing () override
 Checks if the buffer should undergo default processing.
 
bool try_acquire_processing () override
 Try to acquire processing lock for this buffer.
 
void release_processing () override
 Release previously acquired processing lock.
 
bool is_processing () const override
 Query whether the buffer is currently being processed.
 
vk::Buffer & get_buffer ()
 Get VkBuffer handle (VK_NULL_HANDLE if not registered)
 
vk::DeviceSize get_size_bytes () const
 
vk::Format get_format () const
 Convert modality to a recommended VkFormat.
 
bool is_initialized () const
 Check whether Vulkan handles are present (buffer registered)
 
virtual void setup_processors (ProcessingToken token)
 Setup processors with a processing token.
 
Kakshya::DataModality get_modality () const
 Get the buffer's semantic modality.
 
const std::vector< Kakshya::DataDimension > & get_dimensions () const
 Get the inferred data dimensions for the buffer contents.
 
void set_modality (Kakshya::DataModality modality)
 Update the semantic modality and re-infer dimensions.
 
Usage get_usage () const
 Retrieve the declared usage intent.
 
void set_buffer (vk::Buffer buffer)
 Set VkBuffer handle after backend allocation.
 
void set_memory (vk::DeviceMemory memory)
 Set device memory handle after backend allocation.
 
void set_mapped_ptr (void *ptr)
 Set mapped host pointer (for host-visible allocations)
 
void set_buffer_resources (const VKBufferResources &resources)
 Set all buffer resources at once.
 
const VKBufferResourcesget_buffer_resources ()
 Get all buffer resources at once.
 
bool is_host_visible () const
 Whether this VKBuffer should be host-visible.
 
vk::BufferUsageFlags get_usage_flags () const
 Get appropriate VkBufferUsageFlags for creation based on Usage.
 
vk::MemoryPropertyFlags get_memory_properties () const
 Get appropriate VkMemoryPropertyFlags for allocation based on Usage.
 
void * get_mapped_ptr () const
 Get mapped host pointer (nullptr if not host-visible or unmapped)
 
void mark_dirty_range (size_t offset, size_t size)
 Get device memory handle.
 
void mark_invalid_range (size_t offset, size_t size)
 Mark a range as invalid (needs download)
 
std::vector< std::pair< size_t, size_t > > get_and_clear_dirty_ranges ()
 Retrieve and clear all dirty ranges.
 
std::vector< std::pair< size_t, size_t > > get_and_clear_invalid_ranges ()
 Retrieve and clear all invalid ranges.
 
void set_pipeline_window (RenderPipelineID id, const std::shared_ptr< Core::Window > &window)
 Associate this buffer with a window for rendering.
 
std::shared_ptr< Core::Windowget_pipeline_window (RenderPipelineID id) const
 Get the window associated with this buffer.
 
bool has_render_pipeline () const
 Check if this buffer has a rendering pipeline configured.
 
std::unordered_map< RenderPipelineID, std::shared_ptr< Core::Window > > get_render_pipelines () const
 Get all render pipelines associated with this buffer.
 
void set_pipeline_command (RenderPipelineID pipeline_id, CommandBufferID cmd_id)
 Store recorded command buffer for a pipeline.
 
CommandBufferID get_pipeline_command (RenderPipelineID pipeline_id) const
 Get recorded command buffer for a pipeline.
 
void clear_pipeline_commands ()
 Clear all recorded commands (called after presentation)
 
void set_vertex_layout (const Kakshya::VertexLayout &layout)
 Set vertex layout for this buffer.
 
std::optional< Kakshya::VertexLayoutget_vertex_layout () const
 Get vertex layout if set.
 
bool has_vertex_layout () const
 Check if this buffer has vertex layout configured.
 
void clear_vertex_layout ()
 Clear vertex layout.
 
std::shared_ptr< Bufferclone_to (uint8_t dest_desc) override
 Creates a clone of this buffer for a specific channel or usage enum.
 
std::shared_ptr< VKBufferclone_to (Usage usage)
 Create a clone of this buffer with the same data and properties.
 
void force_internal_usage (bool internal) override
 Set whether this buffer is for internal engine usage.
 
bool is_internal_only () const override
 Check whether this buffer is for internal engine usage.
 
- Public Member Functions inherited from MayaFlux::Buffers::Buffer
virtual ~Buffer ()=default
 Virtual destructor for proper resource management.
 

Private Attributes

std::shared_ptr< DescriptorBindingsProcessorm_bindings_processor
 
ShaderProcessorConfig m_config
 

Additional Inherited Members

- Public Types inherited from MayaFlux::Buffers::VKBuffer
enum class  Usage : uint8_t {
  STAGING , DEVICE , COMPUTE , VERTEX ,
  INDEX , UNIFORM
}
 

Detailed Description

Specialized buffer for shader parameter bindings from nodes.

Binds node outputs to shader uniforms and storage buffers. Designed for parameterizing shaders with live data: time, frequencies, control values, arrays, matrices, etc.

Philosophy:

  • Shaders are controlled by DATA, not hardcoded values
  • Any node can drive any shader parameter
  • Cross-domain flow: audio nodes → visual shader parameters

Usage: ShaderProcessorConfig config; config.shader_path = "parametric.comp"; config.bindings = { {"time", {.binding = 0, .type = vk::DescriptorType::eUniformBuffer}}, {"spectrum", {.binding = 1, .type = vk::DescriptorType::eStorageBuffer}} };

auto buffer = std::make_shared<DescriptorBuffer>(config); buffer->bind_scalar("time", time_node, "time"); buffer->bind_vector("spectrum", fft_node, "spectrum");

auto compute = std::make_shared<ComputeProcessor>(config); buffer->add_processor(compute) | Graphics;

Definition at line 36 of file DescriptorBuffer.hpp.


The documentation for this class was generated from the following files: