mirror of
https://github.com/norandom/OpenUltraSAST.git
synced 2026-11-13 01:48:17 +00:00
feat(detection-core): build exact family catalog resolution
This commit is contained in:
@@ -0,0 +1,180 @@
|
||||
"""Exact immutable family catalogs and capability declarations."""
|
||||
|
||||
from __future__ import annotations
|
||||
|
||||
import re
|
||||
from dataclasses import dataclass
|
||||
|
||||
from openultrasast.core import canonical
|
||||
from openultrasast.core.family import FamilyContract, family_contract_digest, validate_contract
|
||||
|
||||
FAMILY_CATALOG_SCHEMA = "openultrasast.detection-core.family-catalog.v2"
|
||||
_DIGEST = re.compile(r"[0-9a-f]{64}")
|
||||
|
||||
|
||||
class FamilyCatalogError(ValueError):
|
||||
"""A family catalog violates its canonical closed-set contract."""
|
||||
|
||||
|
||||
class FamilyCatalogResolutionError(FamilyCatalogError, LookupError):
|
||||
"""No exact family, digest, and ecosystem catalog entry exists."""
|
||||
|
||||
|
||||
class FamilyCapabilityError(FamilyCatalogError):
|
||||
"""A family capability declaration is incomplete or substituted."""
|
||||
|
||||
|
||||
def _text(value: object, field: str, error: type[FamilyCatalogError] = FamilyCatalogError) -> None:
|
||||
if not isinstance(value, str) or not value:
|
||||
raise error(f"{field} must be nonempty text")
|
||||
|
||||
|
||||
def _digest(value: object, field: str, error: type[FamilyCatalogError] = FamilyCatalogError) -> None:
|
||||
if not isinstance(value, str) or _DIGEST.fullmatch(value) is None:
|
||||
raise error(f"{field} must be a full lowercase SHA-256 digest")
|
||||
|
||||
|
||||
@dataclass(frozen=True)
|
||||
class FamilyCatalogEntry:
|
||||
contract: FamilyContract
|
||||
contract_digest: str
|
||||
|
||||
def __post_init__(self) -> None:
|
||||
if not isinstance(self.contract, FamilyContract):
|
||||
raise FamilyCatalogError("entry contract must be a FamilyContract")
|
||||
_digest(self.contract_digest, "contract digest")
|
||||
validate_contract(self.contract)
|
||||
if self.contract_digest != family_contract_digest(self.contract):
|
||||
raise FamilyCatalogError("contract digest does not match contract content")
|
||||
|
||||
|
||||
def catalog_entry_key(entry: FamilyCatalogEntry) -> tuple[str, str, str, str]:
|
||||
if not isinstance(entry, FamilyCatalogEntry):
|
||||
raise FamilyCatalogError("catalog entry must be a FamilyCatalogEntry")
|
||||
contract = entry.contract
|
||||
return (contract.family_id, contract.family_version, contract.schema_version, entry.contract_digest)
|
||||
|
||||
|
||||
def _entry_value(entry: FamilyCatalogEntry) -> canonical.JsonValue:
|
||||
contract = entry.contract
|
||||
return {
|
||||
"family_id": contract.family_id,
|
||||
"family_version": contract.family_version,
|
||||
"schema_version": contract.schema_version,
|
||||
"ecosystems": contract.ecosystems,
|
||||
"contract_digest": entry.contract_digest,
|
||||
}
|
||||
|
||||
|
||||
def derive_catalog_digest(entries: tuple[FamilyCatalogEntry, ...]) -> str:
|
||||
if not isinstance(entries, tuple) or not entries:
|
||||
raise FamilyCatalogError("entries must be a nonempty tuple")
|
||||
if not all(isinstance(entry, FamilyCatalogEntry) for entry in entries):
|
||||
raise FamilyCatalogError("entries must contain FamilyCatalogEntry values")
|
||||
return canonical.content_digest({"schema_version": FAMILY_CATALOG_SCHEMA, "entries": tuple(_entry_value(entry) for entry in entries)})
|
||||
|
||||
|
||||
@dataclass(frozen=True)
|
||||
class FamilyCatalog:
|
||||
entries: tuple[FamilyCatalogEntry, ...]
