MAINT Breaking: Migrating target and scorer registries

build_scripts/evaluate_scorers.py

@@ -61,12 +61,12 @@ async def evaluate_scorers(tags: list[str] | None = None, max_concurrency: int =
     if tags:
         scorer_names: list[str] = []
         for tag in tags:
-            entries = registry.get_by_tag(tag=tag)
+            entries = registry.instances.get_by_tag(tag=tag)
             scorer_names.extend(entry.name for entry in entries if entry.name not in scorer_names)
         scorer_names.sort()
         print(f"\nFiltering by tags: {tags}")
     else:
-        scorer_names = registry.get_names()
+        scorer_names = registry.instances.get_names()
 
     if not scorer_names:
         print("No scorers registered. Check environment variable configuration.")
@@ -85,7 +85,7 @@ async def evaluate_scorers(tags: list[str] | None = None, max_concurrency: int =
 
     # Evaluate each scorer
     for i, scorer_name in scorer_iterator:
-        scorer = registry.get_instance_by_name(scorer_name)
+        scorer = registry.instances.get(scorer_name)
         print(f"\n[{i}/{len(scorer_names)}] Evaluating {scorer_name}...")
         print("  Status: Starting evaluation (this may take several minutes)...")
 

doc/code/registry/2_instance_registry.ipynb

@@ -22,7 +22,7 @@
    "source": [
     "## Listing Available Instances\n",
     "\n",
-    "Use `get_names()` to see registered instances, or `list_metadata()` for details."
+    "Use `instances.get_names()` to see registered instances, or `instances.list_metadata()` for details."
    ]
   },
   {
@@ -69,10 +69,10 @@
     "# Register a scorer instance for demonstration\n",
     "chat_target = OpenAIChatTarget()\n",
     "refusal_scorer = SelfAskRefusalScorer(chat_target=chat_target)\n",
-    "registry.register_instance(refusal_scorer)\n",
+    "registry.instances.register(refusal_scorer)\n",
     "\n",
     "# List what's available\n",
-    "names = registry.get_names()\n",
+    "names = registry.instances.get_names()\n",
     "print(f\"Registered scorers: {names}\")"
    ]
   },
@@ -83,7 +83,7 @@
    "source": [
     "## Getting an Instance\n",
     "\n",
-    "Use `get()` to retrieve a pre-configured instance by name. The instance is ready to use immediately."
+    "Use `instances.get()` to retrieve a pre-configured instance by name. The instance is ready to use immediately."
    ]
   },
   {
@@ -105,7 +105,7 @@
     "# Get the first registered scorer\n",
     "if names:\n",
     "    scorer_name = names[0]\n",
-    "    scorer = registry.get(scorer_name)\n",
+    "    scorer = registry.instances.get(scorer_name)\n",
     "    print(f\"Retrieved scorer: {scorer}\")\n",
     "    print(f\"Scorer type: {type(scorer).__name__}\")"
    ]
@@ -151,7 +151,7 @@
     "from pyrit.output import output_scorer_async\n",
     "\n",
     "# Get metadata for all registered scorers\n",
-    "metadata = registry.list_metadata()\n",
+    "metadata = registry.instances.list_metadata()\n",
     "for item in metadata:\n",
     "    print(f\"\\n{item.unique_name}:\")\n",
     "    print(f\"  Class: {item.class_name}\")\n",
@@ -188,15 +188,15 @@
    ],
    "source": [
     "# Filter by scorer_type (based on isinstance check against TrueFalseScorer/FloatScaleScorer)\n",
-    "true_false_scorers = registry.list_metadata(include_filters={\"scorer_type\": \"true_false\"})\n",
+    "true_false_scorers = registry.instances.list_metadata(include_filters={\"scorer_type\": \"true_false\"})\n",
     "print(f\"True/False scorers: {[m.unique_name for m in true_false_scorers]}\")\n",
     "\n",
     "# Filter by class_name\n",
-    "refusal_scorers = registry.list_metadata(include_filters={\"class_name\": \"SelfAskRefusalScorer\"})\n",
+    "refusal_scorers = registry.instances.list_metadata(include_filters={\"class_name\": \"SelfAskRefusalScorer\"})\n",
     "print(f\"Refusal scorers: {[m.unique_name for m in refusal_scorers]}\")\n",
     "\n",
     "# Combine multiple filters (AND logic)\n",
-    "specific_scorers = registry.list_metadata(\n",
+    "specific_scorers = registry.instances.list_metadata(\n",
     "    include_filters={\"scorer_type\": \"true_false\", \"class_name\": \"SelfAskRefusalScorer\"}\n",
     ")\n",
     "print(f\"True/False refusal scorers: {[m.unique_name for m in specific_scorers]}\")"
@@ -248,7 +248,7 @@
     "# Get the registry singleton\n",
     "registry = TargetRegistry.get_registry_singleton()\n",
     "# List registered targets\n",
-    "target_names = registry.get_names()\n",
+    "target_names = registry.instances.get_names()\n",
     "print(f\"Registered targets after initialization: {target_names}\")"
    ]
   }

