[Config Refactor] sentinel default precedence (#3078)

Signed-off-by: lishunyang <lishunyang12@163.com>
Signed-off-by: xiaohajiayou <923390377@qq.com>
Co-authored-by: lishunyang <lishunyang12@163.com>
This commit is contained in:
Hacoandlishunyang authored and GitHub committed 2026-04-24 20:07:24 +08:00
1 parent b33d536f63
commit 22ffb85063
29 files changed
+458 -188

No files matched your search

+4 -11
View File
@@ -9,7 +9,6 @@ import argparse
import json
import os
import signal
import sys
from types import FrameType
from typing import Any
@@ -20,9 +19,9 @@ from vllm.entrypoints.utils import VLLM_SUBCMD_PARSER_EPILOG
from vllm.logger import init_logger
from vllm.utils.argparse_utils import FlexibleArgumentParser
from vllm_omni.engine.arg_utils import nullify_stage_engine_defaults
from vllm_omni.entrypoints.cli.logo import log_logo
from vllm_omni.entrypoints.openai.api_server import omni_run_server
from vllm_omni.entrypoints.utils import detect_explicit_cli_keys
logger = init_logger(__name__)
@@ -95,10 +94,6 @@ class OmniServeCommand(CLISubcommand):
if hasattr(args, "model_tag") and args.model_tag is not None:
args.model = args.model_tag
# Stash the set of long-option keys the user actually typed so the
# stage-config factory can give YAML precedence over argparse defaults.
args._cli_explicit_keys = detect_explicit_cli_keys(sys.argv[1:], OmniServeCommand._parser)
if args.headless:
run_headless(args)
else:
@@ -485,6 +480,8 @@ class OmniServeCommand(CLISubcommand):
# Stash via type(self) so the docs hook (which execs this function in a
# sandboxed globals dict via ``DummySelf``) doesn't fail on a NameError.
type(self)._parser = serve_parser
nullify_stage_engine_defaults(serve_parser)
return serve_parser
@@ -540,15 +537,11 @@ def run_headless(args: argparse.Namespace) -> None:
raise ValueError("headless mode requires worker_backend=multi_process")
args_dict = vars(args).copy()
# Preserve the explicit-keys set captured at parse time so per-stage yaml
# values (e.g. stage 1's ``gpu_memory_utilization: 0.5``) are not
# overwritten by argparse defaults for flags the user didn't type.
cli_explicit_keys = args_dict.pop("_cli_explicit_keys", None)
args_dict.pop("_cli_explicit_keys", None)
config_path, stage_configs = load_and_resolve_stage_configs(
model,
args_dict.get("stage_configs_path"),
args_dict,
cli_explicit_keys=cli_explicit_keys,
)
# Locate the stage config that matches stage_id.
+16 -30
View File
@@ -17,7 +17,7 @@ from vllm.v1.engine.exceptions import EngineDeadError, EngineGenerateError
from vllm_omni.engine.async_omni_engine import AsyncOmniEngine
from vllm_omni.entrypoints.client_request_state import ClientRequestState
from vllm_omni.entrypoints.pd_utils import PDDisaggregationMixin
from vllm_omni.entrypoints.utils import detect_explicit_cli_keys, get_final_stage_id_for_e2e
from vllm_omni.entrypoints.utils import get_final_stage_id_for_e2e
from vllm_omni.metrics.stats import OrchestratorAggregator as OrchestratorMetrics
from vllm_omni.model_executor.model_loader.weight_utils import download_weights_from_hf_specific
from vllm_omni.outputs import OmniRequestOutput
@@ -96,38 +96,23 @@ class OmniBase(PDDisaggregationMixin):
parser: argparse.ArgumentParser | None = None,
**overrides: Any,
) -> OmniBase:
"""Construct an ``Omni`` / ``AsyncOmni`` from an ``argparse.Namespace``.
"""Build from argparse. If ``parser`` is passed and not yet nullified,
un-typed engine fields are reset to ``None``."""
kwargs: dict[str, Any] = {k: v for k, v in vars(args).items() if not k.startswith("_")}
Mirrors the ``EngineArgs.from_cli_args`` pattern used upstream and in
``OmniEngineArgs.from_cli_args``. This is the recommended entry point
for any argparse-based caller (offline scripts, tests, CI): it
expands ``vars(args)`` into kwargs and automatically captures which
flags the user typed on the command line so that argparse defaults
do not silently override deploy YAML values.
if parser is not None and not getattr(parser, "_omni_nullified", False):
from vllm_omni.engine.arg_utils import (
deploy_override_field_names,
)
from vllm_omni.entrypoints.utils import detect_explicit_cli_keys
Passing ``parser`` is strongly recommended: without it, flag-to-dest
resolution falls back to a name-based heuristic that misidentifies
flags with ``dest=`` overrides, alias flags, and ``--disable-X`` /
``store_false`` pairs. See :func:`detect_explicit_cli_keys`.
explicit = detect_explicit_cli_keys(sys.argv[1:], parser) or set()
override_dests = deploy_override_field_names()
for key in list(kwargs):
if key in override_dests and key not in explicit:
kwargs[key] = None
Args:
args: Parsed argparse namespace from ``parser.parse_args()``.
parser: The argparse parser used to produce ``args``. When
provided, each user-typed flag is resolved to its real
