launcher/src/cli.rs

use clap::{Parser, Subcommand};

#[derive(Parser)]
#[command(author, version, about = "DGX Spark Launcher для vLLM и SGLang")]
pub struct Cli {
    #[arg(long, short, default_value = "config.toml")]
    pub config: String,

    #[arg(long, short, help = "Показать команды без выполнения")]
    pub dry_run: bool,

    #[command(subcommand)]
    pub command: Commands,
}

#[derive(Subcommand)]
pub enum Commands {
    /// Запустить кластер
    Serve {
        #[arg(long, required = true)]
        recipe: String,

        #[arg(long, default_value = "auto")]
        cluster: String,

        /// Keep containers after stop (default: auto-remove on stop)
        #[arg(long, default_value = "false")]
        keep: bool,

        /// Skip pre-serve hooks (cluster-level + recipe-level).
        #[arg(long, default_value = "false")]
        skip_before_serve: bool,
    },

    /// Показать логи контейнеров
    Logs {
        #[arg(long, required = true)]
        recipe: String,

        #[arg(long, action = clap::ArgAction::SetTrue)]
        follow: bool,
    },

    /// Остановить кластер
    Stop {
        #[arg(long, required = true)]
        recipe: String,
    },

    /// Проверить здоровье кластера
    Health {
        #[arg(
            long,
            help = "Имя кластера (например, 4nodes-switch). Если не указано - проверяет все."
        )]
        cluster: Option<String>,
        #[arg(
            long,
            help = "Показать хост-ключи узлов (для добавления в config.toml)"
        )]
        fingerprint: bool,
    },

    /// Проверить и исправить рецепт (doctor mode)
    Doctor {
        #[arg(long, required = true, help = "Имя рецепта для проверки")]
        recipe: String,
        #[arg(
            long,
            default_value = "auto",
            help = "Кластер (если не auto — использует recipe.cluster)"
        )]
        cluster: String,
        #[arg(
            long,
            help = "Автоматически исправить проблемы (создать Dockerfile + собрать образ)"
        )]
        repair: bool,
    },

    /// Управление tmux-сессиями (альтернатива Docker)
    #[command(subcommand)]
    Tmux(TmuxCommands),

    /// Auto-discover host keys and write them into config.toml
    InitHostKeys {
        #[arg(
            long,
            help = "Имя кластера (например, 4nodes-switch). Если не указано — все."
        )]
        cluster: Option<String>,
    },

    /// Run an arbitrary shell command on a cluster or single node
    Exec {
        /// Cluster name (mutually exclusive with --recipe).
        #[arg(long, conflicts_with = "recipe")]
        cluster: Option<String>,

        /// Recipe name; resolves to the recipe's pinned cluster.
        #[arg(long, conflicts_with = "cluster")]
        recipe: Option<String>,

        /// Target a specific node within the cluster (default: all).
        #[arg(long)]
        node: Option<String>,

        /// Name of an alias from `[exec_aliases]` in config.toml.
        /// Mutually exclusive with a trailing command.
        #[arg(long)]
        alias: Option<String>,

        /// Command to run (everything after `--`).
        #[arg(trailing_var_arg = true)]
        command: Vec<String>,
    },

    /// Power on/off nodes via Home Assistant (smart plugs / Wake-on-LAN).
    Power {
        #[command(subcommand)]
        action: PowerAction,
    },
}

#[derive(Subcommand)]
pub enum PowerAction {
    /// Power off: SSH `sudo shutdown now` + wait + HA `turn_off`.
    Off {
        #[arg(long)]
        cluster: String,

        /// Target a specific node within the cluster (default: all).
        #[arg(long)]
        node: Option<String>,
    },
    /// Power on: HA `turn_on` only (no SSH; the smart plug wakes the machine).
    On {
        #[arg(long)]
        cluster: String,

        /// Target a specific node within the cluster (default: all).
        #[arg(long)]
        node: Option<String>,
    },
    /// Switch clusters: disable nodes from source, enable nodes in target.
    Switch {
        /// Current cluster (nodes + entities to shut down).
        #[arg(long)]
        from: String,
        /// Target cluster (nodes + entities to wake up).
        #[arg(long)]
        to: String,
    },
}

#[derive(Subcommand)]
pub enum TmuxCommands {
    /// Запустить кластер через tmux
    Serve {
        #[arg(long, required = true)]
        recipe: String,
        #[arg(long, default_value = "auto")]
        cluster: String,
        /// Skip pre-serve hooks (cluster-level + recipe-level).
        #[arg(long, default_value = "false")]
        skip_before_serve: bool,
    },
    /// Остановить tmux-сессии
    Stop {
        #[arg(long, required = true)]
        recipe: String,
        #[arg(long, default_value = "auto")]
        cluster: String,
    },
    /// Статус tmux-сессий (подробный)
    Status {
        #[arg(long, required = true)]
        recipe: String,
        #[arg(long, default_value = "auto")]
        cluster: String,
    },
    /// Показать логи tmux-сессий
    Logs {
        #[arg(long, required = true)]
        recipe: String,
        #[arg(long, action = clap::ArgAction::SetTrue)]
        follow: bool,
        #[arg(long, default_value = "auto")]
        cluster: String,
    },
    /// Подключиться к tmux-сессии головного узла
    Attach {
        #[arg(long, required = true)]
        recipe: String,
        #[arg(long, default_value = "auto")]
        cluster: String,
    },
}