Add in-toto format as hash of shard hashes

This converts model serialization manifests that record every model file shard hash into an in-toto payload that can then be passed to Sigstore's `sign_intoto` for signing to generate a Sigstore `Bundle` (if using Sigstore). To identify the models, we compute a hash of all hashes of the file shards and use that as the subject. The individual file hashes are used as the payload and we would have the verifier check them as part of the verification process. Signed-off-by: Mihai Maruseac <mihaimaruseac@google.com>
sigstore · Jul 27, 2024 · 51463cd · 51463cd
1 parent 0dd7eb4
commit 51463cd
Show file tree

Hide file tree

Showing 9 changed files with 345 additions and 0 deletions.
diff --git a/model_signing/signing/in_toto.py b/model_signing/signing/in_toto.py
@@ -253,3 +253,101 @@ def from_manifest(cls, manifest: manifest_module.Manifest) -> Self:
             predicate_top_level_name="files",
         )
         return cls(statement)
+
+
+class DigestOfShardDigestsIntotoPayload(IntotoPayload):
+    """In-toto payload where the subject is a digest of digests of file shards.
+
+    This payload is supposed to be used for manifests where every file shard in
+    the model is matched with a digest. Because existing tooling only supports
+    established hashing algorithms, we record every such digest in the predicate
+    part and compute a hash for the subject by using sha256 on the concatenation
+    of the shard hashes. To ensure determinism, the hashes are sorted
+    by file shard (alphabetically by name, then ordered by start offset).
+
+    Example:
+    ```json
+    {
+      "_type": "https://in-toto.io/Statement/v1",
+      "subject": [
+        {
+          "digest": {
+            "sha256": "18b5a4..."
+          }
+        }
+      ],
+      "predicateType": "https://model_signing/DigestOfShardDigests/v0.1",
+      "predicate": {
+        "shards": [
+          {
+            "digest": "6efa14...",
+            "algorithm": "file-sha256-1000000",
+            "name": "d0/d1/d2/d3/d4/f0:0:16"
+          },
+          {
+            "digest": "a9bc14...",
+            "algorithm": "file-sha256-1000000",
+            "name": "d0/d1/d2/d3/d4/f1:0:16"
+          },
+          {
+            "digest": "5f597e...",
+            "algorithm": "file-sha256-1000000",
+            "name": "d0/d1/d2/d3/d4/f2:0:16"
+          },
+          {
+            "digest": "eaf677...",
+            "algorithm": "file-sha256-1000000",
+            "name": "d0/d1/d2/d3/d4/f3:0:16"
+          }
+        ]
+      }
+    }
+    ```
+
+    A missing predicate, or a predicate for which an entry does not have valid
+    name, digest, or algorithm should be considered invalid and fail integrity
+    verification.
+
+    See also https://github.com/sigstore/sigstore-python/issues/1018.
+    """
+
+    predicate_type: Final[str] = (
+        "https://model_signing/DigestOfShardDigests/v0.1"
+    )
+
+    def __init__(self, statement: statement.Statement):
+        """Builds an instance of this in-toto payload.
+
+        Don't call this directly in production. Use `from_manifest()` instead.
+
+        Args:
+            statement: The DSSE statement representing this in-toto payload.
+        """
+        self.statement = statement
+
+    @classmethod
+    @override
+    def from_manifest(cls, manifest: manifest_module.Manifest) -> Self:
+        """Converts a manifest to the signing payload used for signing.
+
+        The manifest must be one where every model shard is paired with its own
+        digest. Currently, this is only `ShardLevelManifest`.
+
+        Args:
+            manifest: the manifest to convert to signing payload.
+
+        Returns:
+            An instance of `DigestOfDigestsIntotoPayload`.
+
+        Raises:
+            TypeError: If the manifest is not `ShardLevelManifest`.
