feat: implement hybrid similarity ranking and search engine modules to improve project retrieval accuracy

src/recommendation_engine/chatbot_engine.py

@@ -1014,9 +1014,8 @@ def chatbot(user_id: str, user_input: str):
 
         if text.startswith(pattern + " "):
 
-            extracted_title = user_input.split(
-                pattern, 1
-            )[1].strip()
+            pattern_idx = user_input.lower().find(pattern)
+            extracted_title = user_input[pattern_idx + len(pattern):].strip() if pattern_idx != -1 else ""
 
             if (
                 extracted_title

src/similarity_model/hybrid_ranker.py

@@ -67,15 +67,11 @@ def compute_hybrid_score(
 
     shared_count = feature_count - unique_query_count
 
-    
-    
-    
+    # Hard duplicate: most features shared, almost nothing unique
     if shared_count >= 6 and unique_query_count <= 1:
         return 0.95
 
-    
-    
-    
+    # Near-total feature coverage with high similarity
     if coverage >= 0.90 and feature_score >= 0.65:
         return round(
             clamp(
@@ -86,15 +82,7 @@ def compute_hybrid_score(
             4
         )
 
-    
-    
-    
-    if feature_score == 0 or coverage == 0:
-        return round(clamp(0.85 * semantic_score), 4)
-
-    
-    
-    
+    # Get adaptive weights regardless of feature availability
     semantic_w, feature_w, coverage_w = get_dynamic_weights(
         feature_count, coverage
     )
@@ -105,11 +93,10 @@ def compute_hybrid_score(
         coverage_w * coverage
     )
 
-    
-    
-    
-    
-    if feature_score < 0.40:
+    # Only penalise low feature score when features actually exist to compare.
+    # Skipping this when feature_score==0 prevents an unfair 50% penalty
+    # on DB projects that have no stored features.
+    if feature_score > 0 and feature_score < 0.40:
         penalty = 0.50 + (feature_score / 0.40) * 0.50
         score *= penalty
 
@@ -121,13 +108,20 @@ def compute_originality(
     total_query_features: int
 ) -> float:
     """
-    Originality Score (0-100)
-    Calibrated piecewise curve to match human perception of similarity.
+    Originality Score (0-100).
+    Base score is (1 - hybrid_score) * 100.
+    Bonus of up to +10 points when the query has many features
+    that are unique vs the candidate (genuine differentiation signal).
     """
     hybrid_score = clamp(hybrid_score)
 
     originality = 100.0 * (1.0 - hybrid_score)
 
+    # Uniqueness bonus: reward projects whose features don't overlap
+    if total_query_features > 0:
+        uniqueness_ratio = unique_query_features / total_query_features
+        originality = min(100.0, originality + (uniqueness_ratio * 10.0))
+
     return round(max(0.0, min(100.0, originality)), 2)
 
 def compute_confidence(

src/similarity_model/similarity_engine.py

@@ -43,7 +43,6 @@ def build_raw_text(
     """
     parts = [
         str(title).strip(),
-        str(title).strip(),   
         str(abstract).strip(),
         str(description).strip()
     ]