using System.Diagnostics; using XFEToolBox.Client.Core.Search; namespace XFEToolBox.Test; public static class FileSearchQueryTests { [Test] public static void FileQueriesCombinePhrasesAlternativesExclusionsAndFilters() { var report = new FileSearchEntry(@"C:\Work\My Report final.CS", false); Ensure(FileSearchQuery.Parse("\"my report\" final ext:cs;json !draft").Matches(report), "词组、AND、扩展名及排除没有正确组合。"); Ensure(FileSearchQuery.Parse("missing | \"my report\" ext:cs").Matches(report), "OR 分支没有匹配。"); Ensure(!FileSearchQuery.Parse("\"my report\" !final").Matches(report), "排除词没有生效。"); Ensure(FileSearchQuery.Parse("folder:work file:report").Matches(report), "文件名和目录筛选没有匹配。"); Ensure(!FileSearchQuery.Parse("folder:").Matches(report), "folder: 应只匹配目录。"); Ensure(FileSearchQuery.Parse("folder:").Matches(new(@"C:\Work", true)), "folder: 没有匹配目录。"); Ensure(FileSearchQuery.Parse("file: path:C:/work").Matches(report), "文件筛选或斜杠标准化失败。"); Ensure(FileSearchQuery.Parse("O'Reilly").Matches(new(@"D:\Books\O'Reilly.txt", false)), "文件名中的撇号应按字面量检索。"); Ensure(FileSearchQuery.Parse("'my report'").Matches(report), "单引号词组没有匹配。"); } [Test] public static void FileQueriesSupportBoundedGlobsAndLiteralBrackets() { var nested = new FileSearchEntry(@"C:\src\Nested\Report7.cs", false); var direct = new FileSearchEntry(@"C:\src\Report7.cs", false); Ensure(FileSearchQuery.Parse("Report?.[ct][sx]").Matches(nested), "? 或字符组没有匹配。"); Ensure(FileSearchQuery.Parse("Report[0-9].cs").Matches(nested), "字符范围没有匹配。"); Ensure(!FileSearchQuery.Parse("Report[!0-9].cs").Matches(nested), "否定字符组没有排除。"); Ensure(!FileSearchQuery.Parse("Report.cs").Matches(nested), "不含通配符的子串检索发生错误匹配。"); Ensure(FileSearchQuery.Parse(@"C:\src\**\*.cs").Matches(nested), "** 没有跨目录匹配。"); Ensure(FileSearchQuery.Parse(@"C:\src\**\*.cs").Matches(direct), "** 没有匹配零层目录。"); Ensure(!FileSearchQuery.Parse(@"C:\src\*.cs").Matches(nested), "单 * 不应跨越路径分隔符。"); Ensure(FileSearchQuery.Parse("Report[[]old[]].txt").Matches(new(@"C:\Report[old].txt", false)), "字面量方括号转义失败。"); Ensure(FileSearchQuery.Parse("needle*omega").MatchesText("prefix needle\nseveral lines\nomega suffix"), "正文通配符应支持跨行子串匹配。"); Ensure(FileSearchQuery.Parse("*a*a*a*a*a*a*a*a*a*a*b").MatchesText(new string('a', 200_000)) == false, "通配符匹配出现错误或回溯问题。"); } [Test] public static void FileQueriesPreserveContentBranchConstraintsAndSafePrefilters() { var query = FileSearchQuery.Parse("ext:cs alpha | ext:txt beta"); Ensure(!query.MatchesContent(new(@"C:\code.cs", false), "beta"), "正文和元数据不能来自不同的 OR 分支。"); Ensure(query.MatchesContent(new(@"C:\code.cs", false), "alpha"), "正文和元数据的同一分支应该匹配。"); Ensure(query.MatchesContent(new(@"C:\notes.txt", false), "beta"), "正文第二分支匹配失败。"); Ensure(query.RequiredLiterals.Count == 0, "OR 分支没有共有词时不得生成 AND 前置过滤。"); var common = FileSearchQuery.Parse("alpha*omega ext:cs | alpha delta !draft"); Ensure(common.RequiredLiterals.SequenceEqual(["alpha"]), "必需正文片段应取 OR 分支交集并忽略筛选器和排除词。"); Ensure(!FileSearchQuery.Parse("needle !secret").MatchesText("needle secret"), "正文排除条件没有生效。"); Ensure(FileSearchQuery.Parse("ext:cs").MatchesContent(new(@"C:\code.cs", false), "anything"), "纯元数据筛选应允许任意正文。"); } [Test] public static void FileQueriesRejectMalformedInputWithoutThrowing() { foreach (var text in new[] { "\"unterminated", "a |", "| a", "a || b", "!", "ext:", "ext:cs;", "[abc", "[z-a]", new