import torch
from typing import List, Optional

from transformers.modeling_utils import PreTrainedModel
from transformers.generation.utils import LogitsProcessorList
from transformers.generation.logits_process import LogitsProcessor

from llmtuner.extras.constants import LAYERNORM_NAMES


class AverageMeter:
    r"""
    Computes and stores the average and current value.
    """
    def __init__(self):
        self.reset()

    def reset(self):
        self.val = 0
        self.avg = 0
        self.sum = 0
        self.count = 0

    def update(self, val, n=1):
        self.val = val
        self.sum += val * n
        self.count += n
        self.avg = self.sum / self.count


# Avoids runtime error in model.generate(do_sample=True).
class InvalidScoreLogitsProcessor(LogitsProcessor):

    def __call__(self, input_ids: torch.LongTensor, scores: torch.FloatTensor) -> torch.FloatTensor:
        if torch.isnan(scores).any() or torch.isinf(scores).any():
            scores.zero_()
            scores[..., 0] = 1.0
        return scores


def get_logits_processor() -> LogitsProcessorList:
    logits_processor = LogitsProcessorList()
    logits_processor.append(InvalidScoreLogitsProcessor())
    return logits_processor


def print_trainable_params(model: torch.nn.Module) -> None:
    trainable_params, all_param = 0, 0
    for param in model.parameters():
        num_params = param.numel()
        # if using DS Zero 3 and the weights are initialized empty
        if num_params == 0 and hasattr(param, "ds_numel"):
            num_params = param.ds_numel
        all_param += num_params
        if param.requires_grad:
            trainable_params += num_params
    print("trainable params: {:d} || all params: {:d} || trainable%: {:.4f}".format(
                trainable_params, all_param, 100 * trainable_params / all_param))


# Includes: (1) cast the layernorm in fp32 (2) make output embedding layer require grads (3) upcast the lm_head to fp32
# Inspired by: https://github.com/huggingface/peft/blob/c0209c35abbf88c63aa267800d98a8e212ed0a42/src/peft/utils/other.py#L35
def prepare_model_for_training(
    model: PreTrainedModel,
    finetuning_type: str,
    output_layer_name: Optional[str] = "lm_head",
    use_gradient_checkpointing: Optional[bool] = True,
    layer_norm_names: Optional[List[str]] = LAYERNORM_NAMES
) -> PreTrainedModel:

    for name, param in model.named_parameters():
        if param.ndim == 1 and any(layer_norm_name in name for layer_norm_name in layer_norm_names):
            param.data = param.data.to(torch.float32)

    if use_gradient_checkpointing:
        if hasattr(model, "enable_input_require_grads"):
            model.enable_input_require_grads()
        else:
            def make_inputs_require_grad(module, input, output):
                output.requires_grad_(True)
            model.get_input_embeddings().register_forward_hook(make_inputs_require_grad)

        model.gradient_checkpointing_enable()
        model.config.use_cache = False # turn off when gradient checkpointing is enabled

    if finetuning_type != "full" and hasattr(model, output_layer_name):
        output_layer: torch.nn.Linear = getattr(model, output_layer_name)
        input_dtype = output_layer.weight.dtype

        class CastOutputToFloat(torch.nn.Sequential):

            def forward(self, x: torch.Tensor) -> torch.Tensor:
                return super().forward(x.to(input_dtype)).to(torch.float32)

        new_output_layer = CastOutputToFloat(output_layer)
        # adapt to LLaMA-2's pretraining_tp (actually LLaMA models can automatically do casting but BLOOM models cannot)
        # (https://github.com/huggingface/transformers/blob/v4.31.0/src/transformers/models/llama/modeling_llama.py#L819)
        setattr(new_output_layer, "weight", output_layer.weight)
        setattr(model, output_layer_name, new_output_layer)
    return model


def torch_gc() -> None:
    r"""
    Collects GPU memory.
    """
    if torch.cuda.is_available():
        torch.cuda.empty_cache()
        torch.cuda.ipc_collect()