from math import ceil
from typing import Optional
import logging
from argparse import ArgumentParser
import sys
import os

class Config:
    def __init__(self):
        self.__logger: Optional[logging.Logger] = None
        self.set_defaults()

    def set_defaults(self):
        self.NUM_TRAIN_EPOCHS = 100
        self.SAVE_EVERY_EPOCHS = 5
        self.TRAIN_BATCH_SIZE = 64

        # model hyper-params
        self.categories = 10
        # self.learning_rate=0.001
        # self.decay_rate=0.9

        self.path_vocab_size = 27500
        self.token_vocab_size = 1500
        self.MAX_CONTEXTS = 200
        self.MAX_TOKEN_VOCAB_SIZE = 1301136
        self.MAX_PATH_VOCAB_SIZE = 911417
        self.DEFAULT_EMBEDDINGS_SIZE = 64
        self.TOKEN_EMBEDDINGS_SIZE = self.DEFAULT_EMBEDDINGS_SIZE
        self.PATH_EMBEDDINGS_SIZE = self.DEFAULT_EMBEDDINGS_SIZE
        self.CODE_VECTOR_SIZE = self.context_vector_size
        self.TARGET_EMBEDDINGS_SIZE = self.CODE_VECTOR_SIZE
        self.DROPOUT_KEEP_RATE = 0.5

    @property
    def context_vector_size(self) -> int:
        # The context vector is actually a concatenation of the embedded
        # source & target vectors and the embedded path vector.
        return self.PATH_EMBEDDINGS_SIZE + 2 * self.TOKEN_EMBEDDINGS_SIZE