🪛 Load image with metadata

Documentation

• Class name: Load image with metadata [Crystools]
• Category: crystools 🪛/Image
• Output node: True

This node is designed to load images along with their associated metadata, facilitating the integration and manipulation of image data within a broader data processing or analysis workflow. It emphasizes the seamless handling of images and their intrinsic information, aiming to streamline tasks that require both visual and contextual data.

Input types

Required

• image
  • The image parameter represents the target image file to be loaded. It is crucial for the node's operation as it determines the source of both the visual content and the associated metadata, which are essential for subsequent processing or analysis.
  • Comfy dtype: COMBO[STRING]
  • Python dtype: str

Output types

• image
  • Comfy dtype: IMAGE
  • The processed image data, ready for further manipulation or analysis.
  • Python dtype: torch.Tensor
• mask
  • Comfy dtype: MASK
  • A mask generated from the image, useful for operations that require differentiation of image regions.
  • Python dtype: torch.Tensor
• prompt
  • Comfy dtype: JSON
  • Extracted prompt information from the image's metadata, providing context or instructions related to the image.
  • Python dtype: str
• Metadata RAW
  • Comfy dtype: METADATA_RAW
  • The comprehensive metadata associated with the image, encapsulating all contextual information extracted during the loading process.
  • Python dtype: dict

Usage tips

• Infra type: GPU
• Common nodes: unknown

Source code

class CImageLoadWithMetadata:
    def __init__(self):
        pass

    @classmethod
    def INPUT_TYPES(cls):
        input_dir = folder_paths.get_input_directory()
        exclude_folders = ["clipspace"]
        file_list = []

        for root, dirs, files in os.walk(input_dir):
            # Exclude specific folders
            dirs[:] = [d for d in dirs if d not in exclude_folders]

            for file in files:
                relpath = os.path.relpath(os.path.join(root, file), start=input_dir)
                # fix for windows
                relpath = relpath.replace("\\", "/")
                file_list.append(relpath)

        return {
            "required": {
                "image": (sorted(file_list), {"image_upload": True})
            },
        }

    CATEGORY = CATEGORY.MAIN.value + CATEGORY.IMAGE.value
    RETURN_TYPES = ("IMAGE", "MASK", "JSON", "METADATA_RAW")
    RETURN_NAMES = ("image", "mask", "prompt", "Metadata RAW")
    OUTPUT_NODE = True

    FUNCTION = "execute"

    def execute(self, image):
        image_path = folder_paths.get_annotated_filepath(image)

        img, prompt, metadata = buildMetadata(image_path)

        img = ImageOps.exif_transpose(img)
        image = img.convert("RGB")
        image = np.array(image).astype(np.float32) / 255.0
        image = torch.from_numpy(image)[None,]
        if 'A' in img.getbands():
            mask = np.array(img.getchannel('A')).astype(np.float32) / 255.0
            mask = 1. - torch.from_numpy(mask)
        else:
            mask = torch.zeros((64, 64), dtype=torch.float32, device="cpu")

        return image, mask.unsqueeze(0), prompt, metadata

    @classmethod
    def IS_CHANGED(cls, image):
        image_path = folder_paths.get_annotated_filepath(image)
        m = hashlib.sha256()
        with open(image_path, 'rb') as f:
            m.update(f.read())
        return m.digest().hex()

    @classmethod
    def VALIDATE_INPUTS(cls, image):
        if not folder_paths.exists_annotated_filepath(image):
            return "Invalid image file: {}".format(image)

        return True