git.stella-ops.org/bench/Scanner.Analyzers/StellaOps.Bench.ScannerAnalyzers/ScenarioRunners.cs

using System.Diagnostics;
using System.Text;
using System.Linq;
using System.Text.Json;
using System.Text.RegularExpressions;
using StellaOps.Scanner.Analyzers.Lang;
using StellaOps.Scanner.Analyzers.Lang.Go;
using StellaOps.Scanner.Analyzers.Lang.Java;
using StellaOps.Scanner.Analyzers.Lang.Node;
using StellaOps.Scanner.Analyzers.Lang.DotNet;

namespace StellaOps.Bench.ScannerAnalyzers.Scenarios;

internal interface IScenarioRunner
{
    Task<ScenarioExecutionResult> ExecuteAsync(string rootPath, int iterations, CancellationToken cancellationToken);
}

internal sealed record ScenarioExecutionResult(double[] Durations, int SampleCount);

internal static class ScenarioRunnerFactory
{
    public static IScenarioRunner Create(BenchmarkScenarioConfig scenario)
    {
        if (scenario.HasAnalyzers)
        {
            return new LanguageAnalyzerScenarioRunner(scenario.Analyzers!);
        }

        if (string.IsNullOrWhiteSpace(scenario.Parser) || string.IsNullOrWhiteSpace(scenario.Matcher))
        {
            throw new InvalidOperationException($"Scenario '{scenario.Id}' missing parser or matcher configuration.");
        }

        return new MetadataWalkScenarioRunner(scenario.Parser, scenario.Matcher);
    }
}

internal sealed class LanguageAnalyzerScenarioRunner : IScenarioRunner
{
    private readonly IReadOnlyList<Func<ILanguageAnalyzer>> _analyzerFactories;

    public LanguageAnalyzerScenarioRunner(IEnumerable<string> analyzerIds)
    {
        if (analyzerIds is null)
        {
            throw new ArgumentNullException(nameof(analyzerIds));
        }

        _analyzerFactories = analyzerIds
            .Where(static id => !string.IsNullOrWhiteSpace(id))
            .Select(CreateFactory)
            .ToArray();

        if (_analyzerFactories.Count == 0)
        {
            throw new InvalidOperationException("At least one analyzer id must be provided.");
        }
    }

    public async Task<ScenarioExecutionResult> ExecuteAsync(string rootPath, int iterations, CancellationToken cancellationToken)
    {
        if (iterations <= 0)
        {
            throw new ArgumentOutOfRangeException(nameof(iterations), iterations, "Iterations must be positive.");
        }

        var analyzers = _analyzerFactories.Select(factory => factory()).ToArray();
        var engine = new LanguageAnalyzerEngine(analyzers);
        var durations = new double[iterations];
        var componentCount = -1;

        for (var i = 0; i < iterations; i++)
        {
            cancellationToken.ThrowIfCancellationRequested();

            var context = new LanguageAnalyzerContext(rootPath, TimeProvider.System);
            var stopwatch = Stopwatch.StartNew();
            var result = await engine.AnalyzeAsync(context, cancellationToken).ConfigureAwait(false);
            stopwatch.Stop();

            durations[i] = stopwatch.Elapsed.TotalMilliseconds;

            var currentCount = result.Components.Count;
            if (componentCount < 0)
            {
                componentCount = currentCount;
            }
            else if (componentCount != currentCount)
            {
                throw new InvalidOperationException($"Analyzer output count changed between iterations ({componentCount} vs {currentCount}).");
            }
        }

        if (componentCount < 0)
        {
            componentCount = 0;
        }

        return new ScenarioExecutionResult(durations, componentCount);
    }

    private static Func<ILanguageAnalyzer> CreateFactory(string analyzerId)
    {
        var id = analyzerId.Trim().ToLowerInvariant();
        return id switch
        {
            "java" => static () => new JavaLanguageAnalyzer(),
            "go" => static () => new GoLanguageAnalyzer(),
            "node" => static () => new NodeLanguageAnalyzer(),
            "dotnet" => static () => new DotNetLanguageAnalyzer(),
            _ => throw new InvalidOperationException($"Unsupported analyzer '{analyzerId}'."),
        };
    }
}

internal sealed class MetadataWalkScenarioRunner : IScenarioRunner
{
    private readonly Regex _matcher;
    private readonly string _parserKind;

    public MetadataWalkScenarioRunner(string parserKind, string globPattern)
    {
        _parserKind = parserKind?.Trim().ToLowerInvariant() ?? throw new ArgumentNullException(nameof(parserKind));
        _matcher = GlobToRegex(globPattern ?? throw new ArgumentNullException(nameof(globPattern)));
    }

