import torch
import torch.nn as nn
import torch.nn.functional as F
import math


class GPTConfig:
    """Configuration for our GPT model."""
    def __init__(
        self,
        vocab_size: int,
        block_size: int,
        n_layer: int = 8,
        n_head: int = 8,
        n_embd: int = 512,
    ):
        self.vocab_size = vocab_size
        self.block_size = block_size
        self.n_layer = n_layer
        self.n_head = n_head
        self.n_embd = n_embd


class CausalSelfAttention(nn.Module):
    """A single multi-head causal self-attention layer."""
    def __init__(self, config: GPTConfig):
        super().__init__()
        assert config.n_embd % config.n_head == 0
        self.key = nn.Linear(config.n_embd, config.n_embd)
        self.query = nn.Linear(config.n_embd, config.n_embd)
        self.value = nn.Linear(config.n_embd, config.n_embd)
        self.proj = nn.Linear(config.n_embd, config.n_embd)
        self.n_head = config.n_head
        self.head_dim = config.n_embd // config.n_head
        # causal mask, buffer not a parameter
        mask = torch.tril(torch.ones(config.block_size, config.block_size))
        self.register_buffer("mask", mask)

    def forward(self, x):
        B, T, C = x.size()
        k = self.key(x).view(B, T, self.n_head, self.head_dim).transpose(1, 2)
        q = self.query(x).view(B, T, self.n_head, self.head_dim).transpose(1, 2)
        # compute attention scores
        att = (q @ k.transpose(-2, -1)) / math.sqrt(self.head_dim)
        att = att.masked_fill(self.mask[:T, :T] == 0, float('-inf'))
        att = F.softmax(att, dim=-1)
        v = self.value(x).view(B, T, self.n_head, self.head_dim).transpose(1, 2)
        out = att @ v
        out = out.transpose(1, 2).contiguous().view(B, T, C)
        return self.proj(out)


class MLP(nn.Module):
    """Feed-forward layer."""
    def __init__(self, config: GPTConfig):
        super().__init__()
        self.fc1 = nn.Linear(config.n_embd, 4 * config.n_embd)
        self.fc2 = nn.Linear(4 * config.n_embd, config.n_embd)

    def forward(self, x):
        return self.fc2(F.gelu(self.fc1(x)))


class Block(nn.Module):
    """Transformer block: attention + feed-forward."""
    def __init__(self, config: GPTConfig):
        super().__init__()
        self.ln1 = nn.LayerNorm(config.n_embd)
        self.ln2 = nn.LayerNorm(config.n_embd)
        self.attn = CausalSelfAttention(config)
        self.mlp = MLP(config)

    def forward(self, x):
        x = x + self.attn(self.ln1(x))
        x = x + self.mlp(self.ln2(x))
        return x


class GPT(nn.Module):
    """GPT language model from scratch."""
    def __init__(self, config: GPTConfig):
        super().__init__()
        self.token_emb = nn.Embedding(config.vocab_size, config.n_embd)
        self.pos_emb = nn.Parameter(torch.zeros(1, config.block_size, config.n_embd))
        self.blocks = nn.ModuleList([Block(config) for _ in range(config.n_layer)])
        self.ln_f = nn.LayerNorm(config.n_embd)
        self.head = nn.Linear(config.n_embd, config.vocab_size, bias=False)
        self.block_size = config.block_size

    def forward(self, idx, targets=None):
        B, T = idx.size()
        tok_emb = self.token_emb(idx)          # (B,T,C)
        pos_emb = self.pos_emb[:, :T, :]      # (1,T,C)
        x = tok_emb + pos_emb
        for block in self.blocks:
            x = block(x)
        x = self.ln_f(x)
        logits = self.head(x)                 # (B,T,vocab)
        loss = None
        if targets is not None:
            # shift logits and targets for next-token prediction
            logits = logits.view(B * T, -1)
            targets = targets.view(B * T)
            loss = F.cross_entropy(logits, targets)
        return logits, loss

    @torch.no_grad()
    def generate(
        self,
        idx,
        max_new_tokens: int,
        temperature: float = 1.0,
        top_k: int = None
    ):
        """
        Iteratively predict next token and append to sequence.
        - idx is (B,T) starting context.
        - Returns (B, T+max_new_tokens).
        """
        for _ in range(max_new_tokens):
            idx_cond = idx[:, -self.block_size :]
            logits, _ = self(idx_cond)
            logits = logits[:, -1, :] / temperature
            if top_k is not None:
                v, _ = torch.topk(logits, top_k)
                logits[logits < v[:, [-1]]] = -float('Inf')
            probs = F.softmax(logits, dim=-1)
            next_token = torch.multinomial(probs, num_samples=1)
            idx = torch.cat([idx, next_token], dim=1)
        return idx