Spaces:
Running
Running
File size: 9,535 Bytes
a721dfa 9734b71 4106e0f a721dfa 4106e0f a721dfa 4106e0f a721dfa 4106e0f 9734b71 a721dfa b9be111 a721dfa ad4554b a721dfa | 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 | from __future__ import annotations
import os
import socket
import sys
from copy import deepcopy
from dataclasses import dataclass
from pathlib import Path
from typing import Any, List, Mapping, MutableMapping, Protocol
import numpy as np
import pandas as pd
DEFAULT_TORCH_NUM_THREADS = 2
DEFAULT_TORCH_NUM_INTEROP_THREADS = 1
DEFAULT_OMP_NUM_THREADS = 2
DEFAULT_MKL_NUM_THREADS = 2
ALLOW_NON_VENV_PYTHON_ENV_KEY = "AIFORECAST_ALLOW_NON_VENV_PYTHON"
@dataclass(frozen=True)
class RuntimePaths:
current_dir: str
project_root: str
@dataclass(frozen=True)
class TorchThreadSettings:
torch_num_threads: int
torch_num_interop_threads: int
omp_num_threads: int
mkl_num_threads: int
class LoggerLike(Protocol):
def info(self, msg: str, *args: Any, **kwargs: Any) -> None: ...
def warning(self, msg: str, *args: Any, **kwargs: Any) -> None: ...
def _read_positive_int(
raw_value: object,
*,
default: int,
) -> int:
try:
value = int(str(raw_value).strip())
except (TypeError, ValueError, AttributeError):
return default
return max(1, value)
def _read_bool(raw_value: object, *, default: bool) -> bool:
normalized = str(raw_value or "").strip().lower()
if not normalized:
return default
if normalized in {"1", "true", "yes", "on"}:
return True
if normalized in {"0", "false", "no", "off"}:
return False
return default
def read_torch_thread_settings(
env: Mapping[str, str] | None = None,
) -> TorchThreadSettings:
runtime_env = os.environ if env is None else env
return TorchThreadSettings(
torch_num_threads=_read_positive_int(
runtime_env.get("TORCH_NUM_THREADS"),
default=DEFAULT_TORCH_NUM_THREADS,
),
torch_num_interop_threads=_read_positive_int(
runtime_env.get("TORCH_NUM_INTEROP_THREADS"),
default=DEFAULT_TORCH_NUM_INTEROP_THREADS,
),
omp_num_threads=_read_positive_int(
runtime_env.get("OMP_NUM_THREADS"),
default=DEFAULT_OMP_NUM_THREADS,
),
mkl_num_threads=_read_positive_int(
runtime_env.get("MKL_NUM_THREADS"),
default=DEFAULT_MKL_NUM_THREADS,
),
)
def _expected_project_venv_python(project_root: Path) -> Path:
if os.name == "nt":
return project_root / "venv" / "Scripts" / "python.exe"
return project_root / "venv" / "bin" / "python"
def _same_path(left: Path, right: Path) -> bool:
try:
if left.exists() and right.exists():
return left.samefile(right)
except OSError:
pass
return str(left.resolve()).lower() == str(right.resolve()).lower()
def describe_runtime_environment(
project_root: str | Path,
*,
env: Mapping[str, str] | None = None,
current_executable: str | None = None,
) -> dict[str, Any]:
runtime_env = os.environ if env is None else env
root_path = Path(project_root).resolve()
current_python = Path(current_executable or sys.executable).resolve()
expected_python = _expected_project_venv_python(root_path)
allow_non_venv_python = _read_bool(
runtime_env.get(ALLOW_NON_VENV_PYTHON_ENV_KEY),
default=False,
)
expected_exists = expected_python.exists()
is_project_venv_python = not expected_exists or _same_path(current_python, expected_python)
settings = read_torch_thread_settings(runtime_env)
return {
"current_executable": str(current_python),
"expected_venv_python": str(expected_python),
"expected_venv_exists": expected_exists,
"is_project_venv_python": is_project_venv_python,
"allow_non_venv_python": allow_non_venv_python,
"thread_caps": {
"torch_num_threads": settings.torch_num_threads,
"torch_num_interop_threads": settings.torch_num_interop_threads,
"omp_num_threads": settings.omp_num_threads,
"mkl_num_threads": settings.mkl_num_threads,
},
}
def prepare_runtime_environment(
project_root: str | Path,
*,
env: MutableMapping[str, str] | None = None,
current_executable: str | None = None,
logger: LoggerLike | None = None,
) -> TorchThreadSettings:
runtime_env = os.environ if env is None else env
settings = read_torch_thread_settings(runtime_env)
runtime_env.setdefault("TORCH_NUM_THREADS", str(settings.torch_num_threads))
runtime_env.setdefault("TORCH_NUM_INTEROP_THREADS", str(settings.torch_num_interop_threads))
runtime_env.setdefault("OMP_NUM_THREADS", str(settings.omp_num_threads))
runtime_env.setdefault("MKL_NUM_THREADS", str(settings.mkl_num_threads))
snapshot = describe_runtime_environment(
project_root,
env=runtime_env,
current_executable=current_executable,
)
if (
snapshot["expected_venv_exists"]
and not snapshot["is_project_venv_python"]
and not snapshot["allow_non_venv_python"]
):
raise RuntimeError(
"Current Python interpreter is outside the project venv. "
f"Expected {snapshot['expected_venv_python']} but got {snapshot['current_executable']}. "
f"Set {ALLOW_NON_VENV_PYTHON_ENV_KEY}=true to override."
