git.stella-ops.org/src/Scanner/__Tests/StellaOps.Scanner.Worker.Tests/CompositeScanAnalyzerDispatcherTests.cs

using System;
using System.Collections.Generic;
using System.Collections.Immutable;
using System.Collections.ObjectModel;
using System.Diagnostics.Metrics;
using System.IO;
using System.Linq;
using System.Text;
using System.Threading;
using Microsoft.Extensions.Configuration;
using Microsoft.Extensions.DependencyInjection;
using Microsoft.Extensions.Logging;
using Microsoft.Extensions.Options;
using StellaOps.Scanner.Analyzers.Lang;
using StellaOps.Scanner.Analyzers.Lang.Plugin;
using StellaOps.Scanner.Analyzers.OS;
using StellaOps.Scanner.Analyzers.OS.Abstractions;
using StellaOps.Scanner.Analyzers.OS.Plugin;
using StellaOps.Scanner.Core.Contracts;
using StellaOps.Scanner.Surface.Env;
using StellaOps.Scanner.Surface.FS;
using StellaOps.Scanner.Surface.Secrets;
using StellaOps.Scanner.Surface.Validation;
using StellaOps.Scanner.Worker.Diagnostics;
using StellaOps.Scanner.Worker.Processing;
using StellaOps.Scanner.Worker.Tests.TestInfrastructure;
using Xunit;
using WorkerOptions = StellaOps.Scanner.Worker.Options.ScannerWorkerOptions;

namespace StellaOps.Scanner.Worker.Tests;

public sealed class CompositeScanAnalyzerDispatcherTests
{
    [Fact]
    public async Task ExecuteAsync_RunsLanguageAnalyzers_StoresResults()
    {
        using var workspace = new TempDirectory();
        using var cacheRoot = new TempDirectory();

        Environment.SetEnvironmentVariable("SCANNER_SURFACE_FS_ENDPOINT", "https://surface.test");
        Environment.SetEnvironmentVariable("SCANNER_SURFACE_FS_BUCKET", "unit-test-bucket");
        Environment.SetEnvironmentVariable("SCANNER_SURFACE_CACHE_ROOT", cacheRoot.Path);
        Environment.SetEnvironmentVariable("SCANNER_SURFACE_SECRETS_PROVIDER", "inline");
        Environment.SetEnvironmentVariable("SCANNER_SURFACE_SECRETS_TENANT", "testtenant");
        Environment.SetEnvironmentVariable(
            "SURFACE_SECRET_TESTTENANT_SCANNERWORKERLANGUAGEANALYZERS_REGISTRY_DEFAULT",
            Convert.ToBase64String(Encoding.UTF8.GetBytes("token-placeholder")));

        var metadata = new Dictionary<string, string>(StringComparer.Ordinal)
        {
            { ScanMetadataKeys.RootFilesystemPath, workspace.Path },
            { ScanMetadataKeys.WorkspacePath, workspace.Path },
        };

        var osCatalog = new FakeOsCatalog();
        var analyzer = new FakeLanguageAnalyzer();
        var languageCatalog = new FakeLanguageCatalog(analyzer);

        long hits = 0;
        long misses = 0;
        MeterListener? meterListener = null;
        ServiceProvider? services = null;
        try
        {
            var serviceCollection = new ServiceCollection();
            serviceCollection.AddSingleton<IConfiguration>(new ConfigurationBuilder().Build());
            serviceCollection.AddLogging(builder => builder.SetMinimumLevel(LogLevel.Debug));
            serviceCollection.AddSingleton(TimeProvider.System);
            serviceCollection.AddSurfaceEnvironment(options => options.ComponentName = "Scanner.Worker");
            serviceCollection.AddSurfaceValidation();
            serviceCollection.AddSingleton<ISurfaceValidatorRunner, NoopSurfaceValidatorRunner>();
            serviceCollection.AddSurfaceFileCache(options => options.RootDirectory = cacheRoot.Path);
            serviceCollection.AddSurfaceSecrets();

