comfyuiDerfuuMathAnd_derfuuNodes02/Derfuu_Nodes/moddedNodes.py

import custom_nodes.Derfuu_Nodes.types as type
import custom_nodes.Derfuu_Nodes.fields as field
from custom_nodes.Derfuu_Nodes.tree import TREE_LATENTS, TREE_IMAGES

import math
import torch

import comfy.utils


class EmptyLatentImage:
    def __init__(self, device="cpu"):
        self.device = device

    @classmethod
    def INPUT_TYPES(cls):
        return {
            "required": {
                "TUPLE": (type.TUPLE,),
                "batch_size": field.INT,
            }
        }

    RETURN_TYPES = (type.LATENT,)
    FUNCTION = "generate"
    CATEGORY = TREE_LATENTS

    def generate(self, TUPLE, batch_size=1):
        width = int(TUPLE[0])
        height = int(TUPLE[1])

        latent = torch.zeros([batch_size, 4, height // 8, width // 8])
        return ({"samples": latent},)


class ImageScale_Ratio:
    upscale_methods = ["nearest-exact", "bilinear", "area"]
    crop_methods = ["disabled", "center"]

    @classmethod
    def INPUT_TYPES(cls):
        return {
            "required": {
                "IMAGE": (type.IMAGE,),
                "TUPLE": (type.TUPLE,),
                "modifier": field.FLOAT,
                "upscale_method": (cls.upscale_methods,),
                "crop": (cls.crop_methods,)}}

    RETURN_TYPES = (type.IMAGE, type.TUPLE,)
    FUNCTION = "upscale"

    CATEGORY = TREE_IMAGES

    def upscale(self, IMAGE, upscale_method, TUPLE, modifier, crop):
        samples = IMAGE.movedim(-1, 1)

        width_B = int(TUPLE[0])
        height_B = int(TUPLE[1])

        height = math.ceil(height_B * modifier)
        width = math.ceil(width_B * modifier)
        cls = comfy.utils.common_upscale(samples, width, height, upscale_method, crop)
        cls = cls.movedim(1, -1)
        return (cls, (width, height),)


class ImageScale_Side:
    upscale_methods = ["nearest-exact", "bilinear", "area"]
    crop_methods = ["disabled", "center"]

    def __init__(self) -> None:
        pass

    @classmethod
    def INPUT_TYPES(cls):
        return {
            "required": {
                "IMAGE": (type.IMAGE,),
                "TUPLE": (type.TUPLE,),
                "side_length": field.INT,
                "side": (["Width", "Height"],),
                "upscale_method": (cls.upscale_methods,),
                "crop": (cls.crop_methods,)}}

    RETURN_TYPES = (type.IMAGE, type.TUPLE,)
    FUNCTION = "upscale"

    CATEGORY = TREE_IMAGES

    def upscale(self, IMAGE, upscale_method, TUPLE, side_length, side, crop):
        samples = IMAGE.movedim(-1, 1)

        width_B = int(TUPLE[0])
        height_B = int(TUPLE[1])

        width = width_B
        height = height_B

        if side == "Width":
            heigh_ratio = height_B / width_B
            width = side_length
            height = heigh_ratio * width
        elif side == "Height":
            width_ratio = width_B / height_B
            height = side_length
            width = width_ratio * height

        width = math.ceil(width)
        height = math.ceil(height)

        cls = comfy.utils.common_upscale(samples, width, height, upscale_method, crop)
        cls = cls.movedim(1, -1)
        return (cls, (width, height), )


class LatentScale_Ratio:
    scale_methods = (["nearest-exact", "bilinear", "area"],)
    crop_methods = (["disabled", "center"],)

    def __init__(self):
        pass

    @classmethod
    def INPUT_TYPES(cls):
        return {
            "required": {
                "LATENT": (type.LATENT,),
                "TUPLE": (type.TUPLE,),
                "modifier": field.FLOAT,
                "scale_method": cls.scale_methods,
                "crop": cls.crop_methods,
            }
        }

    RETURN_TYPES = (type.LATENT, type.TUPLE,)
    FUNCTION = "scale"
    CATEGORY = TREE_LATENTS

    def scale(self, LATENT, scale_method, crop, modifier, TUPLE):

        width = int(TUPLE[0] * modifier)
        height = int(TUPLE[1] * modifier)

        cls = LATENT.copy()
        cls["samples"] = comfy.utils.common_upscale(LATENT["samples"], width // 8, height // 8, scale_method, crop)
        return (cls, (width, height),)


class LatentScale_Side:
    upscale_methods = ["nearest-exact", "bilinear", "area"]
    crop_methods = ["disabled", "center"]

    def __init__(self) -> None:
        pass

    @classmethod
    def INPUT_TYPES(cls):
        return {
            "required": {
                "LATENT": (type.LATENT,),
                "TUPLE": (type.TUPLE,),
                "side_length": field.INT,
                "side": (["Width", "Height"],),
                "scale_method": (cls.upscale_methods,),
                "crop": (cls.crop_methods,)}}

    RETURN_TYPES = (type.LATENT, type.TUPLE,)
    FUNCTION = "upscale"

    CATEGORY = TREE_LATENTS

    def upscale(self, LATENT, scale_method, TUPLE, side_length, side, crop):
        width_B = int(TUPLE[0])
        height_B = int(TUPLE[1])

        width = width_B
        height = height_B

        if side == "Width":
            heigh_ratio = height_B / width_B
            width = side_length
            height = heigh_ratio * width
        elif side == "Height":
            width_ratio = width_B / height_B
            height = side_length
            width = width_ratio * height

        width = math.ceil(width)
        height = math.ceil(height)

        cls = LATENT.copy()
        cls["samples"] = comfy.utils.common_upscale(LATENT["samples"], width // 8, height // 8, scale_method, crop)
        return (cls, (width, height),)