def save_result(n_users, n_items, density, total_fit_time_list):
# Set up writing folder and file
results_path = os.path.join(get_project_root_path(), "report",
"slim_elastic_net_timing_tests")
fd, folder_path_with_date = handle_folder_creation(
result_path=results_path, filename="results.txt")
fd.write("--- SLIM Elastic Net Experiment ---\n")
fd.write(" - Number of experiments: {}\n".format(N_EXPERIMENTS))
fd.write("\n")
fd.write("DATASET INFO\n")
fd.write(" - Dataset name: RANDOMLY DISTRIBUTED SYNTHETIC DATASET\n")
fd.write(" - N_USERS: {}\n".format(n_users))
fd.write(" - N_ITEMS: {}\n".format(n_items))
fd.write(" - DENSITY: {}\n".format(density))
fd.write("\n")
fd.close()
# write data and timing info on csv file
columns = ['n_users', 'n_items', 'density', 'n_exp', 'total_fit_time']
data_info = pd.DataFrame(
[[n_users, n_items, density, N_EXPERIMENTS, total_fit_time_list]],
columns=columns)
data_info.to_csv(os.path.join(folder_path_with_date,
"experiment_info.csv"),
sep=',',
header=True,
index=False)
return folder_path_with_date
def __init__(self, filename, reload_from_original=False):
super().__init__(reload_from_original_data=reload_from_original)
root_path = os.path.join(get_project_root_path(), "data")
self.URM_path = os.path.join(root_path, filename)
self.DATASET_SUBFOLDER = "{}/".format(filename.split(".")[0])
self._LOADED_UCM_DICT = {}
self._LOADED_UCM_MAPPER_DICT = {}
DEFAULT_AGGREGATION = "FIRST"
DEFAULT_FILTER_SAMPLE_METHOD = "NONE"
# FIT DEFAULT VALUES
DEFAULT_TOP_K = 5
DEFAULT_ALPHA_MULTIPLIER = 0.0
DEFAULT_NUM_READS = 50
DEFAULT_MULTIPLIER = 1.0
DEFAULT_UNPOPULAR_THRESHOLD = 0
DEFAULT_QUBO_ROUND_PERCENTAGE = 1.0
DEFAULT_FILTER_ITEM_METHOD = "NONE"
DEFAULT_FILTER_ITEM_NUMBERS = 100
# OTHERS
DEFAULT_CUTOFF = 5
DEFAULT_OUTPUT_FOLDER = os.path.join(get_project_root_path(), "report",
"quantum_slim")
DEFAULT_RESPONSES_CSV_FILENAME = "solver_responses.csv"
DEFAULT_MAPPING_MATRIX_FILENAME = "mapping_matrix.npy"
def get_arguments():
"""
Defining the arguments available for the script
:return: argument parser
"""
parser = argparse.ArgumentParser()
# Data setting
parser.add_argument(
import os
from datetime import datetime
import pandas as pd
from src.utils.utilities import get_project_root_path
REPORT_ROOT_PATH = os.path.join(get_project_root_path(), "report",
"quantum_slim_timing_tests")
FINAL_REPORT_PATH = os.path.join(get_project_root_path(), "report",
"quantum_slim_final_report",
"quantum_slim_timing")
if __name__ == '__main__':
report_files = []
for root, dirs, files in os.walk(REPORT_ROOT_PATH):
for file in files:
if file.endswith(".csv"):
report_files.append(os.path.join(root, file))
report_files = sorted(report_files, key=lambda x: os.path.getmtime(x))
print(report_files)
df = pd.DataFrame()
for path_file in report_files:
df_file = pd.read_csv(path_file, sep=",")
df = df.append(df_file, ignore_index=True)
print(df.head())
date_string = datetime.now().strftime('%b%d_%H-%M-%S')
df.to_csv(os.path.join(FINAL_REPORT_PATH, "{}.csv".format(date_string)),
import os
import pandas as pd
import numpy as np
from src.utils.utilities import get_project_root_path
if __name__ == '__main__':
data_path = os.path.join(get_project_root_path(), "data", "jester100")
input_filepath_list = [
os.path.join(data_path, "jester-data-{}.xls".format(i + 1))
for i in range(3)
]
df = pd.DataFrame(columns=["user", "item", "rating"])
last_user_id = 0
for input_filepath in input_filepath_list:
dataset = pd.read_excel(input_filepath, header=None)
dataset = dataset.drop(columns=0)
mask_df = np.logical_not((dataset == 99) | (dataset <= 0))
counts = np.sum(mask_df.to_numpy(), axis=1)
data_matrix = dataset[mask_df].to_numpy()
index_matrix = np.repeat(np.arange(0, 100).reshape(1, -1),
repeats=dataset.shape[0],
axis=0)
items = index_matrix[mask_df.to_numpy()]
data = np.ones(shape=len(data_matrix[mask_df.to_numpy()]), dtype=int)
users = np.repeat(np.arange(last_user_id,
last_user_id + dataset.shape[0]),
import pandas as pd
import os
import numpy as np
from src.utils.utilities import get_project_root_path
if __name__ == '__main__':
#np.random.seed(1234)
data_path = os.path.join(get_project_root_path(), "data", "ml100k")
input_filepath = os.path.join(data_path, "ml100k.tsv")
df = pd.read_csv(filepath_or_buffer=input_filepath, sep="\t", header=None)
#df[2] = 1
df = df.drop(columns=[3])
output_filepath = os.path.join(data_path, "ml100k.csv")
df.to_csv(output_filepath, header=False, index=False)
"""
item_chosen_filepath = os.path.join(data_path, "{}_item_chosen.txt".format(prefix_name))
with open(item_chosen_filepath, 'w') as f:
for item in item_chosen:
f.write("%s\n" % item)
"""
qpu_times_dict['qpu_anneal_time_per_sample'],
qpu_times_dict['qpu_readout_time_per_sample'],
qpu_times_dict['qpu_programming_time'],
qpu_times_dict['qpu_delay_time_per_sample']
]],
columns=columns)
data_info.to_csv(os.path.join(output_folder, "experiment_info.csv"),
sep=',',
header=True,
index=False)
if __name__ == '__main__':
arguments = get_arguments()
dict_args = vars(arguments).copy()
dict_args["output_folder"] = os.path.join(get_project_root_path(),
"report",
"quantum_slim_timing_tests")
hyperparameters_combinations = list(
itertools.product(*HYPERPARAMETERS.values()))
for i, hyperparameters in enumerate(hyperparameters_combinations):
print("Experiment n.{}/{}: the hyperparameters are {}:{}".format(
i + 1, len(hyperparameters_combinations), HYPERPARAMETERS.keys(),
list(hyperparameters)))
n_users, n_items, density = hyperparameters[:3]
for j, key in enumerate(HYPERPARAMETERS.keys()):
if j <= 2:
continue
dict_args[key] = hyperparameters[j]