+        """
+        if not isinstance(manifest, manifest_module.ShardLevelManifest):
+            raise TypeError("Only ShardLevelManifest is supported")
+
+        statement = _convert_descriptors_to_hashed_statement(
+            manifest,
+            predicate_type=cls.predicate_type,
+            predicate_top_level_name="shards",
+        )
+        return cls(statement)
diff --git a/model_signing/signing/in_toto_test.py b/model_signing/signing/in_toto_test.py
@@ -31,6 +31,7 @@
 from model_signing.hashing import hashing
 from model_signing.manifest import manifest as manifest_module
 from model_signing.serialization import serialize_by_file
+from model_signing.serialization import serialize_by_file_shard
 from model_signing.signing import in_toto
 
 
@@ -145,3 +146,67 @@ def test_only_runs_on_expected_manifest_types(self):
             match="Only FileLevelManifest is supported",
         ):
             in_toto.DigestOfDigestsIntotoPayload.from_manifest(manifest)
+
+
+class TestDigestOfShardDigestsIntotoPayload:
+
+    def _hasher_factory(
+        self, path: pathlib.Path, start: int, end: int
+    ) -> file.ShardedFileHasher:
+        return file.ShardedFileHasher(
+            path, memory.SHA256(), start=start, end=end
+        )
+
+    @pytest.mark.parametrize("model_fixture_name", test_support.all_test_models)
+    def test_known_models(self, request, model_fixture_name):
+        # Set up variables (arrange)
+        testdata_path = request.path.parent / "testdata"
+        test_path = testdata_path / "in_toto"
+        test_class_path = test_path / "TestDigestOfShardDigestsIntotoPayload"
+        golden_path = test_class_path / model_fixture_name
+        should_update = request.config.getoption("update_goldens")
+        model = request.getfixturevalue(model_fixture_name)
+
+        # Compute payload (act)
+        serializer = serialize_by_file_shard.ManifestSerializer(
+            self._hasher_factory, allow_symlinks=True
+        )
+        manifest = serializer.serialize(model)
+        payload = in_toto.DigestOfShardDigestsIntotoPayload.from_manifest(
+            manifest
+        )
+
+        # Compare with golden, or write to golden (approximately "assert")
+        if should_update:
+            with open(golden_path, "w", encoding="utf-8") as f:
+                f.write(f"{json_format.MessageToJson(payload.statement.pb)}\n")
+        else:
+            with open(golden_path, "r", encoding="utf-8") as f:
+                json_contents = f.read()
+                expected_proto = json_format.Parse(
+                    json_contents, statement_pb2.Statement()
+                )
+
+            assert payload.statement.pb == expected_proto
+
+    def test_produces_valid_statements(self, sample_model_folder):
+        serializer = serialize_by_file_shard.ManifestSerializer(
+            self._hasher_factory, allow_symlinks=True
+        )
+        manifest = serializer.serialize(sample_model_folder)
+
+        payload = in_toto.DigestOfShardDigestsIntotoPayload.from_manifest(
+            manifest
+        )
+
+        payload.statement.validate()
+
+    def test_only_runs_on_expected_manifest_types(self):
+        digest = hashing.Digest("test", b"test_digest")
+        manifest = manifest_module.DigestManifest(digest)
+
+        with pytest.raises(
+            TypeError,
