import os
import gym
import torch
import pprint
import argparse
import numpy as np
from torch.utils.tensorboard import SummaryWriter

from tianshou.env import VectorEnv
from tianshou.policy import DQNPolicy
from tianshou.utils.net.common import Net
from tianshou.trainer import offpolicy_trainer
from tianshou.data import Collector, ReplayBuffer, PrioritizedReplayBuffer


def get_args():
    parser = argparse.ArgumentParser()
    parser.add_argument('--task', type=str, default='CartPole-v0')
    parser.add_argument('--seed', type=int, default=1626)
    parser.add_argument('--eps-test', type=float, default=0.05)
    parser.add_argument('--eps-train', type=float, default=0.1)
    parser.add_argument('--buffer-size', type=int, default=20000)
    parser.add_argument('--lr', type=float, default=1e-3)
    parser.add_argument('--gamma', type=float, default=0.9)
    parser.add_argument('--n-step', type=int, default=3)
    parser.add_argument('--target-update-freq', type=int, default=320)
    parser.add_argument('--epoch', type=int, default=10)
    parser.add_argument('--step-per-epoch', type=int, default=1000)
    parser.add_argument('--collect-per-step', type=int, default=10)
    parser.add_argument('--batch-size', type=int, default=64)
    parser.add_argument('--layer-num', type=int, default=3)
    parser.add_argument('--training-num', type=int, default=8)
    parser.add_argument('--test-num', type=int, default=100)
    parser.add_argument('--logdir', type=str, default='log')
    parser.add_argument('--render', type=float, default=0.)
    parser.add_argument('--prioritized-replay', type=int, default=0)
    parser.add_argument('--alpha', type=float, default=0.6)
    parser.add_argument('--beta', type=float, default=0.4)
    parser.add_argument(
        '--device', type=str,
        default='cuda' if torch.cuda.is_available() else 'cpu')
    args = parser.parse_known_args()[0]
    return args


def test_dqn(args=get_args()):
    env = gym.make(args.task)
    args.state_shape = env.observation_space.shape or env.observation_space.n
    args.action_shape = env.action_space.shape or env.action_space.n
    # train_envs = gym.make(args.task)
    # you can also use tianshou.env.SubprocVectorEnv
    train_envs = VectorEnv(
        [lambda: gym.make(args.task) for _ in range(args.training_num)])
    # test_envs = gym.make(args.task)
    test_envs = VectorEnv(
        [lambda: gym.make(args.task) for _ in range(args.test_num)])
    # seed
    np.random.seed(args.seed)
    torch.manual_seed(args.seed)
    train_envs.seed(args.seed)
    test_envs.seed(args.seed)
    # model
    net = Net(args.layer_num, args.state_shape,
              args.action_shape, args.device,  # dueling=(1, 1)
              ).to(args.device)
    optim = torch.optim.Adam(net.parameters(), lr=args.lr)
    policy = DQNPolicy(
        net, optim, args.gamma, args.n_step,
        target_update_freq=args.target_update_freq)
    # buffer
    if args.prioritized_replay > 0:
        buf = PrioritizedReplayBuffer(
            args.buffer_size, alpha=args.alpha, beta=args.beta)
    else:
        buf = ReplayBuffer(args.buffer_size)
    # collector
    train_collector = Collector(policy, train_envs, buf)
    test_collector = Collector(policy, test_envs)
    # policy.set_eps(1)
    train_collector.collect(n_step=args.batch_size)
    # log
    log_path = os.path.join(args.logdir, args.task, 'dqn')
    writer = SummaryWriter(log_path)

    def save_fn(policy):
        torch.save(policy.state_dict(), os.path.join(log_path, 'policy.pth'))

    def stop_fn(x):
        return x >= env.spec.reward_threshold

    def train_fn(x):
        # eps annnealing, just a demo
        if x <= int(0.1 * args.epoch):
            policy.set_eps(args.eps_train)
        elif x <= int(0.5 * args.epoch):
            eps = args.eps_train - (x - 0.1 * args.epoch) / \
                (0.4 * args.epoch) * (0.9 * args.eps_train)
            policy.set_eps(eps)
        else:
            policy.set_eps(0.1 * args.eps_train)

    def test_fn(x):
        policy.set_eps(args.eps_test)

    # trainer
    result = offpolicy_trainer(
        policy, train_collector, test_collector, args.epoch,
        args.step_per_epoch, args.collect_per_step, args.test_num,
        args.batch_size, train_fn=train_fn, test_fn=test_fn,
        stop_fn=stop_fn, save_fn=save_fn, writer=writer)

    assert stop_fn(result['best_reward'])
    train_collector.close()
    test_collector.close()
    if __name__ == '__main__':
        pprint.pprint(result)
        # Let's watch its performance!
        env = gym.make(args.task)
        collector = Collector(policy, env)
        result = collector.collect(n_episode=1, render=args.render)
        print(f'Final reward: {result["rew"]}, length: {result["len"]}')
        collector.close()


def test_pdqn(args=get_args()):
    args.prioritized_replay = 1
    args.gamma = .95
    test_dqn(args)


if __name__ == '__main__':
    test_dqn(get_args())