""" 主程序入口 """ import argparse import sys from pathlib import Path from typing import List, Set from .models import CppcheckIssue from .parsers import parse_cppcheck_xml, parse_cppcheck_text from .analysis import ( analyze_project_structure, filter_and_clean_issues, write_cleaned_report, get_enhanced_issue_analysis ) from .generation import ( generate_test_for_issue, smart_select_issues, write_issue_output ) from .verification import ( auto_verify_tests, generate_verification_report, generate_json_report ) def main(argv: list[str]) -> int: parser = argparse.ArgumentParser(description="根据 cppcheck XML 与源码生成可运行的 C++ 复现用例") parser.add_argument("report", help="cppcheck 报告路径：支持 XML（--xml）或文本日志（自动识别或 --text）") parser.add_argument("--out", default="cppcheck_tests", help="输出目录，默认 cppcheck_tests") parser.add_argument("--model", default="deepseek-chat", help="模型名称，默认 deepseek-chat") parser.add_argument("--emit-runner", action="store_true", help="为每个用例生成一键编译运行的 PowerShell 脚本") parser.add_argument("--text", action="store_true", help="强制按文本日志格式解析") parser.add_argument("--xml", action="store_true", help="强制按 XML 格式解析") parser.add_argument("--max", type=int, default=10, help="最多处理前 N 条问题（默认 10，设为 0 表示不限）") parser.add_argument( "--severities", default="warning,error", help="过滤等级，逗号分隔（如 warning,error,information,note；默认 warning,error）", ) parser.add_argument( "--include-ids", default="", help="仅包含这些 ruleId（逗号分隔，留空表示不限）", ) parser.add_argument( "--exclude-ids", default="missingInclude,missingIncludeSystem,toomanyconfigs,normalCheckLevelMaxBranches,checkLevelNormal,unknown", help="排除这些 ruleId（逗号分隔，默认排除若干低价值项）", ) parser.add_argument( "--smart-select", action="store_true", help="使用AI智能选择最有代表性的测试用例（推荐用于大量问题）", ) parser.add_argument( "--smart-max", type=int, default=10, help="智能选择模式下的最大测试用例数量（默认10）", ) parser.add_argument( "--auto-verify", action="store_true", help="生成测试用例后自动运行验证并生成结果报告", ) parser.add_argument( "--verify-timeout", type=int, default=30, help="验证超时时间（秒，默认30）", ) parser.add_argument( "--verify-tests", action="store_true", help="生成测试用例时立即验证每个测试用例的有效性", ) parser.add_argument( "--use-templates", action="store_true", help="使用预定义的测试用例模板，确保能有效触发cppcheck检测", ) parser.add_argument( "--project-root", help="原始项目根目录路径（用于包含头文件和依赖）", ) parser.add_argument( "--include-dirs", help="额外的头文件包含目录（逗号分隔）", ) parser.add_argument( "--integration-test", action="store_true", help="生成集成测试用例（需要原始项目）", ) parser.add_argument( "--enhanced-analysis", action="store_true", help="启用增强分析模式，基于代码上下文和项目结构进行智能筛选", ) parser.add_argument( "--clean-report", action="store_true", help="生成清理后的cppcheck报告文件，过滤掉不可靠的问题", ) parser.add_argument( "--cleaned-report", help="使用已清理的报告文件（跳过问题过滤步骤）", ) args = parser.parse_args(argv) # 处理报告文件路径 if args.cleaned_report: # 使用已清理的报告文件 report_path = Path(args.cleaned_report).expanduser().resolve() if not report_path.exists(): raise SystemExit(f"找不到已清理的报告文件: {report_path}") print(f"使用已清理的报告文件: {report_path}") else: # 使用原始报告文件 report_path = Path(args.report).expanduser().resolve() if not report_path.exists(): raise SystemExit(f"找不到报告文件: {report_path}") # 解析报告文件 issues: List[CppcheckIssue] = [] if args.xml or (report_path.suffix.lower() in {".xml"} and not args.text): issues = parse_cppcheck_xml(report_path) else: issues = parse_cppcheck_text(report_path) print(f"原始报告包含 {len(issues)} 个问题") # 基本过滤：按严重级别、包含/排除的 ruleId、去重 sev_set: Set[str] = {s.strip().lower() for s in (args.severities or "").split(",") if s.strip()} include_ids: Set[str] = {s.strip() for s in (args.include_ids or "").split(",") if s.strip()} exclude_ids: Set[str] = {s.strip() for s in (args.exclude_ids or "").split(",") if s.strip()} filtered: List[CppcheckIssue] = [] seen: Set[tuple] = set() for iss in issues: if sev_set and iss.severity and iss.severity.lower() not in sev_set: continue if include_ids and iss.id not in include_ids: continue if exclude_ids and iss.id in exclude_ids: continue # 以 (id, first_file, first_line) 去重 key = (iss.id, str(iss.locations[0].file_path) if iss.locations else "", iss.locations[0].line if iss.locations else None) if key in seen: continue