# Copyright 2023 The HuggingFace Team. All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
from dataclasses import dataclass
from typing import Any, Dict, List, Optional, Tuple, Union

import torch
from torch import nn
from torch.nn import functional as F

from diffusers.configuration_utils import ConfigMixin, register_to_config
# from diffusers.loaders import FromOriginalControlnetMixin
from diffusers.utils import BaseOutput, logging
from diffusers.models.attention_processor import (
    ADDED_KV_ATTENTION_PROCESSORS,
    CROSS_ATTENTION_PROCESSORS,
    AttentionProcessor,
    AttnAddedKVProcessor,
    AttnProcessor,
)
from diffusers.models.embeddings import TextImageProjection, TextImageTimeEmbedding, TextTimeEmbedding, TimestepEmbedding, Timesteps
from diffusers.models.modeling_utils import ModelMixin
# from diffusers.models.unet_3d_blocks import get_down_block, get_up_block, UNetMidBlockSpatioTemporal
from .unet_3d_blocks import UNetMidBlockSpatioTemporal, get_down_block, get_up_block
from diffusers.models import UNetSpatioTemporalConditionModel

from huggingface_hub import hf_hub_download
from io import BytesIO
import os
import requests
from pathlib import Path

from huggingface_hub.constants import HF_HUB_OFFLINE, HF_HOME
DIFFUSERS_CACHE = os.path.join(HF_HOME, "diffusers")

logger = logging.get_logger(__name__)  # pylint: disable=invalid-name

class FromOriginalControlnetMixin:
    @classmethod
    def from_single_file(cls, pretrained_model_link_or_path, **kwargs):
        r"""
        Instantiate a [`ControlNetModel`] from pretrained controlnet weights saved in the original `.ckpt` or
        `.safetensors` format. The pipeline is set in evaluation mode (`model.eval()`) by default.

        Parameters:
            pretrained_model_link_or_path (`str` or `os.PathLike`, *optional*):
                Can be either:
                    - A link to the `.ckpt` file (for example
                      `"https://huggingface.co/<repo_id>/blob/main/<path_to_file>.ckpt"`) on the Hub.
                    - A path to a *file* containing all pipeline weights.
            torch_dtype (`str` or `torch.dtype`, *optional*):
                Override the default `torch.dtype` and load the model with another dtype. If `"auto"` is passed, the
                dtype is automatically derived from the model's weights.
            force_download (`bool`, *optional*, defaults to `False`):
                Whether or not to force the (re-)download of the model weights and configuration files, overriding the
                cached versions if they exist.
            cache_dir (`Union[str, os.PathLike]`, *optional*):
                Path to a directory where a downloaded pretrained model configuration is cached if the standard cache
                is not used.
            resume_download (`bool`, *optional*, defaults to `False`):
                Whether or not to resume downloading the model weights and configuration files. If set to `False`, any
                incompletely downloaded files are deleted.
            proxies (`Dict[str, str]`, *optional*):
                A dictionary of proxy servers to use by protocol or endpoint, for example, `{'http': 'foo.bar:3128',
                'http://hostname': 'foo.bar:4012'}`. The proxies are used on each request.
            local_files_only (`bool`, *optional*, defaults to `False`):
                Whether to only load local model weights and configuration files or not. If set to True, the model
                won't be downloaded from the Hub.
            use_auth_token (`str` or *bool*, *optional*):
                The token to use as HTTP bearer authorization for remote files. If `True`, the token generated from
                `diffusers-cli login` (stored in `~/.huggingface`) is used.
            revision (`str`, *optional*, defaults to `"main"`):
                The specific model version to use. It can be a branch name, a tag name, a commit id, or any identifier
                allowed by Git.
            use_safetensors (`bool`, *optional*, defaults to `None`):
                If set to `None`, the safetensors weights are downloaded if they're available **and** if the
                safetensors library is installed. If set to `True`, the model is forcibly loaded from safetensors
                weights. If set to `False`, safetensors weights are not loaded.
            image_size (`int`, *optional*, defaults to 512):
                The image size the model was trained on. Use 512 for all Stable Diffusion v1 models and the Stable
                Diffusion v2 base model. Use 768 for Stable Diffusion v2.
            upcast_attention (`bool`, *optional*, defaults to `None`):
                Whether the attention computation should always be upcasted.
            kwargs (remaining dictionary of keyword arguments, *optional*):
                Can be used to overwrite load and saveable variables (for example the pipeline components of the
                specific pipeline class). The overwritten components are directly passed to the pipelines `__init__`
                method. See example below for more information.

