from pipeline import SelfForcingTrainingPipeline
from typing import Optional, Tuple
import torch

from model.base import SelfForcingModel


class SiD(SelfForcingModel):
    def __init__(self, args, device):
        """
        Initialize the DMD (Distribution Matching Distillation) module.
        This class is self-contained and compute generator and fake score losses
        in the forward pass.
        """
        super().__init__(args, device)
        self.num_frame_per_block = getattr(args, "num_frame_per_block", 1)

        if self.num_frame_per_block > 1:
            self.generator.model.num_frame_per_block = self.num_frame_per_block

        if args.gradient_checkpointing:
            self.generator.enable_gradient_checkpointing()
            self.fake_score.enable_gradient_checkpointing()
            self.real_score.enable_gradient_checkpointing()

        # this will be init later with fsdp-wrapped modules
        self.inference_pipeline: SelfForcingTrainingPipeline = None

        # Step 2: Initialize all dmd hyperparameters
        self.num_train_timestep = args.num_train_timestep
        self.min_step = int(0.02 * self.num_train_timestep)
        self.max_step = int(0.98 * self.num_train_timestep)
        if hasattr(args, "real_guidance_scale"):
            self.real_guidance_scale = args.real_guidance_scale
        else:
            self.real_guidance_scale = args.guidance_scale
        self.timestep_shift = getattr(args, "timestep_shift", 1.0)
        self.sid_alpha = getattr(args, "sid_alpha", 1.0)
        self.ts_schedule = getattr(args, "ts_schedule", True)
        self.ts_schedule_max = getattr(args, "ts_schedule_max", False)

        if getattr(self.scheduler, "alphas_cumprod", None) is not None:
            self.scheduler.alphas_cumprod = self.scheduler.alphas_cumprod.to(device)
        else:
            self.scheduler.alphas_cumprod = None

    def compute_distribution_matching_loss(
        self,
        image_or_video: torch.Tensor,
        conditional_dict: dict,
        unconditional_dict: dict,
        gradient_mask: Optional[torch.Tensor] = None,
        denoised_timestep_from: int = 0,
        denoised_timestep_to: int = 0
    ) -> Tuple[torch.Tensor, dict]:
        """
        Compute the DMD loss (eq 7 in https://arxiv.org/abs/2311.18828).
        Input:
            - image_or_video: a tensor with shape [B, F, C, H, W] where the number of frame is 1 for images.
            - conditional_dict: a dictionary containing the conditional information (e.g. text embeddings, image embeddings).
            - unconditional_dict: a dictionary containing the unconditional information (e.g. null/negative text embeddings, null/negative image embeddings).
            - gradient_mask: a boolean tensor with the same shape as image_or_video indicating which pixels to compute loss .
        Output:
            - dmd_loss: a scalar tensor representing the DMD loss.
            - dmd_log_dict: a dictionary containing the intermediate tensors for logging.
        """
        original_latent = image_or_video

        batch_size, num_frame = image_or_video.shape[:2]

        # Step 1: Randomly sample timestep based on the given schedule and corresponding noise
        min_timestep = denoised_timestep_to if self.ts_schedule and denoised_timestep_to is not None else self.min_score_timestep
        max_timestep = denoised_timestep_from if self.ts_schedule_max and denoised_timestep_from is not None else self.num_train_timestep
        timestep = self._get_timestep(
            min_timestep,
            max_timestep,
            batch_size,
            num_frame,
            self.num_frame_per_block,
            uniform_timestep=True
        )

        if self.timestep_shift > 1:
            timestep = self.timestep_shift * \
                (timestep / 1000) / \
                (1 + (self.timestep_shift - 1) * (timestep / 1000)) * 1000
        timestep = timestep.clamp(self.min_step, self.max_step)

        noise = torch.randn_like(image_or_video)
        noisy_latent = self.scheduler.add_noise(
            image_or_video.flatten(0, 1),
            noise.flatten(0, 1),
            timestep.flatten(0, 1)
        ).unflatten(0, (batch_size, num_frame))

