Update main.py

main.py

@@ -206,18 +206,15 @@ async def extract_resume_and_search(file: UploadFile = File(...)):
             n_results=random.randint(5, 7)  # Match your existing search logic
         )
         
-        # Process results (convert distances to proper similarity scores)
+        # Process results (same as your existing search logic)
         recommendations = []
         ids = results.get('ids', [[]])[0]
         distances = results.get('distances', [[]])[0]
         
-        # Convert cosine distances to similarity scores (0-1 range)
         for i, internship_id in enumerate(ids):
-            # Clamp the score to ensure it's between 0 and 1
-            similarity_score = max(0.0, min(1.0, 1 - distances[i]))
             recommendations.append({
                 "internship_id": internship_id,
-                "score": round(similarity_score, 4)  # Round for cleaner output
+                "score": 1 - distances[i]
             })
         
         print(f"✅ Found {len(recommendations)} recommendations for resume")
@@ -286,29 +283,10 @@ def get_profile_recommendations(profile: UserProfile):
     ids = results.get('ids', [[]])[0]
     distances = results.get('distances', [[]])[0]
     
-    # Convert distances to similarity scores (robust approach)
-    if len(distances) == 0:
-        return {"recommendations": []}
-    
-    # Normalize scores based on the actual range in results
-    min_dist = min(distances)
-    max_dist = max(distances)
-    
     for i, internship_id in enumerate(ids):
-        # DEBUG: Print actual distance values
-        print(f"Debug - Internship {internship_id}: distance = {distances[i]}")
-        
-        # Normalize to 0-1 range where smaller distance = higher score
-        if max_dist == min_dist:
-            # All distances are the same, give them all the same score
-            similarity_score = 0.5
-        else:
-            # Normalize: best match gets 1.0, worst gets 0.0
-            similarity_score = 1.0 - ((distances[i] - min_dist) / (max_dist - min_dist))
-        
         recommendations.append({
             "internship_id": internship_id,
-            "score": round(similarity_score, 4)
+            "score": 1 - distances[i]
         })
         
     return {"recommendations": recommendations}
@@ -329,24 +307,10 @@ def search_internships(search: SearchQuery):
     ids = results.get('ids', [[]])[0]
     distances = results.get('distances', [[]])[0]
     
-    # Convert distances to similarity scores (robust approach)
-    if len(distances) == 0:
-        return {"recommendations": []}
-    
-    # Normalize scores based on the actual range in results
-    min_dist = min(distances)
-    max_dist = max(distances)
-    
     for i, internship_id in enumerate(ids):
-        # Normalize to 0-1 range where smaller distance = higher score
-        if max_dist == min_dist:
-            similarity_score = 0.5
-        else:
-            similarity_score = 1.0 - ((distances[i] - min_dist) / (max_dist - min_dist))
-        
         recommendations.append({
             "internship_id": internship_id,
-            "score": round(similarity_score, 4)
+            "score": 1 - distances[i]
         })
             
     return {"recommendations": recommendations}