Spaces:
Running
Running
| #!/usr/bin/env python3 | |
| """ | |
| 完整整合Batch 04 - 所有剩余292条规则 | |
| 从Google Notebook提取的578条规则中，整合所有未添加的规则 | |
| 基于成功的integrate_all_rules.py模式 | |
| """ | |
| import json | |
| import re | |
| # 读取Google Notebook提取的全部578条规则 | |
| print("=" * 80) | |
| print("📖 Batch 04 规则提取 - 完整版") | |
| print("=" * 80) | |
| print("📖 读取Google Notebook规则文件...") | |
| file_path = '/Volumes/sparksverse/agent-knowledge/mingli-knowledge/eastern/ziwei/interpretations/02-practical-divination rules.md' | |
| with open(file_path, 'r', encoding='utf-8') as f: | |
| content = f.read() | |
| # 从markdown中提取JSON数组 | |
| print("🔄 从Markdown提取JSON数组...") | |
| start_idx = content.find('[') | |
| end_idx = content.rfind(']') | |
| if start_idx == -1 or end_idx == -1: | |
| print("❌ 未找到JSON数组边界") | |
| exit(1) | |
| json_str = content[start_idx:end_idx+1] | |
| print(f"✅ 找到JSON数组 (字符位置: {start_idx} - {end_idx})") | |
| # 修复age_range字段的格式错误 | |
| print("🔧 修复已知格式问题...") | |
| json_str = re.sub(r'"age_range"\s*:\s*,', '"age_range": [20, 70],', json_str) | |
| json_str = re.sub(r'"age_range"\s*:\s*null', '"age_range": [20, 70]', json_str) | |
| json_str = re.sub(r'"age_range"\s*:\s*\n', '"age_range": [20, 70],\n', json_str) | |
| # 解析JSON | |
| print("📊 解析JSON数据...") | |
| try: | |
| all_rules = json.loads(json_str) | |
| print(f"✅ 成功解析 {len(all_rules)} 条规则") | |
| except json.JSONDecodeError as e: | |
| print(f"❌ JSON解析错误: {e}") | |
| print(f"错误位置: 第{e.lineno}行，第{e.colno}列") | |
| exit(1) | |
| # 读取已整合的规则ID | |
| print("\n📊 读取已整合规则...") | |
| import sys | |
| sys.path.insert(0, '.') | |
| try: | |
| from star_combination_rules import STAR_COMBINATION_RULES | |
| integrated_ids = set([rule['id'] for rule in STAR_COMBINATION_RULES]) | |
| print(f"✅ 已整合规则数: {len(integrated_ids)}") | |
| except Exception as e: | |
| print(f"❌ 无法读取现有规则库: {e}") | |
| exit(1) | |
| # 筛选出未整合的规则 | |
| new_rules = [] | |
| for rule in all_rules: | |
| rule_id = rule.get('id', '') | |
| if rule_id not in integrated_ids: | |
| new_rules.append(rule) | |
| print(f"\n📈 统计结果:") | |
| print(f" 总规则数: {len(all_rules)}") | |
| print(f" 已整合: {len(integrated_ids)}") | |
| print(f" 待整合: {len(new_rules)}") | |
| print(f" 覆盖率: {len(integrated_ids)/len(all_rules)*100:.1f}%") | |
| if len(new_rules) == 0: | |
| print("\n✅ 所有规则都已经整合完毕！") | |
| exit(0) | |
| # 为未整合的规则重新分配ID (从SC601开始) | |
| print(f"\n🔢 重新分配规则ID...") | |
| next_id = 601 | |
| id_mapping = {} | |
| for rule in new_rules: | |
| old_id = rule['id'] | |
| new_id = f'SC{next_id:03d}' | |
| id_mapping[old_id] = new_id | |
| rule['id'] = new_id | |
| next_id += 1 | |
| print(f"✅ 重新分配完成: SC601-SC{next_id-1:03d} ({len(new_rules)}条)") | |
| # 修复格式问题 | |
| print("\n🔧 修复格式问题...") | |
| fixed_count = 0 | |
| for rule in new_rules: | |
| # 修复age_range | |
| if 'age_range' not in rule or rule['age_range'] is None or rule['age_range'] == '': | |
| rule['age_range'] = [20, 70] | |
| fixed_count += 1 | |
| # 修复effect_level中的"微WEAK" | |
| if rule.get('effect_level') == '微WEAK': | |
| rule['effect_level'] = 'VERY_WEAK' | |
| fixed_count += 1 | |
| # 确保effect_level是有效值 | |
| valid_levels = ['极强', '强', '中', '弱', '微弱', 'VERY_WEAK', 'STRONG', 'MEDIUM', 'WEAK', 'VERY_STRONG'] | |
| if rule.get('effect_level') not in valid_levels: | |
| rule['effect_level'] = '中' | |
| fixed_count += 1 | |
| print(f"✅ 修复 {fixed_count} 处格式问题") | |
| # 生成Python代码格式的规则 | |
| print("\n📝 生成Python代码...") | |
| def format_rule_to_python(rule): | |
| """将规则转换为Python代码格式，使用正确的Enum语法""" | |
| # Category映射 | |
| category_map = { | |
| '性格特质': 'RuleCategory.PERSONALITY', | |
| '事业财运': 'RuleCategory.CAREER', | |
| '婚姻感情': 'RuleCategory.RELATIONSHIP', | |
| '健康状况': 'RuleCategory.HEALTH', | |
| '运势走向': 'RuleCategory.FORTUNE', | |
| '才华能力': 'RuleCategory.TALENT', | |
| '人际交往': 'RuleCategory.SOCIAL' | |
| } | |
| # EffectLevel映射 | |
| effect_map = { | |
| '极强': 'EffectLevel.VERY_STRONG', | |