    public async Task<ScenarioExecutionResult> ExecuteAsync(string rootPath, int iterations, CancellationToken cancellationToken)
    {
        if (iterations <= 0)
        {
            throw new ArgumentOutOfRangeException(nameof(iterations), iterations, "Iterations must be positive.");
        }

        var durations = new double[iterations];
        var sampleCount = -1;

        for (var i = 0; i < iterations; i++)
        {
            cancellationToken.ThrowIfCancellationRequested();

            var stopwatch = Stopwatch.StartNew();
            var files = EnumerateMatchingFiles(rootPath);
            if (files.Count == 0)
            {
                throw new InvalidOperationException($"Parser '{_parserKind}' matched zero files under '{rootPath}'.");
            }

            foreach (var file in files)
            {
                cancellationToken.ThrowIfCancellationRequested();
                await ParseAsync(file).ConfigureAwait(false);
            }

            stopwatch.Stop();
            durations[i] = stopwatch.Elapsed.TotalMilliseconds;

            if (sampleCount < 0)
            {
                sampleCount = files.Count;
            }
            else if (sampleCount != files.Count)
            {
                throw new InvalidOperationException($"File count changed between iterations ({sampleCount} vs {files.Count}).");
            }
        }

        if (sampleCount < 0)
        {
            sampleCount = 0;
        }

        return new ScenarioExecutionResult(durations, sampleCount);
    }

    private async ValueTask ParseAsync(string filePath)
    {
        switch (_parserKind)
        {
            case "node":
                {
                    using var stream = File.OpenRead(filePath);
                    using var document = await JsonDocument.ParseAsync(stream).ConfigureAwait(false);

                    if (!document.RootElement.TryGetProperty("name", out var name) || name.ValueKind != JsonValueKind.String)
                    {
                        throw new InvalidOperationException($"package.json '{filePath}' missing name.");
                    }

                    if (!document.RootElement.TryGetProperty("version", out var version) || version.ValueKind != JsonValueKind.String)
                    {
                        throw new InvalidOperationException($"package.json '{filePath}' missing version.");
                    }
                }
                break;
            case "python":
                {
                    var (name, version) = await ParsePythonMetadataAsync(filePath).ConfigureAwait(false);
                    if (string.IsNullOrEmpty(name) || string.IsNullOrEmpty(version))
                    {
                        throw new InvalidOperationException($"METADATA '{filePath}' missing Name/Version.");
                    }
                }
                break;
            default:
                throw new InvalidOperationException($"Unknown parser '{_parserKind}'.");
        }
    }

    private static async Task<(string? Name, string? Version)> ParsePythonMetadataAsync(string filePath)
    {
        using var stream = new FileStream(filePath, FileMode.Open, FileAccess.Read, FileShare.Read | FileShare.Delete);
        using var reader = new StreamReader(stream);

        string? name = null;
        string? version = null;

        while (await reader.ReadLineAsync().ConfigureAwait(false) is { } line)
        {
            if (line.StartsWith("Name:", StringComparison.OrdinalIgnoreCase))
            {
                name ??= line[5..].Trim();
            }
            else if (line.StartsWith("Version:", StringComparison.OrdinalIgnoreCase))
            {
                version ??= line[8..].Trim();
            }

            if (!string.IsNullOrEmpty(name) && !string.IsNullOrEmpty(version))
            {
                break;
            }
        }

        return (name, version);
    }

    private IReadOnlyList<string> EnumerateMatchingFiles(string rootPath)
    {
        var files = new List<string>();
        var stack = new Stack<string>();
        stack.Push(rootPath);

        while (stack.Count > 0)
        {
            var current = stack.Pop();
            foreach (var directory in Directory.EnumerateDirectories(current))
            {
                stack.Push(directory);
            }

            foreach (var file in Directory.EnumerateFiles(current))
            {
                var relative = Path.GetRelativePath(rootPath, file).Replace('\\', '/');
                if (_matcher.IsMatch(relative))
                {
                    files.Add(file);
                }
            }
        }

        return files;
    }

    private static Regex GlobToRegex(string pattern)
    {
        if (string.IsNullOrWhiteSpace(pattern))
        {
            throw new ArgumentException("Glob pattern is required.", nameof(pattern));
        }

        var normalized = pattern.Replace("\\", "/");
        normalized = normalized.Replace("**", "\u0001");
        normalized = normalized.Replace("*", "\u0002");

        var escaped = Regex.Escape(normalized);
        escaped = escaped.Replace("\u0001/", "(?:.*/)?", StringComparison.Ordinal);
        escaped = escaped.Replace("\u0001", ".*", StringComparison.Ordinal);
        escaped = escaped.Replace("\u0002", "[^/]*", StringComparison.Ordinal);

        return new Regex("^" + escaped + "$", RegexOptions.Compiled | RegexOptions.CultureInvariant);
    }
}