        Examples:

        ```py
        from diffusers import StableDiffusionControlNetPipeline, ControlNetModel

        url = "https://huggingface.co/lllyasviel/ControlNet-v1-1/blob/main/control_v11p_sd15_canny.pth"  # can also be a local path
        model = ControlNetModel.from_single_file(url)

        url = "https://huggingface.co/runwayml/stable-diffusion-v1-5/blob/main/v1-5-pruned.safetensors"  # can also be a local path
        pipe = StableDiffusionControlNetPipeline.from_single_file(url, controlnet=controlnet)
        ```
        """
        # import here to avoid circular dependency
        from diffusers.pipelines.stable_diffusion.convert_from_ckpt import download_controlnet_from_original_ckpt

        config_file = kwargs.pop("config_file", None)
        cache_dir = kwargs.pop("cache_dir", DIFFUSERS_CACHE)
        resume_download = kwargs.pop("resume_download", False)
        force_download = kwargs.pop("force_download", False)
        proxies = kwargs.pop("proxies", None)
        local_files_only = kwargs.pop("local_files_only", HF_HUB_OFFLINE)
        use_auth_token = kwargs.pop("use_auth_token", None)
        num_in_channels = kwargs.pop("num_in_channels", None)
        use_linear_projection = kwargs.pop("use_linear_projection", None)
        revision = kwargs.pop("revision", None)
        extract_ema = kwargs.pop("extract_ema", False)
        image_size = kwargs.pop("image_size", None)
        upcast_attention = kwargs.pop("upcast_attention", None)

        torch_dtype = kwargs.pop("torch_dtype", None)

        use_safetensors = kwargs.pop("use_safetensors", None)

        file_extension = pretrained_model_link_or_path.rsplit(".", 1)[-1]
        from_safetensors = file_extension == "safetensors"

        if from_safetensors and use_safetensors is False:
            raise ValueError("Make sure to install `safetensors` with `pip install safetensors`.")

        # remove huggingface url
        for prefix in ["https://huggingface.co/", "huggingface.co/", "hf.co/", "https://hf.co/"]:
            if pretrained_model_link_or_path.startswith(prefix):
                pretrained_model_link_or_path = pretrained_model_link_or_path[len(prefix) :]

        # Code based on diffusers.pipelines.pipeline_utils.DiffusionPipeline.from_pretrained
        ckpt_path = Path(pretrained_model_link_or_path)
        if not ckpt_path.is_file():
            # get repo_id and (potentially nested) file path of ckpt in repo
            repo_id = "/".join(ckpt_path.parts[:2])
            file_path = "/".join(ckpt_path.parts[2:])

            if file_path.startswith("blob/"):
                file_path = file_path[len("blob/") :]

            if file_path.startswith("main/"):
                file_path = file_path[len("main/") :]

            pretrained_model_link_or_path = hf_hub_download(
                repo_id,
                filename=file_path,
                cache_dir=cache_dir,
                resume_download=resume_download,
                proxies=proxies,
                local_files_only=local_files_only,
                use_auth_token=use_auth_token,
                revision=revision,
                force_download=force_download,
            )

        if config_file is None:
            config_url = "https://raw.githubusercontent.com/lllyasviel/ControlNet/main/models/cldm_v15.yaml"
            config_file = BytesIO(requests.get(config_url).content)

        image_size = image_size or 512

        controlnet = download_controlnet_from_original_ckpt(
            pretrained_model_link_or_path,
            original_config_file=config_file,
            image_size=image_size,
            extract_ema=extract_ema,
            num_in_channels=num_in_channels,
            upcast_attention=upcast_attention,
            from_safetensors=from_safetensors,
            use_linear_projection=use_linear_projection,
        )

        if torch_dtype is not None:
            controlnet.to(dtype=torch_dtype)

        return controlnet

@dataclass
class ControlNetOutput(BaseOutput):
    """
    The output of [`ControlNetModel`].

    Args:
        down_block_res_samples (`tuple[torch.Tensor]`):
            A tuple of downsample activations at different resolutions for each downsampling block. Each tensor should
            be of shape `(batch_size, channel * resolution, height //resolution, width // resolution)`. Output can be
            used to condition the original UNet's downsampling activations.
        mid_down_block_re_sample (`torch.Tensor`):
            The activation of the midde block (the lowest sample resolution). Each tensor should be of shape
            `(batch_size, channel * lowest_resolution, height // lowest_resolution, width // lowest_resolution)`.
            Output can be used to condition the original UNet's middle block activation.
    """

    down_block_res_samples: Tuple[torch.Tensor]
    mid_block_res_sample: torch.Tensor


class ControlNetConditioningEmbeddingSVD(nn.Module):
    """
    Quoting from https://arxiv.org/abs/2302.05543: "Stable Diffusion uses a pre-processing method similar to VQ-GAN
    [11] to convert the entire dataset of 512 × 512 images into smaller 64 × 64 “latent images” for stabilized
    training. This requires ControlNets to convert image-based conditions to 64 × 64 feature space to match the
    convolution size. We use a tiny network E(·) of four convolution layers with 4 × 4 kernels and 2 × 2 strides
    (activated by ReLU, channels are 16, 32, 64, 128, initialized with Gaussian weights, trained jointly with the full
    model) to encode image-space conditions ... into feature maps ..."
    """

