WP-0001-T009: digest abstraction and content address (ADR-0001)

src/artifactstore/identity/__init__.py:
- Digest: frozen, hashable dataclass (algorithm + lowercase hex), validated.
- ContentAddress: canonical `<algorithm>:<hex>` string form with validating
  parser (to_digest) and emitter (str / from_digest).
- DigestPair: dual-digest result (primary + sha256) from a single hashing pass.
- Algorithm registry: register_algorithm / get_algorithm / list_algorithms
  with name validation `[a-z][a-z0-9_-]*`.
- digest_bytes (sync) and digest_stream (async) — single-pass dual hashing.
- BLAKE3 registered as PRIMARY_ALGORITHM, SHA-256 as INTEROP_ALGORITHM at
  module import.

tests/unit/test_identity.py:
- Hypothesis property test asserts digest_bytes matches hashlib.sha256 and
  blake3.blake3 for random byte sequences up to 4 KiB.
- digest_stream invariants: equivalence with digest_bytes under chunked input;
  defaults to BLAKE3 primary; always computes SHA-256; handles empty input.
- Digest / ContentAddress invariants: rejects uppercase hex, empty fields,
  odd hex length, missing separator; frozen and hashable.

Gates: ruff clean, mypy --strict clean on 21 source files, 18 tests pass.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
This commit is contained in:
2026-05-16 01:34:24 +02:00
parent 6a136dd814
commit c1bfb8b486
3 changed files with 326 additions and 4 deletions

View File

@@ -95,7 +95,7 @@ Acceptance:
```task
id: ARTIFACT-STORE-WP-0001-T009
status: todo
status: done
priority: high
state_hub_task_id: "4dc465c5-5c14-412d-b8c0-aa84076e4560"
```