"""
NanoDiffusionGPT model.
"""

import inspect
import math
from dataclasses import dataclass
from typing import Optional

import torch
import torch.nn as nn
from torch.nn import functional as F


class LayerNorm(nn.Module):
    """LayerNorm with optional bias."""

    def __init__(self, ndim, bias):
        super().__init__()
        self.weight = nn.Parameter(torch.ones(ndim))
        self.bias = nn.Parameter(torch.zeros(ndim)) if bias else None

    def forward(self, input):
        return F.layer_norm(input, self.weight.shape, self.weight, self.bias, 1e-5)


class BidirectionalSelfAttention(nn.Module):
    """Self-attention where every token can look left and right."""

    def __init__(self, config):
        super().__init__()
        assert config.n_embd % config.n_head == 0
        self.c_attn = nn.Linear(config.n_embd, 3 * config.n_embd, bias=config.bias)
        self.c_proj = nn.Linear(config.n_embd, config.n_embd, bias=config.bias)
        self.attn_dropout = nn.Dropout(config.dropout)
        self.resid_dropout = nn.Dropout(config.dropout)
        self.n_head = config.n_head
        self.n_embd = config.n_embd
        self.dropout = config.dropout
        self.flash = hasattr(torch.nn.functional, "scaled_dot_product_attention")
        if not self.flash:
            print("WARNING: using slow attention. Flash Attention requires PyTorch >= 2.0")

    def forward(self, x):
        B, T, C = x.size()

        q, k, v = self.c_attn(x).split(self.n_embd, dim=2)
        k = k.view(B, T, self.n_head, C // self.n_head).transpose(1, 2)
        q = q.view(B, T, self.n_head, C // self.n_head).transpose(1, 2)
        v = v.view(B, T, self.n_head, C // self.n_head).transpose(1, 2)

        if self.flash:
            y = F.scaled_dot_product_attention(
                q,
                k,
                v,
                attn_mask=None,
                dropout_p=self.dropout if self.training else 0,
                is_causal=False,
            )
        else:
            att = (q @ k.transpose(-2, -1)) * (1.0 / math.sqrt(k.size(-1)))
            att = F.softmax(att, dim=-1)
            att = self.attn_dropout(att)
            y = att @ v

        y = y.transpose(1, 2).contiguous().view(B, T, C)
        y = self.resid_dropout(self.c_proj(y))
        return y


class MLP(nn.Module):
    def __init__(self, config):
        super().__init__()
        self.c_fc = nn.Linear(config.n_embd, 4 * config.n_embd, bias=config.bias)
        self.gelu = nn.GELU()
        self.c_proj = nn.Linear(4 * config.n_embd, config.n_embd, bias=config.bias)
        self.dropout = nn.Dropout(config.dropout)

    def forward(self, x):
        x = self.c_fc(x)
        x = self.gelu(x)
        x = self.c_proj(x)
        x = self.dropout(x)
        return x


class Block(nn.Module):
    def __init__(self, config):
        super().__init__()
        self.ln_1 = LayerNorm(config.n_embd, bias=config.bias)
        self.attn = BidirectionalSelfAttention(config)
        self.ln_2 = LayerNorm(config.n_embd, bias=config.bias)
        self.mlp = MLP(config)

    def forward(self, x):
        x = x + self.attn(self.ln_1(x))
        x = x + self.mlp(self.ln_2(x))
        return x


@dataclass
class NanoDiffusionGPTConfig:
    block_size: int = 256
    vocab_size: int = 66
    n_layer: int = 6
    n_head: int = 6
    n_embd: int = 384
    dropout: float = 0.0
    bias: bool = True
    mask_token_id: Optional[int] = None


class NanoDiffusionGPT(nn.Module):
    def __init__(self, config):
        super().__init__()
        assert config.vocab_size is not None
        assert config.block_size is not None
        if config.mask_token_id is None:
            config.mask_token_id = config.vocab_size - 1
        assert 0 <= config.mask_token_id < config.vocab_size
        self.config = config

        self.transformer = nn.ModuleDict(
            dict(
                wte=nn.Embedding(config.vocab_size, config.n_embd),
                wpe=nn.Embedding(config.block_size, config.n_embd),
                drop=nn.Dropout(config.dropout),
                h=nn.ModuleList([Block(config) for _ in range(config.n_layer)]),
                ln_f=LayerNorm(config.n_embd, bias=config.bias),
            )
        )
        self.lm_head = nn.Linear(config.n_embd, config.vocab_size, bias=False)
        self.transformer.wte.weight = self.lm_head.weight

        self.apply(self._init_weights)
        for pn, p in self.named_parameters():
            if pn.endswith("c_proj.weight"):
                torch.nn.init.normal_(p, mean=0.0, std=0.02 / math.sqrt(2 * config.n_layer))

        print("number of parameters: %.2fM" % (self.get_num_params() / 1e6,))

    def get_num_params(self, non_embedding=True):
        n_params = sum(p.numel() for p in self.parameters())
        if non_embedding:
            n_params -= self.transformer.wpe.weight.numel()
        return n_params

