Update model loading and prediction methods to use SVM; enhance label encoding and improve variable naming for clarity

Data/idmapping_2025_06_24_predictions.txt

@@ -1,6 +1,6 @@
 Sequence_ID,Prediction 1,Prediction 2,Prediction 3,Prediction 4,Prediction 5,Prediction 6
-sp|P0A7V8|RS4_ECOLI,Cytoplasmic (0.9020),CytoplasmicMembrane (0.0480),Periplasmic (0.0247),Extracellular (0.0184),OuterMembrane (0.0061),Cellwall (0.0006)
-sp|P0A910|OMPA_ECOLI,OuterMembrane (0.9400),Extracellular (0.0251),Periplasmic (0.0177),CytoplasmicMembrane (0.0124),Cytoplasmic (0.0028),Cellwall (0.0019)
-sp|P0A6F5|CH60_ECOLI,Cytoplasmic (0.9935),CytoplasmicMembrane (0.0059),OuterMembrane (0.0004),Periplasmic (0.0003),Extracellular (0.0000),Cellwall (0.0000)
-sp|P02930|TOLC_ECOLI,OuterMembrane (0.9483),CytoplasmicMembrane (0.0166),Extracellular (0.0154),Periplasmic (0.0150),Cytoplasmic (0.0031),Cellwall (0.0016)
-tr|Q9L1T3|Q9L1T3_STRCO,CytoplasmicMembrane (0.5872),Cytoplasmic (0.1784),Extracellular (0.1599),Periplasmic (0.0445),OuterMembrane (0.0246),Cellwall (0.0053)
+sp|P0A7V8|RS4_ECOLI,Cytoplasmic (0.9860),CytoplasmicMembrane (0.0081),Periplasmic (0.0029),Extracellular (0.0019),OuterMembrane (0.0007),Cellwall (0.0003)
+sp|P0A910|OMPA_ECOLI,OuterMembrane (0.9870),CytoplasmicMembrane (0.0054),Periplasmic (0.0041),Cellwall (0.0021),Extracellular (0.0012),Cytoplasmic (0.0001)
+sp|P0A6F5|CH60_ECOLI,Cytoplasmic (0.9771),CytoplasmicMembrane (0.0103),Periplasmic (0.0063),Extracellular (0.0037),OuterMembrane (0.0016),Cellwall (0.0010)
+sp|P02930|TOLC_ECOLI,OuterMembrane (0.9771),Periplasmic (0.0076),CytoplasmicMembrane (0.0074),Cellwall (0.0034),Extracellular (0.0030),Cytoplasmic (0.0015)
+tr|Q9L1T3|Q9L1T3_STRCO,CytoplasmicMembrane (0.9944),Extracellular (0.0028),Periplasmic (0.0012),Cytoplasmic (0.0007),Cellwall (0.0005),OuterMembrane (0.0004)

src/my_utils.py

@@ -725,25 +725,33 @@ def predict_with_esm(fasta_path : str,
     tk.Button(progress_win, text="Close", command=progress_win.destroy).pack(pady=5)
 
     # Load model
-    messagebox.showinfo("Info", "Loading random forest model for predictions...")
+    messagebox.showinfo("Info", "Loading SVM for predictions...")
     project_root: str = os.path.abspath(os.path.join(os.path.dirname(__file__), '..'))
-    model_path = os.path.join(project_root, 'Models', 'rfESM300.joblib' if model == 'esmc_300m' else 'rfESM600.joblib')
+    model_path = os.path.join(project_root, 'Models/ESMC-300m_svm.joblib'
+                              if 
+                                model == 'esmc_300m' 
+                                    else 'Models/ESMC-600m_svm.joblib')
+    le_path = os.path.join(project_root, 'Models/esm_300m_le_svm.joblib'
+                            if 
+                            model == 'esmc_300m'
+                                else 'Models/ESMC-600m_le_svm.joblib')
 
     try:
         predictor = load(model_path)
+        le: LabelEncoder = load(le_path)
 
     except FileNotFoundError:
         print(f"Error: Could not find the model file '{model_path}'")
         return
     
     sequence_ids = list(embeddings.keys())
-    X = np.array(list(embeddings.values()))# type: ignore
+    x = np.array(list(embeddings.values()))# type: ignore
     messagebox.showinfo("Info", "Making predictions...")
-    y_pred_proba = predictor.predict_proba(X)
+    y_pred_proba = predictor.predict_proba(x)
 
     # Get class names
     if hasattr(predictor, 'classes_'):
-        class_names = [str(cls) for cls in predictor.classes_]
+        class_names = le.inverse_transform(predictor.classes_)
     else:
         class_names = [f"Class_{i}" for i in range(y_pred_proba.shape[1])]
 
@@ -938,25 +946,28 @@ def predict_with_prost(fasta_path: str):
     tk.Button(progress_win, text="Close", command=progress_win.destroy).pack(pady=5)
 
     # Load model
-    messagebox.showinfo("Info", "Loading random forest model for predictions...")
+    messagebox.showinfo("Info", "Loading SVM model for predictions...")
     project_root: str = os.path.abspath(os.path.join(os.path.dirname(__file__), '..'))
-    model_path = os.path.join(project_root, 'Models', 'rfProst.joblib')
+    model_path = os.path.join(project_root, 'Models/Prost T5_svm.joblib')
+    le_path = os.path.join(project_root, 'Models/Prost T5_le_svm.joblib')
 
     try:
         predictor = load(model_path)
+        le : LabelEncoder = load(le_path)
+    
     except FileNotFoundError:
         print(f"Error: Could not find the model file '{model_path}'")
         return
 
     sequence_ids = list(embeddings.keys())
-    X = np.array(list(embeddings.values()))  # type: ignore
+    x = np.array(list(embeddings.values()))  #type: ignore
 
     print("Making predictions...")
-    y_pred_proba = predictor.predict_proba(X)
+    y_pred_proba = predictor.predict_proba(x)
 
     # Get class names
     if hasattr(predictor, 'classes_'):
-        class_names = [str(cls) for cls in predictor.classes_]
+        class_names = le.inverse_transform(predictor.classes_)
     else:
         class_names = [f"Class_{i}" for i in range(y_pred_proba.shape[1])]