def get_data_file_name(basename):
for fn in [
basename, basename + '.gz', 'data/' + basename,
'data/' + basename + '.gz',
utilmlab.get_proj_dir() + '/alg/gcit/ccle_experiments/data/' +
basename,
utilmlab.get_proj_dir() + '/alg/gcit/ccle_experiments/data/' +
basename + '.gz'
]:
if os.path.isfile(fn):
return fn
print(f'***')
print(f'*** Error: cannot find file {basename}, see README.md of how'
'to obtain the required data ***')
print(f'***')
assert 0
def impute_gain(x, odir):
python_exe = 'python3'
script = '{}/gain/gain.py'.format(utilmlab.get_proj_dir())
fn_i = '{}/xmissing.csv.gz'.format(odir)
fn_o = '{}/ximputed.csv'.format(odir)
x.to_csv(fn_i, compression='gzip', index=False)
if os.path.isfile(fn_o):
os.remove(fn_o)
utilmlab.exe_cmd(
logger, '{} {} -i {} -o {} --testall 1'.format(python_exe, script,
fn_i, fn_o))
return pd.read_csv(fn_o)
def init_arg():
parser = argparse.ArgumentParser()
parser.add_argument('--exe', help='python interpreter to use')
parser.add_argument('--it', default=5000, type=int)
parser.add_argument('--projdir')
parser.add_argument("-n", "--numexp", default=10, type=int)
return parser.parse_args()
if __name__ == '__main__':
args = init_arg()
numexp = args.numexp
proj_dir = utilmlab.get_proj_dir() \
if args.projdir is None else args.projdir
alg = 'asac'
if args.exe is not None:
python_exe = args.exe
else:
python_exe = 'python' if sys.version_info[0] < 3 else 'python3'
niter = args.it
version = 1
resdir = '{}/result/{}/v_{}/h_{}'.format(
proj_dir, alg, version,
os.environ['HOSTNAME'] if 'HOSTNAME' in os.environ else 'unknown')
fn_model = get_temp_filename()
verbose = args.verbose
niter = args.it
nstage = args.nstage
if args.exe is not None:
python_exe = args.exe
else:
python_exe = 'python' if sys.version_info[0] < 3 else 'python3'
version = 1
resdir = '{}/result/autoprognosis/v_{}/h_{}'.format(
utilmlab.get_proj_dir(), version,
os.environ['HOSTNAME'] if 'HOSTNAME' in os.environ else 'unknown')
utilmlab.ensure_dir(resdir)
logger = utilmlab.init_logger(resdir, 'log_test_autoprognosis.txt')
result_all_d = dict()
for dataset, nsample, sep in [('bc', 0, ','), ('cover', 5000, ';')]:
odir = '{}/misc/dataset_{}'.format(resdir, dataset)
fn_csv = set_filenames(odir)
proj_dir = utilmlab.get_proj_dir() if args.projdir is None \
model.nmax_model = nmax_model
fn_i = args.i
label = args.target
p_miss = args.pmiss
logger.info('autoprognosis version'.format(version))
logger.info('ds:{} stages:{} n:{} metric:{} pmiss:{} it:{} '
'model:{} aqt:{}'.format(
ds, nmax_model, nsample,
metric, p_miss, niter, fn_model,
acquisition_type))
logger.info('{}'.format(args))
if ds is not None and os.path.isfile(
'{}/util/datasets.py'.format(
utilmlab.get_proj_dir())):
import datasets
X_, Y_ = datasets.load_dataset(ds)
else:
assert fn_i is not None
assert label is not None
logger.info('loading {} lbl:{} sep:{}'.format(
fn_i, label, sep))
df = pd.read_csv(fn_i, sep=sep)
features = list(df.columns)
assert label in features
features.remove(label)
assert len(df)
assert len(features) > 1
X_ = df[features]
Y_ = df[label]
parser.add_argument('--onlynotebook', type=int, default=0)
return parser.parse_args()
if __name__ == '__main__':
args = init_arg()
is_only_notebook = args.onlynotebook
if args.exe is not None:
python_exe = args.exe
else:
python_exe = 'python' if sys.version_info[0] < 3 else 'python3'
proj_dir = utilmlab.get_proj_dir() \
if args.projdir is None else args.projdir
logger = utilmlab.init_logger(
'.',
'log_test_alg_{}.txt'.format(utilmlab.get_hostname()))
if not is_only_notebook:
# execute all unit tests
f_lst = utilmlab.find_file_dir(
'{}/alg'.format(proj_dir),
'test_*.py')
logger.info('Unit tests found:{}'.format(f_lst))