import torch
from torch import nn
import torch.nn.functional as F
import math
from math import sqrt
import numpy as np

# Modified from: https://github.com/thuml/Time-Series-Library
# Modified by Shourya Bose, shbose@ucsc.edu

class PositionalEmbedding(nn.Module):
    def __init__(self, d_model, max_len=5000):
        super(PositionalEmbedding, self).__init__()
        # Compute the positional encodings once in log space.
        pe = torch.zeros(max_len, d_model).float()
        pe.require_grad = False

        position = torch.arange(0, max_len).float().unsqueeze(1)
        div_term = (torch.arange(0, d_model, 2).float()
                    * -(math.log(10000.0) / d_model)).exp()

        pe[:, 0::2] = torch.sin(position * div_term)
        pe[:, 1::2] = torch.cos(position * div_term)

        pe = pe.unsqueeze(0)
        self.register_buffer('pe', pe)

    def forward(self, x):
        return self.pe[:, :x.size(1)]

class PatchEmbedding(nn.Module):
    def __init__(self, d_model, patch_len, stride, padding, dropout):
        super(PatchEmbedding, self).__init__()
        # Patching
        self.patch_len = patch_len
        self.stride = stride
        self.padding_patch_layer = nn.ReplicationPad1d((0, padding))

        # Backbone, Input encoding: projection of feature vectors onto a d-dim vector space
        self.value_embedding = nn.Linear(patch_len, d_model, bias=False)

        # Positional embedding
        self.position_embedding = PositionalEmbedding(d_model)

        # Residual dropout
        self.dropout = nn.Dropout(dropout)

    def forward(self, x):
        # do patching
        n_vars = x.shape[1]
        x = self.padding_patch_layer(x)
        x = x.unfold(dimension=-1, size=self.patch_len, step=self.stride)
        x = torch.reshape(x, (x.shape[0] * x.shape[1], x.shape[2], x.shape[3]))
        # Input encoding
        x = self.value_embedding(x) + self.position_embedding(x)
        return self.dropout(x), n_vars

class AttentionLayer(nn.Module):
    def __init__(self, attention, d_model, n_heads, d_keys=None,
                 d_values=None):
        super(AttentionLayer, self).__init__()

        d_keys = d_keys or (d_model // n_heads)
        d_values = d_values or (d_model // n_heads)

        self.inner_attention = attention
        self.query_projection = nn.Linear(d_model, d_keys * n_heads)
        self.key_projection = nn.Linear(d_model, d_keys * n_heads)
        self.value_projection = nn.Linear(d_model, d_values * n_heads)
        self.out_projection = nn.Linear(d_values * n_heads, d_model)
        self.n_heads = n_heads

    def forward(self, queries, keys, values, attn_mask, tau=None, delta=None):
        B, L, _ = queries.shape
        _, S, _ = keys.shape
        H = self.n_heads

        queries = self.query_projection(queries).view(B, L, H, -1)
        keys = self.key_projection(keys).view(B, S, H, -1)
        values = self.value_projection(values).view(B, S, H, -1)

        out, attn = self.inner_attention(
            queries,
            keys,
            values,
            attn_mask,
            tau=tau,
            delta=delta
        )
        out = out.view(B, L, -1)

        return self.out_projection(out), attn

class FullAttention(nn.Module):
    def __init__(self, mask_flag=True, factor=5, scale=None, attention_dropout=0.1, output_attention=False):
        super(FullAttention, self).__init__()
        self.scale = scale
        self.mask_flag = mask_flag
        self.output_attention = output_attention
        self.dropout = nn.Dropout(attention_dropout)

    def forward(self, queries, keys, values, attn_mask, tau=None, delta=None):
        B, L, H, E = queries.shape
        _, S, _, D = values.shape
        scale = self.scale or 1. / sqrt(E)

        scores = torch.einsum("blhe,bshe->bhls", queries, keys)

        if self.mask_flag:
            if attn_mask is None:
                attn_mask = TriangularCausalMask(B, L, device=queries.device)

            scores.masked_fill_(attn_mask.mask, -np.inf)

        A = self.dropout(torch.softmax(scale * scores, dim=-1))
        V = torch.einsum("bhls,bshd->blhd", A, values)

        if self.output_attention:
            return V.contiguous(), A
        else:
            return V.contiguous(), None

class TriangularCausalMask():
    def __init__(self, B, L, device="cpu"):
        mask_shape = [B, 1, L, L]
        with torch.no_grad():
            self._mask = torch.triu(torch.ones(mask_shape, dtype=torch.bool), diagonal=1).to(device)