string('x', 4097) }) { var query = FileSearchQuery.Parse(text); Ensure(query.Error is not null, $"损坏查询应给出错误:{text[..Math.Min(text.Length, 50)]}"); Ensure(!query.Matches(new(@"C:\a.txt", false)), "损坏查询不能产生匹配。"); } Ensure(FileSearchQuery.Parse(" ").IsEmpty, "空查询没有正确识别。"); } [Test] public static void FileIndexRanksAndPagesWithoutFilesystemAccess() { var index = new FileSearchIndex(); index.AddOrUpdate([ new(@"Z:\report", false), new(@"Z:\report final.txt", false), new(@"Z:\old report.txt", false), new(@"Z:\report archive\unrelated.txt", false), new(@"Z:\work\report", true) ]); var results = index.Search("report"); Ensure(results.TotalMatches == 5, "完整路径关键词匹配丢失结果。"); Ensure(results.Items[0].Entry.FullPath == @"Z:\report", "精确文件名应排在第一位。"); Ensure(results.Items[1].Entry.IsDirectory, "精确目录名应排在非精确文件名之前。"); Ensure(results.Items[2].Entry.Name == "report final.txt", "前缀应排在包含之前。"); var page = index.Search("report", offset: 2, limit: 2); Ensure(page.TotalMatches == 5 && page.Items.SequenceEqual(results.Items.Skip(2).Take(2)), "分页结果与完整排序不一致。"); Ensure(index.Search("report", limit: 0).TotalMatches == 5, "只统计查询丢失总数。"); Ensure(index.Search("").Items.Count == 0, "空查询不应枚举所有文件。"); } [Test] public static void FileIndexHandlesUpdatesSubtreeDeletionUnicodeAndCancellation() { var index = new FileSearchIndex(); index.AddOrUpdate([ new(@"C:\中文项目\配置.json", false), new(@"C:\中文项目\sub\README.md", false), new(@"C:\中文项目2\keep.txt", false), new(@"C:\π\Δelta.txt", false), new("/tmp/example/file.cs", false) ]); Ensure(index.Search("中").TotalMatches == 3, "单个中文字符应匹配目录路径。"); Ensure(index.Search("配置").TotalMatches == 1, "两个中文字符应该匹配文件名。"); Ensure(index.Search("δelta").TotalMatches == 1, "Unicode 大小写匹配失败。"); Ensure(index.Search(@"path:\tmp\example").TotalMatches == 1, "平台无关路径分隔符匹配失败。"); var unix = index.GetEntriesSnapshot().Single(entry => entry.FullPath.StartsWith('/')); Ensure(unix.Name == "file.cs" && unix.DirectoryPath == "/tmp/example", "平台无关文件名/目录提取失败。"); index.AddOrUpdate([new(@"c:/中文项目/配置.JSON", false)]); Ensure(index.Count == 5 && index.Search("配置").Items[0].Entry.Name == "配置.JSON", "大小写/分隔符等价更新产生重复记录。"); Ensure(index.ContainsPath(@"C:\中文项目\配置.json"), "路径存在检查应遵循索引的身份规范化。"); index.RemovePath(@"C:\中文项目", true); Ensure(index.Count == 3 && index.Search("中文项目2").TotalMatches == 1, "子树删除错误地移除了同前缀兄弟目录。"); Ensure(!index.ContainsPath(@"C:\中文项目\配置.json"), "删除的路径不应继续报告存在。"); index.RemovePath(@"C:\中文项目2\keep.txt", false); Ensure(index.Search("keep").TotalMatches == 0, "删除后索引残留结果。"); using var cancellation = new CancellationTokenSource(); cancellation.Cancel(); try { index.Search("anything", cancellationToken: cancellation.Token); throw new InvalidOperationException("搜索应遵循取消令牌。"); } catch (OperationCanceledException) { } } [Test] public static void FileIndexCandidatesAgreeWithFullMatching() { var entries = Enumerable.Range(0, 3000).Select(number => new FileSearchEntry( $@"C:\共享项目\group{number % 37}\{(number % 2 == 0 ? "中文" : "report")}{number:D4}.{(number % 3 == 0 ? "cs" : "json")}", false)).ToArray(); var index = new FileSearchIndex(); index.AddOrUpdate(entries); foreach (var text in new[] { "项目", "中", "中文", "report17*", "path:group12", "group1 ext:cs;json", "report | 中文 ext:cs", "!report ext:json", "*01?.[cj]*", "C:/共享项目/group12/*", "file:report*;中文*", "folder:group1", "report 17", "*.cs !group1", "[[]", "json | !