refactor: improve Hook Hints typing (#285)

* improve Hook Hints typing

Signed-off-by: gruebel <[email protected]>

* ignore lint issue for this line

Signed-off-by: gruebel <[email protected]>

* exclude TYPE_CHECKING from coverage report

Signed-off-by: gruebel <[email protected]>

---------

Signed-off-by: gruebel <[email protected]>

Author: gruebel

openfeature/flag_evaluation.py

@@ -6,8 +6,9 @@
 from openfeature._backports.strenum import StrEnum
 from openfeature.exception import ErrorCode
 
-if typing.TYPE_CHECKING:  # resolves a circular dependency in type annotations
-    from openfeature.hook import Hook
+if typing.TYPE_CHECKING:  # pragma: no cover
+    # resolves a circular dependency in type annotations
+    from openfeature.hook import Hook, HookHints
 
 
 class FlagType(StrEnum):
@@ -48,7 +49,7 @@ class FlagEvaluationDetails(typing.Generic[T_co]):
 @dataclass
 class FlagEvaluationOptions:
     hooks: typing.List[Hook] = field(default_factory=list)
-    hook_hints: dict = field(default_factory=dict)
+    hook_hints: HookHints = field(default_factory=dict)
 
 
 U_co = typing.TypeVar("U_co", covariant=True)

openfeature/hook/__init__.py

@@ -1,7 +1,7 @@
 from __future__ import annotations
 
 import typing
-from dataclasses import dataclass
+from datetime import datetime
 from enum import Enum
 from typing import TYPE_CHECKING
 
@@ -20,24 +20,53 @@ class HookType(Enum):
     ERROR = "error"
 
 
-@dataclass
 class HookContext:
-    flag_key: str
-    flag_type: FlagType
-    default_value: typing.Any
-    evaluation_context: EvaluationContext
-    client_metadata: typing.Optional[ClientMetadata] = None
-    provider_metadata: typing.Optional[Metadata] = None
+    def __init__(  # noqa: PLR0913
+        self,
+        flag_key: str,
+        flag_type: FlagType,
+        default_value: typing.Any,
+        evaluation_context: EvaluationContext,
+        client_metadata: typing.Optional[ClientMetadata] = None,
+        provider_metadata: typing.Optional[Metadata] = None,
+    ):
+        self.flag_key = flag_key
+        self.flag_type = flag_type
+        self.default_value = default_value
+        self.evaluation_context = evaluation_context
+        self.client_metadata = client_metadata
+        self.provider_metadata = provider_metadata
 
     def __setattr__(self, key: str, value: typing.Any) -> None:
-        if hasattr(self, key) and key in ("flag_key", "flag_type", "default_value"):
+        if hasattr(self, key) and key in (
+            "flag_key",
+            "flag_type",
+            "default_value",
+            "client_metadata",
+            "provider_metadata",
+        ):
             raise AttributeError(f"Attribute {key!r} is immutable")
         super().__setattr__(key, value)
 
 
+# https://openfeature.dev/specification/sections/hooks/#requirement-421
+HookHints = typing.Mapping[
+    str,
+    typing.Union[
+        bool,
+        int,
+        float,
+        str,
+        datetime,
+        typing.List[typing.Any],
+        typing.Dict[str, typing.Any],
+    ],
+]
+
+
 class Hook:
     def before(
-        self, hook_context: HookContext, hints: dict
+        self, hook_context: HookContext, hints: HookHints
     ) -> typing.Optional[EvaluationContext]:
         """
         Runs before flag is resolved.
@@ -54,7 +83,7 @@ def after(
         self,
         hook_context: HookContext,
         details: FlagEvaluationDetails[typing.Any],
-        hints: dict,
+        hints: HookHints,
     ) -> None:
         """
         Runs after a flag is resolved.
@@ -67,7 +96,7 @@ def after(
         pass
 
     def error(
-        self, hook_context: HookContext, exception: Exception, hints: dict
+        self, hook_context: HookContext, exception: Exception, hints: HookHints
     ) -> None:
         """
         Run when evaluation encounters an error. Errors thrown will be swallowed.
@@ -78,7 +107,7 @@ def error(
         """
         pass
 