    @property
    def mask(self):
        return self._mask

class FullAttention(nn.Module):
    def __init__(self, mask_flag=True, factor=5, scale=None, attention_dropout=0.1, output_attention=False):
        super(FullAttention, self).__init__()
        self.scale = scale
        self.mask_flag = mask_flag
        self.output_attention = output_attention
        self.dropout = nn.Dropout(attention_dropout)

    def forward(self, queries, keys, values, attn_mask, tau=None, delta=None):
        B, L, H, E = queries.shape
        _, S, _, D = values.shape
        scale = self.scale or 1. / sqrt(E)

        scores = torch.einsum("blhe,bshe->bhls", queries, keys)

        if self.mask_flag:
            if attn_mask is None:
                attn_mask = TriangularCausalMask(B, L, device=queries.device)

            scores.masked_fill_(attn_mask.mask, -np.inf)

        A = self.dropout(torch.softmax(scale * scores, dim=-1))
        V = torch.einsum("bhls,bshd->blhd", A, values)

        if self.output_attention:
            return V.contiguous(), A
        else:
            return V.contiguous(), None
        
class AttentionLayer(nn.Module):
    def __init__(self, attention, d_model, n_heads, d_keys=None,
                 d_values=None):
        super(AttentionLayer, self).__init__()

        d_keys = d_keys or (d_model // n_heads)
        d_values = d_values or (d_model // n_heads)

        self.inner_attention = attention
        self.query_projection = nn.Linear(d_model, d_keys * n_heads)
        self.key_projection = nn.Linear(d_model, d_keys * n_heads)
        self.value_projection = nn.Linear(d_model, d_values * n_heads)
        self.out_projection = nn.Linear(d_values * n_heads, d_model)
        self.n_heads = n_heads

    def forward(self, queries, keys, values, attn_mask, tau=None, delta=None):
        B, L, _ = queries.shape
        _, S, _ = keys.shape
        H = self.n_heads

        queries = self.query_projection(queries).view(B, L, H, -1)
        keys = self.key_projection(keys).view(B, S, H, -1)
        values = self.value_projection(values).view(B, S, H, -1)

        out, attn = self.inner_attention(
            queries,
            keys,
            values,
            attn_mask,
            tau=tau,
            delta=delta
        )
        out = out.view(B, L, -1)

        return self.out_projection(out), attn

class EncoderLayer(nn.Module):
    def __init__(self, attention, d_model, d_ff=None, dropout=0.1, activation="relu"):
        super(EncoderLayer, self).__init__()
        d_ff = d_ff or 4 * d_model
        self.attention = attention
        self.conv1 = nn.Conv1d(in_channels=d_model, out_channels=d_ff, kernel_size=1)
        self.conv2 = nn.Conv1d(in_channels=d_ff, out_channels=d_model, kernel_size=1)
        self.norm1 = nn.LayerNorm(d_model)
        self.norm2 = nn.LayerNorm(d_model)
        self.dropout = nn.Dropout(dropout)
        self.activation = F.relu if activation == "relu" else F.gelu

    def forward(self, x, attn_mask=None, tau=None, delta=None):
        new_x, attn = self.attention(
            x, x, x,
            attn_mask=attn_mask,
            tau=tau, delta=delta
        )
        x = x + self.dropout(new_x)

        y = x = self.norm1(x)
        y = self.dropout(self.activation(self.conv1(y.transpose(-1, 1))))
        y = self.dropout(self.conv2(y).transpose(-1, 1))

        return self.norm2(x + y), attn


class Encoder(nn.Module):
    def __init__(self, attn_layers, conv_layers=None, norm_layer=None):
        super(Encoder, self).__init__()
        self.attn_layers = nn.ModuleList(attn_layers)
        self.conv_layers = nn.ModuleList(conv_layers) if conv_layers is not None else None
        self.norm = norm_layer

    def forward(self, x, attn_mask=None, tau=None, delta=None):
        # x [B, L, D]
        attns = []
        if self.conv_layers is not None:
            for i, (attn_layer, conv_layer) in enumerate(zip(self.attn_layers, self.conv_layers)):
                delta = delta if i == 0 else None
                x, attn = attn_layer(x, attn_mask=attn_mask, tau=tau, delta=delta)
                x = conv_layer(x)
                attns.append(attn)
            x, attn = self.attn_layers[-1](x, tau=tau, delta=None)
            attns.append(attn)
        else:
            for attn_layer in self.attn_layers:
                x, attn = attn_layer(x, attn_mask=attn_mask, tau=tau, delta=delta)
                attns.append(attn)

        if self.norm is not None:
            x = self.norm(x)

