dokimi explore

usage: dokimi explore [-h] {list,run} ...

positional arguments:
  {list,run}
    list      List exploration tasks
    run       Run exploration tasks through arms

options:
  -h, --help  show this help message and exit

dokimi explore list

usage: dokimi explore list [-h] [--tasks-root TASKS_ROOT] [--repo-root REPO_ROOT]

options:
  -h, --help            show this help message and exit
  --tasks-root TASKS_ROOT
                        Directory of exploration task YAML files (default: data/explore_tasks of a
                        source checkout; an installed package has none and needs this flag)
  --repo-root REPO_ROOT
                        Tree a task is scored against when its YAML omits repo_root: (default: the
                        dokimi source checkout, which is only right for the self-dogfooding tasks
                        bundled there; an installed package needs this flag)

dokimi explore run

usage: dokimi explore run [-h] [--tasks TASKS] [--all] [--tasks-root TASKS_ROOT]
                          [--repo-root REPO_ROOT] [--arms ARMS] [--baseline-arm BASELINE_ARM]
                          [--model MODEL] [--output-dir OUTPUT_DIR] [--out-json OUT_JSON]
                          [--out-md OUT_MD]

options:
  -h, --help            show this help message and exit
  --tasks TASKS         Comma-separated task ids
  --all                 Run every discovered task
  --tasks-root TASKS_ROOT
                        Directory of exploration task YAML files (default: data/explore_tasks of a
                        source checkout; an installed package has none and needs this flag)
  --repo-root REPO_ROOT
                        Tree a task is scored against when its YAML omits repo_root: (default: the
                        dokimi source checkout, which is only right for the self-dogfooding tasks
                        bundled there; an installed package needs this flag)
  --arms ARMS           Comma-separated arms: oracle, empty, command:<template>
  --baseline-arm BASELINE_ARM
  --model MODEL
  --output-dir OUTPUT_DIR
  --out-json OUT_JSON
  --out-md OUT_MD       Also render a Markdown report