//! LightningStor object storage server binary

use chainfire_client::Client as ChainFireClient;
use clap::Parser;
use iam_service_auth::AuthService;
use lightningstor_api::{BucketServiceServer, ObjectServiceServer};
use lightningstor_distributed::{
    DistributedConfig, ErasureCodedBackend, RedundancyMode, ReplicatedBackend, RepairQueue,
    StaticNodeRegistry,
};
use lightningstor_server::{
    config::{MetadataBackend, ObjectStorageBackend},
    metadata::MetadataStore,
    repair::{spawn_replicated_repair_worker, MetadataRepairQueue},
    s3, BucketServiceImpl, ObjectServiceImpl, ServerConfig,
};
use lightningstor_storage::{LocalFsBackend, StorageBackend};
use metrics_exporter_prometheus::PrometheusBuilder;
use std::net::SocketAddr;
use std::path::PathBuf;
use std::sync::Arc;
use std::time::{Duration, SystemTime, UNIX_EPOCH};
use tonic::transport::{Certificate, Identity, Server, ServerTlsConfig};
use tonic::{Request, Status};
use tonic_health::server::health_reporter;
use tracing_subscriber::EnvFilter;

const MAX_OBJECT_GRPC_MESSAGE_SIZE: usize = 1024 * 1024 * 1024;
const OBJECT_GRPC_INITIAL_STREAM_WINDOW: u32 = 64 * 1024 * 1024;
const OBJECT_GRPC_INITIAL_CONNECTION_WINDOW: u32 = 512 * 1024 * 1024;
const OBJECT_GRPC_KEEPALIVE_INTERVAL: Duration = Duration::from_secs(30);
const OBJECT_GRPC_KEEPALIVE_TIMEOUT: Duration = Duration::from_secs(10);
const REPLICATED_REPAIR_SCAN_INTERVAL: Duration = Duration::from_secs(5);

struct StorageRuntime {
    backend: Arc<dyn StorageBackend>,
    repair_worker: Option<tokio::task::JoinHandle<()>>,
}

/// LightningStor object storage server
#[derive(Parser, Debug)]
#[command(author, version, about, long_about = None)]
struct Args {
    /// Configuration file path
    #[arg(short, long, default_value = "lightningstor.toml")]
    config: PathBuf,

    /// gRPC address to listen on (overrides config)
    #[arg(long)]
    grpc_addr: Option<String>,

    /// S3 HTTP API address to listen on (overrides config)
    #[arg(long)]
    s3_addr: Option<String>,

    /// Log level (overrides config)
    #[arg(short, long)]
    log_level: Option<String>,

    /// ChainFire endpoint for cluster coordination (overrides config)
    #[arg(long, env = "LIGHTNINGSTOR_CHAINFIRE_ENDPOINT")]
    chainfire_endpoint: Option<String>,

    /// FlareDB endpoint for metadata and tenant data storage (overrides config)
    #[arg(long, env = "LIGHTNINGSTOR_FLAREDB_ENDPOINT")]
    flaredb_endpoint: Option<String>,

    /// Metadata backend (flaredb, postgres, sqlite)
    #[arg(long, env = "LIGHTNINGSTOR_METADATA_BACKEND")]
    metadata_backend: Option<String>,

    /// SQL database URL for metadata (required for postgres/sqlite backend)
    #[arg(long, env = "LIGHTNINGSTOR_METADATA_DATABASE_URL")]
    metadata_database_url: Option<String>,

    /// Run in single-node mode (required when metadata backend is SQLite)
    #[arg(long, env = "LIGHTNINGSTOR_SINGLE_NODE")]
    single_node: bool,

    /// Data directory for object storage (overrides config)
    #[arg(long)]
    data_dir: Option<String>,

    /// Metrics port for Prometheus scraping
    #[arg(long, default_value = "9099")]
    metrics_port: u16,
}

#[tokio::main]
async fn main() -> Result<(), Box<dyn std::error::Error>> {
    let args = Args::parse();