    def __init__(
        self,
        conditioning_embedding_channels: int,
        conditioning_channels: int = 3,
        block_out_channels: Tuple[int, ...] = (16, 32, 96, 256),
    ):
        super().__init__()
        self.conv_in = nn.Conv2d(conditioning_channels, block_out_channels[0], kernel_size=3, padding=1)

        self.blocks = nn.ModuleList([])

        for i in range(len(block_out_channels) - 1):
            channel_in = block_out_channels[i]
            channel_out = block_out_channels[i + 1]
            self.blocks.append(nn.Conv2d(channel_in, channel_in, kernel_size=3, padding=1))
            self.blocks.append(nn.Conv2d(channel_in, channel_out, kernel_size=3, padding=1, stride=2))

        self.conv_out = zero_module(
            nn.Conv2d(block_out_channels[-1], conditioning_embedding_channels, kernel_size=3, padding=1)
        )

    def forward(self, conditioning):
        #this seeems appropriate? idk if i should be applying a more complex setup to handle the frames
        #combine batch and frames dimensions
        batch_size, frames, channels, height, width = conditioning.size()
        conditioning = conditioning.view(batch_size * frames, channels, height, width)

        embedding = self.conv_in(conditioning)
        embedding = F.silu(embedding)

        for block in self.blocks:
            embedding = block(embedding)
            embedding = F.silu(embedding)

        embedding = self.conv_out(embedding)
        
        # # split them apart again
        # # actually not needed
        # new_channels, new_height, new_width = embedding.shape[1], embedding.shape[2], embedding.shape[3]
        # embedding = embedding.view(batch_size, frames, new_channels, new_height, new_width)

        return embedding


class ControlNetSVDModel(ModelMixin, ConfigMixin, FromOriginalControlnetMixin):
    r"""
    A conditional Spatio-Temporal UNet model that takes a noisy video frames, conditional state, and a timestep and returns a sample
    shaped output.

    This model inherits from [`ModelMixin`]. Check the superclass documentation for it's generic methods implemented
    for all models (such as downloading or saving).

    Parameters:
        sample_size (`int` or `Tuple[int, int]`, *optional*, defaults to `None`):
            Height and width of input/output sample.
        in_channels (`int`, *optional*, defaults to 8): Number of channels in the input sample.
        out_channels (`int`, *optional*, defaults to 4): Number of channels in the output.
        down_block_types (`Tuple[str]`, *optional*, defaults to `("CrossAttnDownBlockSpatioTemporal", "CrossAttnDownBlockSpatioTemporal", "CrossAttnDownBlockSpatioTemporal", "DownBlockSpatioTemporal")`):
            The tuple of downsample blocks to use.
        up_block_types (`Tuple[str]`, *optional*, defaults to `("UpBlockSpatioTemporal", "CrossAttnUpBlockSpatioTemporal", "CrossAttnUpBlockSpatioTemporal", "CrossAttnUpBlockSpatioTemporal")`):
            The tuple of upsample blocks to use.
        block_out_channels (`Tuple[int]`, *optional*, defaults to `(320, 640, 1280, 1280)`):
            The tuple of output channels for each block.
        addition_time_embed_dim: (`int`, defaults to 256):
            Dimension to to encode the additional time ids.
        projection_class_embeddings_input_dim (`int`, defaults to 768):
            The dimension of the projection of encoded `added_time_ids`.
        layers_per_block (`int`, *optional*, defaults to 2): The number of layers per block.
        cross_attention_dim (`int` or `Tuple[int]`, *optional*, defaults to 1280):
            The dimension of the cross attention features.
        transformer_layers_per_block (`int`, `Tuple[int]`, or `Tuple[Tuple]` , *optional*, defaults to 1):
            The number of transformer blocks of type [`~models.attention.BasicTransformerBlock`]. Only relevant for
            [`~models.unet_3d_blocks.CrossAttnDownBlockSpatioTemporal`], [`~models.unet_3d_blocks.CrossAttnUpBlockSpatioTemporal`],
            [`~models.unet_3d_blocks.UNetMidBlockSpatioTemporal`].
        num_attention_heads (`int`, `Tuple[int]`, defaults to `(5, 10, 10, 20)`):
            The number of attention heads.
        dropout (`float`, *optional*, defaults to 0.0): The dropout probability to use.
    """

    _supports_gradient_checkpointing = True

    @register_to_config
    def __init__(
        self,
        sample_size: Optional[int] = None,
        in_channels: int = 8,
        out_channels: int = 4,
        down_block_types: Tuple[str] = (
            "CrossAttnDownBlockSpatioTemporal",
            "CrossAttnDownBlockSpatioTemporal",
            "CrossAttnDownBlockSpatioTemporal",
            "DownBlockSpatioTemporal",
        ),
        up_block_types: Tuple[str] = (
            "UpBlockSpatioTemporal",
            "CrossAttnUpBlockSpatioTemporal",
            "CrossAttnUpBlockSpatioTemporal",
            "CrossAttnUpBlockSpatioTemporal",
        ),
        block_out_channels: Tuple[int] = (320, 640, 1280, 1280),
        addition_time_embed_dim: int = 256,
        projection_class_embeddings_input_dim: int = 768,
        layers_per_block: Union[int, Tuple[int]] = 2,
        cross_attention_dim: Union[int, Tuple[int]] = 1024,
        transformer_layers_per_block: Union[int, Tuple[int], Tuple[Tuple]] = 1,
        num_attention_heads: Union[int, Tuple[int]] = (5, 10, 10, 20),
        num_frames: int = 25,
        conditioning_channels: int = 3,
        conditioning_embedding_out_channels: Optional[Tuple[int, ...]] = (16, 32, 96, 256),
        # NOTE: adapter for dift feature
        with_id_feature: bool = False,
        feature_channels: int = 160,
        feature_out_channels: Tuple[int, ...] = (160, 160, 256, 256),
    ):
        super().__init__()
        self.sample_size = sample_size

