Address and Metadata have slots

Default values for frozen dataclass attrs are reused sometimes

Author: software-dov

gapic/schema/metadata.py

@@ -37,21 +37,58 @@
 from gapic.utils import RESERVED_NAMES
 
 
-@dataclasses.dataclass(frozen=True)
+@dataclasses.dataclass(init=False, frozen=True)
 class Address:
-    name: str = ''
-    module: str = ''
-    module_path: Tuple[int, ...] = dataclasses.field(default_factory=tuple)
-    package: Tuple[str, ...] = dataclasses.field(default_factory=tuple)
-    parent: Tuple[str, ...] = dataclasses.field(default_factory=tuple)
-    api_naming: naming.Naming = dataclasses.field(
-        default_factory=naming.NewNaming,
+    # There are lots of addresses and for large APIs they take up a lot of memory.
+    # See
+    # https://stackoverflow.com/questions/50180735/how-can-dataclasses-be-made-to-work-better-with-slots
+    # for the current workaround for using slots with dataclasses.
+    # If this becomes a common pattern (i.e. we need to reduce memory even more)
+    # it may be worthwhile to look at https://github.com/starhel/dataslots
+    __slots__ = (
+        "name",
+        "module",
+        "module_path",
+        "package",
+        "parent",
+        "api_naming",
+        "collisions",
+        "_cached_values",
     )
-    collisions: FrozenSet[str] = dataclasses.field(default_factory=frozenset)
+    name: str
+    module: str
+    module_path: Tuple[int, ...]
+    package: Tuple[str, ...]
+    parent: Tuple[str, ...]
+    api_naming: naming.Naming
+    collisions: FrozenSet[str]
+
+    def __init__(
+            self,
+            name="",
+            module="",
+            module_path=(),
+            package=(),
+            parent=(),
+            api_naming=naming.NewNaming(),
+            collisions=frozenset(),
+    ):
+        super().__init__()
+        super().__setattr__("name", name)
+        super().__setattr__("module", module)
+        super().__setattr__("module_path", module_path)
+        super().__setattr__("package", package)
+        super().__setattr__("parent", parent)
+        super().__setattr__("api_naming", api_naming)
+        super().__setattr__("collisions", collisions)
 
     def __eq__(self, other) -> bool:
-        return all([getattr(self, i) == getattr(other, i) for i
-                    in ('name', 'module', 'module_path', 'package', 'parent')])
+        # Ignore collisions and api_naming.
+        # Collisions are just a name disambiguation mechanism.
+        return all(
+            getattr(self, i) == getattr(other, i)
+            for i in ('name', 'module', 'module_path', 'package', 'parent')
+        )
 
     def __hash__(self):
         # Do NOT include collisions; they are not relevant.
@@ -111,6 +148,8 @@ def module_alias(self) -> str:
         while still providing names that are fundamentally readable
         to users (albeit looking auto-generated).
         """
+        # Don't keep RESERVED_NAMES in self.collisions,
+        # just combine the two when necessary.
         if self.module in self.collisions | RESERVED_NAMES:
             return '_'.join(
                 (
@@ -283,12 +322,23 @@ def with_context(self, *, collisions: FrozenSet[str]) -> 'Address':
         return dataclasses.replace(self, collisions=collisions)
 
 
-@dataclasses.dataclass(frozen=True)
+@dataclasses.dataclass(init=False, frozen=True)
 class Metadata:
-    address: Address = dataclasses.field(default_factory=Address)
-    documentation: descriptor_pb2.SourceCodeInfo.Location = dataclasses.field(
-        default_factory=descriptor_pb2.SourceCodeInfo.Location,
-    )
+    # See the comment for Address that describes the hoops necessary for
+    # __slots__ with dataclasses.
+    __slots__ = ("address", "documentation")
+
+    address: Address
+    documentation: descriptor_pb2.SourceCodeInfo.Location
+
+    def __init__(
+            self,
+            address=Address(),
+            documentation=descriptor_pb2.SourceCodeInfo.Location()
+    ):
+        super().__init__()
+        super().__setattr__("address", address)
+        super().__setattr__("documentation", documentation)
 
     @property
     def doc(self):

gapic/schema/wrappers.py

@@ -1111,7 +1111,7 @@ def with_context(self, *, collisions: FrozenSet[str]) -> 'Service':
             self,
             methods=collections.OrderedDict(
                 (k, v.with_context(
-                    # A methodd's flattened fields create additional names
+                    # A method's flattened fields create additional names
                     # that may conflict with module imports.
                     collisions=collisions | frozenset(v.flattened_fields.keys()))
                  )

tests/unit/schema/wrappers/test_python.py

@@ -22,9 +22,7 @@ def test_python_eq():
     meta = metadata.Metadata(address=metadata.Address(
         name='Foo', module='bar', package=('google', 'api'),
     ))
-    assert wrappers.PythonType(meta=meta) == wrappers.PythonType(
-        meta=copy.copy(meta),
-    )
+    assert wrappers.PythonType(meta=meta) == wrappers.PythonType(meta=meta)
     assert wrappers.PythonType(meta=metadata.Metadata(
         address=metadata.Address(name='Baz', module='bar', package=()),
     )) != wrappers.PythonType(meta=meta)