``dest`` via the parser's action table.
**overrides: Extra keyword arguments that take precedence over
attributes on ``args``.
Example::
parser = FlexibleArgumentParser()
parser.add_argument("--model", required=True)
args = parser.parse_args()
omni = Omni.from_cli_args(args, parser=parser) # preferred
omni = Omni.from_cli_args(args, parser=parser, model="other")
"""
kwargs: dict[str, Any] = {**vars(args), **overrides}
kwargs["_cli_explicit_keys"] = detect_explicit_cli_keys(sys.argv[1:], parser)
kwargs.update(overrides)
return cls(**kwargs)
def __init__(
@@ -136,6 +121,7 @@ class OmniBase(PDDisaggregationMixin):
**kwargs: Any,
) -> None:
engine_args: OmniEngineArgs | None = kwargs.pop("engine_args", None)
stage_init_timeout = kwargs.pop("stage_init_timeout", 300)
init_timeout = kwargs.pop("init_timeout", 600)
log_stats = kwargs.pop("log_stats", False)
@@ -2779,6 +2779,8 @@ if __name__ == "__main__":
from vllm.entrypoints.openai.cli_args import make_arg_parser
from vllm_omni.engine.arg_utils import nullify_stage_engine_defaults
parser = argparse.ArgumentParser(description="vLLM-Omni OpenAI-Compatible REST API server")
parser = make_arg_parser(parser)
registered_flags = set()
@@ -2790,6 +2792,7 @@ if __name__ == "__main__":
parser.add_argument(
"--enable-sleep-mode", action="store_true", default=False, help="Enable GPU memory pool for sleep mode."
)
nullify_stage_engine_defaults(parser)
args = parser.parse_args()
if not hasattr(args, "model_tag"):
setattr(args, "model_tag", args.model)
+19 -15
View File
@@ -21,6 +21,18 @@ PROJECT_ROOT = Path(__file__).parent.parent.parent
logger = init_logger(__name__)
def _warn_deprecated_explicit_keys(kwargs: dict[str, Any]) -> None:
if "cli_explicit_keys" in kwargs:
import warnings
warnings.warn(
"cli_explicit_keys= is deprecated and ignored. Remove the kwarg.",
DeprecationWarning,
stacklevel=3,
)
_DIFFUSERS_CLASS_TO_CONFIG: dict[str, str] = {
"GlmImagePipeline": "glm_image",
}
@@ -343,7 +355,7 @@ def load_stage_configs_from_model(
base_engine_args: dict | None = None,
deploy_config_path: str | None = None,
stage_overrides: dict[str, dict[str, Any]] | None = None,
cli_explicit_keys: set[str] | None = None,
**deprecated_kwargs: Any,
) -> list:
"""Load stage configurations from model's default config file.
@@ -358,33 +370,25 @@ def load_stage_configs_from_model(
base_engine_args: Base engine args to merge as CLI overrides.
deploy_config_path: Optional explicit deploy config path.
stage_overrides: Per-stage overrides from --stage-overrides.
cli_explicit_keys: Set of CLI keys the user actually typed. When
provided, only these keys override deploy YAML; argparse defaults
stay subordinate to YAML. ``None`` means treat every kwarg as
explicit (programmatic ``Omni()`` calls).
Returns:
List of stage configuration dictionaries
"""
_warn_deprecated_explicit_keys(deprecated_kwargs)
if base_engine_args is None:
base_engine_args = {}
cli_overrides = _convert_dataclasses_to_dict(dict(base_engine_args))
# Per-stage JSON overrides are always explicit (the user typed --stage-overrides).
explicit = set(cli_explicit_keys) if cli_explicit_keys is not None else None
if stage_overrides:
for stage_id_str, overrides in stage_overrides.items():
for key, val in overrides.items():
stage_key = f"stage_{stage_id_str}_{key}"
cli_overrides[stage_key] = val
if explicit is not None:
explicit.add(stage_key)
cli_overrides[f"stage_{stage_id_str}_{key}"] = val
stages = StageConfigFactory.create_from_model(
model,
cli_overrides=cli_overrides,
deploy_config_path=deploy_config_path,
cli_explicit_keys=explicit,
)
if stages is not None:
# Convert StageConfig objects to OmegaConf for backward compat
@@ -545,7 +549,7 @@ def load_and_resolve_stage_configs(
default_stage_cfg_factory: Any = None,
deploy_config_path: str | None = None,
stage_overrides: dict[str, dict[str, Any]] | None = None,
cli_explicit_keys: set[str] | None = None,
**deprecated_kwargs: Any,
) -> tuple[str, list]:
"""Load stage configurations from model or YAML file with fallback to defaults.
@@ -589,6 +593,8 @@ def load_and_resolve_stage_configs(
stage_configs_path,
)
_warn_deprecated_explicit_keys(deprecated_kwargs)
if deploy_config_path is not None:
config_path = deploy_config_path
stage_configs = load_stage_configs_from_model(
@@ -596,7 +602,6 @@ def load_and_resolve_stage_configs(
base_engine_args=kwargs,
deploy_config_path=deploy_config_path,
stage_overrides=stage_overrides,
cli_explicit_keys=cli_explicit_keys,
)
if not stage_configs:
if default_stage_cfg_factory is not None:
@@ -610,7 +615,6 @@ def load_and_resolve_stage_configs(
model,
base_engine_args=kwargs,
stage_overrides=stage_overrides,
cli_explicit_keys=cli_explicit_keys,
)
if not stage_configs:
if default_stage_cfg_factory is not None: