from pathlib import Path
from bianca.workflows.bianca import run_bianca
training_data_dir = Path("/home/fliem/lhab_collaboration/WMH/BIANCA/training_data")
base_dir = Path("/home/fliem/lhab_collaboration/WMH/BIANCA/training_subjects")
name = "bianca"
n_cpu = 32
####
# 2D
acq = "2D"
bianca_dir = base_dir / acq / name
wd_dir = base_dir / "_wd" / acq / name
crash_dir = base_dir / "_crash" / acq / name
run_bianca(bianca_dir, wd_dir, crash_dir, n_cpu=n_cpu, save_classifier=True)
####
# 3D
acq = "3D"
bianca_dir = base_dir / acq / name
wd_dir = base_dir / "_wd" / acq / name
crash_dir = base_dir / "_crash" / acq / name
run_bianca(bianca_dir, wd_dir, crash_dir, n_cpu=n_cpu, save_classifier=True)
"/home/fliem/lhab_collaboration/WMH/BIANCA/training_data")
base_dir = Path("/home/fliem/lhab_collaboration/WMH/BIANCA/full_sample")
name = "bianca"
n_cpu = 30
####
# 3D
acq = "3D"
bianca_dir = base_dir / acq / name
wd_dir = Path("/tmp/fl") / "_wd" / acq / name
crash_dir = Path("/tmp/fl") / "_crash" / acq / name
clf_file = bianca_dir / "clf" / f"acq-{acq}_run-1_FLAIR_classifier"
df = pd.read_csv(bianca_dir / "masterfile_wHeader.txt", sep=" ")
test_subs = (df.manual_mask == "XXX")
training_subs = ~test_subs
training_subs_idx = np.where(training_subs)[0]
test_subs_idx = np.where(test_subs)[0]
query_sub_idx = test_subs_idx
run_bianca(bianca_dir,
wd_dir,
crash_dir,
n_cpu=n_cpu,
save_classifier=False,
trained_classifier_file=clf_file,
training_subject_idx=training_subs_idx,
query_subject_idx=query_sub_idx)
acq = "3D"
bianca_dir = base_dir / acq / name
wd_dir = Path("/tmp/fl") / "_wd" / acq / name
crash_dir = Path("/tmp/fl") / "_crash" / acq / name
clf_dir = bianca_dir / "clf"
clf_dir.mkdir(exist_ok=True)
df = pd.read_csv(bianca_dir / "masterfile_wHeader.txt", sep=" ")
test_subs = (df.manual_mask == "XXX")
training_subs = ~test_subs
training_subs_idx = np.where(training_subs)[0]
test_subs_idx = np.where(test_subs)[0]
query_sub_idx = test_subs_idx[:1]
run_bianca(bianca_dir,
wd_dir,
crash_dir,
n_cpu=n_cpu,
save_classifier=True,
training_subject_idx=training_subs_idx,
query_subject_idx=query_sub_idx)
dd = df.iloc[query_sub_idx[0]]
in_dir = bianca_dir / f"sub-{dd.subject}" / f"ses-{dd.session}" / "anat"
in_files = list(in_dir.glob("sub*"))
for f in in_files:
o_name = clf_dir / (f.name.replace(f"sub-{dd.subject}_ses-{dd.session}_",
""))
copyfile(f, o_name)