        print("layers per block is", layers_per_block)
        
        # Check inputs
        if len(down_block_types) != len(up_block_types):
            raise ValueError(
                f"Must provide the same number of `down_block_types` as `up_block_types`. `down_block_types`: {down_block_types}. `up_block_types`: {up_block_types}."
            )

        if len(block_out_channels) != len(down_block_types):
            raise ValueError(
                f"Must provide the same number of `block_out_channels` as `down_block_types`. `block_out_channels`: {block_out_channels}. `down_block_types`: {down_block_types}."
            )

        if not isinstance(num_attention_heads, int) and len(num_attention_heads) != len(down_block_types):
            raise ValueError(
                f"Must provide the same number of `num_attention_heads` as `down_block_types`. `num_attention_heads`: {num_attention_heads}. `down_block_types`: {down_block_types}."
            )

        if isinstance(cross_attention_dim, list) and len(cross_attention_dim) != len(down_block_types):
            raise ValueError(
                f"Must provide the same number of `cross_attention_dim` as `down_block_types`. `cross_attention_dim`: {cross_attention_dim}. `down_block_types`: {down_block_types}."
            )

        if not isinstance(layers_per_block, int) and len(layers_per_block) != len(down_block_types):
            raise ValueError(
                f"Must provide the same number of `layers_per_block` as `down_block_types`. `layers_per_block`: {layers_per_block}. `down_block_types`: {down_block_types}."
            )

        # input
        self.conv_in = nn.Conv2d(
            in_channels,
            block_out_channels[0],
            kernel_size=3,
            padding=1,
        )

        # time
        time_embed_dim = block_out_channels[0] * 4

        self.time_proj = Timesteps(block_out_channels[0], True, downscale_freq_shift=0)
        timestep_input_dim = block_out_channels[0]

        self.time_embedding = TimestepEmbedding(timestep_input_dim, time_embed_dim)

        self.add_time_proj = Timesteps(addition_time_embed_dim, True, downscale_freq_shift=0)
        self.add_embedding = TimestepEmbedding(projection_class_embeddings_input_dim, time_embed_dim)

        self.down_blocks = nn.ModuleList([])
        self.controlnet_down_blocks = nn.ModuleList([])

        if isinstance(num_attention_heads, int):
            num_attention_heads = (num_attention_heads,) * len(down_block_types)

        if isinstance(cross_attention_dim, int):
            cross_attention_dim = (cross_attention_dim,) * len(down_block_types)

        if isinstance(layers_per_block, int):
            layers_per_block = [layers_per_block] * len(down_block_types)

        if isinstance(transformer_layers_per_block, int):
            transformer_layers_per_block = [transformer_layers_per_block] * len(down_block_types)

        blocks_time_embed_dim = time_embed_dim
        self.controlnet_cond_embedding = ControlNetConditioningEmbeddingSVD(
            conditioning_embedding_channels=block_out_channels[0],
            block_out_channels=conditioning_embedding_out_channels,
            conditioning_channels=conditioning_channels,
        )

        # down
        output_channel = block_out_channels[0]
        controlnet_block = nn.Conv2d(output_channel, output_channel, kernel_size=1)
        controlnet_block = zero_module(controlnet_block)
        self.controlnet_down_blocks.append(controlnet_block)

        for i, down_block_type in enumerate(down_block_types):
            input_channel = output_channel
            output_channel = block_out_channels[i]
            is_final_block = i == len(block_out_channels) - 1

            down_block = get_down_block(
                down_block_type,
                num_layers=layers_per_block[i],
                transformer_layers_per_block=transformer_layers_per_block[i],
                in_channels=input_channel,
                out_channels=output_channel,
                temb_channels=blocks_time_embed_dim,
                add_downsample=not is_final_block,
                resnet_eps=1e-5,
                cross_attention_dim=cross_attention_dim[i],
                num_attention_heads=num_attention_heads[i],
                resnet_act_fn="silu",
            )
            self.down_blocks.append(down_block)

            for _ in range(layers_per_block[i]):
                controlnet_block = nn.Conv2d(output_channel, output_channel, kernel_size=1)
                controlnet_block = zero_module(controlnet_block)
                self.controlnet_down_blocks.append(controlnet_block)

            if not is_final_block:
                controlnet_block = nn.Conv2d(output_channel, output_channel, kernel_size=1)
                controlnet_block = zero_module(controlnet_block)
                self.controlnet_down_blocks.append(controlnet_block)

