File size: 16,047 Bytes
6d19a18 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 | #!/usr/bin/env python3
"""Submit OpenBind docking jobs from an experiment YAML configuration.
The positional input is the ``prepared_inputs`` directory. Configured modes must
match child directory names exactly. Software paths and Slurm profiles are read
from ``configs/docking/software_settings.yaml`` by default.
"""
from __future__ import annotations
import argparse
import shlex
import subprocess
import textwrap
from pathlib import Path
from typing import Any
import yaml
SBATCH_TEMPLATE = """\
#!/bin/bash
#SBATCH --job-name={job_name}
{sbatch_lines}
#SBATCH --output={log_dir}/%x_%j.out
#SBATCH --error={log_dir}/%x_%j.err
{command}
"""
SCRIPT_DIR = Path(__file__).resolve().parent
REPO_ROOT = SCRIPT_DIR.parent
DOCK_SCRIPT = SCRIPT_DIR / "dock_complexes.py"
DEFAULT_SOFTWARE_SETTINGS = REPO_ROOT / "configs" / "docking" / "software_settings.yaml"
SUPPORTED_METHODS = {"gnina", "smina", "diffdock"}
def load_yaml(path: Path) -> dict[str, Any]:
"""Load a YAML file and require a mapping at the top level."""
if not path.is_file():
raise FileNotFoundError(f"YAML file not found: {path}")
with path.open("r", encoding="utf-8") as handle:
data = yaml.safe_load(handle)
if not isinstance(data, dict):
raise ValueError(f"YAML root must be a mapping: {path}")
return data
def resolve_repo_path(value: str) -> Path:
"""Resolve a software path relative to the repository root."""
path = Path(value).expanduser()
return (REPO_ROOT / path).resolve() if not path.is_absolute() else path.resolve()
def validate_diffdock_config(executable: Path, num_modes: int) -> None:
"""Require DiffDock's YAML pose count to match the experiment config."""
config_path = executable / "default_inference_args.yaml"
config = load_yaml(config_path)
samples_per_complex = config.get("samples_per_complex")
if not isinstance(samples_per_complex, int) or samples_per_complex < 1:
raise ValueError(
"DiffDock default_inference_args.yaml must define a positive integer "
f"samples_per_complex: {config_path}"
)
if samples_per_complex < num_modes:
raise ValueError(
"DiffDock loads samples_per_complex from default_inference_args.yaml after "
"parsing command-line arguments, so it must match methods.diffdock.num_modes. "
f"Set samples_per_complex: {num_modes} in {config_path} "
f"(currently {samples_per_complex}). See software/README.md."
)
def require_modes(config: dict[str, Any]) -> list[str]:
"""Return the configured docking modes."""
modes = config.get("modes")
if (
not isinstance(modes, list)
or not modes
or not all(isinstance(x, str) for x in modes)
):
raise ValueError(
"Experiment config 'modes' must be a non-empty list of strings"
)
if len(modes) != len(set(modes)):
raise ValueError("Experiment config contains duplicate docking modes")
return modes
def require_methods(config: dict[str, Any]) -> dict[str, dict[str, Any]]:
"""Return and validate experiment method settings."""
methods = config.get("methods")
if not isinstance(methods, dict) or not methods:
raise ValueError("Experiment config 'methods' must be a non-empty mapping")
unknown = set(methods) - SUPPORTED_METHODS
if unknown:
raise ValueError(f"Unsupported docking methods: {sorted(unknown)}")
validated: dict[str, dict[str, Any]] = {}
for name, settings in methods.items():
if not isinstance(settings, dict):
raise ValueError(f"Method '{name}' settings must be a mapping")
if (
"seeds" not in settings
or not isinstance(settings["seeds"], list)
or not settings["seeds"]
):
raise ValueError(f"Method '{name}' must define a non-empty seeds list")
if not all(isinstance(seed, int) for seed in settings["seeds"]):
raise ValueError(f"All seeds for '{name}' must be integers")
required = {"num_modes"}
if name in {"gnina", "smina"}:
required |= {"num_inputs", "exhaustiveness", "box_size"}
missing = required - set(settings)
if missing:
raise ValueError(f"Method '{name}' is missing settings: {sorted(missing)}")
for key in {"num_modes", "num_inputs", "exhaustiveness"} & set(settings):
if not isinstance(settings[key], int) or settings[key] < 1:
raise ValueError(f"{name}.{key} must be a positive integer")
validated[name] = settings
return validated
def require_software_methods(settings: dict[str, Any]) -> dict[str, dict[str, Any]]:
"""Return software method definitions from the software settings file."""
methods = settings.get("methods")
if not isinstance(methods, dict):
raise ValueError("Software settings must contain a 'methods' mapping")
return methods
def require_profiles(settings: dict[str, Any]) -> dict[str, dict[str, Any]]:
"""Return Slurm profiles from the software settings file."""