seen.add(key) filtered.append(iss) print(f"基本过滤后剩余 {len(filtered)} 个问题") if not filtered: print("未在报告中发现问题项。") return 0 # 处理项目上下文 project_root = None include_dirs = [] project_info = None if args.project_root: project_root = Path(args.project_root).expanduser().resolve() if not project_root.exists(): print(f"警告: 项目根目录不存在: {project_root}") project_root = None else: print("正在分析项目结构...") project_info = analyze_project_structure(project_root) print(f"项目分析完成: 发现 {len(project_info['source_files'])} 个源文件, {len(project_info['header_files'])} 个头文件") if args.include_dirs: include_dirs = [d.strip() for d in args.include_dirs.split(",") if d.strip()] valid_include_dirs = [] for include_dir in include_dirs: include_path = Path(include_dir).expanduser().resolve() if include_path.exists(): valid_include_dirs.append(str(include_path)) else: print(f"警告: 头文件目录不存在: {include_path}") include_dirs = valid_include_dirs # 问题过滤和清理 if args.clean_report and not args.cleaned_report: print("\n" + "="*50) print("开始问题过滤和清理...") print("="*50) cleaned_issues = filter_and_clean_issues(filtered, project_info) # 生成清理后的报告文件 cleaned_report_path = Path(args.out) / "cleaned_cppcheck_report.txt" write_cleaned_report(cleaned_issues, cleaned_report_path) print(f"\n清理完成！") print(f"原始问题数量: {len(issues)}") print(f"基本过滤后: {len(filtered)}") print(f"智能清理后: {len(cleaned_issues)}") print(f"清理后的报告已保存: {cleaned_report_path}") # 使用清理后的问题继续处理 filtered = cleaned_issues elif args.enhanced_analysis: # 使用增强分析进行智能筛选 print("\n" + "="*50) print("开始增强分析...") print("="*50) cleaned_issues = filter_and_clean_issues(filtered, project_info) filtered = cleaned_issues # 智能选择模式 if args.smart_select or args.enhanced_analysis: if args.enhanced_analysis: print(f"启用增强分析模式，从 {len(filtered)} 个问题中选择最多 {args.smart_max} 个最有代表性的测试用例...") else: print(f"启用AI智能选择模式，从 {len(filtered)} 个问题中选择最多 {args.smart_max} 个最有代表性的测试用例...") issues = smart_select_issues(filtered, args.smart_max, args.model) else: # 传统模式：简单限制数量 if args.max and args.max > 0: issues = filtered[: args.max] else: issues = filtered output_dir = Path(args.out).expanduser().resolve() # 为每个问题生成增强的测试用例 for idx, issue in enumerate(issues, start=1): print(f"生成测试用例 {idx}/{len(issues)}: {issue.id}") # 获取增强的问题分析 code_context, relevance_analysis = get_enhanced_issue_analysis(issue, project_info) # 显示分析结果 print(f" 相关性分数: {relevance_analysis['relevance_score']}, 置信度: {relevance_analysis['confidence']}%") if code_context.function_name: print(f" 所在函数: {code_context.function_name}") if code_context.class_name: print(f" 所在类: {code_context.class_name}") # 使用AI生成模式（这是核心功能） content = generate_test_for_issue( issue, model=args.model, project_root=project_root, include_dirs=include_dirs, integration_test=args.integration_test, code_context=code_context, relevance_analysis=relevance_analysis ) out_path = write_issue_output(output_dir, idx, issue, content, emit_runner=args.emit_runner, verify=args.verify_tests) print(f" 已生成: {out_path}") print(f"完成，共生成 {len(issues)} 条用例说明。") # 自动验证 if args.auto_verify: print("\n" + "="*50) print("开始自动验证测试用例...") print("="*50) verification_results = auto_verify_tests(output_dir, args.verify_timeout, project_root, include_dirs) # 生成报告 print("\n生成验证报告...") md_report = generate_verification_report(output_dir, verification_results) json_report = generate_json_report(output_dir, verification_results) print(f"Markdown报告: {md_report}") print(f"JSON报告: {json_report}") # 显示汇总 summary = verification_results["summary"] print(f"\n验证汇总:") print(f" 总测试用例: {summary['total']}") print(f" 编译成功: {summary['compiled']}") print(f" 执行成功: {summary['executed']}") print(f" 漏洞确认: {summary['vulnerabilities_confirmed']}") print(f" 验证超时: {summary['timeouts']}") print(f" 验证错误: {summary['errors']}") # 显示确认的漏洞 confirmed_vulns = [r for r in verification_results["results"] if r["vulnerability_confirmed"]] if confirmed_vulns: print(f"\n确认的漏洞 ({len(confirmed_vulns)} 个):") for result in confirmed_vulns: print(f" ✓ {result['file']}: {result['vulnerability_type']}") else: print("\n未确认任何漏洞") return 0 if __name__ == "__main__": raise SystemExit(main(sys.argv[1:]))