from inspect import isfunction
import global_
import math
import torch
import torch.nn.functional as F
from torch import nn, einsum
from einops import rearrange, repeat

from ldm.modules.diffusionmodules.util import checkpoint
from typing import List, Tuple
from confs import *


def exists(val):
    return val is not None


def uniq(arr):
    return{el: True for el in arr}.keys()


def default(val, d):
    if exists(val):
        return val
    return d() if isfunction(d) else d


def max_neg_value(t):
    return -torch.finfo(t.dtype).max


def init_(tensor):
    dim = tensor.shape[-1]
    std = 1 / math.sqrt(dim)
    tensor.uniform_(-std, std)
    return tensor


# feedforward
class GEGLU(nn.Module):
    def __init__(self, dim_in, dim_out):
        super().__init__()
        self.proj = nn.Linear(dim_in, dim_out * 2)

    def forward(self, x):
        x, gate = self.proj(x).chunk(2, dim=-1)
        return x * F.gelu(gate)


class FeedForward(nn.Module):
    def __init__(self, dim, dim_out=None, mult=4, glu=False, dropout=0., inner_dim=None):
        super().__init__()
        if inner_dim is None:
            inner_dim = int(dim * mult)
        dim_out = default(dim_out, dim)
        project_in = nn.Sequential(
            nn.Linear(dim, inner_dim),
            nn.GELU()
        ) if not glu else GEGLU(dim, inner_dim)
        
        self.dim = dim
        self.inner_dim = inner_dim
        self.dim_out = dim_out

        self.net = nn.Sequential(
            project_in,
            nn.Dropout(dropout),
            nn.Linear(inner_dim, dim_out)
        )

    def forward(self, x, token_pos=None):
        return self.net(x)


def zero_module(module):
    """
    Zero out the parameters of a module and return it.
    """
    for p in module.parameters():
        p.detach().zero_()
    return module


def Normalize(in_channels):
    return torch.nn.GroupNorm(num_groups=32, num_channels=in_channels, eps=1e-6, affine=True)


class LinearAttention(nn.Module):
    def __init__(self, dim, heads=4, dim_head=32):
        super().__init__()
        self.heads = heads
        hidden_dim = dim_head * heads
        self.to_qkv = nn.Conv2d(dim, hidden_dim * 3, 1, bias = False)
        self.to_out = nn.Conv2d(hidden_dim, dim, 1)

    def forward(self, x):
        b, c, h, w = x.shape
        qkv = self.to_qkv(x)
        q, k, v = rearrange(qkv, 'b (qkv heads c) h w -> qkv b heads c (h w)', heads = self.heads, qkv=3)
        k = k.softmax(dim=-1)  
        context = torch.einsum('bhdn,bhen->bhde', k, v)
        out = torch.einsum('bhde,bhdn->bhen', context, q)
        out = rearrange(out, 'b heads c (h w) -> b (heads c) h w', heads=self.heads, h=h, w=w)
        return self.to_out(out)


class SpatialSelfAttention(nn.Module):
    def __init__(self, in_channels):
        super().__init__()
        self.in_channels = in_channels

        self.norm = Normalize(in_channels)
        self.q = torch.nn.Conv2d(in_channels,
                                 in_channels,
                                 kernel_size=1,
                                 stride=1,
                                 padding=0)
        self.k = torch.nn.Conv2d(in_channels,
                                 in_channels,
                                 kernel_size=1,
                                 stride=1,
                                 padding=0)
        self.v = torch.nn.Conv2d(in_channels,
                                 in_channels,
                                 kernel_size=1,
                                 stride=1,
                                 padding=0)
        self.proj_out = torch.nn.Conv2d(in_channels,
                                        in_channels,
                                        kernel_size=1,
                                        stride=1,
                                        padding=0)

    def forward(self, x):
        h_ = x
        h_ = self.norm(h_)
        q = self.q(h_)
        k = self.k(h_)
        v = self.v(h_)

