feat: add ComfyUI integration for image generation

- Implemented ComfyUI API for generating images with various modes (text-to-image, image-to-image, inpaint, outpaint).
- Created GenerateSheet and associated controls for user input on generation settings.
- Added subtle scrollbar styles for improved UI experience.
- Enhanced canvas input handling to ignore key events when focused on editable elements.
- Optimized canvas resizing logic to prevent unnecessary dispatches.
- Introduced error handling for generation failures and loading models.
- Added functionality to upload images and masks for inpainting.

input/tool-keybinds.ts

@@ -3,6 +3,7 @@ import type { Dispatch } from "@commands/dispatcher";
 import type { KeybindEvent } from "./keyboard";
 
 const toolKeybinds = {
+  g: "generate",
   b: "brush",
   e: "eraser",
   k: "chromaKey",

input/transform-controls.ts

@@ -16,7 +16,7 @@ import type { PointerInputEvent } from "./pointer";
 
 type TransformHandle = "body" | "nw" | "n" | "ne" | "e" | "se" | "s" | "sw" | "w";
 
-type InputToolId = "select" | "brush" | "eraser" | "chromaKey" | "magicWand" | "pan";
+type InputToolId = "select" | "generate" | "brush" | "eraser" | "chromaKey" | "magicWand" | "pan";
 
 type InputInteractionMode =
   | { type: "tool"; tool: InputToolId }