added benchmarks

ParadoxSaveParser.Benchmarks/BenchData.cs

@@ -0,0 +1,77 @@
+using System;
+using System.IO;
+using System.Text;
+using System.Text.Encodings.Web;
+using System.Text.Json;
+using System.Text.Json.Serialization;
+using ParadoxSaveParser.Lib;
+
+namespace ParadoxSaveParser.Benchmarks;
+
+/// <summary>
+///     Locates and lazily loads the benchmark corpus:
+///     a real EU4 <c>gamestate</c> file and its compact JSON twin (for jq).
+///     Override the directory with the PSP_BENCH_DATA environment variable.
+/// </summary>
+public static class BenchData
+{
+    public const string PdxQuery = "countries.*.technology";
+
+    /// <summary>jq program extracting the same data as <see cref="PdxQuery" />.</summary>
+    public const string JqQuery = ".countries | map_values(.technology)";
+
+    public static string Dir { get; } =
+        Environment.GetEnvironmentVariable("PSP_BENCH_DATA")
+        ?? Path.Combine(FindRepoRoot(), "ParadoxSaveParser.CLI", "bin", "Debug", "net10.0");
+
+    public static string PdxPath => Path.Combine(Dir, "gamestate");
+    public static string JsonPath => Path.Combine(Dir, "gamestate.min.json");
+
+    private static byte[]? _pdxBytes;
+    private static string? _pdxText;
+
+    /// <summary>Raw bytes of the save, kept in memory so benchmarks don't measure disk I/O.</summary>
+    public static byte[] PdxBytes => _pdxBytes ??= File.ReadAllBytes(PdxPath);
+
+    /// <summary>
+    ///     The save decoded as text for the regex engines.
+    ///     Latin1 is used because EU4 saves contain non-UTF8 bytes inside dynasty names,
+    ///     and it maps every byte to exactly one char.
+    /// </summary>
+    public static string PdxText => _pdxText ??= Encoding.Latin1.GetString(PdxBytes);
+
+    private static string FindRepoRoot()
+    {
+        var dir = new DirectoryInfo(AppContext.BaseDirectory);
+        while (dir != null && !File.Exists(Path.Combine(dir.FullName, "ParadoxSaveParser.sln")))
+            dir = dir.Parent;
+        return dir?.FullName ?? Directory.GetCurrentDirectory();
+    }
+
+    /// <summary>
+    ///     Produces the compact JSON twin of the save file, which is what jq is given.
+    ///     Compact (not indented) so jq isn't penalized by parsing whitespace.
+    /// </summary>
+    public static void PrepareJson()
+    {
+        if (File.Exists(JsonPath))
+        {
+            Console.WriteLine($"{JsonPath} already exists, skipping");
+            return;
+        }
+
+        Console.WriteLine($"parsing {PdxPath} ...");
+        using var input = File.OpenRead(PdxPath);
+        var root = new SaveParserEU4(input, null).Parse();
+        Console.WriteLine($"serializing to {JsonPath} ...");
+        using var output = File.Create(JsonPath);
+        JsonSerializer.Serialize(output, root, new JsonSerializerOptions
+        {
+            WriteIndented = false,
+            Encoder = JavaScriptEncoder.UnsafeRelaxedJsonEscaping,
+            MaxDepth = 1024,
+            DefaultIgnoreCondition = JsonIgnoreCondition.WhenWritingNull,
+        });
+        Console.WriteLine($"done: {new FileInfo(JsonPath).Length / 1024 / 1024} MB");
+    }
+}