profiles = settings.get("profiles")
if not isinstance(profiles, dict):
raise ValueError("Software settings must contain a 'profiles' mapping")
for name, profile in profiles.items():
if not isinstance(profile, dict):
raise ValueError(f"Slurm profile '{name}' must be a mapping")
if "cpus-per-task" not in profile:
raise ValueError(f"Slurm profile '{name}' must define cpus-per-task")
return profiles
def validate_centre(value: object) -> tuple[float, float, float] | None:
"""Validate an optional docking-box centre from the experiment config."""
if value is None:
return None
if not isinstance(value, list) or len(value) != 3:
raise ValueError(
"Experiment config 'centre' must contain exactly three numbers"
)
if not all(isinstance(item, (int, float)) for item in value):
raise ValueError("All values in experiment config 'centre' must be numeric")
return float(value[0]), float(value[1]), float(value[2])
def render_sbatch_lines(profile: dict[str, Any]) -> str:
"""Render Slurm profile entries as SBATCH directives."""
lines: list[str] = []
for key, value in profile.items():
if value in (None, "", False):
continue
if value is True:
lines.append(f"#SBATCH --{key}")
else:
lines.append(f"#SBATCH --{key}={value}")
return "\n".join(lines)
def resolve_software_method(
method_name: str,
software_methods: dict[str, dict[str, Any]],
profiles: dict[str, dict[str, Any]],
) -> tuple[Path, Path | None, str, dict[str, Any]]:
"""Resolve executable/container paths and the Slurm profile for one method."""
settings = software_methods.get(method_name)
if not isinstance(settings, dict):
raise ValueError(f"Software settings are missing method '{method_name}'")
executable_value = settings.get("executable")
profile_name = settings.get("profile")
if not isinstance(executable_value, str) or not executable_value:
raise ValueError(f"Software method '{method_name}' must define executable")
if not isinstance(profile_name, str) or profile_name not in profiles:
raise ValueError(
f"Software method '{method_name}' references unknown profile "
f"'{profile_name}'"
)
executable = resolve_repo_path(executable_value)
if method_name == "diffdock":
if not executable.is_dir() or not (executable / "inference.py").is_file():
raise FileNotFoundError(
f"DiffDock directory or inference.py not found: {executable}"
)
elif not executable.is_file():
raise FileNotFoundError(
f"Executable for '{method_name}' not found: {executable}"
)
singularity: Path | None = None
if settings.get("singularity"):
singularity = resolve_repo_path(str(settings["singularity"]))
if not singularity.is_file():
raise FileNotFoundError(
f"Singularity image for '{method_name}' not found: {singularity}"
)
return executable, singularity, profile_name, profiles[profile_name]
def build_docking_command(
mode_dir: Path,
output_dir: Path,
method_name: str,
method_cfg: dict[str, Any],
executable: Path,
mode: str,
seed: int,
centre: tuple[float, float, float] | None,
overwrite: bool,
) -> list[str]:
"""Build one ``dock_complexes.py`` command."""
label = method_name
command = [
"python",
"-u",
str(DOCK_SCRIPT),
str(mode_dir),
"--engine",
method_name,
"--method",
str(executable),
"--method-label",
label,
"--mode-label",
mode,
"--output-dir",
str(output_dir),
"--num-modes",
str(method_cfg["num_modes"]),
"--seed",
str(seed),
]
if method_name in {"gnina", "smina"}:
command += [
"--num-inputs",
str(method_cfg["num_inputs"]),
"--exhaustiveness",
str(method_cfg["exhaustiveness"]),
"--size",
str(method_cfg["box_size"]),
]
if method_cfg.get("scoring") is not None:
command += ["--scoring", str(method_cfg["scoring"])]
if method_name == "gnina" and method_cfg.get("cnn_model") not in (
None,
"default",
):
command += [
"--cnn-model",
str(resolve_repo_path(str(method_cfg["cnn_model"]))),
]
if centre is not None:
command += [
"--x-coord",
str(centre[0]),
"--y-coord",
str(centre[1]),
"--z-coord",
str(centre[2]),
]
if overwrite:
command.append("--overwrite")
return command
def wrap_singularity(
command: list[str],
image: Path | None,
use_gpu: bool,
prepared_inputs_dir: Path,
mode_dir: Path,
output_dir: Path,
) -> list[str]:
"""Wrap a docking command in Singularity when an image is configured."""