| '强': 'EffectLevel.STRONG', | |
| '中': 'EffectLevel.MEDIUM', | |
| '弱': 'EffectLevel.WEAK', | |
| '微弱': 'EffectLevel.VERY_WEAK', | |
| 'VERY_WEAK': 'EffectLevel.VERY_WEAK', | |
| 'VERY_STRONG': 'EffectLevel.VERY_STRONG', | |
| 'STRONG': 'EffectLevel.STRONG', | |
| 'MEDIUM': 'EffectLevel.MEDIUM', | |
| 'WEAK': 'EffectLevel.WEAK' | |
| } | |
| category = category_map.get(rule['category'], 'RuleCategory.PERSONALITY') | |
| effect = effect_map.get(rule['effect_level'], 'EffectLevel.MEDIUM') | |
| # 格式化stars_required | |
| stars_str = str(rule['stars_required']) | |
| # 格式化palace_types | |
| palace_str = str(rule['palace_types']) | |
| # 格式化conditions (使用安全的格式化方式) | |
| conditions = rule['conditions'] | |
| conditions_str = "{\n 'same_palace': " + str(conditions.get('same_palace', False)) + ",\n" | |
| conditions_str += " 'transformations': " + str(conditions.get('transformations')) + ",\n" | |
| conditions_str += " 'brightness': " + str(conditions.get('brightness')) + ",\n" | |
| conditions_str += " 'palace_position': " + str(conditions.get('palace_position')) + "\n }" | |
| # 格式化interpretations | |
| interp = rule['interpretations'] | |
| interp_str = "{\n 'core_meaning': " + repr(interp.get('core_meaning', '')) + ",\n" | |
| interp_str += " 'personality': " + str(interp.get('personality', [])) + ",\n" | |
| interp_str += " 'career': " + str(interp.get('career', [])) + ",\n" | |
| interp_str += " 'relationship': " + str(interp.get('relationship', [])) + ",\n" | |
| interp_str += " 'fortune_timing': " + repr(interp.get('fortune_timing', '')) + ",\n" | |
| interp_str += " 'warnings': " + str(interp.get('warnings', [])) + "\n }" | |
| return f""" {{ | |
| "id": "{rule['id']}", | |
| "name": "{rule['name']}", | |
| "category": {category}, | |
| "stars_required": {stars_str}, | |
| "palace_types": {palace_str}, | |
| "conditions": {conditions_str}, | |
| "interpretations": {interp_str}, | |
| "effect_level": {effect}, | |
| "compatibility_score": {rule['compatibility_score']}, | |
| "age_range": {rule['age_range']}, | |
| "source": "{rule['source']}" | |
| }}""" | |
| # 生成所有新规则的Python代码 | |
| print("🔄 格式化292条规则...") | |
| new_rules_code = ',\n'.join([format_rule_to_python(rule) for rule in new_rules]) | |
| # 保存为Python文件 | |
| output = f'''""" | |
| Google Notebook 提取规则 - Batch 04 (全部剩余规则) | |
| 规则数量: {len(new_rules)}条 (SC601-SC{next_id-1:03d}) | |
| 提取日期: 2025-12-13 | |
| 来源: 南北派实战断诀 (138页专业典籍) | |
| 覆盖率: 新增后将达到 {(len(integrated_ids) + len(new_rules))/len(all_rules)*100:.1f}% | |
| """ | |
| from star_combination_rules import RuleCategory, EffectLevel | |
| BATCH_04_RULES = [ | |
| {new_rules_code} | |
| ] | |
| ''' | |
| output_file = 'rule_extraction_batch_04.py' | |
| with open(output_file, 'w', encoding='utf-8') as f: | |
| f.write(output) | |
| print(f"✅ 已生成: {output_file}") | |
| print(f" 包含 {len(new_rules)} 条新规则") | |
| print(f" ID范围: SC601-SC{next_id-1:03d}") | |
| print(f" 文件大小: {len(output) // 1024}KB") | |
| # 统计信息 | |
| print("\n" + "=" * 80) | |
| print("📊 Batch 04 新增规则分类统计") | |
| print("=" * 80) | |
| categories = {} | |
| for rule in new_rules: | |
| cat = rule['category'] | |
| categories[cat] = categories.get(cat, 0) + 1 | |
| for cat, count in sorted(categories.items(), key=lambda x: -x[1]): | |
| print(f" {cat}: {count}条") | |
| print("\n" + "=" * 80) | |
| print("📈 规则库扩展进度总览") | |
| print("=" * 80) | |
| print(f" 当前版本: v3.11.0 (320条)") | |
| print(f" Batch 04新增: {len(new_rules)}条") | |
| print(f" 新版本: v3.12.0 ({len(integrated_ids) + len(new_rules)}条)") | |
| print(f" Google Notebook总提取: {len(all_rules)}条") | |
| print(f" 新覆盖率: {(len(integrated_ids) + len(new_rules))/len(all_rules)*100:.1f}%") | |
| print(f" 增长率: +{len(new_rules)/len(integrated_ids)*100:.1f}% (相比v3.11.0)") | |
| print("=" * 80) | |
| print("\n✅ Batch 04整合准备完成！") | |
| print(f"📝 下一步: 运行 merge_batch_04.py 将这{len(new_rules)}条规则合并到系统中") | |
| print(f"🚀 预期效果: 匹配率从2.5%提升至8-10% (基于专家调查)") | |