        # Step 2: SiD (May be wrap it?)
        noisy_image_or_video = noisy_latent
        # Step 2.1: Compute the fake score
        _, pred_fake_image = self.fake_score(
            noisy_image_or_video=noisy_image_or_video,
            conditional_dict=conditional_dict,
            timestep=timestep
        )
        # Step 2.2: Compute the real score
        # We compute the conditional and unconditional prediction
        # and add them together to achieve cfg (https://arxiv.org/abs/2207.12598)
        # NOTE: This step may cause OOM issue, which can be addressed by the CFG-free technique

        _, pred_real_image_cond = self.real_score(
            noisy_image_or_video=noisy_image_or_video,
            conditional_dict=conditional_dict,
            timestep=timestep
        )

        _, pred_real_image_uncond = self.real_score(
            noisy_image_or_video=noisy_image_or_video,
            conditional_dict=unconditional_dict,
            timestep=timestep
        )

        pred_real_image = pred_real_image_cond + (
            pred_real_image_cond - pred_real_image_uncond
        ) * self.real_guidance_scale

        # Step 2.3: SiD Loss
        # TODO: Add alpha
        # TODO: Double?
        sid_loss = (pred_real_image.double() - pred_fake_image.double()) * ((pred_real_image.double() - original_latent.double()) - self.sid_alpha * (pred_real_image.double() - pred_fake_image.double()))

        # Step 2.4: Loss normalizer
        with torch.no_grad():
            p_real = (original_latent - pred_real_image)
            normalizer = torch.abs(p_real).mean(dim=[1, 2, 3, 4], keepdim=True)
        sid_loss = sid_loss / normalizer

        sid_loss = torch.nan_to_num(sid_loss)
        num_frame = sid_loss.shape[1]
        sid_loss = sid_loss.mean()

        sid_log_dict = {
            "dmdtrain_gradient_norm": torch.zeros_like(sid_loss),
            "timestep": timestep.detach()
        }

        return sid_loss, sid_log_dict

    def generator_loss(
        self,
        image_or_video_shape,
        conditional_dict: dict,
        unconditional_dict: dict,
        clean_latent: torch.Tensor,
        initial_latent: torch.Tensor = None
    ) -> Tuple[torch.Tensor, dict]:
        """
        Generate image/videos from noise and compute the DMD loss.
        The noisy input to the generator is backward simulated.
        This removes the need of any datasets during distillation.
        See Sec 4.5 of the DMD2 paper (https://arxiv.org/abs/2405.14867) for details.
        Input:
            - image_or_video_shape: a list containing the shape of the image or video [B, F, C, H, W].
            - conditional_dict: a dictionary containing the conditional information (e.g. text embeddings, image embeddings).
            - unconditional_dict: a dictionary containing the unconditional information (e.g. null/negative text embeddings, null/negative image embeddings).
            - clean_latent: a tensor containing the clean latents [B, F, C, H, W]. Need to be passed when no backward simulation is used.
        Output:
            - loss: a scalar tensor representing the generator loss.
            - generator_log_dict: a dictionary containing the intermediate tensors for logging.
        """
        # Step 1: Unroll generator to obtain fake videos
        pred_image, gradient_mask, denoised_timestep_from, denoised_timestep_to = self._run_generator(
            image_or_video_shape=image_or_video_shape,
            conditional_dict=conditional_dict,
            initial_latent=initial_latent
        )