+            match="Only ShardLevelManifest is supported",
+        ):
+            in_toto.DigestOfShardDigestsIntotoPayload.from_manifest(manifest)
diff --git a/..._signing/signing/testdata/in_toto/TestDigestOfShardDigestsIntotoPayload/deep_model_folder b/..._signing/signing/testdata/in_toto/TestDigestOfShardDigestsIntotoPayload/deep_model_folder
@@ -0,0 +1,35 @@
+{
+  "_type": "https://in-toto.io/Statement/v1",
+  "subject": [
+    {
+      "digest": {
+        "sha256": "18b5a45fe7983f7194e8ffd96c80f5f0ec53191bf4a32b6aff293f043e816d7a"
+      }
+    }
+  ],
+  "predicateType": "https://model_signing/DigestOfShardDigests/v0.1",
+  "predicate": {
+    "shards": [
+      {
+        "digest": "6efa14bb03544fcb76045c55f25b9315b6eb5be2d8a85f703193a76b7874c6ff",
+        "algorithm": "file-sha256-1000000",
+        "name": "d0/d1/d2/d3/d4/f0:0:16"
+      },
+      {
+        "digest": "a9bc149b70b9d325cd68d275d582cfdb98c0347d3ce54590aa6533368daed3d2",
+        "algorithm": "file-sha256-1000000",
+        "name": "d0/d1/d2/d3/d4/f1:0:16"
+      },
+      {
+        "digest": "5f597e6a92d1324d9adbed43d527926d11d0131487baf315e65ae1ef3b1ca3c0",
+        "algorithm": "file-sha256-1000000",
+        "name": "d0/d1/d2/d3/d4/f2:0:16"
+      },
+      {
+        "digest": "eaf677c35fec6b87889d9e4563d8bb65dcb9869ca0225697c9cc44cf49dca008",
+        "algorithm": "file-sha256-1000000",
+        "name": "d0/d1/d2/d3/d4/f3:0:16"
+      }
+    ]
+  }
+}
diff --git a/...l_signing/signing/testdata/in_toto/TestDigestOfShardDigestsIntotoPayload/empty_model_file b/...l_signing/signing/testdata/in_toto/TestDigestOfShardDigestsIntotoPayload/empty_model_file
@@ -0,0 +1,14 @@
+{
+  "_type": "https://in-toto.io/Statement/v1",
+  "subject": [
+    {
+      "digest": {
+        "sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855"
+      }
+    }
+  ],
+  "predicateType": "https://model_signing/DigestOfShardDigests/v0.1",
+  "predicate": {
+    "shards": []
+  }
+}
diff --git a/...signing/signing/testdata/in_toto/TestDigestOfShardDigestsIntotoPayload/empty_model_folder b/...signing/signing/testdata/in_toto/TestDigestOfShardDigestsIntotoPayload/empty_model_folder
@@ -0,0 +1,14 @@
+{
+  "_type": "https://in-toto.io/Statement/v1",
+  "subject": [
+    {
+      "digest": {
+        "sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855"
+      }
+    }
+  ],
+  "predicateType": "https://model_signing/DigestOfShardDigests/v0.1",
+  "predicate": {
+    "shards": []
+  }
+}
diff --git a/...gning/testdata/in_toto/TestDigestOfShardDigestsIntotoPayload/model_folder_with_empty_file b/...gning/testdata/in_toto/TestDigestOfShardDigestsIntotoPayload/model_folder_with_empty_file
@@ -0,0 +1,14 @@
+{
+  "_type": "https://in-toto.io/Statement/v1",
+  "subject": [
+    {
+      "digest": {
+        "sha256": "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855"
+      }
+    }
+  ],
+  "predicateType": "https://model_signing/DigestOfShardDigests/v0.1",
+  "predicate": {
+    "shards": []
+  }
+}
diff --git a/..._signing/signing/testdata/in_toto/TestDigestOfShardDigestsIntotoPayload/sample_model_file b/..._signing/signing/testdata/in_toto/TestDigestOfShardDigestsIntotoPayload/sample_model_file
@@ -0,0 +1,20 @@
+{
+  "_type": "https://in-toto.io/Statement/v1",
+  "subject": [
+    {
+      "digest": {
+        "sha256": "002d162f867c5eee944e5080d25829b6625be0e3f081f6fbafc7dd655ca2e178"
