CBDDC/src/ZB.MOM.WW.CBDDC.Persistence/Surreal/CBDDCSurrealSchemaInitializer.cs

using System.Text.RegularExpressions;
using Microsoft.Extensions.Logging;
using Microsoft.Extensions.Logging.Abstractions;

namespace ZB.MOM.WW.CBDDC.Persistence.Surreal;

/// <summary>
///     Initializes Surreal schema objects required by CBDDC persistence stores.
/// </summary>
public sealed class CBDDCSurrealSchemaInitializer : ICBDDCSurrealSchemaInitializer
{
    private static readonly Regex IdentifierRegex = new("^[A-Za-z_][A-Za-z0-9_]*$", RegexOptions.Compiled);
    private readonly SemaphoreSlim _initializeGate = new(1, 1);
    private readonly ICBDDCSurrealEmbeddedClient _surrealClient;
    private readonly ILogger<CBDDCSurrealSchemaInitializer> _logger;
    private readonly string _checkpointTable;
    private readonly string _changefeedRetentionLiteral;
    private bool _initialized;

    /// <summary>
    ///     Initializes a new instance of the <see cref="CBDDCSurrealSchemaInitializer" /> class.
    /// </summary>
    /// <param name="surrealClient">Surreal client abstraction.</param>
    /// <param name="options">Embedded options.</param>
    /// <param name="logger">Optional logger.</param>
    public CBDDCSurrealSchemaInitializer(
        ICBDDCSurrealEmbeddedClient surrealClient,
        CBDDCSurrealEmbeddedOptions options,
        ILogger<CBDDCSurrealSchemaInitializer>? logger = null)
    {
        _surrealClient = surrealClient ?? throw new ArgumentNullException(nameof(surrealClient));
        _logger = logger ?? NullLogger<CBDDCSurrealSchemaInitializer>.Instance;
        if (options == null) throw new ArgumentNullException(nameof(options));
        if (options.Cdc == null) throw new ArgumentException("CDC options are required.", nameof(options));

        _checkpointTable = EnsureValidIdentifier(options.Cdc.CheckpointTable, nameof(options.Cdc.CheckpointTable));
        _changefeedRetentionLiteral = ToSurrealDurationLiteral(
            options.Cdc.RetentionDuration,
            nameof(options.Cdc.RetentionDuration));
    }

    /// <inheritdoc />
    public async Task EnsureInitializedAsync(CancellationToken cancellationToken = default)
    {
        if (_initialized) return;

        await _initializeGate.WaitAsync(cancellationToken);
        try
        {
            if (_initialized) return;

            string schemaSql = BuildSchemaSql();
            await _surrealClient.RawQueryAsync(schemaSql, cancellationToken: cancellationToken);

            _initialized = true;
            _logger.LogInformation(
                "Surreal schema initialized with checkpoint table '{CheckpointTable}'.",
                _checkpointTable);
        }
        finally
        {
            _initializeGate.Release();
        }
    }

    private string BuildSchemaSql()
    {
        return $"""
                DEFINE TABLE OVERWRITE {CBDDCSurrealSchemaNames.OplogEntriesTable} SCHEMAFULL CHANGEFEED {_changefeedRetentionLiteral};
                DEFINE INDEX OVERWRITE {CBDDCSurrealSchemaNames.OplogHashIndex} ON TABLE {CBDDCSurrealSchemaNames.OplogEntriesTable} COLUMNS datasetId, hash UNIQUE;
                DEFINE INDEX OVERWRITE {CBDDCSurrealSchemaNames.OplogHlcIndex} ON TABLE {CBDDCSurrealSchemaNames.OplogEntriesTable} COLUMNS datasetId, timestampPhysicalTime, timestampLogicalCounter, timestampNodeId;
                DEFINE INDEX OVERWRITE {CBDDCSurrealSchemaNames.OplogCollectionIndex} ON TABLE {CBDDCSurrealSchemaNames.OplogEntriesTable} COLUMNS datasetId, collection;

                DEFINE TABLE OVERWRITE {CBDDCSurrealSchemaNames.SnapshotMetadataTable} SCHEMAFULL CHANGEFEED {_changefeedRetentionLiteral};
                DEFINE INDEX OVERWRITE {CBDDCSurrealSchemaNames.SnapshotNodeIdIndex} ON TABLE {CBDDCSurrealSchemaNames.SnapshotMetadataTable} COLUMNS datasetId, nodeId UNIQUE;
                DEFINE INDEX OVERWRITE {CBDDCSurrealSchemaNames.SnapshotHlcIndex} ON TABLE {CBDDCSurrealSchemaNames.SnapshotMetadataTable} COLUMNS datasetId, timestampPhysicalTime, timestampLogicalCounter;

