"""Small but modern decoder-only transformer (~50M params).

Uses RoPE, RMSNorm, SwiGLU FFN, tied embeddings, and PyTorch SDPA
for causal attention (which lights up MPS fast-paths where available).
"""

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

from config import ModelConfig


def precompute_rope(head_dim: int, seq_len: int, theta: float = 10000.0, device=None):
    inv_freq = 1.0 / (theta ** (torch.arange(0, head_dim, 2, device=device).float() / head_dim))
    t = torch.arange(seq_len, device=device).float()
    freqs = torch.outer(t, inv_freq)  # (T, head_dim/2)
    return freqs.cos(), freqs.sin()


def apply_rope(x: torch.Tensor, cos: torch.Tensor, sin: torch.Tensor) -> torch.Tensor:
    # x: (B, H, T, D); cos/sin: (T, D/2)
    x1, x2 = x.chunk(2, dim=-1)
    cos = cos[None, None, :, :]
    sin = sin[None, None, :, :]
    return torch.cat([x1 * cos - x2 * sin, x1 * sin + x2 * cos], dim=-1)


class RMSNorm(nn.Module):
    def __init__(self, d: int, eps: float = 1e-5):
        super().__init__()
        self.weight = nn.Parameter(torch.ones(d))
        self.eps = eps

    def forward(self, x: torch.Tensor) -> torch.Tensor:
        # Always compute the norm in fp32 for stability, then cast back.
        dtype = x.dtype
        x32 = x.float()
        norm = torch.rsqrt(x32.pow(2).mean(-1, keepdim=True) + self.eps)
        return (x32 * norm).to(dtype) * self.weight


class Attention(nn.Module):
    def __init__(self, cfg: ModelConfig):
        super().__init__()
        assert cfg.d_model % cfg.n_heads == 0
        self.n_heads = cfg.n_heads
        self.head_dim = cfg.d_model // cfg.n_heads
        self.qkv = nn.Linear(cfg.d_model, 3 * cfg.d_model, bias=False)
        self.o = nn.Linear(cfg.d_model, cfg.d_model, bias=False)
        self.dropout = cfg.dropout

    def forward(self, x, cos, sin):
        B, T, C = x.shape
        q, k, v = self.qkv(x).chunk(3, dim=-1)
        q = q.view(B, T, self.n_heads, self.head_dim).transpose(1, 2)
        k = k.view(B, T, self.n_heads, self.head_dim).transpose(1, 2)
        v = v.view(B, T, self.n_heads, self.head_dim).transpose(1, 2)
        q = apply_rope(q, cos[:T], sin[:T])
        k = apply_rope(k, cos[:T], sin[:T])
        y = F.scaled_dot_product_attention(
            q, k, v,
            is_causal=True,
            dropout_p=self.dropout if self.training else 0.0,
        )
        y = y.transpose(1, 2).contiguous().view(B, T, C)
        return self.o(y)


class SwiGLU(nn.Module):
    def __init__(self, cfg: ModelConfig):
        super().__init__()
        self.w1 = nn.Linear(cfg.d_model, cfg.d_ff, bias=False)  # gate
        self.w2 = nn.Linear(cfg.d_ff, cfg.d_model, bias=False)  # down
        self.w3 = nn.Linear(cfg.d_model, cfg.d_ff, bias=False)  # up

    def forward(self, x):
        return self.w2(F.silu(self.w1(x)) * self.w3(x))


class Block(nn.Module):
    def __init__(self, cfg: ModelConfig):
        super().__init__()
        self.attn_norm = RMSNorm(cfg.d_model, cfg.norm_eps)
        self.attn = Attention(cfg)
        self.ffn_norm = RMSNorm(cfg.d_model, cfg.norm_eps)
        self.ffn = SwiGLU(cfg)

    def forward(self, x, cos, sin):
        x = x + self.attn(self.attn_norm(x), cos, sin)
        x = x + self.ffn(self.ffn_norm(x))
        return x


class IntelliteGPT(nn.Module):
    def __init__(self, cfg: ModelConfig):
        super().__init__()
        self.cfg = cfg
        self.tok_emb = nn.Embedding(cfg.vocab_size, cfg.d_model)
        self.blocks = nn.ModuleList([Block(cfg) for _ in range(cfg.n_layers)])
        self.norm = RMSNorm(cfg.d_model, cfg.norm_eps)
        self.lm_head = nn.Linear(cfg.d_model, cfg.vocab_size, bias=False)
        if cfg.tie_embeddings:
            self.lm_head.weight = self.tok_emb.weight

        cos, sin = precompute_rope(cfg.d_model // cfg.n_heads, cfg.seq_len, cfg.rope_theta)
        self.register_buffer("cos", cos, persistent=False)
        self.register_buffer("sin", sin, persistent=False)

        self.apply(self._init_weights)
        # GPT-2 style: scale residual projections by 1/sqrt(2*n_layers)
        scale = 0.02 / math.sqrt(2 * cfg.n_layers)
        for n, p in self.named_parameters():
            if n.endswith("attn.o.weight") or n.endswith("ffn.w2.weight"):
                nn.init.normal_(p, mean=0.0, std=scale)

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

    def num_params(self, exclude_embedding: bool = False) -> int:
        n = sum(p.numel() for p in self.parameters())
        if exclude_embedding:
            n -= self.tok_emb.weight.numel()
        return n

    def forward(self, idx: torch.Tensor, targets: torch.Tensor | None = None):
        B, T = idx.shape
        x = self.tok_emb(idx)
        cos, sin = self.cos, self.sin
        for block in self.blocks:
            x = block(x, cos, sin)
        x = self.norm(x)
        logits = self.lm_head(x)
        loss = None
        if targets is not None:
            loss = F.cross_entropy(
                logits.view(-1, logits.size(-1)).float(),
                targets.view(-1),
                ignore_index=-1,
            )
        return logits, loss

    @torch.no_grad()
    def generate(self, idx, max_new_tokens, temperature=1.0, top_k=None):
        for _ in range(max_new_tokens):
            idx_cond = idx[:, -self.cfg.seq_len:]
            logits, _ = self(idx_cond)
            logits = logits[:, -1, :] / max(temperature, 1e-5)
            if top_k is not None:
                v, _ = torch.topk(logits, min(top_k, logits.size(-1)))
                logits[logits < v[:, [-1]]] = -float("inf")
            probs = F.softmax(logits, dim=-1)
            next_tok = torch.multinomial(probs, num_samples=1)
            idx = torch.cat([idx, next_tok], dim=1)
        return idx