Files
Open_Ultra_SAST/src/openultrasast/core/lineage.py
T

274 lines
11 KiB
Python

"""Pure lineage contracts bound to authoritative population custody chains."""
from __future__ import annotations
import re
from dataclasses import dataclass, replace
from openultrasast.core import canonical
from openultrasast.core.custody import (
LineagePurpose,
PopulationCustodyPolicy,
PopulationCustodyPort,
PopulationStatus,
PopulationUseReceipt,
validate_population_custody_log,
)
LINEAGE_MANIFEST_SCHEMA = "openultrasast.detection-core.lineage-manifest.v2"
LINEAGE_DECISION_SCHEMA = "openultrasast.detection-core.lineage-decision.v2"
_DIGEST = re.compile(r"[0-9a-f]{64}")
class LineageValidationError(ValueError):
"""A lineage identity or custody join cannot be reproduced."""
def _require_digest(value: object, field_name: str) -> None:
if not isinstance(value, str) or _DIGEST.fullmatch(value) is None:
raise canonical.DigestFormatError(f"{field_name} must be a full lowercase SHA-256 digest")
def _require_text(value: object, field_name: str) -> None:
if not isinstance(value, str) or not value:
raise ValueError(f"{field_name} must be nonempty text")
def _require_count(value: object, field_name: str) -> None:
if isinstance(value, bool) or not isinstance(value, int) or value < 0:
raise ValueError(f"{field_name} must be a nonnegative integer")
@dataclass(frozen=True)
class LineageUse:
purpose: LineagePurpose
source_population_digest: str
labels_accessed: bool
influenced_artifact_digest: str
population_use_receipt_digest: str
def __post_init__(self) -> None:
if not isinstance(self.purpose, LineagePurpose):
raise TypeError("purpose must be a LineagePurpose")
_require_digest(self.source_population_digest, "source_population_digest")
if not isinstance(self.labels_accessed, bool):
raise TypeError("labels_accessed must be boolean")
_require_digest(self.influenced_artifact_digest, "influenced_artifact_digest")
_require_digest(self.population_use_receipt_digest, "population_use_receipt_digest")
@dataclass(frozen=True)
class LineageManifest:
arm_id: str
subject_digest: str
uses: tuple[LineageUse, ...]
manifest_digest: str
def __post_init__(self) -> None:
_require_text(self.arm_id, "arm_id")
_require_digest(self.subject_digest, "subject_digest")
if not isinstance(self.uses, tuple) or not all(isinstance(use, LineageUse) for use in self.uses):
raise TypeError("uses must be a tuple of LineageUse values")
_require_digest(self.manifest_digest, "manifest_digest")
@dataclass(frozen=True)
class LineageExpectation:
purpose: LineagePurpose
influenced_artifact_digest: str
minimum_uses: int
maximum_uses: int | None
def __post_init__(self) -> None:
if not isinstance(self.purpose, LineagePurpose):
raise TypeError("purpose must be a LineagePurpose")
_require_digest(self.influenced_artifact_digest, "influenced_artifact_digest")
_require_count(self.minimum_uses, "minimum_uses")
if self.maximum_uses is not None:
_require_count(self.maximum_uses, "maximum_uses")
if self.maximum_uses < self.minimum_uses:
raise ValueError("maximum_uses must be at least minimum_uses")
@dataclass(frozen=True)
class LineageDecision:
allowed: bool
spent_population_digests: tuple[str, ...]
reasons: tuple[str, ...]
decision_digest: str
def __post_init__(self) -> None:
if not isinstance(self.allowed, bool):
raise TypeError("allowed must be boolean")
if not isinstance(self.spent_population_digests, tuple):
raise TypeError("spent_population_digests must be a tuple")
for digest in self.spent_population_digests:
_require_digest(digest, "spent_population_digest")
if tuple(sorted(set(self.spent_population_digests))) != self.spent_population_digests:
raise ValueError("spent_population_digests must be sorted and unique")
if not isinstance(self.reasons, tuple) or not all(isinstance(reason, str) and reason for reason in self.reasons):
raise TypeError("reasons must be a tuple of nonempty text")
if tuple(sorted(set(self.reasons))) != self.reasons:
raise ValueError("reasons must be sorted and unique")
_require_digest(self.decision_digest, "decision_digest")
def _use_value(use: LineageUse) -> canonical.JsonValue:
return {
"purpose": use.purpose.value,
"source_population_digest": use.source_population_digest,
"labels_accessed": use.labels_accessed,
"influenced_artifact_digest": use.influenced_artifact_digest,
"population_use_receipt_digest": use.population_use_receipt_digest,
}
def derive_lineage_manifest_digest(arm_id: str, subject_digest: str, uses: tuple[LineageUse, ...]) -> str:
_require_text(arm_id, "arm_id")
_require_digest(subject_digest, "subject_digest")
return canonical.content_digest(
{
"schema_version": LINEAGE_MANIFEST_SCHEMA,
"arm_id": arm_id,
"subject_digest": subject_digest,
"uses": tuple(_use_value(use) for use in uses),
}
)
def _decision_value(decision: LineageDecision) -> canonical.JsonValue:
return {
"schema_version": LINEAGE_DECISION_SCHEMA,
"allowed": decision.allowed,
"spent_population_digests": decision.spent_population_digests,
"reasons": decision.reasons,
}
def derive_lineage_decision_digest(decision: LineageDecision) -> str:
