ComfyUI V3 Custom Nodes

# Custom ComfyUI V3 Node: TextToImageSampler

```python
import torch
from comfy.sd import CLIP
from comfy.samplers import KSampler

class TextToImageSampler:
    def __init__(self):
        self.inputs = {
            "text_prompt": ("STRING", {"default": "a beautiful landscape"}),
            "width": ("INT", {"default": 512, "min": 64, "max": 2048}),
            "height": ("INT", {"default": 512, "min": 64, "max": 2048}),
            "seed": ("INT", {"default": 42, "min": 0, "max": 4294967295})
        }
        self.outputs = ("IMAGE",)

    def process(self, text_prompt, width, height, seed):
        # Initialize CLIP and KSampler
        clip = CLIP()
        latent_image = torch.zeros((1, 4, height // 8, width // 8))
        
        # Encode text prompt
        conditioning = clip.encode(text_prompt)
        
        # Sample latent image
        latent = KSampler.sample(
            latent_image, 
            seed, 
            10,  # steps
            "euler",  # sampler
            "ddim",  # scheduler
            conditioning,
            conditioning
        )
        
        return {"image": latent}
```

## Testing the Node
1. Save the node as `text_to_image_sampler.py` in your ComfyUI custom nodes folder.
2. Restart ComfyUI to load the node.
3. In the UI, search for `TextToImageSampler` and add it to your workflow.
4. Connect a text input and image output node to test the workflow.
5. Run the workflow to generate an image from the text prompt.