# --------------------------------------------------------
# UniFormer
# Copyright (c) 2022 SenseTime X-Lab
# Licensed under The MIT License [see LICENSE for details]
# Written by Kunchang Li
# --------------------------------------------------------


import torch
import torch.nn as nn
import torch.nn.functional as F
import torch.utils.checkpoint as checkpoint

from functools import partial
from collections import OrderedDict
from timm.models.layers import DropPath, to_2tuple, trunc_normal_

try:
    from mmseg.utils import get_root_logger
    from mmseg.models.builder import BACKBONES
except ImportError:
    from annotator.mmpkg.mmseg.utils import get_root_logger
    from annotator.mmpkg.mmseg.models.builder import BACKBONES
    
from annotator.uniformer.mmcv_custom import load_checkpoint


class Mlp(nn.Module):
    def __init__(self, in_features, hidden_features=None, out_features=None, act_layer=nn.GELU, drop=0.):
        super().__init__()
        out_features = out_features or in_features
        hidden_features = hidden_features or in_features
        self.fc1 = nn.Linear(in_features, hidden_features)
        self.act = act_layer()
        self.fc2 = nn.Linear(hidden_features, out_features)
        self.drop = nn.Dropout(drop)

    def forward(self, x):
        x = self.fc1(x)
        x = self.act(x)
        x = self.drop(x)
        x = self.fc2(x)
        x = self.drop(x)
        return x


class CMlp(nn.Module):
    def __init__(self, in_features, hidden_features=None, out_features=None, act_layer=nn.GELU, drop=0.):
        super().__init__()
        out_features = out_features or in_features
        hidden_features = hidden_features or in_features
        self.fc1 = nn.Conv2d(in_features, hidden_features, 1)
        self.act = act_layer()
        self.fc2 = nn.Conv2d(hidden_features, out_features, 1)
        self.drop = nn.Dropout(drop)

    def forward(self, x):
        x = self.fc1(x)
        x = self.act(x)
        x = self.drop(x)
        x = self.fc2(x)
        x = self.drop(x)
        return x


class CBlock(nn.Module):
    def __init__(self, dim, num_heads, mlp_ratio=4., qkv_bias=False, qk_scale=None, drop=0., attn_drop=0.,
                 drop_path=0., act_layer=nn.GELU, norm_layer=nn.LayerNorm):
        super().__init__()
        self.pos_embed = nn.Conv2d(dim, dim, 3, padding=1, groups=dim)
        self.norm1 = nn.BatchNorm2d(dim)
        self.conv1 = nn.Conv2d(dim, dim, 1)
        self.conv2 = nn.Conv2d(dim, dim, 1)
        self.attn = nn.Conv2d(dim, dim, 5, padding=2, groups=dim)
        # NOTE: drop path for stochastic depth, we shall see if this is better than dropout here
        self.drop_path = DropPath(drop_path) if drop_path > 0. else nn.Identity()
        self.norm2 = nn.BatchNorm2d(dim)
        mlp_hidden_dim = int(dim * mlp_ratio)
        self.mlp = CMlp(in_features=dim, hidden_features=mlp_hidden_dim, act_layer=act_layer, drop=drop)

    def forward(self, x):
        x = x + self.pos_embed(x)
        x = x + self.drop_path(self.conv2(self.attn(self.conv1(self.norm1(x)))))
        x = x + self.drop_path(self.mlp(self.norm2(x)))
        return x


class Attention(nn.Module):
    def __init__(self, dim, num_heads=8, qkv_bias=False, qk_scale=None, attn_drop=0., proj_drop=0.):
        super().__init__()
        self.num_heads = num_heads
        head_dim = dim // num_heads
        # NOTE scale factor was wrong in my original version, can set manually to be compat with prev weights
        self.scale = qk_scale or head_dim ** -0.5

        self.qkv = nn.Linear(dim, dim * 3, bias=qkv_bias)
        self.attn_drop = nn.Dropout(attn_drop)
        self.proj = nn.Linear(dim, dim)
        self.proj_drop = nn.Dropout(proj_drop)