    def _init_weights(self, module):
        if isinstance(module, nn.Linear):
            torch.nn.init.normal_(module.weight, mean=0.0, std=0.02)
            if module.bias is not None:
                torch.nn.init.zeros_(module.bias)
        elif isinstance(module, nn.Embedding):
            torch.nn.init.normal_(module.weight, mean=0.0, std=0.02)

    def forward(self, idx, targets=None, loss_mask=None):
        device = idx.device
        b, t = idx.size()
        assert t <= self.config.block_size, (
            f"Cannot forward sequence of length {t}, block size is only {self.config.block_size}"
        )
        pos = torch.arange(0, t, dtype=torch.long, device=device)

        tok_emb = self.transformer.wte(idx)
        pos_emb = self.transformer.wpe(pos)
        x = self.transformer.drop(tok_emb + pos_emb)
        for block in self.transformer.h:
            x = block(x)
        x = self.transformer.ln_f(x)
        logits = self.lm_head(x)

        loss = None
        if targets is not None:
            if loss_mask is None:
                loss_mask = torch.ones_like(targets, dtype=torch.bool)
            logits_for_loss = logits.clone()
            logits_for_loss[..., self.config.mask_token_id] = -float("inf")
            loss = F.cross_entropy(logits_for_loss[loss_mask], targets[loss_mask])

        return logits, loss

    def configure_optimizers(self, weight_decay, learning_rate, betas, device_type):
        param_dict = {pn: p for pn, p in self.named_parameters() if p.requires_grad}
        decay_params = [p for n, p in param_dict.items() if p.dim() >= 2]
        nodecay_params = [p for n, p in param_dict.items() if p.dim() < 2]
        optim_groups = [
            {"params": decay_params, "weight_decay": weight_decay},
            {"params": nodecay_params, "weight_decay": 0.0},
        ]
        num_decay_params = sum(p.numel() for p in decay_params)
        num_nodecay_params = sum(p.numel() for p in nodecay_params)
        print(f"num decayed parameter tensors: {len(decay_params)}, with {num_decay_params:,} parameters")
        print(f"num non-decayed parameter tensors: {len(nodecay_params)}, with {num_nodecay_params:,} parameters")
        fused_available = "fused" in inspect.signature(torch.optim.AdamW).parameters
        use_fused = fused_available and device_type == "cuda"
        extra_args = dict(fused=True) if use_fused else dict()
        optimizer = torch.optim.AdamW(optim_groups, lr=learning_rate, betas=betas, **extra_args)
        print(f"using fused AdamW: {use_fused}")
        return optimizer

    def estimate_mfu(self, fwdbwd_per_iter, dt):
        N = self.get_num_params()
        cfg = self.config
        L, H, Q, T = cfg.n_layer, cfg.n_head, cfg.n_embd // cfg.n_head, cfg.block_size
        flops_per_token = 6 * N + 12 * L * H * Q * T
        flops_per_fwdbwd = flops_per_token * T
        flops_per_iter = flops_per_fwdbwd * fwdbwd_per_iter
        flops_achieved = flops_per_iter * (1.0 / dt)
        flops_promised = 312e12
        return flops_achieved / flops_promised

    @torch.no_grad()
    def generate(self, prompt_ids, max_new_tokens, steps=64, temperature=1.0, top_k=None):
        """
        Start with prompt + [MASK] tokens, then repeatedly fill the most confident masks.
        """
        assert prompt_ids.dim() == 2
        total_len = prompt_ids.size(1) + max_new_tokens
        assert total_len <= self.config.block_size, (
            f"prompt + max_new_tokens is {total_len}, block size is only {self.config.block_size}"
        )

        device = prompt_ids.device
        x = torch.full(
            (prompt_ids.size(0), total_len),
            self.config.mask_token_id,
            dtype=torch.long,
            device=device,
        )
        x[:, : prompt_ids.size(1)] = prompt_ids

        prompt_mask = torch.zeros_like(x, dtype=torch.bool)
        prompt_mask[:, : prompt_ids.size(1)] = True
        steps = max(1, steps)

        for step in range(steps):
            remaining = (x == self.config.mask_token_id) & (~prompt_mask)
            if not remaining.any():
                break

            logits, _ = self(x)
            logits[..., self.config.mask_token_id] = -float("inf")
            logits = logits / max(temperature, 1e-6)

            if top_k is not None:
                v, _ = torch.topk(logits, min(top_k, logits.size(-1)))
                logits = logits.masked_fill(logits < v[..., [-1]], -float("inf"))

            probs = F.softmax(logits, dim=-1)
            pred = torch.multinomial(probs.view(-1, probs.size(-1)), num_samples=1).view(x.shape)
            confidence = probs.gather(-1, pred.unsqueeze(-1)).squeeze(-1)
            confidence = confidence.masked_fill(~remaining, -1.0)

            remaining_count = int(remaining.sum().item())
            steps_left = steps - step
            fill_count = max(1, math.ceil(remaining_count / steps_left))

            selected = torch.topk(confidence.view(-1), fill_count).indices
            flat_x = x.view(-1)
            flat_pred = pred.view(-1)
            flat_x[selected] = flat_pred[selected]

        return x