batter.config.run.RunSection

pydantic model batter.config.run.RunSection

Run-related settings, including where outputs land.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

Config:

• extra: str = forbid

Fields:

• allow_run_id_mismatch (bool)

• batch_gpus (int | None)

• batch_gpus_per_task (int)

• batch_mode (bool)

• batch_srun_extra (List[str])

• clean_failures (bool)

• dry_run (bool)

• email_on_completion (str | None)

• email_sender (str)

• max_active_jobs (int | None)

• max_workers (int | None)

• on_failure (Literal['raise', 'prune', 'retry'])

• only_fe_preparation (bool)

• output_folder (Path)

• remd (Literal['yes', 'no'])

• run_id (str)

• slurm (SlurmConfig)

• slurm_header_dir (Path | None)

• system_type (Literal['MABFE', 'MASFE'] | None)

Validators:

• _coerce_output_folder » output_folder

• _coerce_remd » remd

• _normalize_system_type » system_type

field allow_run_id_mismatch: bool = False

When True, allow reusing an explicit run_id even if the configuration hash differs from the existing execution.

field batch_gpus: int | None = None

GPUs available to the manager process for batch_mode; auto-detected from SLURM env when omitted.

Constraints:

• ge = 0

field batch_gpus_per_task: int = 1

GPUs to assign per task when batch_mode is enabled.

Constraints:

• ge = 1

field batch_mode: bool = False

When true, run SLURM jobs inline via srun inside the manager allocation instead of submitting with sbatch.

field batch_srun_extra: List[str] [Optional]

Extra srun flags appended when launching tasks in batch_mode.

field clean_failures: bool = False

Clear FAILED markers, job_attempt.txt retry counters, and progress caches before rerunning.

field dry_run: bool = False

Force dry-run mode regardless of YAML setting.

field email_on_completion: str | None = None

Email address that should receive a notification once the run finishes or aborts with an uncaught failure.

field email_sender: str = 'nobody@stanford.edu'

Sender address used for BATTER run-status email notifications.

field max_active_jobs: int | None = 1000

Max concurrent SLURM jobs for FE submissions (0 disables throttling).

Constraints:

• ge = 0

field max_workers: int | None = None

Parallel workers for local backend (None = auto, 0 = serial).

field on_failure: Literal['raise', 'prune', 'retry'] = 'raise'

Behavior on ligand failure: ‘raise’, ‘prune’, or ‘retry’ (clear FAILED sentinels and rerun once).

field only_fe_preparation: bool = False

When true, stop the workflow after FE preparation.

field output_folder: Path [Required]

Directory where system artifacts and executions are stored.

Validated by:

• _coerce_output_folder

field remd: Literal['yes', 'no'] = 'no'

Enable REMD execution.

Validated by:

• _coerce_remd

field run_id: str = 'auto'

Run identifier to use (auto picks latest).

field slurm: SlurmConfig [Optional]

field slurm_header_dir: Path | None = None

Optional directory containing user Slurm headers (defaults to ~/.batter).

field system_type: Literal['MABFE', 'MASFE'] | None = None

Optional override for the system builder. When omitted, the orchestrator chooses the builder based on the protocol.

Validated by:

• _normalize_system_type