    def forward(self, x):
        B, N, C = x.shape
        qkv = self.qkv(x).reshape(B, N, 3, self.num_heads, C // self.num_heads).permute(2, 0, 3, 1, 4)
        q, k, v = qkv[0], qkv[1], qkv[2]   # make torchscript happy (cannot use tensor as tuple)

        attn = (q @ k.transpose(-2, -1)) * self.scale
        attn = attn.softmax(dim=-1)
        attn = self.attn_drop(attn)

        x = (attn @ v).transpose(1, 2).reshape(B, N, C)
        x = self.proj(x)
        x = self.proj_drop(x)
        return x


class SABlock(nn.Module):
    def __init__(self, dim, num_heads, mlp_ratio=4., qkv_bias=False, qk_scale=None, drop=0., attn_drop=0.,
                 drop_path=0., act_layer=nn.GELU, norm_layer=nn.LayerNorm):
        super().__init__()
        self.pos_embed = nn.Conv2d(dim, dim, 3, padding=1, groups=dim)
        self.norm1 = norm_layer(dim)
        self.attn = Attention(
            dim,
            num_heads=num_heads, qkv_bias=qkv_bias, qk_scale=qk_scale,
            attn_drop=attn_drop, proj_drop=drop)
        # NOTE: drop path for stochastic depth, we shall see if this is better than dropout here
        self.drop_path = DropPath(drop_path) if drop_path > 0. else nn.Identity()
        self.norm2 = norm_layer(dim)
        mlp_hidden_dim = int(dim * mlp_ratio)
        self.mlp = Mlp(in_features=dim, hidden_features=mlp_hidden_dim, act_layer=act_layer, drop=drop)

    def forward(self, x):
        x = x + self.pos_embed(x)
        B, N, H, W = x.shape
        x = x.flatten(2).transpose(1, 2)
        x = x + self.drop_path(self.attn(self.norm1(x)))
        x = x + self.drop_path(self.mlp(self.norm2(x)))
        x = x.transpose(1, 2).reshape(B, N, H, W)
        return x   


def window_partition(x, window_size):
    """
    Args:
        x: (B, H, W, C)
        window_size (int): window size
    Returns:
        windows: (num_windows*B, window_size, window_size, C)
    """
    B, H, W, C = x.shape
    x = x.view(B, H // window_size, window_size, W // window_size, window_size, C)
    windows = x.permute(0, 1, 3, 2, 4, 5).contiguous().view(-1, window_size, window_size, C)
    return windows


def window_reverse(windows, window_size, H, W):
    """
    Args:
        windows: (num_windows*B, window_size, window_size, C)
        window_size (int): Window size
        H (int): Height of image
        W (int): Width of image
    Returns:
        x: (B, H, W, C)
    """
    B = int(windows.shape[0] / (H * W / window_size / window_size))
    x = windows.view(B, H // window_size, W // window_size, window_size, window_size, -1)
    x = x.permute(0, 1, 3, 2, 4, 5).contiguous().view(B, H, W, -1)
    return x


class SABlock_Windows(nn.Module):
    def __init__(self, dim, num_heads, window_size=14, mlp_ratio=4., qkv_bias=False, qk_scale=None, drop=0., attn_drop=0.,
                 drop_path=0., act_layer=nn.GELU, norm_layer=nn.LayerNorm):
        super().__init__()
        self.window_size=window_size
        self.pos_embed = nn.Conv2d(dim, dim, 3, padding=1, groups=dim)
        self.norm1 = norm_layer(dim)
        self.attn = Attention(
            dim,
            num_heads=num_heads, qkv_bias=qkv_bias, qk_scale=qk_scale,
            attn_drop=attn_drop, proj_drop=drop)
        # NOTE: drop path for stochastic depth, we shall see if this is better than dropout here
        self.drop_path = DropPath(drop_path) if drop_path > 0. else nn.Identity()
        self.norm2 = norm_layer(dim)
        mlp_hidden_dim = int(dim * mlp_ratio)
        self.mlp = Mlp(in_features=dim, hidden_features=mlp_hidden_dim, act_layer=act_layer, drop=drop)

