4. Create camera and transformation system

[quantbagel]

Build camera and matrix transformation utilities:

• Camera class (position, rotation, FOV)
• View matrix generation
• Projection matrix (perspective)
• Model transformation (translation, rotation, scale)
• MVP matrix composition

Dependencies: Issue #1

[quantbagel]

Issue #4 Implementation Complete

I've successfully implemented the camera and transformation system for gtinygrad. Here's what was completed:

📁 Files Created

gtinygrad/camera.py

A comprehensive Camera class featuring:

• Position & Rotation: Configurable position (x, y, z) and rotation using Euler angles (pitch, yaw, roll)
• Field of View: Adjustable FOV with aspect ratio, near/far clipping planes
• View Matrix Generation: Generates view matrices for transforming world space to camera space
• Projection Matrix: Perspective projection matrix generation
• Camera Movement: Helper methods for moving forward/backward, left/right, up/down
• Look-at Functionality: Point the camera at any target position
• Direction Vectors: Get forward, right, and up vectors for the camera

gtinygrad/transform.py

Matrix transformation utilities including:

• Translation Matrix: Create translation transformations
• Rotation Matrices:
  • Individual axis rotations (X, Y, Z)
  • Combined Euler angle rotations
  • Axis-angle rotation support
• Scale Matrices: Uniform and non-uniform scaling
• Projection Matrices:
  • Perspective projection
  • Orthographic projection
• View Matrices: Look-at matrix generation
• Model Matrices: Combined transformation (translate + rotate + scale)
• MVP Composition: Helper to compose Model-View-Projection matrices
• Utility Functions:
  • Point and vector transformation
  • Vector operations (normalize, cross product, dot product)
  • Matrix transpose

✅ Features Implemented

All requested features from the issue are complete:

• Camera class (position, rotation, FOV)
• View matrix generation
• Projection matrix (perspective)
• Model transformation (translation, rotation, scale)
• MVP matrix composition

🔧 Technical Details

• All matrix operations use Tensor from tinygrad for hardware acceleration
• Matrices are 4x4 homogeneous transformation matrices
• All angles are in radians (except FOV input which is in degrees)
• Camera uses right-handed coordinate system
• Standard graphics pipeline conventions (column vectors, post-multiplication)

🎯 Usage Example

from gtinygrad import Camera, transform

# Create a camera
camera = Camera(
    position=(0.0, 5.0, 10.0),
    rotation=(0.0, 0.0, 0.0),
    fov=45.0,
    aspect_ratio=16.0/9.0
)

# Get view and projection matrices
view = camera.get_view_matrix()
projection = camera.get_projection_matrix()

# Create a model transformation
model = transform.model_matrix(
    position=(0.0, 0.0, 0.0),
    rotation=(0.0, math.pi/4, 0.0),  # 45° rotation around Y axis
    scale=(1.0, 1.0, 1.0)
)

# Compose MVP matrix
mvp = transform.mvp_matrix(model, view, projection)

📦 Integration

The new modules are exported in gtinygrad/__init__.py:

from gtinygrad import Camera, transform

The implementation is ready for use with the renderer and can be integrated with the existing Renderer class for complete 3D rendering capabilities.