if image is None:
return command
bind_paths = {
REPO_ROOT.resolve(),
prepared_inputs_dir.resolve(),
mode_dir.resolve(),
output_dir.resolve(),
}
wrapped = ["singularity", "exec"]
if use_gpu:
wrapped.append("--nv")
wrapped += [
"--bind",
",".join(f"{path}:{path}" for path in sorted(bind_paths, key=str)),
"--pwd",
str(REPO_ROOT),
str(image),
*command,
]
return wrapped
def build_parser() -> argparse.ArgumentParser:
"""Build the command-line parser."""
parser = argparse.ArgumentParser(description=__doc__)
parser.add_argument(
"prepared_inputs_dir",
type=Path,
help="Directory containing one prepared-input subdirectory per docking mode.",
)
parser.add_argument(
"--config",
required=True,
type=Path,
help="Docking experiment YAML file.",
)
parser.add_argument(
"--software-settings",
type=Path,
default=DEFAULT_SOFTWARE_SETTINGS,
help=f"Software and Slurm YAML (default: {DEFAULT_SOFTWARE_SETTINGS}).",
)
parser.add_argument(
"--output-dir",
type=Path,
default=None,
help="Docking output directory (default: sibling 'docking' directory).",
)
parser.add_argument(
"--dry-run",
action="store_true",
help="Print jobs without submitting them.",
)
parser.add_argument(
"--overwrite",
action="store_true",
help="Overwrite existing docking outputs.",
)
return parser
def main() -> None:
"""Submit configured docking jobs."""
args = build_parser().parse_args()
if not DOCK_SCRIPT.is_file():
raise FileNotFoundError(f"Docking runner not found: {DOCK_SCRIPT}")
prepared_inputs_dir = args.prepared_inputs_dir.expanduser().resolve()
if not prepared_inputs_dir.is_dir():
raise FileNotFoundError(
f"Prepared-input directory not found: {prepared_inputs_dir}"
)
config_path = args.config.expanduser().resolve()
software_path = args.software_settings.expanduser().resolve()
experiment_cfg = load_yaml(config_path)
software_cfg = load_yaml(software_path)
modes = require_modes(experiment_cfg)
methods = require_methods(experiment_cfg)
software_methods = require_software_methods(software_cfg)
profiles = require_profiles(software_cfg)
centre = validate_centre(experiment_cfg.get("centre"))
output_dir = (
args.output_dir.expanduser().resolve()
if args.output_dir is not None
else prepared_inputs_dir.parent / "docking"
)
# Preflight all configured modes and software before submitting any jobs.
mode_dirs: dict[str, Path] = {}
for mode in modes:
mode_dir = prepared_inputs_dir / mode
if not mode_dir.is_dir():
raise FileNotFoundError(f"Prepared mode directory not found: {mode_dir}")
mode_dirs[mode] = mode_dir
resolved_methods: dict[str, tuple[Path, Path | None, str, dict[str, Any]]] = {}
for method_name, method_cfg in methods.items():
resolved = resolve_software_method(method_name, software_methods, profiles)
executable, _, _, _ = resolved
if method_name == "diffdock":
validate_diffdock_config(executable, method_cfg["num_modes"])
resolved_methods[method_name] = resolved
jobs = 0
for mode in modes:
mode_dir = mode_dirs[mode]
for method_name, method_cfg in methods.items():
executable, image, profile_name, profile = resolved_methods[method_name]
use_gpu = bool(profile.get("gres"))
label = method_name
for seed in method_cfg["seeds"]:
setup_label = f"{label}_{mode}"
job_name = f"{setup_label}_seed{seed}"
job_dir = output_dir / "slurm_jobs" / setup_label
log_dir = output_dir / "slurm_logs" / setup_label
command = build_docking_command(
mode_dir=mode_dir,
output_dir=output_dir,
method_name=method_name,
method_cfg=method_cfg,
executable=executable,
mode=mode,
seed=seed,
centre=centre,
overwrite=args.overwrite,
)
command = wrap_singularity(
command=command,
image=image,
use_gpu=use_gpu,
prepared_inputs_dir=prepared_inputs_dir,
mode_dir=mode_dir,
output_dir=output_dir,
)
script = SBATCH_TEMPLATE.format(
job_name=job_name,
sbatch_lines=render_sbatch_lines(profile),
log_dir=log_dir.resolve(),
command=shlex.join(command),
)
script_path = job_dir / f"{job_name}.sbatch"
if args.dry_run:
print(f"[dry-run] {profile_name}: sbatch {script_path}")
print(textwrap.dedent(script))
else:
job_dir.mkdir(parents=True, exist_ok=True)
log_dir.mkdir(parents=True, exist_ok=True)
script_path.write_text(textwrap.dedent(script), encoding="utf-8")
subprocess.run(["sbatch", str(script_path)], check=True)
print(f"Submitted {job_name}")
jobs += 1
print(f"Total jobs: {jobs}")
if __name__ == "__main__":
main()
|