            var metrics = new ScannerWorkerMetrics();
            serviceCollection.AddSingleton(metrics);

            meterListener = new MeterListener();

            meterListener.InstrumentPublished = (instrument, listener) =>
            {
                if (instrument.Meter.Name == ScannerWorkerInstrumentation.MeterName &&
                    (instrument.Name == "scanner_worker_language_cache_hits_total" || instrument.Name == "scanner_worker_language_cache_misses_total"))
                {
                    listener.EnableMeasurementEvents(instrument);
                }
            };

            meterListener.SetMeasurementEventCallback<long>((instrument, measurement, tags, state) =>
            {
                if (instrument.Name == "scanner_worker_language_cache_hits_total")
                {
                    Interlocked.Add(ref hits, measurement);
                }
                else if (instrument.Name == "scanner_worker_language_cache_misses_total")
                {
                    Interlocked.Add(ref misses, measurement);
                }
            });

            meterListener.Start();

            services = serviceCollection.BuildServiceProvider();

            var scopeFactory = services.GetRequiredService<IServiceScopeFactory>();
            var loggerFactory = services.GetRequiredService<ILoggerFactory>();
            var options = Microsoft.Extensions.Options.Options.Create(new WorkerOptions());
            var dispatcher = new CompositeScanAnalyzerDispatcher(
                scopeFactory,
                osCatalog,
                languageCatalog,
                options,
                loggerFactory.CreateLogger<CompositeScanAnalyzerDispatcher>(),
                metrics,
                new TestCryptoHash());

            var lease = new TestJobLease(metadata);
            var context = new ScanJobContext(lease, TimeProvider.System, TimeProvider.System.GetUtcNow(), CancellationToken.None);

            await dispatcher.ExecuteAsync(context, CancellationToken.None);

            // Re-run with a new context to exercise cache reuse.
            var leaseSecond = new TestJobLease(metadata);
            var contextSecond = new ScanJobContext(leaseSecond, TimeProvider.System, TimeProvider.System.GetUtcNow(), CancellationToken.None);
            await dispatcher.ExecuteAsync(contextSecond, CancellationToken.None);

            meterListener.RecordObservableInstruments();

            Assert.Equal(1, analyzer.InvocationCount);
            Assert.True(context.Analysis.TryGet<ReadOnlyDictionary<string, LanguageAnalyzerResult>>(ScanAnalysisKeys.LanguageAnalyzerResults, out var results));
            Assert.Single(results);
            Assert.True(context.Analysis.TryGet<ImmutableArray<LayerComponentFragment>>(ScanAnalysisKeys.LanguageComponentFragments, out var fragments));
            Assert.False(fragments.IsDefaultOrEmpty);
            Assert.True(context.Analysis.GetLayerFragments().Any(fragment => fragment.Components.Any(component => component.Identity.Name == "demo-package")));
            Assert.Equal(1, hits);
            Assert.Equal(1, misses);
        }
        finally
        {
            Environment.SetEnvironmentVariable("SCANNER_SURFACE_FS_ENDPOINT", null);
            Environment.SetEnvironmentVariable("SCANNER_SURFACE_FS_BUCKET", null);
            Environment.SetEnvironmentVariable("SCANNER_SURFACE_CACHE_ROOT", null);
            Environment.SetEnvironmentVariable("SCANNER_SURFACE_SECRETS_PROVIDER", null);
            Environment.SetEnvironmentVariable("SCANNER_SURFACE_SECRETS_TENANT", null);
            Environment.SetEnvironmentVariable("SURFACE_SECRET_TESTTENANT_SCANNERWORKERLANGUAGEANALYZERS_REGISTRY_DEFAULT", null);
            meterListener?.Dispose();
            services?.Dispose();
        }
    }

    [Fact]
    public async Task ExecuteAsync_RunsOsAnalyzers_UsesSurfaceCache()
    {
        using var rootfs = new TempDirectory();
        using var cacheRoot = new TempDirectory();