    // Load configuration from file or use defaults
    let mut config = if args.config.exists() {
        let contents = tokio::fs::read_to_string(&args.config).await?;
        toml::from_str(&contents)?
    } else {
        tracing::info!(
            "Config file not found: {}, using defaults",
            args.config.display()
        );
        ServerConfig::default()
    };

    // Apply command line overrides
    if let Some(grpc_addr_str) = args.grpc_addr {
        config.grpc_addr = grpc_addr_str.parse()?;
    }
    if let Some(s3_addr_str) = args.s3_addr {
        config.s3_addr = s3_addr_str.parse()?;
    }
    if let Some(log_level) = args.log_level {
        config.log_level = log_level;
    }
    if let Some(chainfire_endpoint) = args.chainfire_endpoint {
        config.chainfire_endpoint = Some(chainfire_endpoint);
    }
    if let Some(flaredb_endpoint) = args.flaredb_endpoint {
        config.flaredb_endpoint = Some(flaredb_endpoint);
    }
    if let Some(metadata_backend) = args.metadata_backend {
        config.metadata_backend = parse_metadata_backend(&metadata_backend)?;
    }
    if let Some(metadata_database_url) = args.metadata_database_url {
        config.metadata_database_url = Some(metadata_database_url);
    }
    if args.single_node {
        config.single_node = true;
    }
    if let Some(data_dir) = args.data_dir {
        config.data_dir = data_dir;
    }

    // Initialize tracing
    tracing_subscriber::fmt()
        .with_env_filter(
            EnvFilter::try_from_default_env().unwrap_or_else(|_| EnvFilter::new(&config.log_level)),
        )
        .init();

    tracing::info!("Starting LightningStor server");
    tracing::info!("  gRPC: {}", config.grpc_addr);
    tracing::info!("  S3 HTTP: {}", config.s3_addr);
    tracing::info!("  Data dir: {}", config.data_dir);

    // Initialize Prometheus metrics exporter
    let metrics_addr = format!("0.0.0.0:{}", args.metrics_port);
    let builder = PrometheusBuilder::new();
    builder
        .with_http_listener(metrics_addr.parse::<std::net::SocketAddr>()?)
        .install()
        .expect("Failed to install Prometheus metrics exporter");

    tracing::info!(
        "Prometheus metrics available at http://{}/metrics",
        metrics_addr
    );

    if let Some(endpoint) = &config.chainfire_endpoint {
        tracing::info!("  Cluster coordination: ChainFire @ {}", endpoint);
        let endpoint = endpoint.clone();
        let addr = config.grpc_addr.to_string();
        tokio::spawn(async move {
            if let Err(error) =
                register_chainfire_membership(&endpoint, "lightningstor", addr).await
            {
                tracing::warn!(error = %error, "ChainFire membership registration failed");
            }
        });
    }

    // Create metadata store from explicitly selected backend.
    let metadata = match config.metadata_backend {
        MetadataBackend::FlareDb => {
            if let Some(flaredb_endpoint) = config.flaredb_endpoint.as_deref() {
                tracing::info!("Metadata backend: FlareDB @ {}", flaredb_endpoint);
            } else {
                tracing::info!("Metadata backend: FlareDB");
            }
            Arc::new(
                MetadataStore::new_flaredb_with_pd(
                    config.flaredb_endpoint.clone(),
                    config.chainfire_endpoint.clone(),
                )
                    .await
                    .map_err(|e| format!("Failed to initialize FlareDB metadata store: {}", e))?,
            )
        }
        MetadataBackend::Postgres | MetadataBackend::Sqlite => {
            let database_url = config
                .metadata_database_url
                .as_deref()
                .ok_or_else(|| {
                    format!(
                        "metadata_database_url is required when metadata_backend={} (env: LIGHTNINGSTOR_METADATA_DATABASE_URL)",
                        metadata_backend_name(config.metadata_backend)
                    )
                })?;
            ensure_sql_backend_matches_url(config.metadata_backend, database_url)?;
            tracing::info!(
                "Metadata backend: {} @ {}",
                metadata_backend_name(config.metadata_backend),
                database_url
            );
            Arc::new(
                MetadataStore::new_sql(database_url, config.single_node)
                    .await
                    .map_err(|e| format!("Failed to initialize SQL metadata store: {}", e))?,
            )
        }
    };

    let storage_runtime = create_storage_backend(&config, metadata.clone()).await?;
    let storage = storage_runtime.backend.clone();
    let _repair_worker = storage_runtime.repair_worker;