    def forward(self, x):
        x = x + self.pos_embed(x)
        x = x.permute(0, 2, 3, 1)
        B, H, W, C = x.shape
        shortcut = x
        x = self.norm1(x)

        pad_l = pad_t = 0
        pad_r = (self.window_size - W % self.window_size) % self.window_size
        pad_b = (self.window_size - H % self.window_size) % self.window_size
        x = F.pad(x, (0, 0, pad_l, pad_r, pad_t, pad_b))
        _, Hp, Wp, _ = x.shape
        
        x_windows = window_partition(x, self.window_size)  # nW*B, window_size, window_size, C
        x_windows = x_windows.view(-1, self.window_size * self.window_size, C)  # nW*B, window_size*window_size, C

        # W-MSA/SW-MSA
        attn_windows = self.attn(x_windows)  # nW*B, window_size*window_size, C

        # merge windows
        attn_windows = attn_windows.view(-1, self.window_size, self.window_size, C)
        x = window_reverse(attn_windows, self.window_size, Hp, Wp)  # B H' W' C

        # reverse cyclic shift
        if pad_r > 0 or pad_b > 0:
            x = x[:, :H, :W, :].contiguous()

        x = shortcut + self.drop_path(x)
        x = x + self.drop_path(self.mlp(self.norm2(x)))
        x = x.permute(0, 3, 1, 2).reshape(B, C, H, W)
        return x 
             

class PatchEmbed(nn.Module):
    """ Image to Patch Embedding
    """
    def __init__(self, img_size=224, patch_size=16, in_chans=3, embed_dim=768):
        super().__init__()
        img_size = to_2tuple(img_size)
        patch_size = to_2tuple(patch_size)
        num_patches = (img_size[1] // patch_size[1]) * (img_size[0] // patch_size[0])
        self.img_size = img_size
        self.patch_size = patch_size
        self.num_patches = num_patches
        self.norm = nn.LayerNorm(embed_dim)
        self.proj = nn.Conv2d(in_chans, embed_dim, kernel_size=patch_size, stride=patch_size)

    def forward(self, x):
        B, _, H, W = x.shape
        x = self.proj(x)
        B, _, H, W = x.shape
        x = x.flatten(2).transpose(1, 2)
        x = self.norm(x)
        x = x.reshape(B, H, W, -1).permute(0, 3, 1, 2).contiguous()
        return x
    

@BACKBONES.register_module()   
class UniFormer(nn.Module):
    """ Vision Transformer
    A PyTorch impl of : `An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale`  -
        https://arxiv.org/abs/2010.11929
    """
    def __init__(self, layers=[3, 4, 8, 3], img_size=224, in_chans=3, num_classes=80, embed_dim=[64, 128, 320, 512],
                 head_dim=64, mlp_ratio=4., qkv_bias=True, qk_scale=None, representation_size=None,
                 drop_rate=0., attn_drop_rate=0., drop_path_rate=0., norm_layer=partial(nn.LayerNorm, eps=1e-6),
                 pretrained_path=None, use_checkpoint=False, checkpoint_num=[0, 0, 0, 0], 
                 windows=False, hybrid=False, window_size=14):
        """
        Args:
            layer (list): number of block in each layer
            img_size (int, tuple): input image size
            in_chans (int): number of input channels
            num_classes (int): number of classes for classification head
            embed_dim (int): embedding dimension
            head_dim (int): dimension of attention heads
            mlp_ratio (int): ratio of mlp hidden dim to embedding dim
            qkv_bias (bool): enable bias for qkv if True
            qk_scale (float): override default qk scale of head_dim ** -0.5 if set
            representation_size (Optional[int]): enable and set representation layer (pre-logits) to this value if set
            drop_rate (float): dropout rate
            attn_drop_rate (float): attention dropout rate
            drop_path_rate (float): stochastic depth rate
            norm_layer (nn.Module): normalization layer
            pretrained_path (str): path of pretrained model
            use_checkpoint (bool): whether use checkpoint
            checkpoint_num (list): index for using checkpoint in every stage
            windows (bool): whether use window MHRA
            hybrid (bool): whether use hybrid MHRA
            window_size (int): size of window (>14)
        """
        super().__init__()
        self.num_classes = num_classes
        self.use_checkpoint = use_checkpoint
        self.checkpoint_num = checkpoint_num
        self.windows = windows
        print(f'Use Checkpoint: {self.use_checkpoint}')
        print(f'Checkpoint Number: {self.checkpoint_num}')
        self.num_features = self.embed_dim = embed_dim  # num_features for consistency with other models
        norm_layer = norm_layer or partial(nn.LayerNorm, eps=1e-6) 
        
