scripts/split_questiondb.py

#!/usr/bin/env python3
"""
QuestionDatabase CSV を教科・ジャンル別に分割してJSONファイルを生成する

入力: knowledge/v1.8.0/超天才クイズv3 - QuestionDatabase.csv
出力: knowledge/v1.8.1/{subject}/{category}.json
"""

import csv
import json
from pathlib import Path
from collections import defaultdict

# パス設定
PROJECT_ROOT = Path(__file__).parent.parent
INPUT_CSV = PROJECT_ROOT / "knowledge" / "v1.8.0" / "超天才クイズv3 - QuestionDatabase.csv"
OUTPUT_DIR = PROJECT_ROOT / "knowledge" / "v1.8.1"

# 教科名マッピング
SUBJECT_NAMES = {
    "jp": "国語",
    "math": "算数",
    "sci": "理科",
    "soc": "社会"
}

# ジャンル名マッピング
CATEGORY_NAMES = {
    # 国語
    "JP01": "漢字の読み書き",
    "JP02": "語句・ことわざ",
    "JP03": "文法",
    "JP04": "読解",
    "JP05": "作文",
    "JP06": "古典",
    # 算数
    "MA01": "計算",
    "MA02": "割合・比",
    "MA03": "速さ",
    "MA04": "図形",
    "MA05": "面積・体積",
    "MA06": "数列・規則性",
    "MA07": "場合の数",
    "MA08": "濃度・仕事算",
    "MA09": "グラフ・表",
    "MA10": "文章題",
    # 理科
    "SC01": "植物",
    "SC02": "動物",
    "SC03": "人体",
    "SC04": "天気",
    "SC05": "天体",
    "SC06": "地層・岩石",
    "SC07": "物質・水溶液",
    "SC08": "燃焼・気体",
    "SC09": "電気・磁石",
    "SC10": "力・運動",
    # 社会
    "SO01": "地理（日本）",
    "SO02": "地理（世界）",
    "SO03": "歴史（古代〜中世）",
    "SO04": "歴史（近世）",
    "SO05": "歴史（近代〜現代）",
    "SO06": "政治",
    "SO07": "経済",
    "SO08": "国際",
    "SO09": "環境・資源",
    "SO10": "時事問題"
}


def main():
    print(f"入力ファイル: {INPUT_CSV}")
    print(f"出力先: {OUTPUT_DIR}")
    print()

    # CSVを読み込み、subject + category でグルーピング
    grouped = defaultdict(list)
    total_count = 0

    with open(INPUT_CSV, 'r', encoding='utf-8') as f:
        reader = csv.DictReader(f)
        for row in reader:
            subject = row['subject']
            category = row['category']
            key = (subject, category)

            # 問題データを抽出（usage_countは除外）
            question = {
                "answer_id": row['answer_id'],
                "answer": row['answer'],
                "question_hint": row['question_hint'],
                "difficulty": row['difficulty'],
                "source_context": row['source_context']
            }
            grouped[key].append(question)
            total_count += 1

    print(f"読み込み完了: {total_count}問")
    print()

    # 各グループをJSONファイルとして出力
    file_count = 0
    for (subject, category), questions in sorted(grouped.items()):
        output_path = OUTPUT_DIR / subject / f"{category}.json"

        data = {
            "subject": subject,
            "subject_name": SUBJECT_NAMES.get(subject, subject),
            "category": category,
            "category_name": CATEGORY_NAMES.get(category, category),
            "question_count": len(questions),
            "questions": questions
        }

        with open(output_path, 'w', encoding='utf-8') as f:
            json.dump(data, f, ensure_ascii=False, indent=2)

        print(f"  {output_path.relative_to(PROJECT_ROOT)}: {len(questions)}問")
        file_count += 1

    print()
    print(f"完了: {file_count}ファイル, {total_count}問")


if __name__ == "__main__":
    main()