        return x, attns

class Transpose(nn.Module):
    def __init__(self, *dims, contiguous=False): 
        super().__init__()
        self.dims, self.contiguous = dims, contiguous
    def forward(self, x):
        if self.contiguous: return x.transpose(*self.dims).contiguous()
        else: return x.transpose(*self.dims)


class FlattenHead(nn.Module):
    def __init__(self, n_vars, nf, target_window, head_dropout=0):
        super().__init__()
        self.n_vars = n_vars
        self.flatten = nn.Flatten(start_dim=-2)
        self.linear = nn.Linear(nf, target_window)
        self.dropout = nn.Dropout(head_dropout)

    def forward(self, x):  # x: [bs x nvars x d_model x patch_num]
        x = self.flatten(x)
        x = self.linear(x)
        x = self.dropout(x)
        return x


class PatchTST(nn.Module):
    """
    Paper link: https://arxiv.org/pdf/2211.14730.pdf
    """

    def __init__(
        self,
        enc_in,
        dec_in, # unused
        c_out, # unused
        pred_len,
        seq_len,
        d_model = 64,
        patch_len = 16,
        stride = 8,
        data_idx = [0,3,4,5,6,7],
        time_idx = [1,2],
        output_attention = False,
        factor = 3,
        n_heads = 4,
        d_ff = 512,
        e_layers = 3,
        activation = 'gelu',
        dropout = 0.1
    ):
    
    #(self, configs, patch_len=16, stride=8):
        """
        patch_len: int, patch len for patch_embedding
        stride: int, stride for patch_embedding
        """
        super().__init__()
        self.seq_len = seq_len
        self.pred_len = pred_len
        self.data_idx = data_idx
        self.time_idx = time_idx
        self.dec_in = dec_in
        padding = stride

        # patching and embedding
        self.patch_embedding = PatchEmbedding(
            d_model, patch_len, stride, padding, dropout)

        # Encoder
        self.encoder = Encoder(
            [
                EncoderLayer(
                    AttentionLayer(
                        FullAttention(False, factor, attention_dropout=dropout,
                                      output_attention=output_attention), d_model, n_heads),
                    d_model,
                    d_ff,
                    dropout=dropout,
                    activation=activation
                ) for l in range(e_layers)
            ],
            norm_layer=nn.Sequential(Transpose(1,2), nn.BatchNorm1d(d_model), Transpose(1,2))
        )

        # Prediction Head
        self.head_nf = d_model * \
                       int((seq_len - patch_len) / stride + 2)
        self.head = FlattenHead(enc_in, self.head_nf,pred_len,
                                head_dropout=dropout)

    def forecast(self, x_enc, x_mark_enc, x_dec, x_mark_dec):
        # Normalization from Non-stationary Transformer
        means = x_enc.mean(1, keepdim=True).detach()
        x_enc = x_enc - means
        stdev = torch.sqrt(
            torch.var(x_enc, dim=1, keepdim=True, unbiased=False) + 1e-5)
        x_enc /= stdev

        # do patching and embedding
        x_enc = x_enc.permute(0, 2, 1)
        # u: [bs * nvars x patch_num x d_model]
        enc_out, n_vars = self.patch_embedding(x_enc)

        # Encoder
        # z: [bs * nvars x patch_num x d_model]
        enc_out, attns = self.encoder(enc_out)
        # z: [bs x nvars x patch_num x d_model]
        enc_out = torch.reshape(
            enc_out, (-1, n_vars, enc_out.shape[-2], enc_out.shape[-1]))
        # z: [bs x nvars x d_model x patch_num]
        enc_out = enc_out.permute(0, 1, 3, 2)

        # Decoder
        dec_out = self.head(enc_out)  # z: [bs x nvars x target_window]
        dec_out = dec_out.permute(0, 2, 1)

        # De-Normalization from Non-stationary Transformer
        dec_out = dec_out * \
                  (stdev[:, 0, :].unsqueeze(1).repeat(1, self.pred_len, 1))
        dec_out = dec_out + \
                  (means[:, 0, :].unsqueeze(1).repeat(1, self.pred_len, 1))
        return dec_out
    
    def forward(self, x, fut_time):
        
        x_enc = x[:,:,self.data_idx]
        x_mark_enc = x[:,:,self.time_idx]
        x_dec = torch.zeros((fut_time.shape[0],fut_time.shape[1],self.dec_in),dtype=fut_time.dtype,device=fut_time.device)
        x_mark_dec = fut_time
        
        return self.forecast(x_enc,x_mark_enc,x_dec,x_mark_dec)[:,-1,[0]]