        self.patch_embed1 = PatchEmbed(
            img_size=img_size, patch_size=4, in_chans=in_chans, embed_dim=embed_dim[0])
        self.patch_embed2 = PatchEmbed(
            img_size=img_size // 4, patch_size=2, in_chans=embed_dim[0], embed_dim=embed_dim[1])
        self.patch_embed3 = PatchEmbed(
            img_size=img_size // 8, patch_size=2, in_chans=embed_dim[1], embed_dim=embed_dim[2])
        self.patch_embed4 = PatchEmbed(
            img_size=img_size // 16, patch_size=2, in_chans=embed_dim[2], embed_dim=embed_dim[3])

        self.pos_drop = nn.Dropout(p=drop_rate)
        dpr = [x.item() for x in torch.linspace(0, drop_path_rate, sum(layers))]  # stochastic depth decay rule
        num_heads = [dim // head_dim for dim in embed_dim]
        self.blocks1 = nn.ModuleList([
            CBlock(
                dim=embed_dim[0], num_heads=num_heads[0], mlp_ratio=mlp_ratio, qkv_bias=qkv_bias, qk_scale=qk_scale,
                drop=drop_rate, attn_drop=attn_drop_rate, drop_path=dpr[i], norm_layer=norm_layer)
            for i in range(layers[0])])
        self.norm1=norm_layer(embed_dim[0])
        self.blocks2 = nn.ModuleList([
            CBlock(
                dim=embed_dim[1], num_heads=num_heads[1], mlp_ratio=mlp_ratio, qkv_bias=qkv_bias, qk_scale=qk_scale,
                drop=drop_rate, attn_drop=attn_drop_rate, drop_path=dpr[i+layers[0]], norm_layer=norm_layer)
            for i in range(layers[1])])
        self.norm2 = norm_layer(embed_dim[1])
        if self.windows:
            print('Use local window for all blocks in stage3')
            self.blocks3 = nn.ModuleList([
            SABlock_Windows(
                dim=embed_dim[2], num_heads=num_heads[2], window_size=window_size, mlp_ratio=mlp_ratio, qkv_bias=qkv_bias, qk_scale=qk_scale,
                drop=drop_rate, attn_drop=attn_drop_rate, drop_path=dpr[i+layers[0]+layers[1]], norm_layer=norm_layer)
            for i in range(layers[2])])
        elif hybrid:
            print('Use hybrid window for blocks in stage3')
            block3 = []
            for i in range(layers[2]):
                if (i + 1) % 4 == 0:
                    block3.append(SABlock(
                    dim=embed_dim[2], num_heads=num_heads[2], mlp_ratio=mlp_ratio, qkv_bias=qkv_bias, qk_scale=qk_scale,
                    drop=drop_rate, attn_drop=attn_drop_rate, drop_path=dpr[i+layers[0]+layers[1]], norm_layer=norm_layer))
                else:
                    block3.append(SABlock_Windows(
                    dim=embed_dim[2], num_heads=num_heads[2], window_size=window_size, mlp_ratio=mlp_ratio, qkv_bias=qkv_bias, qk_scale=qk_scale,
                    drop=drop_rate, attn_drop=attn_drop_rate, drop_path=dpr[i+layers[0]+layers[1]], norm_layer=norm_layer))
            self.blocks3 = nn.ModuleList(block3)
        else:
            print('Use global window for all blocks in stage3')
            self.blocks3 = nn.ModuleList([
            SABlock(
                dim=embed_dim[2], num_heads=num_heads[2], mlp_ratio=mlp_ratio, qkv_bias=qkv_bias, qk_scale=qk_scale,
                drop=drop_rate, attn_drop=attn_drop_rate, drop_path=dpr[i+layers[0]+layers[1]], norm_layer=norm_layer)
            for i in range(layers[2])])
        self.norm3 = norm_layer(embed_dim[2])
        self.blocks4 = nn.ModuleList([
            SABlock(
                dim=embed_dim[3], num_heads=num_heads[3], mlp_ratio=mlp_ratio, qkv_bias=qkv_bias, qk_scale=qk_scale,
                drop=drop_rate, attn_drop=attn_drop_rate, drop_path=dpr[i+layers[0]+layers[1]+layers[2]], norm_layer=norm_layer)
            for i in range(layers[3])])
        self.norm4 = norm_layer(embed_dim[3])
        