)
if logger is not None:
logger.info(
"Prepared runtime thread caps: torch=%s interop=%s omp=%s mkl=%s",
settings.torch_num_threads,
settings.torch_num_interop_threads,
settings.omp_num_threads,
settings.mkl_num_threads,
)
return settings
def apply_torch_thread_settings(
torch_module: Any | None,
env: Mapping[str, str] | None = None,
*,
logger: LoggerLike | None = None,
) -> dict[str, Any]:
settings = read_torch_thread_settings(env)
applied = False
if torch_module is not None:
set_num_threads = getattr(torch_module, "set_num_threads", None)
if callable(set_num_threads):
set_num_threads(settings.torch_num_threads)
applied = True
set_num_interop_threads = getattr(torch_module, "set_num_interop_threads", None)
if callable(set_num_interop_threads):
try:
set_num_interop_threads(settings.torch_num_interop_threads)
except RuntimeError as exc:
if logger is not None:
logger.warning("Torch interop thread cap could not be updated: %s", exc)
else:
applied = True
if logger is not None:
logger.info(
"Applied torch thread caps: torch=%s interop=%s omp=%s mkl=%s applied=%s",
settings.torch_num_threads,
settings.torch_num_interop_threads,
settings.omp_num_threads,
settings.mkl_num_threads,
applied,
)
return {
"torch_num_threads": settings.torch_num_threads,
"torch_num_interop_threads": settings.torch_num_interop_threads,
"omp_num_threads": settings.omp_num_threads,
"mkl_num_threads": settings.mkl_num_threads,
"applied": applied,
}
def can_bind_tcp_port(host: str, port: int) -> bool:
with socket.socket(socket.AF_INET, socket.SOCK_STREAM) as server_socket:
try:
server_socket.bind((host, port))
except OSError:
return False
return True
def find_free_tcp_port(host: str = "127.0.0.1") -> int:
with socket.socket(socket.AF_INET, socket.SOCK_STREAM) as server_socket:
server_socket.bind((host, 0))
server_socket.listen(1)
return int(server_socket.getsockname()[1])
def parse_cors_origins(raw_origins: str) -> List[str]:
origins = [origin.strip() for origin in raw_origins.split(",") if origin.strip()]
return origins or ["*"]
def clone_cache_payload(payload: Any) -> Any:
if payload is None or isinstance(payload, (str, int, float, bool)):
return payload
if isinstance(payload, list):
return [clone_cache_payload(item) for item in payload]
if isinstance(payload, tuple):
return tuple(clone_cache_payload(item) for item in payload)
if isinstance(payload, dict):
return {clone_cache_payload(key): clone_cache_payload(value) for key, value in payload.items()}
if isinstance(payload, set):
return {clone_cache_payload(item) for item in payload}
try:
return deepcopy(payload)
except Exception:
return payload
def make_json_compatible(value: Any) -> Any:
if isinstance(value, dict):
return {str(key): make_json_compatible(item) for key, item in value.items()}
if isinstance(value, (list, tuple, set)):
return [make_json_compatible(item) for item in value]
if isinstance(value, np.ndarray):
return [make_json_compatible(item) for item in value.tolist()]
if isinstance(value, np.generic):
return make_json_compatible(value.item())
if isinstance(value, pd.Timestamp):
return value.isoformat()
if isinstance(value, float):
if not np.isfinite(value):
return None
return value
return value
def resolve_runtime_paths(
module_file: str,
is_frozen: bool,
bundle_dir: str,
) -> RuntimePaths:
current_dir = os.path.dirname(os.path.abspath(module_file))
if is_frozen:
return RuntimePaths(current_dir=bundle_dir, project_root=bundle_dir)
return RuntimePaths(
current_dir=current_dir,
project_root=os.path.dirname(current_dir),
)
|