fujammy
0a761ad09a
增强 DynamicImageCombiner 的元配置并更新 README 文件。在 DynamicImageCombiner 中添加默认宽度和尺寸维护选项。更新 README_CN.md 和 README.md，以包含一张图片并改进了 Dynamic Image Input 节点的描述。
57 lines
1.8 KiB
Python
57 lines
1.8 KiB
Python
# -*- coding: utf-8 -*-
|
|
# Filename: dynamic_image_input.py
|
|
# Developer: jammyfu
|
|
# Category: 🎨Painting👓Coder/🖼️Image
|
|
|
|
import torch
|
|
|
|
class DynamicImageCombiner:
|
|
@classmethod
|
|
def INPUT_TYPES(s):
|
|
return {
|
|
"required": {}, # 不需要默认输入
|
|
"optional": {},
|
|
"_meta": {
|
|
"preferred_width": 300, # 设置默认宽度为300
|
|
"maintain_dimensions": True # 保持刷新时的尺寸
|
|
}
|
|
}
|
|
|
|
RETURN_TYPES = ("IMAGE",)
|
|
RETURN_NAMES = ("images",)
|
|
OUTPUT_IS_LIST = (True,) # 标记输出为列表
|
|
FUNCTION = "combine_images"
|
|
CATEGORY = "🎨Painting👓Coder/🖼️Image"
|
|
|
|
def combine_images(self, **image_inputs):
|
|
try:
|
|
# 收集所有非空图像
|
|
images = []
|
|
for i in range(1, len(image_inputs) + 1):
|
|
key = f"image_{i}"
|
|
if key in image_inputs and image_inputs[key] is not None:
|
|
if isinstance(image_inputs[key], torch.Tensor):
|
|
images.append(image_inputs[key])
|
|
|
|
# 如果没有图像，返回一个包含空白图像的列表
|
|
if not images:
|
|
empty_image = torch.zeros((1, 512, 512, 3))
|
|
return ([empty_image],)
|
|
|
|
# 直接返回图像列表
|
|
return (images,)
|
|
|
|
except Exception as e:
|
|
print(f"Error in DynamicImageCombiner: {str(e)}")
|
|
# 发生错误时返回包含空白图像的列表
|
|
empty_image = torch.zeros((1, 512, 512, 3))
|
|
return ([empty_image],)
|
|
|
|
# 添加到 ComfyUI 节点注册
|
|
NODE_CLASS_MAPPINGS = {
|
|
"DynamicImageCombiner": DynamicImageCombiner,
|
|
}
|
|
|
|
NODE_DISPLAY_NAME_MAPPINGS = {
|
|
"DynamicImageCombiner": "Dynamic Image Input 🖼️",
|
|
} |