Add ROI name column to clinical files

preprocessing/prepare_clinical.py

@@ -96,6 +96,14 @@ if DATASET_NAME == 'HN_val':
         lambda x: grading_dict_grouped[x]
     )
 
+    old_clinical_len = len(clinical)
+
+    # Get ROI name for each segmentation mask
+    ROI_INFO_FILENAME = 'INFO_GTVcontours_HN.csv'
+    roi_info = pd.read_csv(RAW_DATA_PATH / ROI_INFO_FILENAME)
+    clinical = clinical.merge(roi_info, on='patient')
+    assert old_clinical_len == len(clinical)
+
     clinical.to_csv(PROCESSED_DATA_PATH / CLINICAL_DATA_FILENAME_CLEAN)
 
     ################### clean metadata file to keep only CT scans, PT scans and 1 segmentation mask
@@ -332,7 +340,6 @@ if DATASET_NAME == 'HN_val':
 
         dicom_folder = str(path_to_study / folder)
         dicom_folders.append(dicom_folder)
-    # TODO: finish to reconstruct path
 
     metadata['dicom_folder'] = dicom_folders
 
@@ -347,7 +354,8 @@ elif DATASET_NAME == 'HN_BZ':
     clinical_no_na[LABEL_COL_LOCOREGIONAL] = clinical_no_na[
         LABEL_COL_LOCOREGIONAL
     ].astype(np.uint8)
-    # clinical_no_na.sort_values(by=PATIENT_COL, inplace=True)
+
+    clinical['ROI_name'] = 'lesione'
 
     clinical_no_na.to_csv(
         PROCESSED_DATA_PATH / CLINICAL_DATA_FILENAME_CLEAN, index=False