import torch
from torch.optim.optimizer import Optimizer

from .types import OptFloat, OptLossClosure, Params


class SGDLRD(Optimizer):
    r"""Implements stochastic gradient descent (optionally with momentum).

    Nesterov momentum is based on the formula from
    `On the importance of initialization and momentum in deep learning`__.

    Args:
        params (iterable): iterable of parameters to optimize or dicts defining
            parameter groups
        lr (float): learning rate
        momentum (float, optional): momentum factor (default: 0)
        weight_decay (float, optional): weight decay (L2 penalty) (default: 0)
        dampening (float, optional): dampening for momentum (default: 0)

    Example:
        >>> optimizer = SGD(model.parameters(), lr=0.1, momentum=0.9)
        >>> optimizer.zero_grad()
        >>> loss_fn(model(input), target).backward()
        >>> optimizer.step()

    __ http://www.cs.toronto.edu/%7Ehinton/absps/momentum.pdf
    """

    def __init__(
        self,
        params: Params,
        lr: float = 1e-2,
        momentum: float = 0,
        dampening: float = 0,
        weight_decay: float = 0,
        dropout: float = 0.0,
    ) -> None:
        if lr <= 0.0:
            raise ValueError('Invalid learning rate: {}'.format(lr))
        if momentum < 0.0:
            raise ValueError('Invalid momentum value: {}'.format(momentum))
        if dampening < 0.0:
            raise ValueError('Invalid dampening value: {}'.format(dampening))
        if weight_decay < 0:
            raise ValueError(
                'Invalid weight_decay value: {}'.format(weight_decay)
            )

        defaults = dict(
            lr=lr,
            momentum=momentum,
            dampening=dampening,
            weight_decay=weight_decay,
            dropout=dropout,
        )
        super(SGDLRD, self).__init__(params, defaults)

    def step(self, closure: OptLossClosure = None) -> OptFloat:
        r"""Performs a single optimization step.

        Arguments:
            closure: A closure that reevaluates the model and returns the loss.
        """
        loss = None
        if closure is not None:
            loss = closure()

        for group in self.param_groups:
            weight_decay = group['weight_decay']
            momentum = group['momentum']
            dampening = group['dampening']

            for p in group['params']:
                if p.grad is None:
                    continue
                d_p = p.grad.data

                # mask
                m = torch.ones_like(p.data) * group['dropout']
                mask = torch.bernoulli(m)

                if weight_decay != 0:
                    d_p.add_(weight_decay, p.data)
                if momentum != 0:
                    param_state = self.state[p]
                    if 'momentum_buffer' not in param_state:
                        buf = param_state['momentum_buffer'] = torch.clone(
                            d_p
                        ).detach()
                    else:
                        buf = param_state['momentum_buffer']
                        buf.mul_(momentum).add_(1 - dampening, d_p)

                # dropout learning rate
                lr_dropout = group['lr'] * mask
                I_buf = lr_dropout * buf.clone()

                p.data.add_(-1, I_buf)

        return loss