import logging import os import pprint import pickle from pathlib import Path import numpy as np import zmq from data_class import opunit_data from mini_trainer import MiniTrainer from util import logging_util from type import OpUnit from info import data_info logging_util.init_logging('info') class Callback(IntEnum): """ ModelServerManager <==> ModelServer callback Id. Needs to be kept consistent with ModelServerManager.h's Callback Enum """ NOOP = 0 CONNECTED = 1 class Command(Enum): """ Command enum for actions to take from the manager. This has to be kept consistent with the C++ ModelServerManager.
help='OLTPBench warmup period') aparser.add_argument('--tpcc_hack', default=False, help='Should do feature correction for TPCC') aparser.add_argument('--ee_sample_interval', type=int, default=9, help='Sampling interval for the execution engine OUs') aparser.add_argument('--txn_sample_interval', type=int, default=0, help='Sampling interval for the transaction OUs') aparser.add_argument('--log', default='info', help='The logging level') args = aparser.parse_args() logging_util.init_logging(args.log) logging.info("Global trainer starts.") with open(args.mini_model_file, 'rb') as pickle_file: model_map = pickle.load(pickle_file) trainer = GlobalTrainer(args.input_path, args.model_results_path, args.ml_models, args.test_ratio, args.impact_model_ratio, model_map, args.warmup_period, args.tpcc_hack, args.ee_sample_interval, args.txn_sample_interval) resource_model, impact_model, direct_model = trainer.train() with open(args.save_path + '/global_resource_model.pickle', 'wb') as file: pickle.dump(resource_model, file) with open(args.save_path + '/global_impact_model.pickle', 'wb') as file: pickle.dump(impact_model, file)
from Constants import mem_per_1vCPU from TranslationService import translate, genitive_case from util.constants_util import get_logs_root_path, get_fn_tag from util.file_util import read_file, read_csv_to_dataframe, write_to_csv from util.subprocess_util import run_executable from util.logging_util import log, init_logging env_config = dotenv_values() REPOSITORY_NAME = env_config["AWS_ECR_REPOSITORY_NAME"] describe_image_exec_path = './script/aws/environment/describe-image.sh' results_header = ["image_size_mb", "memory_size_mb", "init_duration_ms", "artificial_init_duration", "duration", "available_threads"] init_logging() parser = argparse.ArgumentParser(description="Results Parser") parser.add_argument("-t", "--test", type=str) def main(): args = parser.parse_args() test_num = args.test parse_logs(test_num) def parse_logs(test_num): log(f'Parsing logs for test: {test_num}')