-    def finally_after(self, hook_context: HookContext, hints: dict) -> None:
+    def finally_after(self, hook_context: HookContext, hints: HookHints) -> None:
         """
         Run after flag evaluation, including any error processing.
         This will always run. Errors will be swallowed.

openfeature/hook/hook_support.py

@@ -4,7 +4,7 @@
 
 from openfeature.evaluation_context import EvaluationContext
 from openfeature.flag_evaluation import FlagEvaluationDetails, FlagType
-from openfeature.hook import Hook, HookContext, HookType
+from openfeature.hook import Hook, HookContext, HookHints, HookType
 
 logger = logging.getLogger("openfeature")
 
@@ -14,7 +14,7 @@ def error_hooks(
     hook_context: HookContext,
     exception: Exception,
     hooks: typing.List[Hook],
-    hints: typing.Optional[typing.Mapping] = None,
+    hints: typing.Optional[HookHints] = None,
 ) -> None:
     kwargs = {"hook_context": hook_context, "exception": exception, "hints": hints}
     _execute_hooks(
@@ -26,7 +26,7 @@ def after_all_hooks(
     flag_type: FlagType,
     hook_context: HookContext,
     hooks: typing.List[Hook],
-    hints: typing.Optional[typing.Mapping] = None,
+    hints: typing.Optional[HookHints] = None,
 ) -> None:
     kwargs = {"hook_context": hook_context, "hints": hints}
     _execute_hooks(
@@ -39,7 +39,7 @@ def after_hooks(
     hook_context: HookContext,
     details: FlagEvaluationDetails[typing.Any],
     hooks: typing.List[Hook],
-    hints: typing.Optional[typing.Mapping] = None,
+    hints: typing.Optional[HookHints] = None,
 ) -> None:
     kwargs = {"hook_context": hook_context, "details": details, "hints": hints}
     _execute_hooks_unchecked(
@@ -51,7 +51,7 @@ def before_hooks(
     flag_type: FlagType,
     hook_context: HookContext,
     hooks: typing.List[Hook],
-    hints: typing.Optional[typing.Mapping] = None,
+    hints: typing.Optional[HookHints] = None,
 ) -> EvaluationContext:
     kwargs = {"hook_context": hook_context, "hints": hints}
     executed_hooks = _execute_hooks_unchecked(

tests/hook/test_hook_support.py

@@ -40,10 +40,14 @@ def test_hook_context_has_immutable_and_mutable_fields():
 
     4.1.3 - The "flag key", "flag type", and "default value" properties MUST be immutable.
     4.1.4.1 - The evaluation context MUST be mutable only within the before hook.
+    4.2.2.2 - The client "metadata" field in the "hook context" MUST be immutable.
+    4.2.2.3 - The provider "metadata" field in the "hook context" MUST be immutable.
     """
 
     # Given
-    hook_context = HookContext("flag_key", FlagType.BOOLEAN, True, EvaluationContext())
+    hook_context = HookContext(
+        "flag_key", FlagType.BOOLEAN, True, EvaluationContext(), ClientMetadata("name")
+    )
 
     # When
     with pytest.raises(AttributeError):
@@ -52,18 +56,20 @@ def test_hook_context_has_immutable_and_mutable_fields():
         hook_context.flag_type = FlagType.STRING
     with pytest.raises(AttributeError):
         hook_context.default_value = "new_value"
+    with pytest.raises(AttributeError):
+        hook_context.client_metadata = ClientMetadata("new_name")
+    with pytest.raises(AttributeError):
+        hook_context.provider_metadata = Metadata("name")
 
     hook_context.evaluation_context = EvaluationContext("targeting_key")
-    hook_context.client_metadata = ClientMetadata("name")
-    hook_context.provider_metadata = Metadata("name")
 
     # Then
     assert hook_context.flag_key == "flag_key"
     assert hook_context.flag_type is FlagType.BOOLEAN
     assert hook_context.default_value is True
     assert hook_context.evaluation_context.targeting_key == "targeting_key"
     assert hook_context.client_metadata.name == "name"
-    assert hook_context.provider_metadata.name == "name"
+    assert hook_context.provider_metadata is None
 
 
 def test_error_hooks_run_error_method(mock_hook):