        # Step 2: Compute the DMD loss
        dmd_loss, dmd_log_dict = self.compute_distribution_matching_loss(
            image_or_video=pred_image,
            conditional_dict=conditional_dict,
            unconditional_dict=unconditional_dict,
            gradient_mask=gradient_mask,
            denoised_timestep_from=denoised_timestep_from,
            denoised_timestep_to=denoised_timestep_to
        )

        return dmd_loss, dmd_log_dict

    def critic_loss(
        self,
        image_or_video_shape,
        conditional_dict: dict,
        unconditional_dict: dict,
        clean_latent: torch.Tensor,
        initial_latent: torch.Tensor = None
    ) -> Tuple[torch.Tensor, dict]:
        """
        Generate image/videos from noise and train the critic with generated samples.
        The noisy input to the generator is backward simulated.
        This removes the need of any datasets during distillation.
        See Sec 4.5 of the DMD2 paper (https://arxiv.org/abs/2405.14867) for details.
        Input:
            - image_or_video_shape: a list containing the shape of the image or video [B, F, C, H, W].
            - conditional_dict: a dictionary containing the conditional information (e.g. text embeddings, image embeddings).
            - unconditional_dict: a dictionary containing the unconditional information (e.g. null/negative text embeddings, null/negative image embeddings).
            - clean_latent: a tensor containing the clean latents [B, F, C, H, W]. Need to be passed when no backward simulation is used.
        Output:
            - loss: a scalar tensor representing the generator loss.
            - critic_log_dict: a dictionary containing the intermediate tensors for logging.
        """

        # Step 1: Run generator on backward simulated noisy input
        with torch.no_grad():
            generated_image, _, denoised_timestep_from, denoised_timestep_to = self._run_generator(
                image_or_video_shape=image_or_video_shape,
                conditional_dict=conditional_dict,
                initial_latent=initial_latent
            )

        # Step 2: Compute the fake prediction
        min_timestep = denoised_timestep_to if self.ts_schedule and denoised_timestep_to is not None else self.min_score_timestep
        max_timestep = denoised_timestep_from if self.ts_schedule_max and denoised_timestep_from is not None else self.num_train_timestep
        critic_timestep = self._get_timestep(
            min_timestep,
            max_timestep,
            image_or_video_shape[0],
            image_or_video_shape[1],
            self.num_frame_per_block,
            uniform_timestep=True
        )

        if self.timestep_shift > 1:
            critic_timestep = self.timestep_shift * \
                (critic_timestep / 1000) / (1 + (self.timestep_shift - 1) * (critic_timestep / 1000)) * 1000

        critic_timestep = critic_timestep.clamp(self.min_step, self.max_step)

        critic_noise = torch.randn_like(generated_image)
        noisy_generated_image = self.scheduler.add_noise(
            generated_image.flatten(0, 1),
            critic_noise.flatten(0, 1),
            critic_timestep.flatten(0, 1)
        ).unflatten(0, image_or_video_shape[:2])

        _, pred_fake_image = self.fake_score(
            noisy_image_or_video=noisy_generated_image,
            conditional_dict=conditional_dict,
            timestep=critic_timestep
        )

        # Step 3: Compute the denoising loss for the fake critic
        if self.args.denoising_loss_type == "flow":
            from utils.wan_wrapper import WanDiffusionWrapper
            flow_pred = WanDiffusionWrapper._convert_x0_to_flow_pred(
                scheduler=self.scheduler,
                x0_pred=pred_fake_image.flatten(0, 1),
                xt=noisy_generated_image.flatten(0, 1),
                timestep=critic_timestep.flatten(0, 1)
            )
            pred_fake_noise = None
        else:
            flow_pred = None
            pred_fake_noise = self.scheduler.convert_x0_to_noise(
                x0=pred_fake_image.flatten(0, 1),
                xt=noisy_generated_image.flatten(0, 1),
                timestep=critic_timestep.flatten(0, 1)
            ).unflatten(0, image_or_video_shape[:2])

        denoising_loss = self.denoising_loss_func(
            x=generated_image.flatten(0, 1),
            x_pred=pred_fake_image.flatten(0, 1),
            noise=critic_noise.flatten(0, 1),
            noise_pred=pred_fake_noise,
            alphas_cumprod=self.scheduler.alphas_cumprod,
            timestep=critic_timestep.flatten(0, 1),
            flow_pred=flow_pred
        )

        # Step 5: Debugging Log
        critic_log_dict = {
            "critic_timestep": critic_timestep.detach()
        }

        return denoising_loss, critic_log_dict