        Environment.SetEnvironmentVariable("SCANNER_SURFACE_FS_ENDPOINT", "https://surface.test");
        Environment.SetEnvironmentVariable("SCANNER_SURFACE_FS_BUCKET", "unit-test-bucket");
        Environment.SetEnvironmentVariable("SCANNER_SURFACE_CACHE_ROOT", cacheRoot.Path);
        Environment.SetEnvironmentVariable("SCANNER_SURFACE_SECRETS_PROVIDER", "inline");
        Environment.SetEnvironmentVariable("SCANNER_SURFACE_SECRETS_TENANT", "testtenant");
        Environment.SetEnvironmentVariable(
            "SURFACE_SECRET_TESTTENANT_SCANNERWORKEROSANALYZERS_REGISTRY_DEFAULT",
            Convert.ToBase64String(Encoding.UTF8.GetBytes("token-placeholder")));

        var dpkgStatusPath = Path.Combine(rootfs.Path, "var", "lib", "dpkg", "status");
        Directory.CreateDirectory(Path.GetDirectoryName(dpkgStatusPath)!);
        await File.WriteAllTextAsync(dpkgStatusPath, "Package: demo\nStatus: install ok installed\n", CancellationToken.None);

        var metadata = new Dictionary<string, string>(StringComparer.Ordinal)
        {
            { ScanMetadataKeys.RootFilesystemPath, rootfs.Path },
            { ScanMetadataKeys.WorkspacePath, rootfs.Path },
        };

        var analyzer = new FakeOsAnalyzer();
        var osCatalog = new FakeOsCatalog(analyzer);
        var languageCatalog = new FakeLanguageCatalog();

        long hits = 0;
        long misses = 0;
        MeterListener? meterListener = null;
        ServiceProvider? services = null;
        try
        {
            var serviceCollection = new ServiceCollection();
            serviceCollection.AddSingleton<IConfiguration>(new ConfigurationBuilder().Build());
            serviceCollection.AddLogging(builder => builder.SetMinimumLevel(LogLevel.Debug));
            serviceCollection.AddSingleton(TimeProvider.System);
            serviceCollection.AddSurfaceEnvironment(options => options.ComponentName = "Scanner.Worker");
            serviceCollection.AddSurfaceValidation();
            serviceCollection.AddSingleton<ISurfaceValidatorRunner, NoopSurfaceValidatorRunner>();
            serviceCollection.AddSurfaceFileCache(options => options.RootDirectory = cacheRoot.Path);
            serviceCollection.AddSurfaceSecrets();

            var metrics = new ScannerWorkerMetrics();
            serviceCollection.AddSingleton(metrics);

            meterListener = new MeterListener();

            meterListener.InstrumentPublished = (instrument, listener) =>
            {
                if (instrument.Meter.Name == ScannerWorkerInstrumentation.MeterName &&
                    (instrument.Name == "scanner_worker_os_cache_hits_total" || instrument.Name == "scanner_worker_os_cache_misses_total"))
                {
                    listener.EnableMeasurementEvents(instrument);
                }
            };

            meterListener.SetMeasurementEventCallback<long>((instrument, measurement, tags, state) =>
            {
                if (instrument.Name == "scanner_worker_os_cache_hits_total")
                {
                    Interlocked.Add(ref hits, measurement);
                }
                else if (instrument.Name == "scanner_worker_os_cache_misses_total")
                {
                    Interlocked.Add(ref misses, measurement);
                }
            });

            meterListener.Start();

            services = serviceCollection.BuildServiceProvider();

            var scopeFactory = services.GetRequiredService<IServiceScopeFactory>();
            var loggerFactory = services.GetRequiredService<ILoggerFactory>();
            var options = Microsoft.Extensions.Options.Options.Create(new WorkerOptions());
            var dispatcher = new CompositeScanAnalyzerDispatcher(
                scopeFactory,
                osCatalog,
                languageCatalog,
                options,
                loggerFactory.CreateLogger<CompositeScanAnalyzerDispatcher>(),
                metrics,
                new TestCryptoHash());

