pipelines/scoring.py

import re
from typing import Dict

def _coverage_flags(normalized: Dict) -> Dict[str, bool]:
    return {
        "has_work": bool(normalized.get("work_experience")),
        "has_edu": bool(normalized.get("education")),
        "has_certs": bool(normalized.get("certifications")),
        "has_skills": bool(normalized.get("skills")),
    }

def _blank_gaps(normalized: Dict) -> int:
    periods = [w.get("period", "") for w in normalized.get("work_experience", [])]
    return max(0, 0 if not periods else len([p for p in periods if ("~" in p or "-" in p) and "現在" not in p]))

def _lexical_diversity(text: str) -> float:
    tokens = re.findall(r"\w+", text.lower())
    if not tokens:
        return 0.0
    uniq = len(set(tokens))
    return round(uniq / len(tokens), 4)

def compute_quality_score(text: str, normalized: Dict) -> Dict:
    flags = _coverage_flags(normalized)
    coverage = sum(1 for v in flags.values() if v) / 4.0
    gaps = _blank_gaps(normalized)
    lexdiv = _lexical_diversity(text)
    total = round(0.5 * coverage + 0.1 * max(0.0, 0.5 - min(0.5, gaps * 0.05)) + 0.4 * lexdiv, 4)
    return {"coverage_flags": flags, "gap_count": gaps, "lexical_diversity": lexdiv, "total_score": total}