new scoring fin

embeddings.py

@@ -611,9 +611,9 @@ class EmbeddingManager:
             return RiskLevel.LOW_RISK  # Different country is low risk per requirements
         return RiskLevel.NO_RISK
     def calculate_similarity(self, job_embeddings: Dict[str, np.ndarray],
-                           seeker_embeddings: Dict[str, np.ndarray],
-                           job_posting: JobPosting,
-                           processed_seeker: IndependentJobSeekerAssessmentRDS) -> MatchResult:
+                       seeker_embeddings: Dict[str, np.ndarray],
+                       job_posting: JobPosting,
+                       processed_seeker: IndependentJobSeekerAssessmentRDS) -> MatchResult:
         """Calculate similarity with comprehensive scoring system"""
         # Calculate embedding-based similarity scores
         field_scores = {}
@@ -640,6 +640,7 @@ class EmbeddingManager:
                 
                 field_pair = f"{job_field}{seeker_field}"
                 field_scores[field_pair] = field_score
+
         # Calculate all risk scores
         risk_scores = {
             'job_stability': self.calculate_job_stability_risk(processed_seeker.experiences),
@@ -664,6 +665,7 @@ class EmbeddingManager:
             'soft_skills': self.evaluate_soft_skills(processed_seeker.experiences),
             'certifications': self.evaluate_certifications(processed_seeker.certifications, job_posting)
         }
+
         # Calculate final score using the formula:
         # Total Score = ∑(Risk Weight × Risk Level) + ∑(Bonus Weight × Bonus Level)
         total_score = 0.0
@@ -672,26 +674,34 @@ class EmbeddingManager:
         for category, score in risk_scores.items():
             weight = FIELD_WEIGHTS.get(category, 0)
             total_score += weight * score
-            if score != RiskLevel.NO_RISK:
-                explanation_parts.append(
-                    f"{category.replace('', ' ').title()}: "
-                    f"{'High' if score == RiskLevel.HIGH_RISK else 'Medium' if score == RiskLevel.MEDIUM_RISK else 'Low'} Risk"
-                )
+            category_name = category.replace('_', '*').upper()
+            if score == RiskLevel.NO_RISK:
+                explanation_parts.append(f"{category_name}: No Risk")
+            elif score == RiskLevel.LOW_RISK:
+                explanation_parts.append(f"{category_name}: Low Risk")
+            elif score == RiskLevel.MEDIUM_RISK:
+                explanation_parts.append(f"{category_name}: Medium Risk")
+            elif score == RiskLevel.HIGH_RISK:
+                explanation_parts.append(f"{category_name}: High Risk")
         
         # Add weighted bonus scores
         for category, score in bonus_scores.items():
             weight = FIELD_WEIGHTS.get(category, 0)
             total_score += weight * score
-            if score != BonusLevel.NO_BONUS:
-                explanation_parts.append(
-                    f"{category.replace('', ' ').title()}: "
-                    f"{'Best' if score == BonusLevel.BEST else 'Better' if score == BonusLevel.BETTER else 'Good'} Bonus"
-                )
+            category_name = category.replace('_', '*').upper()
+            if score == BonusLevel.BEST:
+                explanation_parts.append(f"{category_name}: Best Bonus")
+            elif score == BonusLevel.BETTER:
+                explanation_parts.append(f"{category_name}: Better Bonus")
+            elif score == BonusLevel.GOOD:
+                explanation_parts.append(f"{category_name}: Good Bonus")
+
         # Normalize to 0-100 scale
         min_possible = sum(w * RiskLevel.HIGH_RISK for w in FIELD_WEIGHTS.values())
         max_possible = sum(w * max(RiskLevel.NO_RISK, BonusLevel.BEST) for w in FIELD_WEIGHTS.values())
         normalized_score = ((total_score - min_possible) / (max_possible - min_possible)) * 100
         final_score = max(0, min(100, normalized_score))
+
         return MatchResult(
             similarity_score=final_score,
             field_scores={**field_scores, **risk_scores, **bonus_scores},