import spaces
import os
import random
import sys
from typing import Sequence, Mapping, Any, Union
import torch
import numpy as np
from PIL import Image
# from comfy import model_management
from nodes import NODE_CLASS_MAPPINGS as NODE_CLASS_MAPPINGS_1
from comfy_extras.nodes_custom_sampler import NODE_CLASS_MAPPINGS as NODE_CLASS_MAPPINGS_2
from custom_nodes.ComfyUI_Comfyroll_CustomNodes.node_mappings import NODE_CLASS_MAPPINGS as NODE_CLASS_MAPPINGS_3
from custom_nodes.ComfyUI_Comfyroll_CustomNodes.node_mappings import NODE_CLASS_MAPPINGS as NODE_CLASS_MAPPINGS_4
from comfy_extras.nodes_model_advanced import NODE_CLASS_MAPPINGS as NODE_CLASS_MAPPINGS_5
from comfy_extras.nodes_flux import NODE_CLASS_MAPPINGS as NODE_CLASS_MAPPINGS_6
from torchvision.transforms.functional import to_pil_image
from PIL import Image
import numpy as np
import time

from huggingface_hub import hf_hub_download
token =  os.getenv("HF_TKN")
# Merge both mappings
NODE_CLASS_MAPPINGS = {**NODE_CLASS_MAPPINGS_1, **NODE_CLASS_MAPPINGS_2, **NODE_CLASS_MAPPINGS_3, **NODE_CLASS_MAPPINGS_4, **NODE_CLASS_MAPPINGS_5, **NODE_CLASS_MAPPINGS_6}
hf_hub_download(repo_id="black-forest-labs/FLUX.1-dev", filename="flux1-dev.safetensors", local_dir="models/unet", token = token)
hf_hub_download(repo_id="black-forest-labs/FLUX.1-dev", filename="ae.safetensors", local_dir="models/vae", token = token)
hf_hub_download(repo_id="comfyanonymous/flux_text_encoders", filename="clip_l.safetensors", local_dir="models/text_encoders", token = token)
hf_hub_download(repo_id="comfyanonymous/flux_text_encoders", filename="t5xxl_fp16.safetensors", local_dir="models/text_encoders", token = token)

def preprocess_image_tensor(image):
    # If image has a batch dimension (shape: [1, C, H, W]), remove it.
    if image.ndim == 4 and image.shape[0] == 1:
        image = image.squeeze(0)
    # If image is in channels-first format (i.e. [C, H, W]) and has 3 or 4 channels,
    # convert it to channels-last format ([H, W, C]).
    if image.ndim == 3 and image.shape[0] in [1, 3, 4]:
        image = image.permute(1, 2, 0)
    # Ensure the image values are between 0 and 1. Then scale them to [0, 255].
    image = image.detach().cpu().numpy()
    image = np.clip(image, 0, 1) * 255
    # Convert to unsigned 8-bit integer type.
    image = image.astype(np.uint8)
    return image


def get_value_at_index(obj: Union[Sequence, Mapping], index: int) -> Any:
    """Returns the value at the given index of a sequence or mapping.

    If the object is a sequence (like list or string), returns the value at the given index.
    If the object is a mapping (like a dictionary), returns the value at the index-th key.

    Some return a dictionary, in these cases, we look for the "results" key

    Args:
        obj (Union[Sequence, Mapping]): The object to retrieve the value from.
        index (int): The index of the value to retrieve.

    Returns:
        Any: The value at the given index.

    Raises:
        IndexError: If the index is o of bounds for the object and the object is not a mapping.
    """
    try:
        return obj[index]
    except KeyError:
        return obj["result"][index]


def find_path(name: str, path: str = None) -> str:
    """
    Recursively looks at parent folders starting from the given path until it finds the given name.
    Returns the path as a Path object if found, or None otherwise.
    """
    # If no path is given, use the current working directory
    if path is None:
        path = os.getcwd()

    # Check if the current directory contains the name
    if name in os.listdir(path):
        path_name = os.path.join(path, name)
        print(f"{name} found: {path_name}")
        return path_name

    # Get the parent directory
    parent_directory = os.path.dirname(path)

    # If the parent directory is the same as the current directory, we've reached the root and stop the search
    if parent_directory == path:
        return None

