import torch
import torch.nn as nn
import torch.nn.functional as F
import mmap
import random

device = torch.device("cuda" if torch.cuda.is_available() else "cpu")

# Hyperparameters
batch_size = 64
block_size = 256
max_iters = 200
learning_rate = 2e-5
eval_iters = 100
num_embed = 384  # Ensure consistency in naming
num_heads = 8
num_layers = 8
dropout = 0.2

chars = ""
with open("vocab.txt", "r", encoding="utf-8") as f:
    text = f.read()
    chars = sorted(list(set(text)))
vocab_size = len(chars)

string_to_int = {ch: i for i, ch in enumerate(chars)}
int_to_string = {i: ch for i, ch in enumerate(chars)}


def encode(s):
    return [string_to_int[c] for c in s]


def decode(lst):
    return "".join([int_to_string[i] for i in lst])


def get_random_chunk(split):
    filename = "train_split.txt" if split == "train" else "eval_split.txt"
    with open(filename, "rb") as f:
        with mmap.mmap(f.fileno(), length=0, access=mmap.ACCESS_READ) as mm:
            file_size = len(mm)
            start = random.randint(0, file_size - block_size * batch_size)
            mm.seek(start)
            block = mm.read(block_size * batch_size - 1)
            decoded_block = block.decode("utf-8", errors="ignore").replace(
                "\r", ""
            )
            data = torch.tensor(encode(decoded_block), dtype=torch.long)
            return data


def get_batch(split):
    data = get_random_chunk(split)
    ix = torch.randint(len(data) - block_size, (batch_size,))
    x = torch.stack([data[i : i + block_size] for i in ix])
    y = torch.stack([data[i + 1 : i + block_size + 1] for i in ix])
    x, y = x.to(device), y.to(device)
    return x, y


class Head(nn.Module):
    def __init__(self, head_size):
        super().__init__()
        self.key = nn.Linear(num_embed, head_size)
        self.query = nn.Linear(num_embed, head_size)
        self.value = nn.Linear(num_embed, head_size)
        self.register_buffer(
            "tril", torch.tril(torch.ones(block_size, block_size))
        )
        self.dropout = nn.Dropout(dropout)

    def forward(self, x):
        B, T, C = x.shape
        k = self.key(x)
        q = self.query(x)
        wei = q @ k.transpose(-2, -1) * C**-0.5
        wei = wei.masked_fill(self.tril[:T, :T] == 0, float("-inf"))
        wei = F.softmax(wei, dim=-1)
        wei = self.dropout(wei)
        v = self.value(x)
        out = wei @ v
        return out


class MultiHeadAttention(nn.Module):
    def __init__(self, num_heads, head_size):
        super().__init__()
        self.heads = nn.ModuleList([Head(head_size) for _ in range(num_heads)])
        self.proj = nn.Linear(num_embed, num_embed)
        self.dropout = nn.Dropout(dropout)

    def forward(self, x):
        out = torch.cat([h(x) for h in self.heads], dim=-1)
        out = self.dropout(self.proj(out))
        return out


class FeedForward(nn.Module):
    def __init__(self, num_embed):
        super().__init__()
        self.net = nn.Sequential(
            nn.Linear(num_embed, 4 * num_embed),
            nn.ReLU(),
            nn.Linear(4 * num_embed, num_embed),
            nn.Dropout(dropout),
        )

    def forward(self, x):
        return self.net(x)


class Block(nn.Module):
    def __init__(self, num_embed, num_head):
        super().__init__()
        head_size = num_embed // num_head
        self.sa = MultiHeadAttention(num_head, head_size)
        self.ff = FeedForward(num_embed)
        self.ln1 = nn.LayerNorm(num_embed)
        self.ln2 = nn.LayerNorm(num_embed)

    def forward(self, x):
        y = self.sa(x)
        x = self.ln1(x + y)
        y = self.ff(x)
        x = self.ln2(x + y)
        return x


class GPT(nn.Module):
    def __init__(self):
        super().__init__()
        self.token_embedding_table = nn.Embedding(vocab_size, num_embed)
        self.position_embedding_table = nn.Embedding(block_size, num_embed)
        self.blocks = nn.Sequential(
            *[Block(num_embed, num_heads) for _ in range(num_layers)]
        )
        self.ln = nn.LayerNorm(num_embed)
        self.lm_head = nn.Linear(num_embed, vocab_size)
        self.apply(self._init_weights)

    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):
        B, T = idx.shape
        tok_emb = self.token_embedding_table(idx)
        pos_emb = self.position_embedding_table(
            torch.arange(T, device=idx.device)
        )
        x = tok_emb + pos_emb
        x = self.blocks(x)
        x = self.ln(x)
        logits = self.lm_head(x)

        if targets is None:
            loss = None
        else:
            B, T, C = logits.shape
            logits = logits.view(B * T, C)
            targets = targets.view(B * T)
            loss = F.cross_entropy(logits, targets)
        return logits, loss

    def generate(self, idx, max_new_tokens):
        for _ in range(max_new_tokens):
            idx_cond = idx[:, -block_size:]
            logits, loss = self(idx_cond)
            logits = logits[:, -1, :]
            probs = F.softmax(logits, dim=-1)
            idx_next = torch.multinomial(probs, num_samples=1)
            idx = torch.cat((idx, idx_next), dim=1)
        return idx


model = GPT().to(device)


@torch.no_grad()
def estimate_loss():
    out = {}
    model.eval()
    for split in ["train", "val"]:
        losses = torch.zeros(eval_iters)
        for k in range(eval_iters):
            X, Y = get_batch(split)
            logits, loss = model(X, Y)
            losses[k] = loss.item()
        out[split] = losses.mean().item()
    model.train()
    return out


optimizer = torch.optim.AdamW(model.parameters(), lr=learning_rate)
for iter in range(max_iters):
    if iter % eval_iters == 0:
        losses = estimate_loss()
        print(
            f"step {iter}: train loss {losses['train']:.3f}, "
            f"val loss {losses['val']:.3f}"
        )
    xb, yb = get_batch("train")
    logits, loss = model(xb, yb)
    optimizer.zero_grad(set_to_none=True)
    loss.backward()
    optimizer.step()

print(loss.item())

torch.save(model.state_dict(), "phoebe_model.pt")
print("Model Saved!")