Added backups of all data, restorable on another computer (#1008)

Co-authored-by: Thorsten Sommer <SommerEngineering@users.noreply.github.com>

app/MindWork AI Studio/Tools/DataBackup/DataBackupFormat.cs

@@ -0,0 +1,86 @@
+using System.Text.Json;
+using System.Text.Json.Serialization;
+
+namespace AIStudio.Tools.DataBackup;
+
+/// <summary>
+/// The format of a backup: a ZIP file which holds the data folder below DATA_ROOT and the
+/// configuration folder below CONFIG_ROOT, each in the layout AI Studio keeps on the disk, and the
+/// manifest as its last entry.
+/// </summary>
+/// <remarks>
+/// What goes in follows DataBackupContents. The paths below the roots are the relative paths of
+/// that list, so restoring a backup is unpacking it. The manifest comes last: an archive which lacks
+/// it was not written completely.
+/// </remarks>
+public static class DataBackupFormat
+{
+    /// <summary>
+    /// The version of the format this version of AI Studio writes. A restore refuses a backup of a
+    /// later version, since it cannot know what such a backup holds.
+    /// </summary>
+    public const int FORMAT_VERSION = 1;
+
+    /// <summary>
+    /// The file extension of a backup, named like the one of plugin archives.
+    /// </summary>
+    public const string FILE_EXTENSION = ".mwbackup";
+
+    /// <summary>
+    /// The manifest, the last entry of every backup.
+    /// </summary>
+    public const string MANIFEST_ENTRY_NAME = "manifest.json";
+
+    /// <summary>
+    /// Where the data folder lies within a backup.
+    /// </summary>
+    public const string DATA_ROOT = "data/";
+
+    /// <summary>
+    /// Where the configuration folder lies within a backup.
+    /// </summary>
+    public const string CONFIG_ROOT = "config/";
+
+    /// <summary>
+    /// The directory in the data folder where the copies of the vector stores wait while a backup runs.
+    /// The runtime creates it; keep the name in sync with BACKUP_STAGING_DIRECTORY there.
+    /// </summary>
+    public const string STAGING_DIRECTORY = ".backup-staging/";
+
+    /// <summary>
+    /// How the manifest and the plan of a restore are written: readable for anybody who opens them,
+    /// and for the runtime, which applies the plan.
+    /// </summary>
+    public static readonly JsonSerializerOptions JSON_OPTIONS = new()
+    {
+        WriteIndented = true,
+        PropertyNamingPolicy = JsonNamingPolicy.SnakeCaseLower,
+        Converters = { new JsonStringEnumConverter() },
+    };
+
+    /// <summary>
+    /// File types which are compressed already. Compressing them once more costs time and makes the
+    /// backup no smaller.
+    /// </summary>
+    public static readonly IReadOnlySet<string> ALREADY_COMPRESSED_EXTENSIONS = new HashSet<string>(StringComparer.OrdinalIgnoreCase)
+    {
+        ".jpg", ".jpeg", ".png", ".gif", ".webp", ".heic", ".tiff",
+        ".mp3", ".m4a", ".m4b", ".aac", ".ogg", ".opus", ".flac", ".wma",
+        ".mp4", ".m4v", ".mov", ".mkv", ".webm", ".avi", ".wmv", ".flv",
+        ".zip", ".7z", ".gz", ".bz2", ".xz", ".rar",
+        ".pdf", ".docx", ".xlsx", ".pptx", ".odt", ".odp",
+        ".mwplugin", FILE_EXTENSION,
+    };
+
+    /// <summary>
+    /// The name of the entry which holds a file of the data or configuration folder.
+    /// </summary>
+    /// <param name="location">The folder of the file.</param>
+    /// <param name="relativePath">The path of the file within its folder, with forward slashes.</param>
+    /// <returns>The name of the entry.</returns>
+    public static string EntryName(DataBackupLocation location, string relativePath) => location switch
+    {
+        DataBackupLocation.CONFIG_DIRECTORY => CONFIG_ROOT + relativePath,
+        _ => DATA_ROOT + relativePath,
+    };
+}