from pathlib import Path from datetime import datetime, timezone, timedelta import click from tqdm import tqdm import numpy as np import pandas as pd from service.App import * from common.classifiers import * from common.feature_generation import * from common.model_store import * """ Train models for all target labels and all algorithms declared in the configuration using the specified features. """ # # Parameters # class P: in_nrows = 100_000_000 # For debugging tail_rows = 0 # How many last rows to select (for debugging) # Whether to store file with predictions store_predictions = True @click.command() @click.option('--config_file', '-c', type=click.Path(), default='', help='Configuration file name') def main(config_file): load_config(config_file) time_column = App.config["time_column"] now = datetime.now() # # Load feature matrix # symbol = App.config["symbol"] data_path = Path(App.config["data_folder"]) / symbol file_path = (data_path / App.config.get("matrix_file_name")).with_suffix(".csv") if not file_path.is_file(): print(f"ERROR: Input file does not exist: {file_path}") return print(f"Loading data from source data file {file_path}...") df = pd.read_csv(file_path, parse_dates=[time_column], nrows=P.in_nrows) print(f"Finished loading {len(df)} records with {len(df.columns)} columns.") df = df.iloc[-P.tail_rows:] df = df.reset_index(drop=True) # # Prepare data by selecting columns and rows # label_horizon = App.config["label_horizon"] # Labels are generated from future data and hence we might want to explicitly remove some tail rows train_length = App.config.get("train_length") train_features = App.config.get("train_features") labels = App.config["labels"] algorithms = App.config.get("algorithms") # Select necessary features and label out_columns = ['timestamp', 'open', 'high', 'low', 'close', 'volume', 'close_time'] out_columns = [x for x in out_columns if x in df.columns] all_features = train_features + labels df = df[out_columns + all_features] for label in labels: # "category" NN does not work without this (note that we assume a classification task here) df[label] = df[label].astype(int) # Spot and futures have different available histories. If we drop nans in all of them, then we get a very short data frame (corresponding to futureus which have little data) # So we do not drop data here but rather when we select necessary input features # Nans result in constant accuracy and nan loss. MissingValues procedure does not work and produces exceptions pd.set_option('use_inf_as_na', True) #in_df = in_df.dropna(subset=labels) df = df.reset_index(drop=True) # We must reset index after removing rows to remove gaps # Remove the tail data for which no labels are available # The reason is that these labels are computed from future which is not available if label_horizon: df = df.head(-label_horizon) # Limit maximum length train_df = df.tail(train_length) train_df = train_df.dropna(subset=train_features) if len(train_df) == 0: print(f"ERROR: Empty data set after removing NULLs in feature columns. Some features might have all NULL values.") #print(train_df.isnull().sum().sort_values(ascending=False)) return models = dict() scores = dict() out_df = pd.DataFrame() # Collect predictions for label in tqdm(labels, desc="LABELS", colour='red', position=0): for algo_name in tqdm(algorithms, desc="ALGORITHMS", colour='red', leave=False, position=1): model_config = get_model(algo_name) # Get algorithm description from the algo store algo_type = model_config.get("algo") algo_train_length = model_config.get("train", {}).get("length") score_column_name = label + label_algo_separator + algo_name # Limit length according to the algorith parameters if algo_train_length and algo_train_length < train_length: train_df_2 = train_df.iloc[-algo_train_length:] else: train_df_2 = train_df df_X = train_df_2[train_features] df_y = train_df_2[label] print(f"Train '{score_column_name}'. Train length {len(df_X)}. Train columns {len(df_X.columns)}. Algorithm {algo_name}") if algo_type == "gb": model_pair = train_gb(df_X, df_y, model_config) models[score_column_name] = model_pair df_y_hat = predict_gb(model_pair, df_X, model_config) elif algo_type == "nn": model_pair = train_nn(df_X, df_y, model_config) models[score_column_name] = model_pair df_y_hat = predict_nn(model_pair, df_X, model_config) elif algo_type == "lc": model_pair = train_lc(df_X, df_y, model_config) models[score_column_name] = model_pair df_y_hat = predict_lc(model_pair, df_X, model_config) elif algo_type == "svc": model_pair = train_svc(df_X, df_y, model_config) models[score_column_name] = model_pair df_y_hat = predict_svc(model_pair, df_X, model_config) else: print(f"ERROR: Unknown algorithm type {algo_type}. Check algorithm list.") return scores[score_column_name] = compute_scores(df_y, df_y_hat) out_df[score_column_name] = df_y_hat # # Store all collected models in files # model_path = data_path / "MODELS" model_path.mkdir(parents=True, exist_ok=True) # Ensure that folder exists for score_column_name, model_pair in models.items(): save_model_pair(model_path, score_column_name, model_pair) print(f"Models stored in path: {model_path.absolute()}") # # Store scores # lines = list() for score_column_name, score in scores.items(): line = score_column_name + ", " + str(score) lines.append(line) metrics_file_name = f"prediction-metrics.txt" metrics_path = (data_path / metrics_file_name).resolve() with open(metrics_path, 'a+') as f: f.write("\n".join(lines) + "\n\n") print(f"Metrics stored in path: {metrics_path.absolute()}") # # Store predictions if necessary # if P.store_predictions: # Store only selected original data, labels, and their predictions out_df = out_df.join(df[out_columns + labels]) out_path = data_path / App.config.get("predict_file_name") print(f"Storing output file...") out_df.to_csv(out_path.with_suffix(".csv"), index=False, float_format='%.4f') print(f"Predictions stored in file: {out_path}. Length: {len(out_df)}. Columns: {len(out_df.columns)}") # # End # elapsed = datetime.now() - now print(f"Finished training models in {str(elapsed).split('.')[0]}") if __name__ == '__main__': main()