def _train_val_dataset_from_data_path(project_parameters):
data, label = [], []
for stage in ['train', 'val']:
for c in project_parameters.classes:
files = get_files(filepath=join(
project_parameters.data_path, '{}/{}'.format(stage, c)), file_type=['wav'])
data += sorted(files)
label += [project_parameters.class_to_idx[c]]*len(files)
return {'data': np.array(data), 'label': np.array(label)}
from src.utils import get_files, get_platform_selector, read_json_file
from os import path
import pandas as pd
dirname = path.dirname(__file__)
results_dir = path.join(dirname, "../tests/network/output")
output_dir = path.join(dirname, "extracted")
bandwidth_file = "bandwidth_results.csv"
jitter_file = "jitter_results.csv"
results = [f for f in get_files(results_dir) if 'server' not in f]
b_results = [f for f in results if 'jitter' not in f]
j_results = [f for f in results if 'jitter' in f]
df = pd.DataFrame(columns=['platform', 'sent_bps', 'received_bps', 'cpu_client', 'cpu_server'])
for offset, f in enumerate(b_results):
selector = get_platform_selector(f)
iterations = read_json_file(path.join(results_dir, f))
for i, iteration in enumerate(iterations):
end = iteration['end']
sent_bps = end['sum_sent']['bits_per_second']
received_bps = end['sum_received']['bits_per_second']
cpu_local = end['cpu_utilization_percent']['host_total']
cpu_remote = end['cpu_utilization_percent']['remote_total']
index = offset * 10 + i
# Store from the platform perspecvive, hence column "sent_bps" means bits sent by the server, not client
df.loc[index] = [selector, received_bps, sent_bps, cpu_local, cpu_remote]
df.to_csv(path.join(output_dir, bandwidth_file), index=False)
from src.utils import get_files, get_platform_selector, read_file_per_line
from os import path
import re
import pandas as pd
dirname = path.dirname(__file__)
results_dir = path.join(dirname, "../tests/case-app/http-benchmark/output")
output_dir = path.join(dirname, "extracted")
output_file = "http_benchmark_results.csv"
results = get_files(results_dir)
df = pd.DataFrame(columns=[
'platform', 'concurrency', 'duration', 'completed_req', 'failed_req',
'bytes_transferred', 'req_per_sec', 'mean_time_per_req'
])
concurrency = 0
duration = 0
completed_req = 0
failed_req = 0
bytes_transferred = 0
req_per_sec = 0
mean_time_per_req = 0
def clear_row():
global concurrency, duration, completed_req, failed_req, bytes_transferred, req_per_sec, mean_time_per_req
concurrency = 0
duration = 0
completed_req = 0
types_datasets_dict = pickle.load(open(parser.columns, "rb"))
datasets_types_dict = inverse_type_dict(types_datasets_dict)
DATA_TYPES = sorted(list(types_datasets_dict.keys()))
else:
types_datasets_dict = None
if parser.detective:
detective_df = pd.read_csv(parser.detective, sep="\t")
else:
detective_df = None
with_excel = parser.excel
n_jobs = int(parser.cores)
files = get_files(files_path)
if n_jobs == 1:
list_dfs = column_analysis_single(
files,
detective_df=detective_df,
datasets_types_dict=datasets_types_dict)
else:
list_dfs = column_analysis(files,
n_jobs=n_jobs,
detective_df=detective_df,
datasets_types_dict=datasets_types_dict)
pass
final_df = pd.concat(list_dfs)