return canonical.content_digest(_decision_value(decision))
def _validate_manifest(lineage: LineageManifest, expected_subject_digest: str) -> None:
_require_digest(expected_subject_digest, "expected_subject_digest")
expected = derive_lineage_manifest_digest(lineage.arm_id, lineage.subject_digest, lineage.uses)
if lineage.manifest_digest != expected:
raise LineageValidationError("lineage manifest digest cannot be re-derived")
if lineage.subject_digest != expected_subject_digest:
raise LineageValidationError("lineage subject does not match expected subject")
receipt_digests = tuple(use.population_use_receipt_digest for use in lineage.uses)
if len(set(receipt_digests)) != len(receipt_digests):
raise LineageValidationError("duplicate lineage receipt")
def _expectation_index(
expectations: tuple[LineageExpectation, ...],
) -> dict[tuple[LineagePurpose, str], LineageExpectation]:
if not isinstance(expectations, tuple) or not all(isinstance(item, LineageExpectation) for item in expectations):
raise TypeError("expectations must be a tuple of LineageExpectation values")
indexed: dict[tuple[LineagePurpose, str], LineageExpectation] = {}
for expectation in expectations:
key = (expectation.purpose, expectation.influenced_artifact_digest)
if key in indexed:
raise LineageValidationError("duplicate lineage expectation")
indexed[key] = expectation
return indexed
def _receipt_for_use(
use: LineageUse,
arm_id: str,
custody: PopulationCustodyPort,
policy: PopulationCustodyPolicy,
) -> PopulationUseReceipt:
log = custody.current_log(use.source_population_digest)
validate_population_custody_log(log, policy, current=custody)
matches = tuple(receipt for receipt in log.receipts if receipt.receipt_digest == use.population_use_receipt_digest)
if len(matches) != 1:
raise LineageValidationError("lineage receipt is absent from the current complete custody chain")
receipt = matches[0]
request = receipt.request
if request.arm_id != arm_id:
raise LineageValidationError("lineage arm does not match its current-chain receipt")
if (
request.population_digest != use.source_population_digest
or request.purpose is not use.purpose
or request.labels_accessed is not use.labels_accessed
or request.influenced_artifact_digest != use.influenced_artifact_digest
):
raise LineageValidationError("lineage use does not match its current-chain receipt")
if use.labels_accessed and receipt.status is not PopulationStatus.SPENT:
raise LineageValidationError("labelled lineage receipt did not spend its population")
return receipt
def _count_reasons(
uses: tuple[LineageUse, ...],
expected: dict[tuple[LineagePurpose, str], LineageExpectation],
) -> list[str]:
counts = {key: 0 for key in expected}
reasons: list[str] = []
for use in uses:
key = (use.purpose, use.influenced_artifact_digest)
if key not in expected:
reasons.append(f"unexpected_use:{use.purpose.value}:{use.influenced_artifact_digest}")
else:
counts[key] += 1
for key, expectation in expected.items():
count = counts[key]
label = f"{key[0].value}:{key[1]}"
if count < expectation.minimum_uses:
reasons.append(f"missing_required_use:{label}")
if expectation.maximum_uses is not None and count > expectation.maximum_uses:
reasons.append(f"too_many_uses:{label}")
return reasons
def _decision(allowed: bool, spent: tuple[str, ...], reasons: tuple[str, ...]) -> LineageDecision:
provisional = LineageDecision(allowed, spent, reasons, "0" * 64)
return replace(provisional, decision_digest=derive_lineage_decision_digest(provisional))
def validate_lineage(
lineage: LineageManifest,
expected_subject_digest: str,
expectations: tuple[LineageExpectation, ...],
protected_population_digest: str | None,
forbid_label_access: bool,
custody: PopulationCustodyPort,
policy: PopulationCustodyPolicy,
) -> LineageDecision:
"""Bind every lineage use to an exact expected artifact and current receipt."""
_validate_manifest(lineage, expected_subject_digest)
if protected_population_digest is not None:
_require_digest(protected_population_digest, "protected_population_digest")
if not isinstance(forbid_label_access, bool):
raise TypeError("forbid_label_access must be boolean")
expected = _expectation_index(expectations)
reasons = _count_reasons(lineage.uses, expected)
spent: set[str] = set()
for use in lineage.uses:
_receipt_for_use(use, lineage.arm_id, custody, policy)
if use.labels_accessed:
spent.add(use.source_population_digest)
if forbid_label_access:
reasons.append(f"label_access_forbidden:{use.source_population_digest}")
if use.source_population_digest == protected_population_digest:
reasons.append(f"protected_population_labels:{use.source_population_digest}")
normalized_reasons = tuple(sorted(set(reasons)))
return _decision(not normalized_reasons, tuple(sorted(spent)), normalized_reasons)
__all__ = [
"LINEAGE_DECISION_SCHEMA",
"LINEAGE_MANIFEST_SCHEMA",
"LineageDecision",
"LineageExpectation",
"LineageManifest",
"LineageUse",
"LineageValidationError",
"derive_lineage_decision_digest",
"derive_lineage_manifest_digest",
"validate_lineage",
]