| 4 | |
| 5 | |
| 6 | class DataUtils: |
| 7 | def __init__(self, ticker, dataset, experiment_id, horizons, normalization_window): |
| 8 | self.ticker = ticker # Ticker of the stock to be processed. |
| 9 | self.dataset = dataset # Dataset to be used. |
| 10 | self.experiment_id = experiment_id # Experiment ID. |
| 11 | self.horizons = horizons # Horizons to be used when computing labels. |
| 12 | self.normalization_window = normalization_window # Normalization window to be used when normalizing data. |
| 13 | |
| 14 | self.__raw_data_path = None # Path containing the raw LOB data. |
| 15 | self.__processed_data_path_unscaled_data = ( |
| 16 | None # Path containing the processed but unscaled LOB data. |
| 17 | ) |
| 18 | self.__processed_data_path_scaled_data = ( |
| 19 | None # Path containing the processed and scaled LOB data. |
| 20 | ) |
| 21 | self.__logs_path = None # Path containing the logs of the data processing. |
| 22 | |
| 23 | def __set_raw_data_path(self): |
| 24 | # Set the raw data path according to the dataset. |
| 25 | if self.dataset == "nasdaq": |
| 26 | self.__raw_data_path = f"./data/{self.dataset}/raw/{self.ticker}" |
| 27 | |
| 28 | def __set_processed_data_path_unscaled_data(self): |
| 29 | # Set the path containing the processed but unscaled LOB data according to the dataset. |
| 30 | if self.dataset == "nasdaq": |
| 31 | self.__processed_data_path_unscaled_data = ( |
| 32 | f"./data/{self.dataset}/unscaled_data/{self.ticker}" |
| 33 | ) |
| 34 | |
| 35 | def __set_processed_data_path_scaled_data(self): |
| 36 | # Set the path containing the processed and scaled LOB data according to the dataset. |
| 37 | if self.dataset == "nasdaq": |
| 38 | self.__processed_data_path_scaled_data = ( |
| 39 | f"./data/{self.dataset}/scaled_data/{self.ticker}" |
| 40 | ) |
| 41 | |
| 42 | def __set_logs_path(self): |
| 43 | # Set the path containing the logs of the data processing according to the experiment ID. |
| 44 | self.__logs_path = ( |
| 45 | f"./loggers/results/{self.experiment_id}/data_processing_logs" |
| 46 | ) |
| 47 | |
| 48 | def generate_data_folders(self): |
| 49 | self.__set_raw_data_path() # Set the raw data path. |
| 50 | self.__set_processed_data_path_unscaled_data() # Set the path containing the processed but unscaled LOB data. |
| 51 | self.__set_processed_data_path_scaled_data() # Set the path containing the processed and scaled LOB data. |
| 52 | self.__set_logs_path() # Set the path containing the logs of the data processing. |
| 53 | |
| 54 | # Create the folder for the processed but unscaled LOB data if it does not exist. |
| 55 | if not os.path.exists(self.__processed_data_path_unscaled_data): |
| 56 | os.makedirs(self.__processed_data_path_unscaled_data) |
| 57 | |
| 58 | # Create the folder for the processed and scaled LOB data if it does not exist. |
| 59 | if not os.path.exists(self.__processed_data_path_scaled_data): |
| 60 | os.makedirs(self.__processed_data_path_scaled_data) |
| 61 | |
| 62 | # Create the folder for the logs of the data processing if it does not exist. |
| 63 | if not os.path.exists(self.__logs_path): |
nothing calls this directly
no outgoing calls
no test coverage detected