        # compute attention
        b,c,h,w = q.shape
        q = rearrange(q, 'b c h w -> b (h w) c')
        k = rearrange(k, 'b c h w -> b c (h w)')
        w_ = torch.einsum('bij,bjk->bik', q, k)

        w_ = w_ * (int(c)**(-0.5))
        w_ = torch.nn.functional.softmax(w_, dim=2)

        # attend to values
        v = rearrange(v, 'b c h w -> b c (h w)')
        w_ = rearrange(w_, 'b i j -> b j i')
        h_ = torch.einsum('bij,bjk->bik', v, w_)
        h_ = rearrange(h_, 'b c (h w) -> b c h w', h=h)
        h_ = self.proj_out(h_)

        return x+h_


class CrossAttention(nn.Module):
    def __init__(self, query_dim, context_dim=None, heads=8, dim_head=64, dropout=0.,sep_head_att=False):
        super().__init__()
        inner_dim = dim_head * heads
        context_dim = default(context_dim, query_dim)

        self.scale = dim_head ** -0.5
        self.heads = heads  # 8
        self.dim_head=dim_head #40
        self.to_q = nn.Linear(query_dim, inner_dim, bias=False)
        # self.to_k = nn.Linear(context_dim, inner_dim, bias=False)
        # self.to_v = nn.Linear(context_dim, inner_dim, bias=False)

        head_splits=[6,2]
        self.head_splits=head_splits
        # if sep_head_att:
        #     self.to_k = nn.ModuleList([nn.Linear(context_dim, dim_head*head_splits[i], bias=False) for i in range(len(head_splits))])
        #     self.to_v = nn.ModuleList([nn.Linear(context_dim, dim_head*head_splits[i], bias=False) for i in range(len(head_splits))])
        # else:
        self.to_k = nn.Linear(context_dim, inner_dim, bias=False)
        self.to_v = nn.Linear(context_dim, inner_dim, bias=False)

        self.to_out = nn.Sequential(
            nn.Linear(inner_dim, query_dim),
            nn.Dropout(dropout)
        )

    def forward(self, x, context=None, mask=None):
        h = self.heads

        q = self.to_q(x)        # 2,4096,320
        context = default(context, x) #2,4096,320
        if context.shape[-1]==768*2:
            # this is for different attention heads
            context1,context2=torch.chunk(context,2,dim=-1) # clip/id context1, landmark context2
            k1=self.to_k(context1)
            k2=self.to_k(context2)
            v1=self.to_v(context1)
            v2=self.to_v(context2)
            
            k=torch.cat([k1[:,:,:self.head_splits[0]*self.dim_head],k2[:,:,-self.head_splits[1]*self.dim_head:]],dim=-1)
            v=torch.cat([v1[:,:,:self.head_splits[0]*self.dim_head],v2[:,:,-self.head_splits[1]*self.dim_head:]],dim=-1)
            # head_splits=[6,2]
            # k1 = self.to_k[0](context1)
            # v1 = self.to_v[0](context1)
            # k2 = self.to_k[1](context2)
            # v2 = self.to_v[1](context2)
            # k=torch.cat([k1,k2],dim=-1)
            # v=torch.cat([v1,v2],dim=-1)
            
        else:
            k = self.to_k(context)
            v = self.to_v(context)

        q, k, v = map(lambda t: rearrange(t, 'b n (h d) -> (b h) n d', h=h), (q, k, v))

        sim = einsum('b i d, b j d -> b i j', q, k) * self.scale

        if exists(mask):
            mask = rearrange(mask, 'b ... -> b (...)')
            max_neg_value = -torch.finfo(sim.dtype).max
            mask = repeat(mask, 'b j -> (b h) () j', h=h)
            sim.masked_fill_(~mask, max_neg_value)