*.cs" }) { var parsed = FileSearchQuery.Parse(text); Ensure(parsed.Error is null, $"有效查询错误:{text},{parsed.Error}"); var expected = entries.Where(parsed.Matches).Select(entry => entry.FullPath).Order(StringComparer.OrdinalIgnoreCase).ToArray(); var actual = index.Search(text, limit: entries.Length); Ensure(actual.TotalMatches == expected.Length, $"候选过滤丢失/增加匹配:{text},期望 {expected.Length},实际 {actual.TotalMatches}。"); Ensure(actual.Items.Select(match => match.Entry.FullPath).Order(StringComparer.OrdinalIgnoreCase).SequenceEqual(expected), $"候选过滤结果错误:{text}。"); } } [Test] public static void FileIndexCompactsTombstonesAndSupportsConcurrentReadsAndUpdates() { var index = new FileSearchIndex(); index.AddOrUpdate(Enumerable.Range(0, 9000).Select(number => new FileSearchEntry($@"C:\batch\item{number:D5}.txt", false))); var reader = Task.Run(() => { for (var iteration = 0; iteration < 12; iteration++) { var result = index.Search("item", limit: 5); Ensure(result.TotalMatches is >= 4500 and <= 9000 && result.Items.Count == 5, "并发读取看到了不完整的索引代次。"); } }); for (var number = 0; number < 4500; number++) index.RemovePath($@"C:\batch\item{number:D5}.txt", false); reader.GetAwaiter().GetResult(); Ensure(index.Count == 4500 && index.Search("item").TotalMatches == 4500, "墓碑压缩后匹配数错误。"); Ensure(index.Search("item00001").TotalMatches == 0 && index.Search("item08999").TotalMatches == 1, "墓碑压缩后倒排 ID 错位。"); index.AddOrUpdate([new(@"C:\batch\item00001.txt", false)]); Ensure(index.Search("item00001").TotalMatches == 1, "重新添加已删除条目失败。"); } [Test] public static void FileIndexSearchesOneHundredThousandNamesQuickly() { const int count = 100_000; var index = new FileSearchIndex(); var watch = Stopwatch.StartNew(); index.AddOrUpdate(Enumerable.Range(0, count).Select(number => new FileSearchEntry( $@"C:\Workspace\Project{number % 250:D3}\{(number % 100 == 0 ? "中文配置" : "document")}_{number:D6}.{(number % 4 == 0 ? "cs" : "json")}", false))); var buildTime = watch.Elapsed; var timings = new List(); for (var iteration = 0; iteration < 30; iteration++) { var number = 49_001 + iteration; var result = index.Search($"document_{number:D6}"); Ensure(result.TotalMatches == 1, "十万文件精确片段搜索错误。"); timings.Add(result.Elapsed.TotalMilliseconds); } var directory = index.Search("Project123"); Ensure(directory.TotalMatches == 400, "共享目录倒排索引匹配丢失。"); Ensure(index.Search("中文").TotalMatches == 1000, "大索引短中文关键词匹配丢失。"); using var cancellation = new CancellationTokenSource(); cancellation.CancelAfter(TimeSpan.FromMilliseconds(1)); try { // A faster matcher may finish before the timer callback is scheduled. Completion // is valid in that race; only a partial successful result would be a regression. var completed = index.Search("*", cancellationToken: cancellation.Token); Ensure(completed.TotalMatches == count, "取消与完成竞争时返回了不完整的大范围结果。"); } catch (OperationCanceledException) { } timings.Sort(); Ensure(timings[^1] < 3000, $"十万文件选择性查询严重退化:{timings[^1]:N1} ms。"); Console.WriteLine($"文件索引:{count:N0} 项,构建 {buildTime.TotalMilliseconds:N0} ms,选择性搜索中位 {timings[timings.Count / 2]:N2} ms,最大 {timings[^1]:N2} ms,目录查询 {directory.Elapsed.TotalMilliseconds:N2} ms。"); } private static void Ensure(bool condition, string message) { if (!condition) throw new InvalidOperationException(message); } }