import torch
import torch.nn as nn
from transformers import PreTrainedModel, GenerationMixin
from transformers.modeling_outputs import CausalLMOutput
from .configuration_chess import ChessConfig 

class Block(nn.Module):
    def __init__(self, config):
        super().__init__()
        self.ln1 = nn.LayerNorm(config.n_embd)
        self.attn = nn.MultiheadAttention(config.n_embd, config.n_head, batch_first=True)
        self.ln2 = nn.LayerNorm(config.n_embd)
        self.mlp = nn.Sequential(
            nn.Linear(config.n_embd, config.n_inner), 
            nn.GELU(), 
            nn.Linear(config.n_inner, config.n_embd)
        )
        
    def forward(self, x, mask=None):
        attn_out, _ = self.attn(self.ln1(x), self.ln1(x), self.ln1(x), attn_mask=mask, need_weights=False)
        return x + attn_out + self.mlp(self.ln2(x + attn_out))

class ChessForCausalLM(PreTrainedModel, GenerationMixin):
    config_class = ChessConfig
    
    def __init__(self, config):
        super().__init__(config)
        self.config = config
        self.token_emb = nn.Embedding(config.vocab_size, config.n_embd)
        self.pos_emb = nn.Embedding(config.n_ctx, config.n_embd)
        self.blocks = nn.ModuleList([Block(config) for _ in range(config.n_layer)])
        self.ln_f = nn.LayerNorm(config.n_embd)
        self.lm_head = nn.Linear(config.n_embd, config.vocab_size, bias=False)
        
        # Weight Tying (Important pour le comptage param)
        self.lm_head.weight = self.token_emb.weight
        
        self.apply(self._init_weights)
            
    def _init_weights(self, module):
        if isinstance(module, (nn.Linear, nn.Embedding)): 
            torch.nn.init.normal_(module.weight, mean=0.0, std=0.02)

    def prepare_inputs_for_generation(self, input_ids, **kwargs):
        return {"input_ids": input_ids}
            
    def forward(self, input_ids, attention_mask=None, labels=None, **kwargs):
        B, T = input_ids.shape
        x = self.token_emb(input_ids) + self.pos_emb(torch.arange(T, device=input_ids.device))
        mask = torch.triu(torch.ones(T, T, device=input_ids.device) * float('-inf'), diagonal=1)
        for block in self.blocks: x = block(x, mask=mask)
        logits = self.lm_head(self.ln_f(x))
        loss = None
        if labels is not None:
            loss = nn.CrossEntropyLoss(ignore_index=-100)(logits.view(-1, logits.size(-1)), labels.view(-1))
        return CausalLMOutput(loss=loss, logits=logits)