            var lease = new TestJobLease(metadata);
            var context = new ScanJobContext(lease, TimeProvider.System, TimeProvider.System.GetUtcNow(), CancellationToken.None);

            await dispatcher.ExecuteAsync(context, CancellationToken.None);

            var leaseSecond = new TestJobLease(metadata);
            var contextSecond = new ScanJobContext(leaseSecond, TimeProvider.System, TimeProvider.System.GetUtcNow(), CancellationToken.None);
            await dispatcher.ExecuteAsync(contextSecond, CancellationToken.None);

            Assert.Equal(1, analyzer.InvocationCount);
            Assert.True(context.Analysis.TryGet<Dictionary<string, OSPackageAnalyzerResult>>(ScanAnalysisKeys.OsPackageAnalyzers, out var results));
            Assert.Single(results);
            Assert.True(context.Analysis.TryGet<ImmutableArray<LayerComponentFragment>>(ScanAnalysisKeys.OsComponentFragments, out var fragments));
            Assert.False(fragments.IsDefaultOrEmpty);

            Assert.Equal(1, hits);
            Assert.Equal(1, misses);
        }
        finally
        {
            Environment.SetEnvironmentVariable("SCANNER_SURFACE_FS_ENDPOINT", null);
            Environment.SetEnvironmentVariable("SCANNER_SURFACE_FS_BUCKET", null);
            Environment.SetEnvironmentVariable("SCANNER_SURFACE_CACHE_ROOT", null);
            Environment.SetEnvironmentVariable("SCANNER_SURFACE_SECRETS_PROVIDER", null);
            Environment.SetEnvironmentVariable("SCANNER_SURFACE_SECRETS_TENANT", null);
            Environment.SetEnvironmentVariable("SURFACE_SECRET_TESTTENANT_SCANNERWORKEROSANALYZERS_REGISTRY_DEFAULT", null);
            meterListener?.Dispose();
            services?.Dispose();
        }
    }

    private sealed class FakeOsCatalog : IOSAnalyzerPluginCatalog
    {
        private readonly IReadOnlyList<IOSPackageAnalyzer> _analyzers;

        public FakeOsCatalog(params IOSPackageAnalyzer[] analyzers)
        {
            _analyzers = analyzers ?? Array.Empty<IOSPackageAnalyzer>();
        }

        public IReadOnlyCollection<IOSAnalyzerPlugin> Plugins => Array.Empty<IOSAnalyzerPlugin>();

        public void LoadFromDirectory(string directory, bool seal = true)
        {
        }

        public IReadOnlyList<IOSPackageAnalyzer> CreateAnalyzers(IServiceProvider services) => _analyzers;
    }

    private sealed class FakeLanguageCatalog : ILanguageAnalyzerPluginCatalog
    {
        private readonly IReadOnlyList<ILanguageAnalyzer> _analyzers;

        public FakeLanguageCatalog(params ILanguageAnalyzer[] analyzers)
        {
            _analyzers = analyzers ?? Array.Empty<ILanguageAnalyzer>();
        }

        public IReadOnlyCollection<ILanguageAnalyzerPlugin> Plugins => Array.Empty<ILanguageAnalyzerPlugin>();

        public void LoadFromDirectory(string directory, bool seal = true)
        {
        }

        public IReadOnlyList<ILanguageAnalyzer> CreateAnalyzers(IServiceProvider services) => _analyzers;
    }

    private sealed class NoopSurfaceValidatorRunner : ISurfaceValidatorRunner
    {
        public ValueTask<SurfaceValidationResult> RunAllAsync(SurfaceValidationContext context, CancellationToken cancellationToken = default)
            => ValueTask.FromResult(SurfaceValidationResult.Success());