    # Recursively call the function with the parent directory
    return find_path(name, parent_directory)


def add_comfyui_directory_to_sys_path() -> None:
    """
    Add 'ComfyUI' to the sys.path
    """
    comfyui_path = find_path("ComfyUI")
    if comfyui_path is not None and os.path.isdir(comfyui_path):
        sys.path.append(comfyui_path)
        print(f"'{comfyui_path}' added to sys.path")


def add_extra_model_paths() -> None:
    """
    Parse the optional extra_model_paths.yaml file and add the parsed paths to the sys.path.
    """
    try:
        from main import load_extra_path_config
    except ImportError:
        print(
            "Could not import load_extra_path_config from main.py. Looking in utils.extra_config instead."
        )
        from utils.extra_config import load_extra_path_config

    extra_model_paths = find_path("extra_model_paths.yaml")

    if extra_model_paths is not None:
        load_extra_path_config(extra_model_paths)
    else:
        print("Could not find the extra_model_paths config file.")


def import_custom_nodes() -> None:
    """Find all custom nodes in the custom_nodes folder and add those node objects to NODE_CLASS_MAPPINGS

    This function sets up a new asyncio event loop, initializes the PromptServer,
    creates a PromptQueue, and initializes the custom nodes.
    """
    import asyncio
    import execution
    from nodes import init_extra_nodes
    import server

    # Creating a new event loop and setting it as the default loop
    loop = asyncio.new_event_loop()
    asyncio.set_event_loop(loop)

    # Creating an instance of PromptServer with the loop
    server_instance = server.PromptServer(loop)
    execution.PromptQueue(server_instance)

    # Initializing custom nodes
    init_extra_nodes()

def preprocess_image_tensor(image):
    # If image has a batch dimension (shape: [1, C, H, W]), remove it.
    if image.ndim == 4 and image.shape[0] == 1:
        image = image.squeeze(0)
    # If image is in channels-first format (i.e. [C, H, W]) and has 3 or 4 channels,
    # convert it to channels-last format ([H, W, C]).
    if image.ndim == 3 and image.shape[0] in [1, 3, 4]:
        image = image.permute(1, 2, 0)
    # Ensure the image values are between 0 and 1. Then scale them to [0, 255].
    image = image.detach().cpu().numpy()
    image = np.clip(image, 0, 1) * 255
    # Convert to unsigned 8-bit integer type.
    image = image.astype(np.uint8)
    return image



add_comfyui_directory_to_sys_path()
import_custom_nodes()
# add_extra_model_paths()
dualcliploader = NODE_CLASS_MAPPINGS["DualCLIPLoader"]()
dualcliploader_11 = dualcliploader.load_clip(
    clip_name1="t5xxl_fp16.safetensors",
    clip_name2="clip_l.safetensors",
    type="flux",
    device="default",
)

cliptextencode = NODE_CLASS_MAPPINGS["CLIPTextEncode"]()
cliptextencode_6 = cliptextencode.encode(
    text="Photo on a small glass panel. Color. Photo of trees with a body of water in the front and moutain in the background.",
    clip=get_value_at_index(dualcliploader_11, 0),
)

vaeloader = NODE_CLASS_MAPPINGS["VAELoader"]()
vaeloader_10 = vaeloader.load_vae(vae_name="ae.safetensors")

unetloader = NODE_CLASS_MAPPINGS["UNETLoader"]()
unetloader_12 = unetloader.load_unet(
    unet_name="flux1-dev.safetensors", weight_dtype="default"
)

ksamplerselect = NODE_CLASS_MAPPINGS["KSamplerSelect"]()
ksamplerselect_16 = ksamplerselect.get_sampler(sampler_name="dpmpp_2m")

# randomnoise = NODE_CLASS_MAPPINGS["RandomNoise"]()
# randomnoise_25 = randomnoise.get_noise(noise_seed='42')

loraloadermodelonly = NODE_CLASS_MAPPINGS["LoraLoaderModelOnly"]()
loraloadermodelonly_72 = loraloadermodelonly.load_lora_model_only(
    lora_name='lora_weight_rank_32_alpha_32.safetensors',
    strength_model=1,
    model=get_value_at_index(unetloader_12, 0),
)

cr_sdxl_aspect_ratio = NODE_CLASS_MAPPINGS["CR SDXL Aspect Ratio"]()
cr_sdxl_aspect_ratio_85 = cr_sdxl_aspect_ratio.Aspect_Ratio(
    width=1024,
    height=1024,
    aspect_ratio="4:3 landscape 1152x896",
    swap_dimensions="Off",
    upscale_factor=1.5,
    batch_size=1,
)