        # attention, what we cannot get enough of
        attn = sim.softmax(dim=-1)

        out = einsum('b i j, b j d -> b i d', attn, v)
        out = rearrange(out, '(b h) n d -> b n (h d)', h=h)
        return self.to_out(out)


class BasicTransformerBlock(nn.Module):
    def __init__(self, dim, n_heads, d_head, dropout=0., context_dim=None, gated_ff=True, checkpoint=True,sep_head_att=False):
        super().__init__()
        self.attn1 = CrossAttention(query_dim=dim, heads=n_heads, dim_head=d_head, dropout=dropout,sep_head_att=False)  # is a self-attention
        self.ff = FeedForward(dim, dropout=dropout, glu=gated_ff)
        self.attn2 = CrossAttention(query_dim=dim, context_dim=context_dim,
                                    heads=n_heads, dim_head=d_head, dropout=dropout,sep_head_att=sep_head_att)  # is self-attn if context is none
        self.norm1 = nn.LayerNorm(dim)
        self.norm2 = nn.LayerNorm(dim)
        self.norm3 = nn.LayerNorm(dim)
        self.checkpoint = checkpoint

    def forward(self, x, context=None, token_pos=None):
        inputs = (x, context, token_pos, )
        if hasattr(self,'name4bank') and REFNET.task2layerNum[global_.task]>0:
            if self.isReader_4bank:
                inputs = (x, context, token_pos, self.bank.get(self.name4bank) ) # x, context, x_refNet
            else:
                self.bank.set(self.name4bank, x)
        return checkpoint(self._forward, inputs, self.parameters(), self.checkpoint)

    def _forward(self, x, context=None, token_pos=None, x_refNet=None):# x, x_refNet: before LN
        if x_refNet is None:
            x = self.attn1(self.norm1(x)) + x
        else:
            x_norm = self.norm1(x)
            x_norm_cat = torch.cat( [ x_norm, self.norm1(x_refNet) ] , dim=1 )
            x = self.attn1(x_norm, context=x_norm_cat) + x
            del x_norm,x_norm_cat
        x = self.attn2(self.norm2(x), context=context) + x
        # This ff might be modified into an MoE module, so pass token_pos
        x = self.ff(self.norm3(x), token_pos) + x
        return x


class SpatialTransformer(nn.Module):
    """
    Transformer block for image-like data.
    First, project the input (aka embedding)
    and reshape to b, t, d.
    Then apply standard transformer action.
    Finally, reshape to image
    """
    def __init__(self, in_channels, n_heads, d_head,
                 depth=1, dropout=0., context_dim=None,sep_head_att=False,head_splits=None):
        super().__init__()
        self.in_channels = in_channels
        inner_dim = n_heads * d_head
        self.norm = Normalize(in_channels)

        self.proj_in = nn.Conv2d(in_channels,
                                 inner_dim,
                                 kernel_size=1,
                                 stride=1,
                                 padding=0)

        self.transformer_blocks = nn.ModuleList(
            [BasicTransformerBlock(inner_dim, n_heads, d_head, dropout=dropout, context_dim=context_dim,sep_head_att=sep_head_att)
                for d in range(depth)]
        )

        self.proj_out = zero_module(nn.Conv2d(inner_dim,
                                              in_channels,
                                              kernel_size=1,
                                              stride=1,
                                              padding=0))

    def forward(self, x, context=None):
        # note: if no context is given, cross-attention defaults to self-attention
        b, c, h, w = x.shape
        x_in = x
        x = self.norm(x)
        x = self.proj_in(x)
        x = rearrange(x, 'b c h w -> b (h w) c')
        if 1: # set token position grid (normalized centers) for gating/router use
            num_tokens = h * w
            y_coords = torch.arange(h, device=x.device, dtype=x.dtype)
            x_coords = torch.arange(w, device=x.device, dtype=x.dtype)
            yy, xx = torch.meshgrid(y_coords, x_coords, indexing='ij')
            pos = torch.stack([(xx + 0.5) / float(w), (yy + 0.5) / float(h)], dim=-1)  # [h,w,2]
            pos = pos.reshape(1, num_tokens, 2).expand(b, -1, -1).contiguous()  # b, n, 2
        for block in self.transformer_blocks:
            x = block(x, context=context, token_pos=pos)
        x = rearrange(x, 'b (h w) c -> b c h w', h=h, w=w)
        x = self.proj_out(x)
        return x + x_in