                DEFINE TABLE OVERWRITE {CBDDCSurrealSchemaNames.RemotePeerConfigurationsTable} SCHEMAFULL CHANGEFEED {_changefeedRetentionLiteral};
                DEFINE INDEX OVERWRITE {CBDDCSurrealSchemaNames.PeerNodeIdIndex} ON TABLE {CBDDCSurrealSchemaNames.RemotePeerConfigurationsTable} COLUMNS nodeId UNIQUE;
                DEFINE INDEX OVERWRITE {CBDDCSurrealSchemaNames.PeerEnabledIndex} ON TABLE {CBDDCSurrealSchemaNames.RemotePeerConfigurationsTable} COLUMNS isEnabled;

                DEFINE TABLE OVERWRITE {CBDDCSurrealSchemaNames.DocumentMetadataTable} SCHEMAFULL CHANGEFEED {_changefeedRetentionLiteral};
                DEFINE INDEX OVERWRITE {CBDDCSurrealSchemaNames.DocumentMetadataCollectionKeyIndex} ON TABLE {CBDDCSurrealSchemaNames.DocumentMetadataTable} COLUMNS datasetId, collection, key UNIQUE;
                DEFINE INDEX OVERWRITE {CBDDCSurrealSchemaNames.DocumentMetadataHlcIndex} ON TABLE {CBDDCSurrealSchemaNames.DocumentMetadataTable} COLUMNS datasetId, hlcPhysicalTime, hlcLogicalCounter, hlcNodeId;
                DEFINE INDEX OVERWRITE {CBDDCSurrealSchemaNames.DocumentMetadataCollectionIndex} ON TABLE {CBDDCSurrealSchemaNames.DocumentMetadataTable} COLUMNS datasetId, collection;

                DEFINE TABLE OVERWRITE {CBDDCSurrealSchemaNames.PeerOplogConfirmationsTable} SCHEMAFULL CHANGEFEED {_changefeedRetentionLiteral};
                DEFINE INDEX OVERWRITE {CBDDCSurrealSchemaNames.PeerConfirmationPairIndex} ON TABLE {CBDDCSurrealSchemaNames.PeerOplogConfirmationsTable} COLUMNS datasetId, peerNodeId, sourceNodeId UNIQUE;
                DEFINE INDEX OVERWRITE {CBDDCSurrealSchemaNames.PeerConfirmationActiveIndex} ON TABLE {CBDDCSurrealSchemaNames.PeerOplogConfirmationsTable} COLUMNS datasetId, isActive;
                DEFINE INDEX OVERWRITE {CBDDCSurrealSchemaNames.PeerConfirmationSourceHlcIndex} ON TABLE {CBDDCSurrealSchemaNames.PeerOplogConfirmationsTable} COLUMNS datasetId, sourceNodeId, confirmedWall, confirmedLogic;

                DEFINE TABLE OVERWRITE {_checkpointTable} SCHEMAFULL;
                DEFINE INDEX OVERWRITE {CBDDCSurrealSchemaNames.CdcCheckpointConsumerIndex} ON TABLE {_checkpointTable} COLUMNS datasetId, consumerId UNIQUE;
                DEFINE INDEX OVERWRITE {CBDDCSurrealSchemaNames.CdcCheckpointVersionstampIndex} ON TABLE {_checkpointTable} COLUMNS datasetId, versionstampCursor;
                """;
    }

    private static string EnsureValidIdentifier(string? identifier, string argumentName)
    {
        if (string.IsNullOrWhiteSpace(identifier))
            throw new ArgumentException("Surreal identifier is required.", argumentName);

        if (!IdentifierRegex.IsMatch(identifier))
            throw new ArgumentException(
                $"Invalid Surreal identifier '{identifier}'. Use letters, numbers, and underscores only.",
                argumentName);

        return identifier;
    }

    private static string ToSurrealDurationLiteral(TimeSpan duration, string argumentName)
    {
        if (duration <= TimeSpan.Zero)
            throw new ArgumentOutOfRangeException(argumentName, "Surreal changefeed retention duration must be positive.");

        if (duration.TotalDays >= 1 && duration.TotalDays == Math.Truncate(duration.TotalDays))
            return $"{(long)duration.TotalDays}d";

        if (duration.TotalHours >= 1 && duration.TotalHours == Math.Truncate(duration.TotalHours))
            return $"{(long)duration.TotalHours}h";

        if (duration.TotalMinutes >= 1 && duration.TotalMinutes == Math.Truncate(duration.TotalMinutes))
            return $"{(long)duration.TotalMinutes}m";

        if (duration.TotalSeconds >= 1 && duration.TotalSeconds == Math.Truncate(duration.TotalSeconds))
            return $"{(long)duration.TotalSeconds}s";

        long totalMs = checked((long)Math.Ceiling(duration.TotalMilliseconds));
        return $"{totalMs}ms";
    }
}