modelsamplingflux = NODE_CLASS_MAPPINGS["ModelSamplingFlux"]()
fluxguidance = NODE_CLASS_MAPPINGS["FluxGuidance"]()
basicguider = NODE_CLASS_MAPPINGS["BasicGuider"]()
basicscheduler = NODE_CLASS_MAPPINGS["BasicScheduler"]()
samplercustomadvanced = NODE_CLASS_MAPPINGS["SamplerCustomAdvanced"]()
vaedecode = NODE_CLASS_MAPPINGS["VAEDecode"]()
saveimage = NODE_CLASS_MAPPINGS["SaveImage"]()

# def model_loading():
#     model_loaders = [dualcliploader_11, vaeloader_10, unetloader_12, loraloadermodelonly_72]
#     valid_models = [
#     getattr(loader[0], 'patcher', loader[0]) 
#     for loader in model_loaders
#     if not isinstance(loader[0], dict) and not isinstance(getattr(loader[0], 'patcher', None), dict)
# ]
#     #Load the models
#     # model_management.load_models_gpu(valid_models)
    

def generate_image(prompt,
        guidance_scale,
        aspect_ratio,
        seed,
        num_inference_steps,
        lora_weight,
        ):
    # print(seed)
    cliptextencode = NODE_CLASS_MAPPINGS["CLIPTextEncode"]()
    cliptextencode_6 = cliptextencode.encode(
        text=prompt,
        clip=get_value_at_index(dualcliploader_11, 0),
    )
    cr_sdxl_aspect_ratio = NODE_CLASS_MAPPINGS["CR SDXL Aspect Ratio"]()
    cr_sdxl_aspect_ratio_85 = cr_sdxl_aspect_ratio.Aspect_Ratio(
        width=1024,
        height=1024,
        aspect_ratio=aspect_ratio,
        swap_dimensions="Off",
        upscale_factor=1.5,
        batch_size=1,
    )
    loraloadermodelonly = NODE_CLASS_MAPPINGS["LoraLoaderModelOnly"]()
    loraloadermodelonly_72 = loraloadermodelonly.load_lora_model_only(
        lora_name=lora_weight,
        strength_model=1,
        model=get_value_at_index(unetloader_12, 0),
    )
    randomnoise = NODE_CLASS_MAPPINGS["RandomNoise"]()
    randomnoise_25 = randomnoise.get_noise(noise_seed=str(seed))
    with torch.inference_mode():
        for q in range(1):
            modelsamplingflux_61 = modelsamplingflux.patch(
                max_shift=1.15,
                base_shift=0.5,
                width=get_value_at_index(cr_sdxl_aspect_ratio_85, 0),
                height=get_value_at_index(cr_sdxl_aspect_ratio_85, 1),
                model=get_value_at_index(loraloadermodelonly_72, 0),
            )

            fluxguidance_60 = fluxguidance.append(
                guidance=guidance_scale, conditioning=get_value_at_index(cliptextencode_6, 0)
            )

            basicguider_22 = basicguider.get_guider(
                model=get_value_at_index(modelsamplingflux_61, 0),
                conditioning=get_value_at_index(fluxguidance_60, 0),
            )

            basicscheduler_17 = basicscheduler.get_sigmas(
                scheduler="sgm_uniform",
                steps=num_inference_steps,
                denoise=1,
                model=get_value_at_index(modelsamplingflux_61, 0),
            )

            samplercustomadvanced_13 = samplercustomadvanced.sample(
                noise=get_value_at_index(randomnoise_25, 0),
                guider=get_value_at_index(basicguider_22, 0),
                sampler=get_value_at_index(ksamplerselect_16, 0),
                sigmas=get_value_at_index(basicscheduler_17, 0),
                latent_image=get_value_at_index(cr_sdxl_aspect_ratio_85, 4),
            )

            vaedecode_8 = vaedecode.decode(
                samples=get_value_at_index(samplercustomadvanced_13, 0),
                vae=get_value_at_index(vaeloader_10, 0),
            )
            # saveimage_9 = saveimage.save_images(
            #     filename_prefix="image", images=get_value_at_index(vaedecode_8, 0)
            # )
            image_tensor = get_value_at_index(vaedecode_8, 0)
            preprocessed_image = preprocess_image_tensor(image_tensor)
            pil_image = Image.fromarray(preprocessed_image)
            return pil_image, seed