Update app.py

app.py

@@ -51,6 +51,183 @@ Authors: Masoud Samaei1, Roohollah Shirani Faradonbeh1,*, Morteza Alinejad Omran
 2.	Department of Civil Engineering, Shahrood University of Technology, Shahrood, Iran
 3.	Faculty of Engineering, University of Tabriz, Tabriz, Iran
 * Corresponding Author—Email: roohollah.shiranifaradonbeh@curtin.edu.au
+""",
+                   live=False)
+app.launch()
+
+
+
+import numpy as np
+import matplotlib.pyplot as plt
+
+
+def make_prediction2(parameter, Type, dr, sn, percent, tmax):
+    with open("Decision_Tree.pkl", "rb") as f1:
+        clf1  = pickle.load(f1)
+    with open("random_forest.pkl", "rb") as f2:
+        clf2  = pickle.load(f2)
+    with open("XGBoost.pkl", "rb") as f3:
+        clf3  = pickle.load(f3)
+    with open("AdaBoost.pkl", "rb") as f4:
+        clf4  = pickle.load(f4)
+
+    predses1=np.array([])
+    predses2=np.array([])
+    predses3=np.array([])
+    predses4=np.array([])
+    
+    # fig = plt.figure()
+    
+    if parameter=="PET Type":
+        Types = np.array([])
+        for Type in range(1,4):
+            Types = np.append(Types, Type)
+          
+            preds1 = clf1.predict([[Type, dr, sn, percent]])
+            preds1 =  (round((preds1.squeeze() * tmax),2))
+            predses1 = np.append(predses1, preds1)
+            
+            preds2 = clf2.predict([[Type, dr, sn, percent]])
+            preds2 =(round((preds2.squeeze() * tmax),2))
+            predses2= np.append(predses2, preds2)
+            
+            preds3 = clf3.predict([[Type, dr, sn, percent]])
+            preds3 =(round((preds3.squeeze() * tmax),2))
+            predses3= np.append(predses3, preds3)
+            
+            preds4 = clf4.predict([[Type, dr, sn, percent]])
+            preds4 =(round((preds4.squeeze() * tmax),2))
+            predses4= np.append(predses4, preds4)
+            
+        plt.xlim(0, 4)     # set the xlim to left, right
+        plt.ylim(tmax*0.8, 1.5*tmax)
+        plt.scatter(Types, predses1, color='k', label='Decision Tree')
+        plt.scatter(Types, predses2, color='b', label='Random Forest')
+        plt.scatter(Types, predses3, color='r', label='XGBoost')
+        plt.scatter(Types, predses4, color='y', label='AdaBoost')
+        plt.ylabel("Shear Strength", fontweight='bold')
+        plt.axhline(y=tmax, color='r', linestyle='-')
+        plt.xlabel("PET Type", fontweight='bold')
+        plt.legend(loc='lower right')
+        plt.show()
+
+
+    elif parameter=="Dr":
+        drs = np.array([])
+        for dr in np.arange(0.55,1,0.2):
+            drs = np.append(drs, dr)
+          
+            preds1 = clf1.predict([[Type, dr, sn, percent]])
+            preds1 =  (round((preds1.squeeze() * tmax),2))
+            predses1 = np.append(predses1, preds1)
+            
+            preds2 = clf2.predict([[Type, dr, sn, percent]])
+            preds2 =(round((preds2.squeeze() * tmax),2))
+            predses2= np.append(predses2, preds2)
+            
+            preds3 = clf3.predict([[Type, dr, sn, percent]])
+            preds3 =(round((preds3.squeeze() * tmax),2))
+            predses3= np.append(predses3, preds3)
+            
+            preds4 = clf4.predict([[Type, dr, sn, percent]])
+            preds4 =(round((preds4.squeeze() * tmax),2))
+            predses4= np.append(predses4, preds4)
+            
+        plt.xlim(0, 1.5)     # set the xlim to left, right
+        plt.ylim(tmax*0.9, 1.1*tmax)
+        plt.scatter(drs, predses1, color='k', label='Decision Tree')
+        plt.scatter(drs, predses2, color='b', label='Random Forest')
+        plt.scatter(drs, predses3, color='r', label='XGBoost')
+        plt.scatter(drs, predses4, color='y', label='AdaBoost')
+        plt.ylabel("Shear Strength", fontweight='bold')
+        plt.xlabel("Dr", fontweight='bold')
+        plt.axhline(y=tmax, color='r', linestyle='-')
+        plt.legend(loc='lower right')
+        plt.show()
+
+
+    elif parameter=="Sn":