+      }
+    }
+  ],
+  "predicateType": "https://model_signing/DigestOfShardDigests/v0.1",
+  "predicate": {
+    "shards": [
+      {
+        "digest": "3aab065c7181a173b5dd9e9d32a9f79923440b413be1e1ffcdba26a7365f719b",
+        "algorithm": "file-sha256-1000000",
+        "name": ".:0:22"
+      }
+    ]
+  }
+}
diff --git a/...igning/signing/testdata/in_toto/TestDigestOfShardDigestsIntotoPayload/sample_model_folder b/...igning/signing/testdata/in_toto/TestDigestOfShardDigestsIntotoPayload/sample_model_folder
@@ -0,0 +1,65 @@
+{
+  "_type": "https://in-toto.io/Statement/v1",
+  "subject": [
+    {
+      "digest": {
+        "sha256": "143cc682e555951649f18e2761c3d526d2502996f5e32dc187ef7f8a614f8df7"
+      }
+    }
+  ],
+  "predicateType": "https://model_signing/DigestOfShardDigests/v0.1",
+  "predicate": {
+    "shards": [
+      {
+        "digest": "fdd8925354242a7fd1515e79534317b800015607a609cd306e0b4dcfe6c92249",
+        "algorithm": "file-sha256-1000000",
+        "name": "d0/f00:0:23"
+      },
+      {
+        "digest": "e16940b5e44ce981150bda37c4ba95881a749a521b4a297c5cdf97bdcfe965e6",
+        "algorithm": "file-sha256-1000000",
+        "name": "d0/f01:0:23"
+      },
+      {
+        "digest": "407822246ea8f9e26380842c3f4cd10d7b23e78f1fe7c74c293608682886a426",
+        "algorithm": "file-sha256-1000000",
+        "name": "d0/f02:0:23"
+      },
+      {
+        "digest": "6a3b08b5df77c4d418ceee1ac136a9ad49fc7c41358b5e82c1176daccb21ff3f",
+        "algorithm": "file-sha256-1000000",
+        "name": "d1/f10:0:23"
+      },
+      {
+        "digest": "a484b3d8ea5e99b75f9f123f9a42c882388693edc7d85d82ccba54834712cadf",
+        "algorithm": "file-sha256-1000000",
+        "name": "d1/f11:0:23"
+      },
+      {
+        "digest": "8f577930f5f40c2c2133cb299d36f9527fde98c1608569017cae6b5bcd01abb3",
+        "algorithm": "file-sha256-1000000",
+        "name": "d1/f12:0:23"
+      },
+      {
+        "digest": "997b37cc51f1ca1c7a270466607e26847429cd7264c30148c1b9352e224083fc",
+        "algorithm": "file-sha256-1000000",
+        "name": "f0:0:24"
+      },
+      {
+        "digest": "c88a04d48353133fb065ba2c8ab369abab21395b9526aa20373ad828915fa7ae",
+        "algorithm": "file-sha256-1000000",
+        "name": "f1:0:24"
+      },
+      {
+        "digest": "700e3ba5065d8dd47e41fd928ea086670d628f891ba363be0ca3c31d20d7d719",
+        "algorithm": "file-sha256-1000000",
+        "name": "f2:0:24"
+      },
+      {
+        "digest": "912bcf5ebdf44dc7b4085b07940e0a81d157fba24b276e73fd911121d4544c4a",
+        "algorithm": "file-sha256-1000000",
+        "name": "f3:0:24"
+      }
+    ]
+  }
+}
diff --git a/...gning/signing/testdata/in_toto/TestDigestOfShardDigestsIntotoPayload/symlink_model_folder b/...gning/signing/testdata/in_toto/TestDigestOfShardDigestsIntotoPayload/symlink_model_folder
@@ -0,0 +1,20 @@
+{
+  "_type": "https://in-toto.io/Statement/v1",
+  "subject": [
+    {
+      "digest": {
+        "sha256": "002d162f867c5eee944e5080d25829b6625be0e3f081f6fbafc7dd655ca2e178"
+      }
+    }
+  ],
+  "predicateType": "https://model_signing/DigestOfShardDigests/v0.1",
+  "predicate": {
+    "shards": [
+      {
+        "digest": "3aab065c7181a173b5dd9e9d32a9f79923440b413be1e1ffcdba26a7365f719b",
+        "algorithm": "file-sha256-1000000",
+        "name": "symlink_file:0:22"
+      }
+    ]
+  }
+}