        # mid
        mid_block_channel = block_out_channels[-1]
        controlnet_block = nn.Conv2d(mid_block_channel, mid_block_channel, kernel_size=1)
        controlnet_block = zero_module(controlnet_block)
        self.controlnet_mid_block = controlnet_block

        self.mid_block = UNetMidBlockSpatioTemporal(
            block_out_channels[-1],
            temb_channels=blocks_time_embed_dim,
            transformer_layers_per_block=transformer_layers_per_block[-1],
            cross_attention_dim=cross_attention_dim[-1],
            num_attention_heads=num_attention_heads[-1],
        )

        # # out
        # self.conv_norm_out = nn.GroupNorm(num_channels=block_out_channels[0], num_groups=32, eps=1e-5)
        # self.conv_act = nn.SiLU()

        # self.conv_out = nn.Conv2d(
        #    block_out_channels[0],
        #    out_channels,
        #    kernel_size=3,
        #    padding=1,
        # )

    @property
    def attn_processors(self) -> Dict[str, AttentionProcessor]:
        r"""
        Returns:
            `dict` of attention processors: A dictionary containing all attention processors used in the model with
            indexed by its weight name.
        """
        # set recursively
        processors = {}

        def fn_recursive_add_processors(
            name: str,
            module: torch.nn.Module,
            processors: Dict[str, AttentionProcessor],
        ):
            if hasattr(module, "get_processor"):
                processors[f"{name}.processor"] = module.get_processor(return_deprecated_lora=True)

            for sub_name, child in module.named_children():
                fn_recursive_add_processors(f"{name}.{sub_name}", child, processors)

            return processors

        for name, module in self.named_children():
            fn_recursive_add_processors(name, module, processors)

        return processors

    def set_attn_processor(self, processor: Union[AttentionProcessor, Dict[str, AttentionProcessor]]):
        r"""
        Sets the attention processor to use to compute attention.

        Parameters:
            processor (`dict` of `AttentionProcessor` or only `AttentionProcessor`):
                The instantiated processor class or a dictionary of processor classes that will be set as the processor
                for **all** `Attention` layers.

                If `processor` is a dict, the key needs to define the path to the corresponding cross attention
                processor. This is strongly recommended when setting trainable attention processors.

        """
        count = len(self.attn_processors.keys())

        if isinstance(processor, dict) and len(processor) != count:
            raise ValueError(
                f"A dict of processors was passed, but the number of processors {len(processor)} does not match the"
                f" number of attention layers: {count}. Please make sure to pass {count} processor classes."
            )

        def fn_recursive_attn_processor(name: str, module: torch.nn.Module, processor):
            if hasattr(module, "set_processor"):
                if not isinstance(processor, dict):
                    module.set_processor(processor)
                else:
                    module.set_processor(processor.pop(f"{name}.processor"))

            for sub_name, child in module.named_children():
                fn_recursive_attn_processor(f"{name}.{sub_name}", child, processor)

        for name, module in self.named_children():
            fn_recursive_attn_processor(name, module, processor)

    def set_default_attn_processor(self):
        """
        Disables custom attention processors and sets the default attention implementation.
        """
        if all(proc.__class__ in CROSS_ATTENTION_PROCESSORS for proc in self.attn_processors.values()):
            processor = AttnProcessor()
        else:
            raise ValueError(
                f"Cannot call `set_default_attn_processor` when attention processors are of type {next(iter(self.attn_processors.values()))}"
            )

        self.set_attn_processor(processor)

    def _set_gradient_checkpointing(self, module, value=False):
        if hasattr(module, "gradient_checkpointing"):
            module.gradient_checkpointing = value

    # Copied from diffusers.models.unet_3d_condition.UNet3DConditionModel.enable_forward_chunking
    def enable_forward_chunking(self, chunk_size: Optional[int] = None, dim: int = 0) -> None:
        """
        Sets the attention processor to use [feed forward
        chunking](https://huggingface.co/blog/reformer#2-chunked-feed-forward-layers).

        Parameters:
            chunk_size (`int`, *optional*):
                The chunk size of the feed-forward layers. If not specified, will run feed-forward layer individually
                over each tensor of dim=`dim`.
            dim (`int`, *optional*, defaults to `0`):
                The dimension over which the feed-forward computation should be chunked. Choose between dim=0 (batch)
                or dim=1 (sequence length).
        """
        if dim not in [0, 1]:
            raise ValueError(f"Make sure to set `dim` to either 0 or 1, not {dim}")

        # By default chunk size is 1
        chunk_size = chunk_size or 1

        def fn_recursive_feed_forward(module: torch.nn.Module, chunk_size: int, dim: int):
            if hasattr(module, "set_chunk_feed_forward"):
                module.set_chunk_feed_forward(chunk_size=chunk_size, dim=dim)

            for child in module.children():
                fn_recursive_feed_forward(child, chunk_size, dim)

        for module in self.children():
            fn_recursive_feed_forward(module, chunk_size, dim)

    def forward(
        self,
        sample: torch.FloatTensor,
        timestep: Union[torch.Tensor, float, int],
        encoder_hidden_states: torch.Tensor,
        added_time_ids: torch.Tensor,
        controlnet_cond: torch.FloatTensor = None,
        image_only_indicator: Optional[torch.Tensor] = None,
        return_dict: bool = True,
        guess_mode: bool = False,
        conditioning_scale: float = 1.0,
    ) -> Union[ControlNetOutput, Tuple]:
        r"""
        The [`UNetSpatioTemporalConditionModel`] forward method.

        Args:
            sample (`torch.FloatTensor`):
                The noisy input tensor with the following shape `(batch, num_frames, channel, height, width)`.
            timestep (`torch.FloatTensor` or `float` or `int`): The number of timesteps to denoise an input.
            encoder_hidden_states (`torch.FloatTensor`):
                The encoder hidden states with shape `(batch, sequence_length, cross_attention_dim)`.
            added_time_ids: (`torch.FloatTensor`):
                The additional time ids with shape `(batch, num_additional_ids)`. These are encoded with sinusoidal
                embeddings and added to the time embeddings.
            return_dict (`bool`, *optional*, defaults to `True`):
                Whether or not to return a [`~models.unet_slatio_temporal.UNetSpatioTemporalConditionOutput`] instead of a plain
                tuple.
        Returns:
            [`~models.unet_slatio_temporal.UNetSpatioTemporalConditionOutput`] or `tuple`:
                If `return_dict` is True, an [`~models.unet_slatio_temporal.UNetSpatioTemporalConditionOutput`] is returned, otherwise
                a `tuple` is returned where the first element is the sample tensor.
        """
        # 1. time
        timesteps = timestep
        if not torch.is_tensor(timesteps):
            # TODO: this requires sync between CPU and GPU. So try to pass timesteps as tensors if you can
            # This would be a good case for the `match` statement (Python 3.10+)
            is_mps = sample.device.type == "mps"
            if isinstance(timestep, float):
                dtype = torch.float32 if is_mps else torch.float64
            else:
                dtype = torch.int32 if is_mps else torch.int64
            timesteps = torch.tensor([timesteps], dtype=dtype, device=sample.device)
        elif len(timesteps.shape) == 0:
            timesteps = timesteps[None].to(sample.device)

        # broadcast to batch dimension in a way that's compatible with ONNX/Core ML
        batch_size, num_frames = sample.shape[:2]
        timesteps = timesteps.expand(batch_size)

        t_emb = self.time_proj(timesteps)

        # `Timesteps` does not contain any weights and will always return f32 tensors
        # but time_embedding might actually be running in fp16. so we need to cast here.
        # there might be better ways to encapsulate this.
        t_emb = t_emb.to(dtype=sample.dtype)

        emb = self.time_embedding(t_emb)

        time_embeds = self.add_time_proj(added_time_ids.flatten())
        time_embeds = time_embeds.reshape((batch_size, -1))
        time_embeds = time_embeds.to(emb.dtype)
        aug_emb = self.add_embedding(time_embeds)
        emb = emb + aug_emb

        # Flatten the batch and frames dimensions
        # sample: [batch, frames, channels, height, width] -> [batch * frames, channels, height, width]
        sample = sample.flatten(0, 1)
        # Repeat the embeddings num_video_frames times
        # emb: [batch, channels] -> [batch * frames, channels]
        emb = emb.repeat_interleave(num_frames, dim=0)
        # encoder_hidden_states: [batch, 1, channels] -> [batch * frames, 1, channels]
        encoder_hidden_states = encoder_hidden_states.repeat_interleave(num_frames, dim=0)

        # 2. pre-process
        sample = self.conv_in(sample)
        
        # controlnet cond
        if controlnet_cond != None:
            controlnet_cond = self.controlnet_cond_embedding(controlnet_cond)
            sample = sample + controlnet_cond

        image_only_indicator = torch.zeros(batch_size, num_frames, dtype=sample.dtype, device=sample.device)

        down_block_res_samples = (sample,)
        for downsample_block in self.down_blocks:
            if hasattr(downsample_block, "has_cross_attention") and downsample_block.has_cross_attention:
                # print('has_cross_attention', type(downsample_block))
                # models_diffusers.unet_3d_blocks.CrossAttnDownBlockSpatioTemporal

                sample, res_samples = downsample_block(
                    hidden_states=sample,
                    temb=emb,
                    encoder_hidden_states=encoder_hidden_states,
                    image_only_indicator=image_only_indicator,
                )
            else:
                # print('no_cross_attention', type(downsample_block))
                # models_diffusers.unet_3d_blocks.DownBlockSpatioTemporal

                sample, res_samples = downsample_block(
                    hidden_states=sample,
                    temb=emb,
                    image_only_indicator=image_only_indicator,
                )

            down_block_res_samples += res_samples

        # 4. mid
        sample = self.mid_block(
            hidden_states=sample,
            temb=emb,
            encoder_hidden_states=encoder_hidden_states,
            image_only_indicator=image_only_indicator,
        )

        controlnet_down_block_res_samples = ()

        for down_block_res_sample, controlnet_block in zip(down_block_res_samples, self.controlnet_down_blocks):
            down_block_res_sample = controlnet_block(down_block_res_sample)
            controlnet_down_block_res_samples = controlnet_down_block_res_samples + (down_block_res_sample,)

        down_block_res_samples = controlnet_down_block_res_samples

        mid_block_res_sample = self.controlnet_mid_block(sample)

        # 6. scaling

        down_block_res_samples = [sample * conditioning_scale for sample in down_block_res_samples]
        mid_block_res_sample = mid_block_res_sample * conditioning_scale

        if not return_dict:
            return (down_block_res_samples, mid_block_res_sample)

        return ControlNetOutput(
            down_block_res_samples=down_block_res_samples, mid_block_res_sample=mid_block_res_sample
        )

    @classmethod
    def from_unet(
        cls,
        unet: UNetSpatioTemporalConditionModel,
        # controlnet_conditioning_channel_order: str = "rgb",
        conditioning_embedding_out_channels: Optional[Tuple[int, ...]] = (16, 32, 96, 256),
        load_weights_from_unet: bool = True,
        conditioning_channels: int = 3,
    ):
        r"""
        Instantiate a [`ControlNetModel`] from [`UNet2DConditionModel`].

        Parameters:
            unet (`UNet2DConditionModel`):
                The UNet model weights to copy to the [`ControlNetModel`]. All configuration options are also copied
                where applicable.
        """

        # transformer_layers_per_block = (
        #     unet.config.transformer_layers_per_block if "transformer_layers_per_block" in unet.config else 1
        # )
        # encoder_hid_dim = unet.config.encoder_hid_dim if "encoder_hid_dim" in unet.config else None
        # encoder_hid_dim_type = unet.config.encoder_hid_dim_type if "encoder_hid_dim_type" in unet.config else None
        # addition_embed_type = unet.config.addition_embed_type if "addition_embed_type" in unet.config else None
        # addition_time_embed_dim = (
        #     unet.config.addition_time_embed_dim if "addition_time_embed_dim" in unet.config else None
        # )
        print(unet.config)
        controlnet = cls(
            in_channels=unet.config.in_channels,
            down_block_types=unet.config.down_block_types,
            block_out_channels=unet.config.block_out_channels,
            addition_time_embed_dim=unet.config.addition_time_embed_dim,
            transformer_layers_per_block=unet.config.transformer_layers_per_block,
            cross_attention_dim=unet.config.cross_attention_dim,
            num_attention_heads=unet.config.num_attention_heads,
            num_frames=unet.config.num_frames,
            sample_size=unet.config.sample_size,  # Added based on the dict
            layers_per_block=unet.config.layers_per_block,
            projection_class_embeddings_input_dim=unet.config.projection_class_embeddings_input_dim,
            conditioning_channels = conditioning_channels,
            conditioning_embedding_out_channels = conditioning_embedding_out_channels,
        )
        # controlnet rgb channel order ignored, set to not makea  difference by default

        if load_weights_from_unet:
            controlnet.conv_in.load_state_dict(unet.conv_in.state_dict())
            controlnet.time_proj.load_state_dict(unet.time_proj.state_dict())
            controlnet.time_embedding.load_state_dict(unet.time_embedding.state_dict())

           # if controlnet.class_embedding:
           #     controlnet.class_embedding.load_state_dict(unet.class_embedding.state_dict())

            controlnet.down_blocks.load_state_dict(unet.down_blocks.state_dict())
            controlnet.mid_block.load_state_dict(unet.mid_block.state_dict())

        return controlnet

    @property
    # Copied from diffusers.models.unet_2d_condition.UNet2DConditionModel.attn_processors
    def attn_processors(self) -> Dict[str, AttentionProcessor]:
        r"""
        Returns:
            `dict` of attention processors: A dictionary containing all attention processors used in the model with
            indexed by its weight name.
        """
        # set recursively
        processors = {}

        def fn_recursive_add_processors(name: str, module: torch.nn.Module, processors: Dict[str, AttentionProcessor]):
            if hasattr(module, "get_processor"):
                processors[f"{name}.processor"] = module.get_processor(return_deprecated_lora=True)

            for sub_name, child in module.named_children():
                fn_recursive_add_processors(f"{name}.{sub_name}", child, processors)

            return processors

        for name, module in self.named_children():
            fn_recursive_add_processors(name, module, processors)

        return processors

    # Copied from diffusers.models.unet_2d_condition.UNet2DConditionModel.set_attn_processor
    def set_attn_processor(
        self, processor: Union[AttentionProcessor, Dict[str, AttentionProcessor]], _remove_lora=False
    ):
        r"""
        Sets the attention processor to use to compute attention.