|
||||
catalog_digest: str
|
||||
|
||||
def __post_init__(self) -> None:
|
||||
validate_catalog(self)
|
||||
|
||||
|
||||
def _validate_unique_entries(entries: tuple[FamilyCatalogEntry, ...]) -> None:
|
||||
identities = tuple((item.contract.family_id, item.contract.family_version, item.contract.schema_version) for item in entries)
|
||||
if len(set(identities)) != len(identities):
|
||||
raise FamilyCatalogError("duplicate family contract identity")
|
||||
digests = tuple(item.contract_digest for item in entries)
|
||||
if len(set(digests)) != len(digests):
|
||||
raise FamilyCatalogError("duplicate contract digest")
|
||||
|
||||
|
||||
def validate_catalog(value: FamilyCatalog) -> None:
|
||||
if not isinstance(value, FamilyCatalog):
|
||||
raise FamilyCatalogError("catalog must be a FamilyCatalog")
|
||||
if not isinstance(value.entries, tuple) or not value.entries:
|
||||
raise FamilyCatalogError("entries must be a nonempty tuple")
|
||||
if not all(isinstance(entry, FamilyCatalogEntry) for entry in value.entries):
|
||||
raise FamilyCatalogError("entries must contain FamilyCatalogEntry values")
|
||||
_digest(value.catalog_digest, "catalog digest")
|
||||
if value.entries != tuple(sorted(value.entries, key=catalog_entry_key)):
|
||||
raise FamilyCatalogError("catalog entries must be in canonical order")
|
||||
_validate_unique_entries(value.entries)
|
||||
for entry in value.entries:
|
||||
validate_contract(entry.contract)
|
||||
if entry.contract_digest != family_contract_digest(entry.contract):
|
||||
raise FamilyCatalogError("contract digest does not match contract content")
|
||||
if value.catalog_digest != derive_catalog_digest(value.entries):
|
||||
raise FamilyCatalogError("catalog digest does not match canonical entries")
|
||||
|
||||
|
||||
def resolve_contract(
|
||||
value: FamilyCatalog,
|
||||
family_id: str,
|
||||
contract_digest: str,
|
||||
ecosystem: str,
|
||||
) -> FamilyContract:
|
||||
validate_catalog(value)
|
||||
_text(family_id, "family_id", FamilyCatalogResolutionError)
|
||||
_digest(contract_digest, "contract digest", FamilyCatalogResolutionError)
|
||||
_text(ecosystem, "ecosystem", FamilyCatalogResolutionError)
|
||||
matches = tuple(
|
||||
entry.contract
|
||||
for entry in value.entries
|
||||
if entry.contract.family_id == family_id and entry.contract_digest == contract_digest and ecosystem in entry.contract.ecosystems
|
||||
)
|
||||
if len(matches) != 1:
|
||||
raise FamilyCatalogResolutionError("family, digest, and ecosystem must resolve one exact contract")
|
||||
return matches[0]
|
||||
|
||||
|
||||
@dataclass(frozen=True)
|
||||
class FamilyCapabilityDeclaration:
|
||||
arm_id: str
|
||||
family_contract_digest: str
|
||||
readiness_fixture_digest: str
|
||||
development_population_digest: str
|
||||
admission_boundary_digest: str
|
||||
|
||||
def __post_init__(self) -> None:
|
||||
_validate_capability_fields(self)
|
||||
|
||||
|
||||
def _validate_capability_fields(declaration: FamilyCapabilityDeclaration) -> None:
|
||||
_text(declaration.arm_id, "arm_id", FamilyCapabilityError)
|
||||
for value, field in (
|
||||
(declaration.family_contract_digest, "family contract digest"),
|
||||
(declaration.readiness_fixture_digest, "readiness fixture digest"),
|
||||
(declaration.development_population_digest, "development population digest"),
|
||||
(declaration.admission_boundary_digest, "admission boundary digest"),
|
||||
):
|
||||
_digest(value, field, FamilyCapabilityError)
|
||||
|
||||
|
||||
def validate_capability(declaration: FamilyCapabilityDeclaration, value: FamilyCatalog) -> None:
|
||||
if not isinstance(declaration, FamilyCapabilityDeclaration):
|
||||
raise FamilyCapabilityError("declaration must be a FamilyCapabilityDeclaration")
|
||||
_validate_capability_fields(declaration)
|
||||
validate_catalog(value)
|
||||
matches = tuple(entry for entry in value.entries if entry.contract_digest == declaration.family_contract_digest)
|
||||
if len(matches) != 1:
|
||||
raise FamilyCapabilityError("capability contract digest must identify one exact catalog entry")
|
||||
|
||||
|
||||
__all__ = [
|
||||
"FAMILY_CATALOG_SCHEMA",
|
||||
"FamilyCapabilityDeclaration",
|
||||
"FamilyCapabilityError",
|
||||
"FamilyCatalog",
|
||||
"FamilyCatalogEntry",
|
||||
"FamilyCatalogError",
|
||||
"FamilyCatalogResolutionError",
|
||||
"catalog_entry_key",
|
||||
"derive_catalog_digest",
|
||||
"resolve_contract",
|
||||
"validate_capability",
|
||||
"validate_catalog",
|
||||
]
|
||||
Reference in New Issue
Block a user