Metadata API: Make Metadata Generic

When we use Metadata, it is helpful if the specific signed type (and all of
the signed types attribute types are correctly annotated. Currently this is
not possible.

Making Metadata Generic with constraint T, where

    T = TypeVar("T", "Root", "Timestamp", "Snapshot", "Targets")

allows these annotations. Using Generic annotations is completely
optional so all existing code still works -- the changes in test code
are done to make IDE annotations more useful in the test code, not
because they are required.

Examples:

    md = Metadata[Root].from_bytes(data)
    md:Metadata[Root] = Metadata.from_bytes(data)

In both examples md.signed is now statically typed as "Root" allowing IDE
annotations and static type checking by mypy.

Note that it's not possible to validate that "data" actually contains a
root metadata at runtime in these examples as the annotations are _not_
visible at runtime at all: new constructors would have to be added for that.

Partially fixes theupdateframework#1433

Signed-off-by: Jussi Kukkonen <[email protected]>

tests/test_api.py

@@ -183,7 +183,7 @@ def test_to_from_bytes(self):
 
     def test_sign_verify(self):
         root_path = os.path.join(self.repo_dir, 'metadata', 'root.json')
-        root:Root = Metadata.from_file(root_path).signed
+        root = Metadata[Root].from_file(root_path).signed
 
         # Locate the public keys we need from root
         targets_keyid = next(iter(root.roles["targets"].keyids))
@@ -302,7 +302,7 @@ def test_metadata_base(self):
     def test_metadata_snapshot(self):
         snapshot_path = os.path.join(
                 self.repo_dir, 'metadata', 'snapshot.json')
-        snapshot = Metadata.from_file(snapshot_path)
+        snapshot = Metadata[Snapshot].from_file(snapshot_path)
 
         # Create a MetaFile instance representing what we expect
         # the updated data to be.
@@ -321,7 +321,7 @@ def test_metadata_snapshot(self):
     def test_metadata_timestamp(self):
         timestamp_path = os.path.join(
                 self.repo_dir, 'metadata', 'timestamp.json')
-        timestamp = Metadata.from_file(timestamp_path)
+        timestamp = Metadata[Timestamp].from_file(timestamp_path)
 
         self.assertEqual(timestamp.signed.version, 1)
         timestamp.signed.bump_version()
@@ -358,19 +358,19 @@ def test_metadata_timestamp(self):
 
     def test_metadata_verify_delegate(self):
         root_path = os.path.join(self.repo_dir, 'metadata', 'root.json')
-        root = Metadata.from_file(root_path)
+        root = Metadata[Root].from_file(root_path)
         snapshot_path = os.path.join(
                 self.repo_dir, 'metadata', 'snapshot.json')
-        snapshot = Metadata.from_file(snapshot_path)
+        snapshot = Metadata[Snapshot].from_file(snapshot_path)
         targets_path = os.path.join(
                 self.repo_dir, 'metadata', 'targets.json')
-        targets = Metadata.from_file(targets_path)
+        targets = Metadata[Targets].from_file(targets_path)
         role1_path = os.path.join(
                 self.repo_dir, 'metadata', 'role1.json')
-        role1 = Metadata.from_file(role1_path)
+        role1 = Metadata[Targets].from_file(role1_path)
         role2_path = os.path.join(
                 self.repo_dir, 'metadata', 'role2.json')
-        role2 = Metadata.from_file(role2_path)
+        role2 = Metadata[Targets].from_file(role2_path)
 
         # test the expected delegation tree
         root.verify_delegate('root', root)
@@ -468,7 +468,7 @@ def test_role_class(self):
     def test_metadata_root(self):
         root_path = os.path.join(
                 self.repo_dir, 'metadata', 'root.json')
-        root = Metadata.from_file(root_path)
+        root = Metadata[Root].from_file(root_path)
 
         # Add a second key to root role
         root_key2 =  import_ed25519_publickey_from_file(
@@ -530,7 +530,7 @@ def test_delegation_class(self):
     def test_metadata_targets(self):
         targets_path = os.path.join(
                 self.repo_dir, 'metadata', 'targets.json')
-        targets = Metadata.from_file(targets_path)
+        targets = Metadata[Targets].from_file(targets_path)
 
