81 lines
2.6 KiB
Python
81 lines
2.6 KiB
Python
import torch
|
|
import hashlib
|
|
from pathlib import Path
|
|
from typing import Iterable
|
|
from PIL import Image, ImageOps
|
|
import numpy as np
|
|
import folder_paths
|
|
import os # 引入 os 库用于创建目录
|
|
|
|
class Snapload:
|
|
@classmethod
|
|
def INPUT_TYPES(s):
|
|
return {"required":
|
|
{"image": ("STRING", {"default": r"请转换为输入并SnapCanvas的string输出"})},
|
|
}
|
|
|
|
CATEGORY = "Snap Processing"
|
|
|
|
RETURN_TYPES = ("IMAGE", "MASK")
|
|
FUNCTION = "load_image"
|
|
|
|
def load_image(self, image):
|
|
# 调用 _resolve_path 来获取图像路径
|
|
image_path = Snapload._resolve_path(image)
|
|
|
|
# 打开并处理图像
|
|
i = Image.open(image_path)
|
|
i = ImageOps.exif_transpose(i)
|
|
image = i.convert("RGB")
|
|
image = np.array(image).astype(np.float32) / 255.0
|
|
image = torch.from_numpy(image)[None,]
|
|
|
|
# 如果有 Alpha 通道，则生成掩码
|
|
if 'A' in i.getbands():
|
|
mask = np.array(i.getchannel('A')).astype(np.float32) / 255.0
|
|
mask = 1. - torch.from_numpy(mask)
|
|
else:
|
|
# 否则生成一个全为零的掩码
|
|
mask = torch.zeros((64, 64), dtype=torch.float32, device="cpu")
|
|
|
|
return (image, mask)
|
|
|
|
# 将图像路径解析为相对路径，默认为指定的路径
|
|
def _resolve_path(image) -> Path:
|
|
# 如果未提供图像路径，则使用默认路径
|
|
if not image:
|
|
default_path = Path("custom_nodes/ComfyUI-Snap_Processing/save/output.png").resolve()
|
|
else:
|
|
default_path = Path(folder_paths.get_annotated_filepath(image)).resolve()
|
|
|
|
# 创建默认保存目录，如果不存在
|
|
save_directory = default_path.parent
|
|
if not save_directory.exists():
|
|
os.makedirs(save_directory)
|
|
|
|
# 返回默认路径或解析后的路径
|
|
return default_path
|
|
|
|
@classmethod
|
|
def IS_CHANGED(s, image):
|
|
# 获取图像路径并计算其哈希值，判断是否改变
|
|
image_path = Snapload._resolve_path(image)
|
|
m = hashlib.sha256()
|
|
with open(image_path, 'rb') as f:
|
|
m.update(f.read())
|
|
return m.digest().hex()
|
|
|
|
@classmethod
|
|
def VALIDATE_INPUTS(s, image):
|
|
# 如果图像为空，则返回 True
|
|
if image is None:
|
|
return True
|
|
|
|
# 验证图像路径是否存在
|
|
image_path = Snapload._resolve_path(image)
|
|
if not image_path.exists():
|
|
return "Invalid image path: {}".format(image_path)
|
|
|
|
return True
|
|
|