tidy: clean up old benchmark and add gym (#7081)

crates/goose-cli/src/cli.rs

@@ -10,7 +10,6 @@ use goose_mcp::{
     AutoVisualiserRouter, ComputerControllerServer, DeveloperServer, MemoryServer, TutorialServer,
 };
 
-use crate::commands::bench::agent_generator;
 use crate::commands::configure::{configure_telemetry_consent_dialog, handle_configure};
 use crate::commands::info::handle_info;
 use crate::commands::project::{handle_project_default, handle_projects_interactive};
@@ -31,11 +30,6 @@ use crate::session::{build_session, SessionBuilderConfig};
 use goose::agents::Container;
 use goose::session::session_manager::SessionType;
 use goose::session::SessionManager;
-use goose_bench::bench_config::BenchRunConfig;
-use goose_bench::runners::bench_runner::BenchRunner;
-use goose_bench::runners::eval_runner::EvalRunner;
-use goose_bench::runners::metric_aggregator::MetricAggregator;
-use goose_bench::runners::model_runner::ModelRunner;
 use std::io::Read;
 use std::path::PathBuf;
 use tracing::warn;
@@ -598,60 +592,6 @@ enum SchedulerCommand {
     CronHelp {},
 }
 
-#[derive(Subcommand)]
-pub enum BenchCommand {
-    #[command(name = "init-config", about = "Create a new starter-config")]
-    InitConfig {
-        #[arg(short, long, help = "filename with extension for generated config")]
-        name: String,
-    },
-
-    #[command(about = "Run all benchmarks from a config")]
-    Run {
-        #[arg(
-            short,
-            long,
-            help = "A config file generated by the config-init command"
-        )]
-        config: PathBuf,
-    },
-
-    #[command(about = "List all available selectors")]
-    Selectors {
-        #[arg(
-            short,
-            long,
-            help = "A config file generated by the config-init command"
-        )]
-        config: Option<PathBuf>,
-    },
-
-    #[command(name = "eval-model", about = "Run an eval of model")]
-    EvalModel {
-        #[arg(short, long, help = "A serialized config file for the model only.")]
-        config: String,
-    },
-
-    #[command(name = "exec-eval", about = "run a single eval")]
-    ExecEval {
-        #[arg(short, long, help = "A serialized config file for the eval only.")]
-        config: String,
-    },
-
-    #[command(
-        name = "generate-leaderboard",
-        about = "Generate a leaderboard CSV from benchmark results"
-    )]
-    GenerateLeaderboard {
-        #[arg(
-            short,
-            long,
-            help = "Path to the benchmark directory containing model evaluation results"
-        )]
-        benchmark_dir: PathBuf,
-    },
-}
-
 #[derive(Subcommand)]
 enum RecipeCommand {
     /// Validate a recipe file
@@ -862,13 +802,6 @@ enum Command {
         reconfigure: bool,
     },
 
-    /// Evaluate system configuration across a range of practical tasks
-    #[command(about = "Evaluate system configuration across a range of practical tasks")]
-    Bench {
-        #[command(subcommand)]
-        cmd: BenchCommand,
-    },
-
     /// Start a web server with a chat interface
     #[command(about = "Experimental: Start a web server with a chat interface")]
     Web {
@@ -1018,7 +951,6 @@ fn get_command_name(command: &Option<Command>) -> &'static str {
         Some(Command::Run { .. }) => "run",
         Some(Command::Schedule { .. }) => "schedule",
         Some(Command::Update { .. }) => "update",
-        Some(Command::Bench { .. }) => "bench",
         Some(Command::Recipe { .. }) => "recipe",
         Some(Command::Web { .. }) => "web",
         Some(Command::Term { .. }) => "term",
@@ -1029,7 +961,7 @@ fn get_command_name(command: &Option<Command>) -> &'static str {
 
 async fn handle_mcp_command(server: McpCommand) -> Result<()> {
     let name = server.name();
-    crate::logging::setup_logging(Some(&format!("mcp-{name}")), None)?;
+    let _ = crate::logging::setup_logging(Some(&format!("mcp-{name}")));
     match server {
         McpCommand::AutoVisualiser => serve(AutoVisualiserRouter::new()).await?,
         McpCommand::ComputerController => serve(ComputerControllerServer::new()).await?,
@@ -1426,25 +1358,6 @@ async fn handle_schedule_command(command: SchedulerCommand) -> Result<()> {
     }
 }
 
-async fn handle_bench_command(cmd: BenchCommand) -> Result<()> {
-    match cmd {
-        BenchCommand::Selectors { config } => BenchRunner::list_selectors(config)?,
-        BenchCommand::InitConfig { name } => {
-            let mut config = BenchRunConfig::default();
-            let cwd = std::env::current_dir()?;
-            config.output_dir = Some(cwd);
-            config.save(name);
-        }
-        BenchCommand::Run { config } => BenchRunner::new(config)?.run()?,
-        BenchCommand::EvalModel { config } => ModelRunner::from(config)?.run()?,
-        BenchCommand::ExecEval { config } => EvalRunner::from(config)?.run(agent_generator).await?,
-        BenchCommand::GenerateLeaderboard { benchmark_dir } => {
-            MetricAggregator::generate_csv_from_benchmark_dir(&benchmark_dir)?
-        }
-    }
-    Ok(())
-}
-
 fn handle_recipe_subcommand(command: RecipeCommand) -> Result<()> {
     match command {
         RecipeCommand::Validate { recipe_name } => handle_validate(&recipe_name),
@@ -1597,7 +1510,6 @@ pub async fn cli() -> anyhow::Result<()> {
             crate::commands::update::update(canary, reconfigure)?;
             Ok(())
         }
-        Some(Command::Bench { cmd }) => handle_bench_command(cmd).await,
         Some(Command::Recipe { command }) => handle_recipe_subcommand(command),
         Some(Command::Web {
             port,