    // Initialize IAM authentication service
    tracing::info!(
        "Connecting to IAM server at {}",
        config.auth.iam_server_addr
    );
    let auth_service = AuthService::new(&config.auth.iam_server_addr)
        .await
        .map_err(|e| format!("Failed to connect to IAM server: {}", e))?;
    let auth_service = Arc::new(auth_service);

    // Dedicated runtime for auth interceptors to avoid blocking the main async runtime
    let auth_runtime = Arc::new(tokio::runtime::Runtime::new()?);
    let make_interceptor = |auth: Arc<AuthService>| {
        let rt = auth_runtime.clone();
        move |mut req: Request<()>| -> Result<Request<()>, Status> {
            let auth = auth.clone();
            tokio::task::block_in_place(|| {
                rt.block_on(async move {
                    let tenant_context = auth.authenticate_request(&req).await?;
                    req.extensions_mut().insert(tenant_context);
                    Ok(req)
                })
            })
        }
    };

    // Create services
    let object_service =
        ObjectServiceImpl::new(storage.clone(), metadata.clone(), auth_service.clone())
            .await
            .expect("Failed to create ObjectService");
    let bucket_service = BucketServiceImpl::new(metadata.clone(), auth_service.clone())
        .await
        .expect("Failed to create BucketService");

    // Setup health service
    let (mut health_reporter, health_service) = health_reporter();
    health_reporter
        .set_serving::<ObjectServiceServer<ObjectServiceImpl>>()
        .await;
    health_reporter
        .set_serving::<BucketServiceServer<BucketServiceImpl>>()
        .await;

    // Parse addresses
    let grpc_addr: SocketAddr = config.grpc_addr;
    let s3_addr: SocketAddr = config.s3_addr;

    // Start S3 HTTP server with shared state
    let s3_router = s3::create_router_with_auth(
        storage.clone(),
        metadata.clone(),
        Some(config.auth.iam_server_addr.clone()),
    );
    let s3_server = tokio::spawn(async move {
        tracing::info!("S3 HTTP server listening on {}", s3_addr);
        let listener = tokio::net::TcpListener::bind(s3_addr).await.unwrap();
        axum::serve(listener, s3_router).await.unwrap();
    });

    // Configure TLS if enabled
    let mut server = Server::builder()
        .tcp_nodelay(true)
        .initial_stream_window_size(OBJECT_GRPC_INITIAL_STREAM_WINDOW)
        .initial_connection_window_size(OBJECT_GRPC_INITIAL_CONNECTION_WINDOW)
        .http2_keepalive_interval(Some(OBJECT_GRPC_KEEPALIVE_INTERVAL))
        .http2_keepalive_timeout(Some(OBJECT_GRPC_KEEPALIVE_TIMEOUT));

    if let Some(tls_config) = &config.tls {
        tracing::info!("TLS enabled, loading certificates...");
        let cert = tokio::fs::read(&tls_config.cert_file).await?;
        let key = tokio::fs::read(&tls_config.key_file).await?;
        let server_identity = Identity::from_pem(cert, key);

        let tls = if tls_config.require_client_cert {
            tracing::info!("mTLS enabled");
            let ca_cert = tokio::fs::read(
                tls_config
                    .ca_file
                    .as_ref()
                    .ok_or("ca_file required for mTLS")?,
            )
            .await?;
            let ca = Certificate::from_pem(ca_cert);
            ServerTlsConfig::new()
                .identity(server_identity)
                .client_ca_root(ca)
        } else {
            ServerTlsConfig::new().identity(server_identity)
        };

