Datasets:
Formats:
parquet
Size:
100K - 1M
Tags:
population-genetics
demographic-inference
effective-population-size
coalescent
msprime
simulation
License:
File size: 8,156 Bytes
ee5b44c | 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 | from __future__ import annotations
import argparse
import dataclasses
import sys
from pathlib import Path
from .config import Config, apply_preset, validate_config
from .dataset_files import write_dataset_files
from .dataset_state import DatasetConfigMismatchError
from .empirical_maps import load_empirical_maps
from .shards import (
DatasetGenerationIncompleteError,
DatasetGenerationLockedError,
DatasetSourceUnavailableError,
UnsafeOutputDirectoryError,
dataset_output_lock,
finalize_dataset,
generate_dataset,
generate_dataset_partition,
prepare_dataset,
reuse_completed_dataset_if_compatible,
)
from .slurm import SlurmEnvironmentError, distributed_context
def parse_args() -> Config:
base = apply_preset(Config())
parser = argparse.ArgumentParser(formatter_class=argparse.ArgumentDefaultsHelpFormatter)
bool_fields = {"force", "compression", "include_truth", "enable_stdpopsim", "mut_map_scale_to_baseline"}
choice_fields = {"recomb_map_unit", "mut_map_unit"}
for f in dataclasses.fields(Config):
if f.name in bool_fields or f.name in choice_fields:
continue
default = getattr(base, f.name)
parser.add_argument("--" + f.name.replace("_", "-"), dest=f.name, type=type(default), default=None)
parser.add_argument("--n", dest="n_samples", type=int, default=None, help="alias for --n-samples")
parser.add_argument("--recomb-map-unit", choices=["per_bp", "cM_per_Mb"], default=None)
parser.add_argument("--mut-map-unit", choices=["per_bp", "relative", "roulette_raw"], default=None)
parser.add_argument("--force", action="store_true", help="overwrite an existing simulator output directory")
parser.add_argument("--include-truth", action="store_true", help="store true maps for analysis; not model inputs")
parser.add_argument("--enable-stdpopsim", action="store_true", help="include optional stdpopsim anchors")
parser.add_argument(
"--mut-map-scale-to-baseline",
dest="mut_map_scale_to_baseline",
action="store_true",
help="normalize relative or raw mutation maps to baseline_mu after unit conversion",
)
parser.add_argument(
"--no-mut-map-scale-to-baseline",
dest="mut_map_scale_to_baseline",
action="store_false",
help="preserve the absolute scale implied by --mut-map-unit",
)
parser.add_argument("--no-compression", dest="compression", action="store_false", help="write uncompressed npz shards")
parser.set_defaults(compression=None)
parser.set_defaults(mut_map_scale_to_baseline=None)
generation_mode = parser.add_mutually_exclusive_group()
generation_mode.add_argument(
"--prepare-only",
dest="_generation_mode",
action="store_const",
const="prepare",
help="create or validate the resumable generation plan, then exit",
)
generation_mode.add_argument(
"--generate-only",
dest="_generation_mode",
action="store_const",
const="generate",
help="generate only this rank's deterministic partition, then exit",
)
generation_mode.add_argument(
"--finalize-only",
dest="_generation_mode",
action="store_const",
const="finalize",
help="merge all committed fragments and publish the completed dataset",
)
parser.add_argument(
"--rank",
dest="_distributed_rank",
type=int,
metavar="RANK",
default=None,
help="explicit zero-based generation rank (normally inferred from SLURM_PROCID)",
)
parser.add_argument(
"--world-size",
dest="_distributed_world_size",
type=int,
metavar="N",
default=None,
help="explicit number of generation ranks (normally inferred from SLURM_NTASKS)",
)
parser.set_defaults(_generation_mode="auto")
ns = parser.parse_args()
preset = ns.preset if ns.preset is not None else "mini"
try:
cfg = apply_preset(Config(preset=preset))
except ValueError as exc:
parser.error(str(exc))
for f in dataclasses.fields(Config):
val = getattr(ns, f.name, None)
if val is not None:
setattr(cfg, f.name, val)
if cfg.enable_stdpopsim and cfg.p_stdpopsim_anchor == 0.0:
cfg.p_stdpopsim_anchor = 0.10
cfg.p_custom_hetero = max(0.0, cfg.p_custom_hetero - 0.10)
try:
validate_config(cfg)
except ValueError as exc:
parser.error(str(exc))
cfg._generation_mode = ns._generation_mode
cfg._distributed_rank = ns._distributed_rank
cfg._distributed_world_size = ns._distributed_world_size
return cfg
def _load_maps(cfg: Config):
try:
return load_empirical_maps(cfg)
except (OSError, ValueError) as exc:
raise SystemExit(f"[error] failed to load empirical maps: {exc}") from None
def _write_completed_dataset(cfg: Config, sample_manifest: dict) -> None:
write_dataset_files(cfg, sample_manifest)
print("[done] dataset written to", cfg.out_dir)
print(" manifest:", Path(cfg.out_dir) / "manifest.json")
print(" loader: ", Path(cfg.out_dir) / "scripts" / "loader.py")
def main() -> None:
cfg = parse_args()
mode = getattr(cfg, "_generation_mode", "auto")
explicit_rank = getattr(cfg, "_distributed_rank", None)
explicit_world_size = getattr(cfg, "_distributed_world_size", None)
print("=" * 80)
print(f"DLCoalSim generator: {cfg.dataset_name}")
print(f"out_dir={cfg.out_dir}")
print(
f"preset={cfg.preset} seq_len={cfg.seq_len:,} H={cfg.n_haplotypes} "
f"target_T={cfg.time_bins} history_tol={cfg.demography_resolution_log10}"
)
print("=" * 80)
sys.stdout.flush()
try:
if mode in {"prepare", "finalize"}:
if explicit_rank is not None or explicit_world_size is not None:
raise ValueError(
"--rank/--world-size apply only to generation, not "
"--prepare-only or --finalize-only"
)
if mode == "prepare":
prepare_dataset(cfg, _load_maps(cfg))
print("[prepared] resumable generation plan:", Path(cfg.out_dir) / "samples" / "PLAN.json")
return
with dataset_output_lock(cfg.out_dir, blocking=True):
sample_manifest = finalize_dataset(
cfg,
_lock_already_held=True,
)
_write_completed_dataset(cfg, sample_manifest)
return
context = distributed_context(
rank=explicit_rank,
world_size=explicit_world_size,
)
if mode == "generate" or context.distributed:
rank_result = generate_dataset_partition(
cfg,
_load_maps(cfg),
rank=context.rank,
world_size=context.world_size,
)
print(
f"[rank done] rank={context.rank}/{context.world_size} "
f"generated_chunks={rank_result.get('generated_chunks', 0)} "
f"reused_chunks={rank_result.get('reused_chunks', 0)}"
)
return
with dataset_output_lock(cfg.out_dir):
sample_manifest = reuse_completed_dataset_if_compatible(
cfg,
_lock_already_held=True,
)
if sample_manifest is None:
sample_manifest = generate_dataset(
cfg,
_load_maps(cfg),
_lock_already_held=True,
)
_write_completed_dataset(cfg, sample_manifest)
except (
DatasetConfigMismatchError,
DatasetGenerationIncompleteError,
DatasetGenerationLockedError,
DatasetSourceUnavailableError,
SlurmEnvironmentError,
UnsafeOutputDirectoryError,
ValueError,
) as exc:
raise SystemExit(f"[error] {exc}") from None
except OSError as exc:
raise SystemExit(f"[error] filesystem operation failed: {exc}") from None
|