         # Create a fileinfo dict representing what we expect the updated data to be
         filename = 'file2.txt'
@@ -560,7 +560,7 @@ def  test_length_and_hash_validation(self):
         # for untrusted metadata file to verify.
         timestamp_path = os.path.join(
             self.repo_dir, 'metadata', 'timestamp.json')
-        timestamp = Metadata.from_file(timestamp_path)
+        timestamp = Metadata[Timestamp].from_file(timestamp_path)
         snapshot_metafile = timestamp.signed.meta["snapshot.json"]
 
         snapshot_path = os.path.join(
@@ -603,7 +603,7 @@ def  test_length_and_hash_validation(self):
         # Test target files' hash and length verification
         targets_path = os.path.join(
             self.repo_dir, 'metadata', 'targets.json')
-        targets = Metadata.from_file(targets_path)
+        targets = Metadata[Targets].from_file(targets_path)
         file1_targetfile = targets.signed.targets['file1.txt']
         filepath = os.path.join(
             self.repo_dir, 'targets', 'file1.txt')

tuf/api/metadata.py

@@ -26,12 +26,15 @@
     BinaryIO,
     ClassVar,
     Dict,
+    Generic,
     List,
     Mapping,
     Optional,
     Tuple,
     Type,
+    TypeVar,
     Union,
+    cast,
 )
 
 from securesystemslib import exceptions as sslib_exceptions
@@ -43,6 +46,7 @@
 
 from tuf import exceptions
 from tuf.api.serialization import (
+    DeserializationError,
     MetadataDeserializer,
     MetadataSerializer,
     SignedSerializer,
@@ -56,24 +60,40 @@
 # files to have the same major version (the first number) as ours.
 SPECIFICATION_VERSION = ["1", "0", "19"]
 
+# T is a Generic type constraint for Metadata.signed
+T = TypeVar("T", "Root", "Timestamp", "Snapshot", "Targets")
 
-class Metadata:
+
+class Metadata(Generic[T]):
     """A container for signed TUF metadata.
 
     Provides methods to convert to and from dictionary, read and write to and
     from file and to create and verify metadata signatures.
 
+    Metadata[T] is a generic container type where T can be any one type of
+    [Root, Timestamp, Snapshot, Targets]. The purpose of this is to allow
+    static type checking of the signed attribute in code using Metadata::
+
+        root_md = Metadata[Root].from_file("root.json")
+        # root_md type is now Metadata[Root]. This means signed and its
+        # attributes like consistent_snapshot are now statically typed and the
+        # types can be verified by static type checkers and shown by IDEs
+        print(root_md.signed.consistent_snapshot)
+
+    Using a type constraint is not required but not doing so means T is not a
+    specific type so static typing cannot happen. Note that the type constraint
+    "[Root]" is not validated at runtime (as pure annotations are not available
+    then).
+
     Attributes:
         signed: A subclass of Signed, which has the actual metadata payload,
             i.e. one of Targets, Snapshot, Timestamp or Root.
         signatures: An ordered dictionary of keyids to Signature objects, each
             signing the canonical serialized representation of 'signed'.
     """
 
-    def __init__(
-        self, signed: "Signed", signatures: "OrderedDict[str, Signature]"
-    ):
-        self.signed = signed
+    def __init__(self, signed: T, signatures: OrderedDict[str, Signature]):
+        self.signed: T = signed
         self.signatures = signatures
 
     @classmethod
@@ -119,7 +139,8 @@ def from_dict(cls, metadata: Dict[str, Any]) -> "Metadata":
             signatures[sig.keyid] = sig
 
         return cls(
-            signed=inner_cls.from_dict(metadata.pop("signed")),
+            # Specific type T is not known at static type check time: use cast
+            signed=cast(T, inner_cls.from_dict(metadata.pop("signed"))),
             signatures=signatures,
         )
 
@@ -129,7 +150,7 @@ def from_file(
         filename: str,
         deserializer: Optional[MetadataDeserializer] = None,
         storage_backend: Optional[StorageBackendInterface] = None,
-    ) -> "Metadata":
+    ) -> "Metadata[T]":
         """Loads TUF metadata from file storage.
 
         Arguments:
@@ -160,7 +181,7 @@ def from_file(
     def from_bytes(
         data: bytes,
         deserializer: Optional[MetadataDeserializer] = None,
-    ) -> "Metadata":
+    ) -> "Metadata[T]":
         """Loads TUF metadata from raw data.
 
         Arguments: