import math
import numpy as np
import torch

from typing import List
from torch import nn

import torch
from torch import nn
from torch.nn.utils.parametrizations import weight_norm
from einops import rearrange


# conv modules
def WNConv1d(*args, **kwargs):
    return weight_norm(nn.Conv1d(*args, **kwargs))


def WNConvTranspose1d(*args, **kwargs):
    return weight_norm(nn.ConvTranspose1d(*args, **kwargs))


@torch.jit.script
def snake(x, alpha):
    shape = x.shape
    x = x.reshape(shape[0], shape[1], -1)
    x = x + (alpha + 1e-9).reciprocal() * torch.sin(alpha * x).pow(2)
    x = x.reshape(shape)
    return x


class Snake1d(nn.Module):
    def __init__(self, channels):
        super().__init__()
        self.alpha = nn.Parameter(torch.ones(1, channels, 1))

    def forward(self, x):
        return snake(x, self.alpha)


class ConvNeXtUnit(nn.Module):
    def __init__(self, dim: int = 16, dilation: int = 1):
        super().__init__()
        self.dwconv = nn.Conv1d(
            dim, dim, kernel_size=7, padding=3 * dilation, groups=dim, dilation=dilation
        )
        self.norm = nn.LayerNorm(dim)
        self.pwconv1 = nn.Linear(dim, 4 * dim)
        self.act = nn.GELU()
        self.pwconv2 = nn.Linear(4 * dim, dim)
        self.gamma = nn.Parameter(torch.ones(dim, 1) * 1e-6)

    def forward(self, x):
        input = x
        x = self.dwconv(x)
        x = x.permute(0, 2, 1)  # (N, C, L) -> (N, L, C)
        x = self.norm(x)
        x = self.pwconv1(x)
        x = self.act(x)
        x = self.pwconv2(x)
        x = x.permute(0, 2, 1)  # (N, L, C) -> (N, C, L)


        # Padding adjustment
        pad = (input.shape[-1] - x.shape[-1]) // 2
        if pad > 0:
            input = input[..., pad:-pad]
        elif pad < 0:
            x = x[..., -pad:pad]


        x = input + self.gamma * x
        return x


def vae_sample(mean, scale):
    stdev = nn.functional.softplus(scale) + 1e-4
    var = stdev * stdev
    logvar = torch.log(var)
    latents = torch.randn_like(mean) * stdev + mean

    kl = (mean * mean + var - logvar - 1).sum(1).mean()

    return latents, kl


# VAE module
class VAEBottleneck(nn.Module):
    def __init__(self, latent_dim, dim, is_discrete=False, **kwargs):
        super().__init__()
        self.project_in = nn.Linear(latent_dim, dim * 2)
        self.project_out = nn.Linear(dim, latent_dim) if latent_dim != dim else nn.Identity()
        self.is_discrete = is_discrete

    def forward(self, x, transpose=True, **kwargs):
        if transpose:
            x = rearrange(x, "b d n -> b n d")
        x = self.project_in(x)
        mean, scale = x.chunk(2, dim=-1)

        z, kl = vae_sample(mean, scale)
        x = self.project_out(z)

        if transpose:
            x = rearrange(x, "b n d -> b d n")
            z = rearrange(z, "b n d -> b d n")

        return {
            "z": z,
            "x": x,
            "kl": kl,
            "mean": mean,
            "scale": scale,
        }

    def encode(self, x, return_info=False, **kwargs):
        if return_info:
            out = self.forward(x)
            latents = out["z"]
            # latents = rearrange(latents, "b n d -> b d n")
            bottleneck_info = {"kl": float(out["kl"].item())}
            return latents, bottleneck_info
        return self.forward(x)

    def decode(self, x, **kwargs):
        return x


class EncoderBlock(nn.Module):
    def __init__(self, dim: int = 16, stride: int = 1):
        super().__init__()
        self.block = nn.Sequential(
            ConvNeXtUnit(dim // 2, dilation=1),
            ConvNeXtUnit(dim // 2, dilation=3),
            ConvNeXtUnit(dim // 2, dilation=9),
            Snake1d(dim // 2),
            WNConv1d(
                dim // 2,
                dim,
                kernel_size=2 * stride,
                stride=stride,
                padding=math.ceil(stride / 2),
            ),
        )

    def forward(self, x):
        return self.block(x)


class Encoder(nn.Module):
    def __init__(
        self,
        d_model: int = 64,
        strides: list = [2, 4, 8, 8],
        d_latent: int = 64,
    ):
        super().__init__()
        # Create first convolution
        self.block = [WNConv1d(2, d_model, kernel_size=7, padding=3)]

        # Create EncoderBlocks that double channels as they downsample by `stride`
        for stride in strides:
            d_model *= 2
            self.block += [EncoderBlock(d_model, stride=stride)]

        # Create last convolution and groupnorm
        self.block += [
            Snake1d(d_model),
            WNConv1d(d_model, d_latent, kernel_size=3, padding=1),
        ]

        # Wrap black into nn.Sequential
        self.block = nn.Sequential(*self.block)
        self.enc_dim = d_model

    def forward(self, x):
        return self.block(x)


class DecoderBlock(nn.Module):
    def __init__(self, input_dim: int = 16, output_dim: int = 8, stride: int = 1):
        super().__init__()
        self.block = nn.Sequential(
            Snake1d(input_dim),
            WNConvTranspose1d(
                input_dim,
                output_dim,
                kernel_size=2 * stride,
                stride=stride,
                padding=math.floor(stride / 2),
            ),
            ConvNeXtUnit(output_dim, dilation=1),
            ConvNeXtUnit(output_dim, dilation=3),
            ConvNeXtUnit(output_dim, dilation=9),
        )

    def forward(self, x):
        return self.block(x)


class Decoder(nn.Module):
    def __init__(
        self,
        input_channel,
        channels,
        rates,
        d_out: int = 2,
    ):
        super().__init__()

        # Add first conv layer
        layers = [WNConv1d(input_channel, channels, kernel_size=7, padding=3)]

        # Add upsampling + MRF blocks
        for i, stride in enumerate(rates):
            input_dim = channels // 2**i
            output_dim = channels // 2 ** (i + 1)
            layers += [DecoderBlock(input_dim, output_dim, stride)]

        # Add final conv layer
        layers += [
            Snake1d(output_dim),
            WNConv1d(output_dim, d_out, kernel_size=7, padding=3),
            nn.Tanh(),
        ]

        self.model = nn.Sequential(*layers)

    def forward(self, x):
        return self.model(x)


# convnext codec
class ConvNextVAE(nn.Module):
    def __init__(
        self,
        encoder_dim: int = 128,
        encoder_rates: List[int] = [2, 3, 5, 8, 8],
        latent_dim: int = 128,
        decoder_dim: int = 1536,
        decoder_rates: List[int] = [8, 8, 5, 3, 2],
        vae_dim: int = 128,
        sample_rate: int = 48000,
        **kwargs,
    ):
        super().__init__()
        self.encoder_dim = encoder_dim
        self.encoder_rates = encoder_rates
        self.latent_dim = latent_dim
        self.decoder_dim = decoder_dim
        self.decoder_rates = decoder_rates
        self.vae_dim = vae_dim
        self.sample_rate = sample_rate

        self.hop_length = np.prod(self.encoder_rates)
        self.encoder = Encoder(self.encoder_dim, self.encoder_rates, self.latent_dim)

        # bottleneck
        self.quantizer = VAEBottleneck(latent_dim=self.latent_dim, dim=self.vae_dim)

        self.decoder = Decoder(
            self.vae_dim,
            self.decoder_dim,
            self.decoder_rates,
        )

        # S = torch.load("/home/minz/logs/sac_vae_encodecP_peaq_5e5_stereo/last.ckpt")
        # relevant_weights = {k[6:]: v for k, v in S["state_dict"].items() if k.split(".")[1] in ["encoder", "decoder", "quantizer"]}
        # self.load_state_dict(relevant_weights, strict=True)

        # Freeze encoder and quantizer parameters
        # for param in self.encoder.parameters():
        #     param.requires_grad = False
        # for param in self.quantizer.parameters():
        #     param.requires_grad = False

    def get_model_hyperparameters(self):
        return {
            "kwargs": {
                "encoder_dim": self.encoder_dim,
                "encoder_rates": self.encoder_rates,
                "latent_dim": self.latent_dim,
                "decoder_dim": self.decoder_dim,
                "decoder_rates": self.decoder_rates,
                "vae_dim": self.vae_dim,
                "sample_rate": self.sample_rate,
            }
        }

    def preprocess(self, audio_data, sample_rate):
        if sample_rate is None:
            sample_rate = self.sample_rate
        assert sample_rate == self.sample_rate

        length = audio_data.shape[-1]
        right_pad = math.ceil(length / self.hop_length) * self.hop_length - length
        audio_data = nn.functional.pad(audio_data, (0, right_pad))

        return audio_data

    def encode(
        self,
        audio_data: torch.Tensor,
    ):
        """Encode given audio data and return quantized latent codes

        Parameters
        ----------
        audio_data : Tensor[B x 1 x T]
            Audio data to encode
        n_quantizers : int, optional
            Number of quantizers to use, by default None
            If None, all quantizers are used.

        Returns
        -------
        dict
            A dictionary with the following keys:
            "z" : Tensor[B x D x T]
                Quantized continuous representation of input
            "codes" : Tensor[B x N x T]
                Codebook indices for each codebook
                (quantized discrete representation of input)
            "latents" : Tensor[B x N*D x T]
                Projected latents (continuous representation of input before quantization)
            "vq/commitment_loss" : Tensor[1]
                Commitment loss to train encoder to predict vectors closer to codebook
                entries
            "vq/codebook_loss" : Tensor[1]
                Codebook loss to update the codebook
            "length" : int
                Number of samples in input audio
        """
        z = self.encoder(audio_data)
        return self.quantizer(z)

    def decode(self, z: torch.Tensor):
        """Decode given latent codes and return audio data

        Parameters
        ----------
        z : Tensor[B x D x T]
            Quantized continuous representation of input
        length : int, optional
            Number of samples in output audio, by default None

        Returns
        -------
        dict
            A dictionary with the following keys:
            "audio" : Tensor[B x 1 x length]
                Decoded audio data.
        """

        return self.decoder(z)

    def forward(
        self,
        audio_data: torch.Tensor,
        sample_rate: int = None,
    ):
        """Model forward pass

        Parameters
        ----------
        audio_data : Tensor[B x 2 x T]
            Audio data to encode
        sample_rate : int, optional
            Sample rate of audio data in Hz, by default None
            If None, defaults to `self.sample_rate`

        Returns
        -------
        dict
            A dictionary with the following keys:
            "z" : Tensor[B x D x T]
                Quantized continuous representation of input
            "audio" : Tensor[B x 1 x length]
                Decoded audio data.
        """
        length = audio_data.shape[-1]
        audio_data = self.preprocess(audio_data, sample_rate)
        q_res = self.encode(audio_data)

        x = self.decode(q_res["z"])
        return {
            "audio": x[..., :length],
            **q_res,
        }

    def get_num_params(self):
        return sum(p.numel() for p in self.parameters() if p.requires_grad)