Update Mimic4Dataset.py
Browse files- Mimic4Dataset.py +26 -10
Mimic4Dataset.py
CHANGED
@@ -361,6 +361,11 @@ def generate_split_deep(path,task,feat_cond,feat_chart,feat_proc, feat_meds, fea
|
|
361 |
taskf=task.replace(" ","_")
|
362 |
for hid, data in tqdm(X.iterrows(),desc='Encoding Splits Data for '+task+' task'):
|
363 |
stat, demo, meds, chart, out, proc, lab, y = getXY_deep(data, taskf, feat_cond, feat_proc, feat_out, feat_chart,feat_meds)
|
|
|
|
|
|
|
|
|
|
|
364 |
X_dict[hid] = {'stat': stat, 'demo': demo, 'meds': meds, 'chart': chart, 'out': out, 'proc': proc, 'lab': lab, 'y': y}
|
365 |
|
366 |
return X_dict
|
@@ -742,8 +747,19 @@ class Mimic4Dataset(datasets.GeneratorBasedBuilder):
|
|
742 |
yield i, row.to_dict()
|
743 |
######################################################DEEP###############################################################
|
744 |
def _info_deep(self,X_encoded):
|
745 |
-
|
746 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
747 |
return datasets.DatasetInfo(
|
748 |
description=_DESCRIPTION,
|
749 |
features=features,
|
@@ -774,14 +790,14 @@ class Mimic4Dataset(datasets.GeneratorBasedBuilder):
|
|
774 |
lab=data['lab']
|
775 |
|
776 |
yield int(key), {
|
777 |
-
'
|
778 |
-
'
|
779 |
-
'
|
780 |
-
'
|
781 |
-
'
|
782 |
-
'
|
783 |
-
'
|
784 |
-
'
|
785 |
}
|
786 |
|
787 |
#############################################################################################################################
|
|
|
361 |
taskf=task.replace(" ","_")
|
362 |
for hid, data in tqdm(X.iterrows(),desc='Encoding Splits Data for '+task+' task'):
|
363 |
stat, demo, meds, chart, out, proc, lab, y = getXY_deep(data, taskf, feat_cond, feat_proc, feat_out, feat_chart,feat_meds)
|
364 |
+
meds=np.nan_to_num(meds, copy=False)
|
365 |
+
chart=np.nan_to_num(chart, copy=False)
|
366 |
+
out=np.nan_to_num(out, copy=False)
|
367 |
+
proc=np.nan_to_num(proc, copy=False)
|
368 |
+
lab=np.nan_to_num(lab, copy=False)
|
369 |
X_dict[hid] = {'stat': stat, 'demo': demo, 'meds': meds, 'chart': chart, 'out': out, 'proc': proc, 'lab': lab, 'y': y}
|
370 |
|
371 |
return X_dict
|
|
|
747 |
yield i, row.to_dict()
|
748 |
######################################################DEEP###############################################################
|
749 |
def _info_deep(self,X_encoded):
|
750 |
+
features = datasets.Features(
|
751 |
+
{
|
752 |
+
"label": datasets.ClassLabel(num_classes=2,names=["0", "1"]),
|
753 |
+
"DEMO": datasets.Sequence(datasets.Value("int32")),
|
754 |
+
"COND" : datasets.Sequence(datasets.Sequence(datasets.Value("float32"))) ,
|
755 |
+
"MEDS" : datasets.Sequence(datasets.Sequence(datasets.Value("float32"))) ,
|
756 |
+
"PROC" : datasets.Sequence(datasets.Sequence(datasets.Value("float32"))) ,
|
757 |
+
"CHART" : datasets.Sequence(datasets.Sequence(datasets.Value("float32"))) ,
|
758 |
+
"OUT" : datasets.Sequence(datasets.Sequence(datasets.Value("float32"))) ,
|
759 |
+
"LAB" : datasets.Sequence(datasets.Sequence(datasets.Value("float32"))) ,
|
760 |
+
|
761 |
+
}
|
762 |
+
)
|
763 |
return datasets.DatasetInfo(
|
764 |
description=_DESCRIPTION,
|
765 |
features=features,
|
|
|
790 |
lab=data['lab']
|
791 |
|
792 |
yield int(key), {
|
793 |
+
'label': label,
|
794 |
+
'DEMO': demo,
|
795 |
+
'COND': cond_features,
|
796 |
+
'MEDS': meds_features,
|
797 |
+
'PROC': proc_features,
|
798 |
+
'CHART': chart_features,
|
799 |
+
'OUT': out_features,
|
800 |
+
'LAB': lab,
|
801 |
}
|
802 |
|
803 |
#############################################################################################################################
|