adding script to form a proper label for training images

vinsentp · vinsentp · commit dc2a25865c3d · 2019-02-14T15:29:25.000+05:30
diff --git a/scripts/dataframepdf/multiclass_label creation.py b/scripts/dataframepdf/multiclass_label creation.py
@@ -0,0 +1,55 @@
+# https://stackoverflow.com/questions/36909977/update-row-values-where-certain-condition-is-met-in-pandas
+
+## converting the dataframe into required format...
+import pandas as pd
+df = pd.read_csv('/root/sharedfolder/CheXpert-v1.0-small/train_v1.csv')
+df.head()
+df1 = df.drop(['Sex', 'Age', 'Frontal/Lateral','AP/PA'], axis=1)
+df1.shape
+
+df_rename = df1.rename(columns={'No Finding':'NoFinding','Enlarged Cardiomediastinum': 'EnlargedCardiomediastinum', 'Lung Opacity': 'LungOpacity',
+                        'Lung Lesion':'LungLesion', 'Pleural Effusion':'PleuralEffusion', 'Pleural Other':'PleuralOther','Support Devices':'SupportDevices'
+                        
+                       })
+list(df_rename.columns)
+
+df_rename=df_rename.dropna(subset=[ 'NoFinding', 'EnlargedCardiomediastinum', 'Cardiomegaly', 'LungOpacity', 'LungLesion', 'Edema', 'Consolidation', 'Pneumonia', 'Atelectasis', 'Pneumothorax', 'PleuralEffusion', 'PleuralOther', 'Fracture', 'SupportDevices'], how='all')
+
+
+df_rename.columns
+#df_rename.dropna(thresh=2)
+
+
+new_cols = ['v1','v2','v3','v4','v5','v6','v7','v8','v9','v10','v11','v12','v13','v14']
+df_rename = df_rename.reindex(df_rename.columns.union(new_cols), axis=1)
+
+df_rename.loc[df_rename['NoFinding']==1, 'v1'] = 'NoFinding'
+df_rename.loc[df_rename['Atelectasis']==1, 'v2'] = 'Atelectasis'
+df_rename.loc[df_rename['Cardiomegaly']==1, 'v3'] = 'Cardiomegaly'
+df_rename.loc[df_rename['Consolidation']==1, 'v4'] = 'Consolidation'
+df_rename.loc[df_rename['Edema']==1, 'v5'] = 'Edema'
+df_rename.loc[df_rename['EnlargedCardiomediastinum']==1, 'v6'] = 'EnlargedCardiomediastinum'
+df_rename.loc[df_rename['Fracture']==1, 'v7'] = 'Fracture'
+df_rename.loc[df_rename['LungLesion']==1, 'v8'] = 'LungLesion'
+df_rename.loc[df_rename['LungOpacity']==1, 'v9'] = 'LungOpacity'
+df_rename.loc[df_rename['PleuralEffusion']==1, 'v10'] = 'PleuralEffusion'
+df_rename.loc[df_rename['PleuralOther']==1, 'v11'] = 'PleuralOther'
+df_rename.loc[df_rename['Pneumonia']==1, 'v12'] = 'Pneumonia'
+df_rename.loc[df_rename['Pneumothorax']==1, 'v13'] = 'Pneumothorax'
+df_rename.loc[df_rename['Support'Path',Devices']==1, 'v14'] = 'SupportDevices'
+
+
+
+df_final = df_rename[['Path','v1','v2','v3','v4','v5','v6','v7','v8','v9','v10','v11','v12','v13','v14']]
+
+
+# replace all NaN
+df_final = df_final.fillna('')
+df_final['tags'] =  df_final[['v1','v2','v3','v4','v5','v6','v7','v8','v9','v10','v11','v12','v13','v14']].apply(lambda x: ' '.join(x), axis=1)
+df_final01 = df_final[['Path','tags']]
+
+df_final02.dropna(subset=['tags'])
+df_final02.to_csv('/root/sharedfolder/train_v01.csv', sep=",", encoding='utf-8',index=False)
+
+train_df = pd.read_csv('/root/sharedfolder/CheXpert-v1.0-small/train_v03.csv')
+train_df