        # Representation layer
        if representation_size:
            self.num_features = representation_size
            self.pre_logits = nn.Sequential(OrderedDict([
                ('fc', nn.Linear(embed_dim, representation_size)),
                ('act', nn.Tanh())
            ]))
        else:
            self.pre_logits = nn.Identity()
        
        self.apply(self._init_weights)
        self.init_weights(pretrained=pretrained_path)
        
    def init_weights(self, pretrained):
        if isinstance(pretrained, str):
            logger = get_root_logger()
            load_checkpoint(self, pretrained, map_location='cpu', strict=False, logger=logger)
            print(f'Load pretrained model from {pretrained}')
    def _init_weights(self, m):
        if isinstance(m, nn.Linear):
            trunc_normal_(m.weight, std=.02)
            if isinstance(m, nn.Linear) and m.bias is not None:
                nn.init.constant_(m.bias, 0)
        elif isinstance(m, nn.LayerNorm):
            nn.init.constant_(m.bias, 0)
            nn.init.constant_(m.weight, 1.0)

    @torch.jit.ignore
    def no_weight_decay(self):
        return {'pos_embed', 'cls_token'}

    def get_classifier(self):
        return self.head

    def reset_classifier(self, num_classes, global_pool=''):
        self.num_classes = num_classes
        self.head = nn.Linear(self.embed_dim, num_classes) if num_classes > 0 else nn.Identity()

    def forward_features(self, x):
        out = []
        x = self.patch_embed1(x)
        x = self.pos_drop(x)
        for i, blk in enumerate(self.blocks1):
            if self.use_checkpoint and i < self.checkpoint_num[0]:
                x = checkpoint.checkpoint(blk, x)
            else:
                x = blk(x)
        x_out = self.norm1(x.permute(0, 2, 3, 1))
        out.append(x_out.permute(0, 3, 1, 2).contiguous())
        x = self.patch_embed2(x)
        for i, blk in enumerate(self.blocks2):
            if self.use_checkpoint and i < self.checkpoint_num[1]:
                x = checkpoint.checkpoint(blk, x)
            else:
                x = blk(x)
        x_out = self.norm2(x.permute(0, 2, 3, 1))
        out.append(x_out.permute(0, 3, 1, 2).contiguous())
        x = self.patch_embed3(x)
        for i, blk in enumerate(self.blocks3):
            if self.use_checkpoint and i < self.checkpoint_num[2]:
                x = checkpoint.checkpoint(blk, x)
            else:
                x = blk(x)
        x_out = self.norm3(x.permute(0, 2, 3, 1))
        out.append(x_out.permute(0, 3, 1, 2).contiguous())
        x = self.patch_embed4(x)
        for i, blk in enumerate(self.blocks4):
            if self.use_checkpoint and i < self.checkpoint_num[3]:
                x = checkpoint.checkpoint(blk, x)
            else:
                x = blk(x)
        x_out = self.norm4(x.permute(0, 2, 3, 1))
        out.append(x_out.permute(0, 3, 1, 2).contiguous())
        return tuple(out)

    def forward(self, x):
        x = self.forward_features(x)
        return x