"Preview image format" setting does not work

### Custom Node Testing

- [x] I have tried disabling custom nodes and the issue persists

### Expected Behavior

When that setting is set to, for example, `.jpeg`, the images in the `temp` folder should be in `.jpeg` format, not `.png`.

### Actual Behavior

The setting is ignored and preview images are stored in the default `.png` format.

### Steps to Reproduce

1. Set `Preview image format` to `jpeg` (Located under: `Settings > Lite Graph > Node Widget`)
2. Generate an image and preview it with the `Preview Image` node
3. Observe the `..\ComfyUI\temp` folder and see that the generated image is still in the default, `.png`, format and no compression has been applied.

### Debug Logs

```powershell
F:\StabilityMatrix\Data\Packages\ComfyUI\venv\Lib\site-packages\torch\cuda\__init__.py:63: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
import pynvml # type: ignore[import]
Checkpoint files will always be loaded safely.
Total VRAM 24564 MB, total RAM 32679 MB
pytorch version: 2.9.1+cu130
Set vram state to: HIGH_VRAM
Device: cuda:0 NVIDIA GeForce RTX 3090 Ti : cudaMallocAsync
Enabled pinned memory 14705.0
working around nvidia conv3d memory bug.
Using pytorch attention
Python version: 3.12.10 (main, May 30 2025, 05:39:07) [MSC v.1943 64 bit (AMD64)]
ComfyUI version: 0.3.75
ComfyUI frontend version: 1.30.6
[Prompt Server] web root: F:\StabilityMatrix\Data\Packages\ComfyUI\venv\Lib\site-packages\comfyui_frontend_package\static
Total VRAM 24564 MB, total RAM 32679 MB
pytorch version: 2.9.1+cu130
Set vram state to: HIGH_VRAM
Device: cuda:0 NVIDIA GeForce RTX 3090 Ti : cudaMallocAsync
Enabled pinned memory 14705.0
Skipping loading of custom nodes
Context impl SQLiteImpl.
Will assume non-transactional DDL.
No target revision found.
Starting server

To see the GUI go to: http://127.0.0.1:8188
got prompt
model weight dtype torch.float16, manual cast: None
model_type EPS
Using pytorch attention in VAE
Using pytorch attention in VAE
VAE load device: cuda:0, offload device: cpu, dtype: torch.bfloat16
Requested to load SDXLClipModel
loaded completely; 95367431640625005117571072.00 MB usable, 1560.80 MB loaded, full load: True
CLIP/text encoder model load device: cuda:0, offload device: cpu, current: cuda:0, dtype: torch.float16
loaded diffusion model directly to GPU
Requested to load SDXL
loaded completely; 95367431640625005117571072.00 MB usable, 4897.05 MB loaded, full load: True
100%|██████████| 20/20 [00:02<00:00, 8.44it/s]
Requested to load AutoencoderKL
loaded completely; 95367431640625005117571072.00 MB usable, 159.56 MB loaded, full load: True
Prompt executed in 6.24 seconds
got prompt
100%|██████████| 20/20 [00:02<00:00, 9.48it/s]
Prompt executed in 2.28 seconds
```

### Other

- Comfy v0.3.75 (Stability Matrix installation)
- Chromium & Librewolf
- Windows 10
- Default workflow except the previewing instead of saving
- The format does not matter nor the quality percentage, it's the same if set to `webp;50`