        server = server.tls_config(tls)?;
    }

    // Start gRPC server
    tracing::info!("gRPC server listening on {}", grpc_addr);
    let grpc_server = server
        .add_service(health_service)
        .add_service(tonic::codegen::InterceptedService::new(
            ObjectServiceServer::new(object_service)
                .max_decoding_message_size(MAX_OBJECT_GRPC_MESSAGE_SIZE)
                .max_encoding_message_size(MAX_OBJECT_GRPC_MESSAGE_SIZE),
            make_interceptor(auth_service.clone()),
        ))
        .add_service(tonic::codegen::InterceptedService::new(
            BucketServiceServer::new(bucket_service)
                .max_decoding_message_size(MAX_OBJECT_GRPC_MESSAGE_SIZE)
                .max_encoding_message_size(MAX_OBJECT_GRPC_MESSAGE_SIZE),
            make_interceptor(auth_service.clone()),
        ))
        .serve(grpc_addr);

    // Run both servers
    tokio::select! {
        result = grpc_server => {
            if let Err(e) = result {
                tracing::error!("gRPC server error: {}", e);
            }
        }
        _ = s3_server => {
            tracing::error!("S3 server unexpectedly terminated");
        }
    }

    Ok(())
}

fn parse_metadata_backend(value: &str) -> Result<MetadataBackend, Box<dyn std::error::Error>> {
    match value.trim().to_ascii_lowercase().as_str() {
        "flaredb" => Ok(MetadataBackend::FlareDb),
        "postgres" => Ok(MetadataBackend::Postgres),
        "sqlite" => Ok(MetadataBackend::Sqlite),
        other => Err(format!(
            "invalid metadata backend '{}'; expected one of: flaredb, postgres, sqlite",
            other
        )
        .into()),
    }
}

fn metadata_backend_name(backend: MetadataBackend) -> &'static str {
    match backend {
        MetadataBackend::FlareDb => "flaredb",
        MetadataBackend::Postgres => "postgres",
        MetadataBackend::Sqlite => "sqlite",
    }
}

fn ensure_sql_backend_matches_url(
    backend: MetadataBackend,
    database_url: &str,
) -> Result<(), Box<dyn std::error::Error>> {
    let normalized = database_url.trim().to_ascii_lowercase();
    match backend {
        MetadataBackend::Postgres => {
            if normalized.starts_with("postgres://") || normalized.starts_with("postgresql://") {
                Ok(())
            } else {
                Err("metadata_backend=postgres requires postgres:// or postgresql:// URL".into())
            }
        }
        MetadataBackend::Sqlite => {
            if normalized.starts_with("sqlite:") {
                Ok(())
            } else {
                Err("metadata_backend=sqlite requires sqlite: URL".into())
            }
        }
        MetadataBackend::FlareDb => Ok(()),
    }
}

async fn register_chainfire_membership(
    endpoint: &str,
    service: &str,
    addr: String,
) -> Result<(), Box<dyn std::error::Error>> {
    let node_id =
        std::env::var("HOSTNAME").unwrap_or_else(|_| format!("{}-{}", service, std::process::id()));
    let ts = SystemTime::now()
        .duration_since(UNIX_EPOCH)
        .unwrap_or_default()
        .as_secs();
    let key = format!("/cluster/{}/members/{}", service, node_id);
    let value = format!(r#"{{"addr":"{}","ts":{}}}"#, addr, ts);
    let deadline = tokio::time::Instant::now() + std::time::Duration::from_secs(120);
    let mut attempt = 0usize;
    let mut last_error = String::new();

    loop {
        attempt += 1;
        match ChainFireClient::connect(endpoint).await {
            Ok(mut client) => match client.put_str(&key, &value).await {
                Ok(_) => return Ok(()),
                Err(error) => last_error = format!("put failed: {}", error),
            },
            Err(error) => last_error = format!("connect failed: {}", error),
        }

        if tokio::time::Instant::now() >= deadline {
            break;
        }

        tracing::warn!(
            attempt,
            endpoint,
            service,
            error = %last_error,
            "retrying ChainFire membership registration"
        );
        tokio::time::sleep(std::time::Duration::from_secs(2)).await;
    }

