Spaces:
Runtime error
Runtime error
| """ | |
| HireSense AI — Scoring Engine Orchestrator | |
| Combines all ML scoring services into a unified candidate score. | |
| """ | |
| from backend.services.skill_matcher import SkillMatcher | |
| from backend.services.learning_predictor import LearningPredictor | |
| from backend.services.inflation_detector import InflationDetector | |
| class ScoringEngine: | |
| """Orchestrates all scoring services to produce a unified candidate assessment.""" | |
| WEIGHTS = { | |
| "match_score": 0.40, | |
| "github_score": 0.20, | |
| "leetcode_score": 0.15, | |
| "learning_score": 0.10, | |
| "credibility_score": 0.15, | |
| } | |
| RECOMMENDATION_THRESHOLDS = { | |
| 80: "Strong Hire", | |
| 60: "Hire", | |
| 40: "Maybe", | |
| 0: "Pass", | |
| } | |
| def __init__(self): | |
| self.skill_matcher = SkillMatcher() | |
| self.learning_predictor = LearningPredictor() | |
| self.inflation_detector = InflationDetector() | |
| def score( | |
| self, | |
| parsed_resume: dict, | |
| job_description: str, | |
| github_data: dict = None, | |
| leetcode_data: dict = None, | |
| ) -> dict: | |
| """ | |
| Run all scoring services and produce a unified assessment. | |
| """ | |
| candidate_skills = parsed_resume.get("skills", []) | |
| # 1. Skill matching | |
| github_languages = None | |
| github_repo_context = None | |
| if github_data and not github_data.get("user_not_found"): | |
| github_languages = github_data.get("details", {}).get("languages") | |
| github_repo_context = github_data.get("repo_context") | |
| skill_match_result = self.skill_matcher.match( | |
| candidate_skills=candidate_skills, | |
| job_description=job_description, | |
| github_languages=github_languages, | |
| github_repo_context=github_repo_context, | |
| ) | |
| # 2. Learning ability prediction | |
| missing_skills = skill_match_result.get("missing_skills", []) | |
| learning_result = self.learning_predictor.predict( | |
| candidate_skills=candidate_skills, | |
| missing_skills=missing_skills, | |
| github_data=github_data if github_data and not github_data.get("user_not_found") else None, | |
| leetcode_data=leetcode_data if leetcode_data and not leetcode_data.get("user_not_found") else None, | |
| ) | |
| # 3. Inflation detection | |
| inflation_result = self.inflation_detector.detect( | |
| resume_skills=candidate_skills, | |
| github_data=github_data if github_data and not github_data.get("user_not_found") else None, | |
| leetcode_data=leetcode_data if leetcode_data and not leetcode_data.get("user_not_found") else None, | |
| ) | |
| # 4. Extract sub-scores | |
| match_score = skill_match_result.get("match_score", 0) | |
| github_score = ( | |
| github_data.get("scores", {}).get("overall", 0) | |
| if github_data and not github_data.get("user_not_found") | |
| else 0 | |
| ) | |
| leetcode_score = ( | |
| leetcode_data.get("scores", {}).get("overall", 0) | |
| if leetcode_data and not leetcode_data.get("user_not_found") | |
| else 0 | |
| ) | |
| learning_score = learning_result.get("learning_score", 0) | |
| credibility_score = inflation_result.get("credibility_score", 0) | |
| # 5. Compute dynamic weights (adjust if data is missing) | |
| weights = dict(self.WEIGHTS) | |
| if not github_data or github_data.get("user_not_found"): | |
| # Redistribute GitHub weight | |
| weights["github_score"] = 0 | |
| weights["match_score"] += 0.10 | |
| weights["credibility_score"] += 0.05 | |
| weights["learning_score"] += 0.05 | |
| if not leetcode_data or leetcode_data.get("user_not_found"): | |
| # Redistribute LeetCode weight | |
| weights["leetcode_score"] = 0 | |
| weights["match_score"] += 0.08 | |
| weights["learning_score"] += 0.07 | |
| # 6. Weighted overall score | |
| overall_score = ( | |
| match_score * weights["match_score"] | |
| + github_score * weights["github_score"] | |
| + leetcode_score * weights["leetcode_score"] | |
| + learning_score * weights["learning_score"] | |
| + credibility_score * weights["credibility_score"] | |
| ) | |
| # 7. Generate recommendation | |
| recommendation = self._get_recommendation(overall_score) | |
| # 8. Overall explanation | |
| explanation = self._generate_explanation( | |
| overall_score, recommendation, match_score, github_score, | |
| leetcode_score, learning_score, credibility_score, | |
| ) | |
| return { | |
| "overall_score": round(overall_score, 1), | |
| "recommendation": recommendation, | |
| "scores": { | |
| "match_score": round(match_score, 1), | |
| "github_score": round(github_score, 1), | |
| "leetcode_score": round(leetcode_score, 1), | |
| "learning_score": round(learning_score, 1), | |
| "credibility_score": round(credibility_score, 1), | |
| }, | |
| "skill_match": skill_match_result, | |
| "learning_analysis": learning_result, | |
| "credibility_analysis": inflation_result, | |
| "explanation": explanation, | |
| } | |
| def _get_recommendation(self, score: float) -> str: | |
| """Get recommendation label based on overall score.""" | |
| for threshold in sorted(self.RECOMMENDATION_THRESHOLDS.keys(), reverse=True): | |
| if score >= threshold: | |
| return self.RECOMMENDATION_THRESHOLDS[threshold] | |
| return "Pass" | |
| def _generate_explanation( | |
| self, | |
| overall: float, | |
| recommendation: str, | |
| match: float, | |
| github: float, | |
| leetcode: float, | |
| learning: float, | |
| credibility: float, | |
| ) -> str: | |
| """Generate narrative summary of the candidate assessment.""" | |
| parts = [f"Overall assessment: {recommendation}."] | |
| # Highlight strengths | |
| strengths = [] | |
| if match >= 70: | |
| strengths.append("strong skill match") | |
| if github >= 70: | |
| strengths.append("solid GitHub presence") | |
| if leetcode >= 70: | |
| strengths.append("strong problem-solving ability") | |
| if learning >= 80: | |
| strengths.append("high learning potential") | |
| if credibility >= 85: | |
| strengths.append("high resume credibility") | |
| if strengths: | |
| parts.append(f"Key strengths: {', '.join(strengths)}.") | |
| # Highlight concerns | |
| concerns = [] | |
| if match < 40: | |
| concerns.append("significant skill gaps") | |
| if github > 0 and github < 30: | |
| concerns.append("limited GitHub activity") | |
| if credibility < 50: | |
| concerns.append("credibility concerns") | |
| if concerns: | |
| parts.append(f"Areas of concern: {', '.join(concerns)}.") | |
| return " ".join(parts) | |