// Each #kernel tells which function to compile; you can have many kernels
#pragma kernel CSMain

RWStructuredBuffer<float4> KinectPositionBuffer;
RWStructuredBuffer<float4> KinectVelocityBuffer;
RWStructuredBuffer<float4> KinectGridBuffer;

Texture2D<float4> PositionTexture;
Texture2D<float4> VelocityTexture;

float4 ClipMin;
float4 ClipMax;
float4x4 ModelMatrix;
float4 FluidSize;
float4 FluidRoot;
float4 FluidDim;
int flipX;

[numthreads(512,1,1)]
void CSMain (uint3 id : SV_DispatchThreadID)
{
	uint2 tex_coord = uint2(id.x % 424, id.x / 424);

	float4 pos = float4(PositionTexture[tex_coord].xyz, 1);
	//pos.y *= -1;
	pos.z *= -1;
	pos.x = lerp(pos.x, -pos.x, flipX);

	float pos_mag = length(pos.xyz);

	{
		pos = mul(pos, ModelMatrix);

		float4 vel = VelocityTexture[tex_coord];
		//vel.y *= -1;
		vel.z *= -1;
		vel.x = lerp(vel.x, -vel.x, flipX);

		float3 grid_pos = (pos.xyz - FluidRoot.xyz) / FluidSize.xyz;
		int grid_index = -1;
		if (grid_pos.x > 0 && grid_pos.x < 1 &&
			grid_pos.y > 0 && grid_pos.y < 1 &&
			grid_pos.z > 0 && grid_pos.z < 1 &&
			pos_mag > 0.0f)// && length(pos) > 10.0f)
		{
			grid_pos *= FluidDim.xyz;
			grid_index = (int)grid_pos.x + (int)grid_pos.y * FluidDim.x + (int)grid_pos.z * FluidDim.x * FluidDim.y;
			KinectGridBuffer[grid_index] += float4(vel.xyz, 1);

			//const int layer = 0;
			//for (int i = 0; i < layer; i++)
			//{
			//	grid_index += FluidDim.x * FluidDim.y;
			//	KinectGridBuffer[grid_index] += float4(vel.xyz, 1);
			//}

			///PositionBuffer.Append(float4(pos.xyz, grid_index));
			///VelocityBuffer.Append(float4(vel.xyz, 0));
		}

		KinectPositionBuffer[id.x] = float4(pos.xyz, grid_index);
		KinectVelocityBuffer[id.x] = float4((pos_mag == 0.0f) ? float3(0, 0, 0) : vel.xyz, 0);

		//FullPositionBuffer[id.x] = float4(pos.xyz, grid_index);
		//FullVelocityBuffer[id.x] = float4((pos_mag == 0.0f) ? float3(0, 0, 0) : vel.xyz, 0);
	}
	
	//PositionBuffer.Append(float4(0, 0, 0, 1));
	//Result[id.xy] = float4(id.x & id.y, (id.x & 15)/15.0, (id.y & 15)/15.0, 0.0); 
}