    Err(std::io::Error::other(format!(
        "failed to register ChainFire membership for {} via {} after {} attempts: {}",
        service, endpoint, attempt, last_error
    ))
    .into())
}

async fn create_storage_backend(
    config: &ServerConfig,
    metadata: Arc<MetadataStore>,
) -> Result<StorageRuntime, Box<dyn std::error::Error>> {
    match config.object_storage_backend {
        ObjectStorageBackend::LocalFs => {
            tracing::info!("Object storage backend: local_fs");
            Ok(StorageRuntime {
                backend: Arc::new(LocalFsBackend::new(&config.data_dir, config.sync_on_write).await?),
                repair_worker: None,
            })
        }
        ObjectStorageBackend::Distributed => {
            tracing::info!("Object storage backend: distributed");
            create_distributed_storage_backend(&config.distributed, metadata).await
        }
    }
}

async fn create_distributed_storage_backend(
    config: &DistributedConfig,
    metadata: Arc<MetadataStore>,
) -> Result<StorageRuntime, Box<dyn std::error::Error>> {
    let endpoints: Vec<String> = config
        .node_endpoints
        .iter()
        .map(|endpoint| endpoint.trim().to_string())
        .filter(|endpoint| !endpoint.is_empty())
        .collect();

    if endpoints.is_empty() {
        return Err(std::io::Error::other(
            "distributed object storage requires at least one node endpoint",
        )
        .into());
    }

    let min_nodes = config.redundancy.min_nodes();
    if endpoints.len() < min_nodes {
        return Err(std::io::Error::other(format!(
            "distributed object storage requires at least {} node endpoints for the configured redundancy mode, got {}",
            min_nodes,
            endpoints.len()
        ))
        .into());
    }

    if let Some(registry_endpoint) = config.registry_endpoint.as_deref() {
        tracing::warn!(
            registry_endpoint,
            "registry_endpoint is not implemented yet; using static node_endpoints only"
        );
    }

    tracing::info!(
        node_count = endpoints.len(),
        min_nodes,
        fault_tolerance = config.redundancy.fault_tolerance(),
        connection_timeout_ms = config.connection_timeout_ms,
        request_timeout_ms = config.request_timeout_ms,
        "Initializing LightningStor distributed object storage"
    );

    let registry = Arc::new(
        StaticNodeRegistry::new_with_timeouts(
            &endpoints,
            std::time::Duration::from_millis(config.connection_timeout_ms),
            std::time::Duration::from_millis(config.request_timeout_ms),
        )
        .await?,
    );

    match &config.redundancy {
        RedundancyMode::Replicated {
            replica_count,
            read_quorum,
            write_quorum,
        } => {
            tracing::info!(
                replica_count,
                read_quorum,
                write_quorum,
                "Using replicated LightningStor storage backend"
            );
            let repair_queue: Arc<dyn RepairQueue> =
                Arc::new(MetadataRepairQueue::new(metadata.clone()));
            let backend = Arc::new(
                ReplicatedBackend::new_with_repair_queue(
                    config.clone(),
                    registry,
                    Some(repair_queue),
                )
                .await?,
            );
            let repair_worker = Some(spawn_replicated_repair_worker(
                metadata,
                backend.clone(),
                REPLICATED_REPAIR_SCAN_INTERVAL,
            ));
            Ok(StorageRuntime {
                backend,
                repair_worker,
            })
        }
        RedundancyMode::ErasureCoded {
            data_shards,
            parity_shards,
        } => {
            tracing::info!(
                data_shards,
                parity_shards,
                "Using erasure-coded LightningStor storage backend"
            );
            Ok(StorageRuntime {
                backend: Arc::new(ErasureCodedBackend::new(config.clone(), registry).await?),
                repair_worker: None,
            })
        }
        RedundancyMode::None => Err(std::io::Error::other(
            "distributed object storage does not support redundancy.type=none; use object_storage_backend=local_fs instead",
        )
        .into()),
    }
}