        Parameters:
            processor (`dict` of `AttentionProcessor` or only `AttentionProcessor`):
                The instantiated processor class or a dictionary of processor classes that will be set as the processor
                for **all** `Attention` layers.

                If `processor` is a dict, the key needs to define the path to the corresponding cross attention
                processor. This is strongly recommended when setting trainable attention processors.

        """
        count = len(self.attn_processors.keys())

        if isinstance(processor, dict) and len(processor) != count:
            raise ValueError(
                f"A dict of processors was passed, but the number of processors {len(processor)} does not match the"
                f" number of attention layers: {count}. Please make sure to pass {count} processor classes."
            )

        def fn_recursive_attn_processor(name: str, module: torch.nn.Module, processor):
            if hasattr(module, "set_processor"):
                if not isinstance(processor, dict):
                    module.set_processor(processor, _remove_lora=_remove_lora)
                else:
                    module.set_processor(processor.pop(f"{name}.processor"), _remove_lora=_remove_lora)

            for sub_name, child in module.named_children():
                fn_recursive_attn_processor(f"{name}.{sub_name}", child, processor)

        for name, module in self.named_children():
            fn_recursive_attn_processor(name, module, processor)

    # Copied from diffusers.models.unet_2d_condition.UNet2DConditionModel.set_default_attn_processor
    def set_default_attn_processor(self):
        """
        Disables custom attention processors and sets the default attention implementation.
        """
        if all(proc.__class__ in ADDED_KV_ATTENTION_PROCESSORS for proc in self.attn_processors.values()):
            processor = AttnAddedKVProcessor()
        elif all(proc.__class__ in CROSS_ATTENTION_PROCESSORS for proc in self.attn_processors.values()):
            processor = AttnProcessor()
        else:
            raise ValueError(
                f"Cannot call `set_default_attn_processor` when attention processors are of type {next(iter(self.attn_processors.values()))}"
            )

        self.set_attn_processor(processor, _remove_lora=True)

    # Copied from diffusers.models.unet_2d_condition.UNet2DConditionModel.set_attention_slice
    def set_attention_slice(self, slice_size: Union[str, int, List[int]]) -> None:
        r"""
        Enable sliced attention computation.

        When this option is enabled, the attention module splits the input tensor in slices to compute attention in
        several steps. This is useful for saving some memory in exchange for a small decrease in speed.

        Args:
            slice_size (`str` or `int` or `list(int)`, *optional*, defaults to `"auto"`):
                When `"auto"`, input to the attention heads is halved, so attention is computed in two steps. If
                `"max"`, maximum amount of memory is saved by running only one slice at a time. If a number is
                provided, uses as many slices as `attention_head_dim // slice_size`. In this case, `attention_head_dim`
                must be a multiple of `slice_size`.
        """
        sliceable_head_dims = []

        def fn_recursive_retrieve_sliceable_dims(module: torch.nn.Module):
            if hasattr(module, "set_attention_slice"):
                sliceable_head_dims.append(module.sliceable_head_dim)

            for child in module.children():
                fn_recursive_retrieve_sliceable_dims(child)

        # retrieve number of attention layers
        for module in self.children():
            fn_recursive_retrieve_sliceable_dims(module)

        num_sliceable_layers = len(sliceable_head_dims)

        if slice_size == "auto":
            # half the attention head size is usually a good trade-off between
            # speed and memory
            slice_size = [dim // 2 for dim in sliceable_head_dims]
        elif slice_size == "max":
            # make smallest slice possible
            slice_size = num_sliceable_layers * [1]

        slice_size = num_sliceable_layers * [slice_size] if not isinstance(slice_size, list) else slice_size

        if len(slice_size) != len(sliceable_head_dims):
            raise ValueError(
                f"You have provided {len(slice_size)}, but {self.config} has {len(sliceable_head_dims)} different"
                f" attention layers. Make sure to match `len(slice_size)` to be {len(sliceable_head_dims)}."
            )

        for i in range(len(slice_size)):
            size = slice_size[i]
            dim = sliceable_head_dims[i]
            if size is not None and size > dim:
                raise ValueError(f"size {size} has to be smaller or equal to {dim}.")

        # Recursively walk through all the children.
        # Any children which exposes the set_attention_slice method
        # gets the message
        def fn_recursive_set_attention_slice(module: torch.nn.Module, slice_size: List[int]):
            if hasattr(module, "set_attention_slice"):
                module.set_attention_slice(slice_size.pop())

            for child in module.children():
                fn_recursive_set_attention_slice(child, slice_size)

        reversed_slice_size = list(reversed(slice_size))
        for module in self.children():
            fn_recursive_set_attention_slice(module, reversed_slice_size)

    # def _set_gradient_checkpointing(self, module, value: bool = False) -> None:
    #     if isinstance(module, (CrossAttnDownBlock2D, DownBlock2D)):
    #         module.gradient_checkpointing = value

    
def zero_module(module):
    for p in module.parameters():
        nn.init.zeros_(p)
    return module