        public ValueTask EnsureAsync(SurfaceValidationContext context, CancellationToken cancellationToken = default)
            => ValueTask.CompletedTask;
    }

    private sealed class FakeOsAnalyzer : IOSPackageAnalyzer
    {
        public string AnalyzerId => "dpkg";

        public int InvocationCount { get; private set; }

        private int _invocationCount;

        public ValueTask<OSPackageAnalyzerResult> AnalyzeAsync(OSPackageAnalyzerContext context, CancellationToken cancellationToken)
        {
            Interlocked.Increment(ref _invocationCount);
            InvocationCount = _invocationCount;

            var package = new OSPackageRecord(
                analyzerId: AnalyzerId,
                packageUrl: "pkg:deb/debian/demo@1.0?arch=amd64",
                name: "demo",
                version: "1.0",
                architecture: "amd64",
                evidenceSource: PackageEvidenceSource.DpkgStatus,
                files: new[] { new OSPackageFileEvidence("/usr/bin/demo") });

            var telemetry = new OSAnalyzerTelemetry(TimeSpan.Zero, 1, 1);
            return ValueTask.FromResult(new OSPackageAnalyzerResult(AnalyzerId, new[] { package }, telemetry));
        }
    }

    private sealed class FakeLanguageAnalyzer : ILanguageAnalyzer
    {
        public string Id => "lang.fake";

        public string DisplayName => "Fake Language Analyzer";

        public int InvocationCount { get; private set; }

        public ValueTask AnalyzeAsync(LanguageAnalyzerContext context, LanguageComponentWriter writer, CancellationToken cancellationToken)
        {
            Interlocked.Increment(ref _invocationCount);
            InvocationCount = _invocationCount;
            writer.AddFromPurl(
                analyzerId: Id,
                purl: "pkg:npm/demo-package@1.0.0",
                name: "demo-package",
                version: "1.0.0",
                type: "npm");
            return ValueTask.CompletedTask;
        }

        private int _invocationCount;
    }

    private sealed class TestJobLease : IScanJobLease
    {
        private readonly Dictionary<string, string> _metadata;

        public TestJobLease(Dictionary<string, string> metadata)
        {
            _metadata = metadata;
            JobId = Guid.NewGuid().ToString("n");
            ScanId = $"scan-{Guid.NewGuid():n}";
            Attempt = 1;
            EnqueuedAtUtc = DateTimeOffset.UtcNow.AddSeconds(-1);
            LeasedAtUtc = DateTimeOffset.UtcNow;
            LeaseDuration = TimeSpan.FromMinutes(5);
        }

        public string JobId { get; }

        public string ScanId { get; }

        public int Attempt { get; }

        public DateTimeOffset EnqueuedAtUtc { get; }

        public DateTimeOffset LeasedAtUtc { get; }

        public TimeSpan LeaseDuration { get; }

        public IReadOnlyDictionary<string, string> Metadata => _metadata;

        public ValueTask RenewAsync(CancellationToken cancellationToken) => ValueTask.CompletedTask;

        public ValueTask CompleteAsync(CancellationToken cancellationToken) => ValueTask.CompletedTask;

        public ValueTask AbandonAsync(string reason, CancellationToken cancellationToken) => ValueTask.CompletedTask;

        public ValueTask PoisonAsync(string reason, CancellationToken cancellationToken) => ValueTask.CompletedTask;

        public ValueTask DisposeAsync() => ValueTask.CompletedTask;
    }

    private sealed class TempDirectory : IDisposable
    {
        public TempDirectory()
        {
            Path = System.IO.Path.Combine(System.IO.Path.GetTempPath(), $"stellaops-test-{Guid.NewGuid():n}");
            Directory.CreateDirectory(Path);
        }

        public string Path { get; }

        public void Dispose()
        {
            try
            {
                if (Directory.Exists(Path))
                {
                    Directory.Delete(Path, recursive: true);
                }
            }
            catch
            {
            }
        }
    }
}