+        snss = np.array([])
+        for sn in np.arange(50,155,50):  
+            snss = np.append(snss, sn)
+            preds1 = clf1.predict([[Type, dr, sn, percent]])
+            preds1 =  (round((preds1.squeeze() * tmax),2))
+            predses1 = np.append(predses1, preds1)
+            
+            preds2 = clf2.predict([[Type, dr, sn, percent]])
+            preds2 =(round((preds2.squeeze() * tmax),2))
+            predses2= np.append(predses2, preds2)
+            
+            preds3 = clf3.predict([[Type, dr, sn, percent]])
+            preds3 =(round((preds3.squeeze() * tmax),2))
+            predses3= np.append(predses3, preds3)
+            
+            preds4 = clf4.predict([[Type, dr, sn, percent]])
+            preds4 =(round((preds4.squeeze() * tmax),2))
+            predses4= np.append(predses4, preds4)
+        
+        plt.xlim(40, 160)     # set the xlim to left, right
+        plt.ylim(tmax*0.8, 1.5*tmax)
+        plt.scatter(snss, predses1, color='k', label='Decision Tree')
+        plt.scatter(snss, predses2, color='b', label='Random Forest')
+        plt.scatter(snss, predses3, color='r', label='XGBoost')
+        plt.scatter(snss, predses4, color='y', label='AdaBoost')
+        plt.ylabel("Shear Strength", fontweight='bold')
+        plt.xlabel("Normal Stress (Sn)", fontweight='bold')
+        plt.axhline(y=tmax, color='r', linestyle='-')
+        plt.legend(loc='lower right')
+        plt.show()  
+        
+    elif parameter=="PET percantage":
+        percents= np.array([])
+        for percent in np.arange(0.1,2.05,0.1):          
+            percents = np.append(percents, percent)
+            preds1 = clf1.predict([[Type, dr, sn, percent]])
+            preds1 =  (round((preds1.squeeze() * tmax),2))
+            predses1 = np.append(predses1, preds1)
+            
+            preds2 = clf2.predict([[Type, dr, sn, percent]])
+            preds2 =(round((preds2.squeeze() * tmax),2))
+            predses2= np.append(predses2, preds2)
+            
+            preds3 = clf3.predict([[Type, dr, sn, percent]])
+            preds3 =(round((preds3.squeeze() * tmax),2))
+            predses3= np.append(predses3, preds3)
+            
+            preds4 = clf4.predict([[Type, dr, sn, percent]])
+            preds4 =(round((preds4.squeeze() * tmax),2))
+            predses4= np.append(predses4, preds4)
+        
+        plt.xlim(0, 2.5)     # set the xlim to left, right
+        plt.ylim(tmax*0.6, 1.5*tmax)
+        plt.scatter(percents, predses1, color='k', label='Decision Tree')
+        plt.scatter(percents, predses2, color='b', label='Random Forest')
+        plt.scatter(percents, predses3, color='r', label='XGBoost')
+        plt.scatter(percents, predses4, color='y', label='AdaBoost')
+        plt.ylabel("Shear Strength", fontweight='bold')
+        plt.xlabel("PET percentage", fontweight='bold')
+        plt.axhline(y=tmax, color='r', linestyle='-')
+        plt.legend(loc='lower right')
+        plt.show()  
+        
+    return plt
+
+
+
+#
+    
+parameter = gr.Radio(["PET Type", "Dr", "Sn", "PET percantage"], value="PET Type", label="Select the parameter")
+Type = gr.Slider(1, 3, value=1, step=1, label="PET type", info="Choose 1 for 1*1 PET, 2 for 1*5 PET, and 3 for fiber PET")
+dr = gr.Slider(0.55, 0.95, step=0.2, label="Relative Density (Dr)", info="Range between 0.55 to 0.95")
+sn = gr.Slider(50, 150, step=50, label="Normal Stresses (Sn)", info="Range between 50 to 150")
+percent= gr.Slider(0, 2, step=0.1, label="PET Percent", info="Range between 0 to 2")
+tmax= gr.Slider(36.5, 144, step=0.1, label="Shear Strength (plain soil without PET)", info="Range between 36 to 144")
+
+# We create the output
+
+app = gr.Interface(fn = make_prediction2, inputs=[parameter, Type, dr, sn, percent, tmax], outputs=['plot'], 
+                   title= "Parametric Analysis",
+                   description = """ abc
 """,
                    live=False)
 app.launch()