simready-validate 2026.4.8

SimReady Validation Library

simready.validate

Standalone Python library for validating USD assets against SimReady profiles. Reports which features passed or failed, with per-requirement detail on failures.

Part of the SimReady Python Library Suite — no Omniverse or Kit installation required.

---

Installation

pip install simready-validate

Quick start

import simready.validate as sv
from pathlib import Path

# Load profiles from your spec data (omit when running inside Kit)
specs = Path("/path/to/specs")
sv.initialize(
    rules_and_requirements_paths=[specs / "capabilities"],
    features_paths=[specs / "features"],
    profiles_paths=[specs / "profiles/profiles.toml"],
)

result = sv.validate_asset(sv.AssetValidationConfig(
    asset_path="/assets/props/crate_01/crate_01.usd",
    profile_id="Prop-Robotics-Neutral",
    profile_version="2.0.0",
))

if result is None:
    print("Could not validate — file missing, wrong extension, or no profile found.")
else:
    passed = all(f["passed"] for f in result.features_summary.values())
    print(f"{result.profile_id} v{result.profile_version}: {'PASSED' if passed else 'FAILED'}")
    for feat_id, data in result.features_summary.items():
        if not data["passed"]:
            print(f"  {feat_id}: failing {data.get('failing requirements')}")

sv.destroy()

---

Key concepts

Profiles

A profile is a named set of features (e.g. Prop-Robotics-Neutral, Prop-Robotics-Physx). Each feature groups a set of validation requirements. Passing a profile means every requirement of every feature in that profile passed.

Profile data (rules, features, profile TOML) is loaded via initialize(). When running inside a Kit environment, profiles are already registered and initialize() can be omitted.

Profile inference

If profile_id is omitted from AssetValidationConfig, the library reads the profile from SimReady_Metadata.validation.profile stamped in the asset's USD customLayerData. This allows previously-stamped assets to be re-validated without specifying the profile explicitly.

Stamping results into USD

Setting write_metadata=True on AssetValidationConfig writes the validation outcome back into the USD file's customLayerData["SimReady_Metadata"]["validation"].

---

Python API

Initialization

sv.initialize(
    rules_and_requirements_paths: list[Path],
    features_paths:               list[Path],
    profiles_paths:               list[Path],
) -> None

sv.destroy() -> None

Validate a single asset

sv.validate_asset(config: AssetValidationConfig) -> AssetValidationResult | None

Returns None if the file is missing, has the wrong extension (.usd / .usda only), or has no profile in metadata and none was given.

Validate multiple assets

sv.validate_asset_list(
    configs:          list[AssetValidationConfig],
    report_file_path: str | None = None,          # merged JSON report
) -> list[AssetValidationResult | None]

Configuration and result types

@dataclass
class AssetValidationConfig:
    asset_path:      str       # path to .usd / .usda file
    profile_id:      str | None = None   # None → infer from USD metadata
    profile_version: str | None = None   # None → any registered version
    write_metadata:  bool = False         # stamp results into USD on full pass

@dataclass
class AssetValidationResult:
    asset_path:       str
    profile_id:       str
    profile_version:  str
    features_summary: dict[str, dict]   # feature_id → {version, passed, failing requirements}
    issues:           list              # raw failing issues from the engine

Batch validation with JSON report

configs = [
    sv.AssetValidationConfig(asset_path=str(p), profile_id="Prop-Robotics-Neutral")
    for p in Path("/assets").rglob("*.usd")
]

results = sv.validate_asset_list(configs, report_file_path="/reports/validation.json")

n_passed = sum(
    1 for r in results
    if r and all(f["passed"] for f in r.features_summary.values())
)
print(f"{n_passed}/{len(configs)} passed")

The JSON report is merged — new results for the same asset path override existing entries; other entries are preserved. The parent directory must exist before calling.

---

CLI

The simready-validate command is included with the package.

# Validate a single asset
simready-validate \
  --rules-path /specs/capabilities \
  --features-path /specs/features \
  --profiles-path /specs/profiles/profiles.toml \
  --profile Prop-Robotics-Neutral --version 2.0.0 \
  /assets/crate_01.usd

# Validate a list of assets and save a JSON report
simready-validate \
  --rules-path /specs/capabilities \
  --features-path /specs/features \
  --profiles-path /specs/profiles/profiles.toml \
  --profile Prop-Robotics-Neutral \
  --output /reports/results.json \
  --asset_list assets.txt

# Stamp results into the USD on a full pass
simready-validate ... --stamp-asset-validation /assets/crate_01.usd

Flag	Description
ASSET_PATH	Single .usd / .usda file to validate
--asset_list FILE	Text file, one asset path per line
--profile ID	Profile to validate against (omit to infer from USD metadata)
--version VER	Profile version (requires --profile)
--stamp-asset-validation	Write results into USD customLayerData on full pass
--output JSON	Append results to a JSON report file
--rules-path / --features-path / --profiles-path	Load spec data from local paths (repeatable)
-v / --verbose	Per-issue detail on failure + DEBUG logging

Exit code: 0 = all passed, 1 = any failed or not found.

---

Metadata ownership

simready.validate owns only the validation sub-key of SimReady_Metadata in the USD customLayerData. It never reads or writes runtime_tests, asset_id, or any other section.

customLayerData["SimReady_Metadata"]
├── asset_id        ← simready.create
├── validation      ← simready.validate  ✓
└── runtime_tests   ← simready.test

---

License

Copyright (c) 2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
SPDX-License-Identifier: LicenseRef-NvidiaProprietary