doc/code/registry/2_instance_registry.py

@@ -21,7 +21,7 @@
 # %% [markdown]
 # ## Listing Available Instances
 #
-# Use `get_names()` to see registered instances, or `list_metadata()` for details.
+# Use `instances.get_names()` to see registered instances, or `instances.list_metadata()` for details.
 
 # %%
 from pyrit.prompt_target import OpenAIChatTarget
@@ -37,22 +37,22 @@
 # Register a scorer instance for demonstration
 chat_target = OpenAIChatTarget()
 refusal_scorer = SelfAskRefusalScorer(chat_target=chat_target)
-registry.register_instance(refusal_scorer)
+registry.instances.register(refusal_scorer)
 
 # List what's available
-names = registry.get_names()
+names = registry.instances.get_names()
 print(f"Registered scorers: {names}")
 
 # %% [markdown]
 # ## Getting an Instance
 #
-# Use `get()` to retrieve a pre-configured instance by name. The instance is ready to use immediately.
+# Use `instances.get()` to retrieve a pre-configured instance by name. The instance is ready to use immediately.
 
 # %%
 # Get the first registered scorer
 if names:
     scorer_name = names[0]
-    scorer = registry.get(scorer_name)
+    scorer = registry.instances.get(scorer_name)
     print(f"Retrieved scorer: {scorer}")
     print(f"Scorer type: {type(scorer).__name__}")
 
@@ -65,7 +65,7 @@
 from pyrit.output import output_scorer_async
 
 # Get metadata for all registered scorers
-metadata = registry.list_metadata()
+metadata = registry.instances.list_metadata()
 for item in metadata:
     print(f"\n{item.unique_name}:")
     print(f"  Class: {item.class_name}")
@@ -80,15 +80,15 @@
 
 # %%
 # Filter by scorer_type (based on isinstance check against TrueFalseScorer/FloatScaleScorer)
-true_false_scorers = registry.list_metadata(include_filters={"scorer_type": "true_false"})
+true_false_scorers = registry.instances.list_metadata(include_filters={"scorer_type": "true_false"})
 print(f"True/False scorers: {[m.unique_name for m in true_false_scorers]}")
 
 # Filter by class_name
-refusal_scorers = registry.list_metadata(include_filters={"class_name": "SelfAskRefusalScorer"})
+refusal_scorers = registry.instances.list_metadata(include_filters={"class_name": "SelfAskRefusalScorer"})
 print(f"Refusal scorers: {[m.unique_name for m in refusal_scorers]}")
 
 # Combine multiple filters (AND logic)
-specific_scorers = registry.list_metadata(
+specific_scorers = registry.instances.list_metadata(
     include_filters={"scorer_type": "true_false", "class_name": "SelfAskRefusalScorer"}
 )
 print(f"True/False refusal scorers: {[m.unique_name for m in specific_scorers]}")
@@ -111,5 +111,5 @@
 # Get the registry singleton
 registry = TargetRegistry.get_registry_singleton()
 # List registered targets
-target_names = registry.get_names()
+target_names = registry.instances.get_names()
 print(f"Registered targets after initialization: {target_names}")

doc/code/scenarios/0_scenarios.ipynb

@@ -336,8 +336,8 @@
       "      ``supported_parameters``). Each target must already be registered in\n",
       "      ``TargetRegistry`` — typically by ``TargetInitializer`` from\n",
       "      ``ADVERSARIAL_CHAT_*`` env vars, or programmatically via\n",
-      "      ``TargetRegistry.register_instance``. At run time,\n",
-      "      ``_get_atomic_attacks_async`` performs the ``(technique ×\n",
+      "      ``TargetRegistry.get_registry_singleton().instances.register``. At run\n",
+      "      time, ``_get_atomic_attacks_async`` performs the ``(technique ×\n",
       "      adversarial_target × dataset)`` cross-product: for each selected\n",
       "      adversarial-capable ``core`` factory in the ``AttackTechniqueRegistry``\n",
       "      and each requested target, it calls\n",
@@ -356,7 +356,7 @@
       "    Default Datasets (1, max 8 per dataset):\n",
       "      harmbench\n",
       "    Supported Parameters:\n",
-      "      - adversarial_targets (list[str]): Registry names of adversarial chat targets to benchmark. Each name must already be registered in TargetRegistry (via TargetInitializer or TargetRegistry.register_instance). Use 'pyrit_scan list-targets' to see registered targets. Settable via --adversarial-targets <name> [<name> ...] on the CLI, or scenario.args.adversarial_targets in .pyrit_conf.\n",
+      "      - adversarial_targets (list[str]): Registry names of adversarial chat targets to benchmark. Each name must already be registered in TargetRegistry (via TargetInitializer or TargetRegistry instance registration). Use 'pyrit_scan list-targets' to see registered targets. Settable via --adversarial-targets <name> [<name> ...] on the CLI, or scenario.args.adversarial_targets in .pyrit_conf.\n",
       "\u001b[1m\u001b[36m\n",
       "  foundry.red_team_agent\u001b[0m\n",
       "    Class: RedTeamAgent\n",

doc/code/scenarios/1_common_scenario_parameters.ipynb

@@ -85,7 +85,7 @@
     "\n",
     "await initialize_from_config_async(config_path=Path(\"../../scanner/pyrit_conf.yaml\"))  # type: ignore\n",
     "\n",
-    "objective_target = TargetRegistry.get_registry_singleton().get_instance_by_name(\"openai_chat\")"
+    "objective_target = TargetRegistry.get_registry_singleton().instances.get(\"openai_chat\")"
    ]
   },
   {

doc/code/scenarios/1_common_scenario_parameters.py

@@ -35,7 +35,7 @@
 
 await initialize_from_config_async(config_path=Path("../../scanner/pyrit_conf.yaml"))  # type: ignore
 
-objective_target = TargetRegistry.get_registry_singleton().get_instance_by_name("openai_chat")
+objective_target = TargetRegistry.get_registry_singleton().instances.get("openai_chat")
 # %% [markdown]
 # ## Dataset Configuration
 #

doc/code/scenarios/3_adaptive_scenarios.ipynb

@@ -115,7 +115,7 @@
     "\n",
     "await initialize_from_config_async(config_path=Path(\"../../scanner/pyrit_conf.yaml\"))  # type: ignore\n",
     "\n",
-    "objective_target = TargetRegistry.get_registry_singleton().get_instance_by_name(\"openai_chat\")\n",
+    "objective_target = TargetRegistry.get_registry_singleton().instances.get(\"openai_chat\")\n",
     "printer = ConsoleScenarioResultPrinter()"
    ]
   },

doc/code/scenarios/3_adaptive_scenarios.py

@@ -54,7 +54,7 @@
 
 await initialize_from_config_async(config_path=Path("../../scanner/pyrit_conf.yaml"))  # type: ignore
 
-objective_target = TargetRegistry.get_registry_singleton().get_instance_by_name("openai_chat")
+objective_target = TargetRegistry.get_registry_singleton().instances.get("openai_chat")
 printer = ConsoleScenarioResultPrinter()
 
 # %% [markdown]

doc/scanner/foundry.ipynb

@@ -50,7 +50,7 @@
     "\n",
     "await initialize_from_config_async(config_path=Path(\"pyrit_conf.yaml\"))  # type: ignore\n",
     "\n",
-    "objective_target = TargetRegistry.get_registry_singleton().get_instance_by_name(\"openai_chat\")"
+    "objective_target = TargetRegistry.get_registry_singleton().instances.get(\"openai_chat\")"
    ]
   },
   {

doc/scanner/foundry.py

@@ -30,7 +30,7 @@
 
 await initialize_from_config_async(config_path=Path("pyrit_conf.yaml"))  # type: ignore
 
-objective_target = TargetRegistry.get_registry_singleton().get_instance_by_name("openai_chat")
+objective_target = TargetRegistry.get_registry_singleton().instances.get("openai_chat")
 # %% [markdown]
 # ## RedTeamAgent
 #

doc/scanner/garak.ipynb

@@ -51,7 +51,7 @@
     "\n",
     "await initialize_from_config_async(config_path=Path(\"pyrit_conf.yaml\"))  # type: ignore\n",
     "\n",
-    "objective_target = TargetRegistry.get_registry_singleton().get_instance_by_name(\"openai_chat\")"
+    "objective_target = TargetRegistry.get_registry_singleton().instances.get(\"openai_chat\")"
    ]
   },
   {

doc/scanner/garak.py

@@ -31,7 +31,7 @@
 
 await initialize_from_config_async(config_path=Path("pyrit_conf.yaml"))  # type: ignore
 
-objective_target = TargetRegistry.get_registry_singleton().get_instance_by_name("openai_chat")
+objective_target = TargetRegistry.get_registry_singleton().instances.get("openai_chat")
 # %% [markdown]
 # ## Encoding
 #

pyrit/backend/services/scenario_run_service.py

@@ -243,9 +243,9 @@ def _resolve_target(self, *, request: RunScenarioRequest) -> "PromptTarget":
             ValueError: If the target is not found in the registry.
         """
         target_registry = TargetRegistry.get_registry_singleton()
-        objective_target = target_registry.get_instance_by_name(request.target_name)
+        objective_target = target_registry.instances.get(request.target_name)
         if objective_target is None:
-            available_names = target_registry.get_names()
+            available_names = target_registry.instances.get_names()
             if not available_names:
                 raise ValueError(
                     f"Target '{request.target_name}' not found. The target registry is empty. "

pyrit/backend/services/target_service.py

@@ -31,7 +31,7 @@
 from pyrit.prompt_target.azure_ml_chat_target import AzureMLChatTarget
 from pyrit.prompt_target.openai.openai_target import OpenAITarget
 from pyrit.prompt_target.round_robin_target import RoundRobinTarget
-from pyrit.registry.object_registries import TargetRegistry
+from pyrit.registry import TargetRegistry
 
 logger = logging.getLogger(__name__)
 
@@ -192,7 +192,7 @@ async def list_targets_async(
         """
         items = [
             self._build_instance_from_object(target_registry_name=entry.name, target_obj=entry.instance)
-            for entry in self._registry.get_all_instances()
+            for entry in self._registry.instances.get_all_instances()
         ]
         page, has_more = self._paginate(items=items, cursor=cursor, limit=limit)
         next_cursor = page[-1].target_registry_name if has_more and page else None
@@ -227,7 +227,7 @@ async def get_target_async(self, *, target_registry_name: str) -> TargetInstance
         Returns:
             TargetInstance if found, None otherwise.
         """
-        obj = self._registry.get_instance_by_name(target_registry_name)
+        obj = self._registry.instances.get(target_registry_name)
         if obj is None:
             return None
         return self._build_instance_from_object(target_registry_name=target_registry_name, target_obj=obj)
@@ -239,7 +239,7 @@ def get_target_object(self, *, target_registry_name: str) -> Any | None:
         Returns:
             The PromptTarget object if found, None otherwise.
         """
-        return self._registry.get_instance_by_name(target_registry_name)
+        return self._registry.instances.get(target_registry_name)
 
     async def create_target_async(self, *, request: CreateTargetRequest) -> TargetInstance:
         """
@@ -281,7 +281,7 @@ async def create_target_async(self, *, request: CreateTargetRequest) -> TargetIn
 
             target_obj = target_class(**params)
 
-        self._registry.register_instance(target_obj)
+        self._registry.instances.register(target_obj)
 
         target_registry_name = target_obj.get_identifier().unique_name
         return self._build_instance_from_object(target_registry_name=target_registry_name, target_obj=target_obj)
@@ -334,7 +334,7 @@ def _create_round_robin_target(self, *, params: dict[str, Any]) -> RoundRobinTar
         resolved_weights: list[int] = []
         duplicates: list[str] = []
         for idx, name in enumerate(registry_names):
-            target_obj = self._registry.get_instance_by_name(name)
+            target_obj = self._registry.instances.get(name)
             if target_obj is None:
                 raise ValueError(f"Target '{name}' not found in the registry.")
             target_hash = target_obj.get_identifier().hash

pyrit/models/identifiers/scorer_identifier.py

@@ -11,6 +11,7 @@
 
 from pyrit.models.identifiers.component_identifier import ComponentIdentifier
 from pyrit.models.identifiers.evaluation_markers import Evaluate
+from pyrit.models.identifiers.param_markers import Param
 from pyrit.models.identifiers.target_identifier import (  # noqa: TC001
     TargetIdentifier,  # runtime-required by Pydantic field annotations
 )
@@ -24,6 +25,12 @@ class ScorerIdentifier(ComponentIdentifier):
     Promotes the ``scorer_type`` discriminator, the ``score_aggregator`` name, and
     the scorer's own child slots — ``prompt_target`` (an LLM target) and
     ``sub_scorers`` (nested scorers).
+
+    Build markers (``Param.*``) declare how the child slots map to the scorer's
+    constructor: ``prompt_target`` is an included parameter aliased to the
+    ``chat_target`` constructor arg, and ``sub_scorers`` is an included parameter
+    aliased to the composite scorer's ``scorers`` arg. Their identifier types make
+    them references resolved by name from the target and scorer registries.
     """
 
     component_type: ClassVar[ComponentType] = ComponentType.SCORER
@@ -32,7 +39,11 @@ class ScorerIdentifier(ComponentIdentifier):
     scorer_type: Annotated[str | None, Evaluate.Include()] = None
     #: Name of the aggregator function combining sub-scores (e.g., ``"AND_"``).
     score_aggregator: Annotated[str | None, Evaluate.Include()] = None
-    #: Target an LLM-backed scorer calls (e.g., ``SelfAskScaleScorer``).
-    prompt_target: Annotated[TargetIdentifier | None, Evaluate.Include()] = None
-    #: Nested scorers a composite wraps, typed recursively.
-    sub_scorers: Annotated[list[ScorerIdentifier], Evaluate.Include()] = Field(default_factory=list)
+    #: Target an LLM-backed scorer calls (e.g., ``SelfAskScaleScorer``). The
+    #: constructor arg is ``chat_target``, so the build marker aliases it.
+    prompt_target: Annotated[TargetIdentifier | None, Evaluate.Include(), Param.Include(alias="chat_target")] = None
+    #: Nested scorers a composite wraps, typed recursively. The composite
+    #: constructor arg is ``scorers`` (a list), so the build marker aliases it.
+    sub_scorers: Annotated[list[ScorerIdentifier], Evaluate.Include(), Param.Include(alias="scorers")] = Field(
+        default_factory=list
+    )

pyrit/models/identifiers/target_identifier.py

@@ -11,6 +11,7 @@
 
 from pyrit.models.identifiers.component_identifier import ComponentIdentifier
 from pyrit.models.identifiers.evaluation_markers import Evaluate
+from pyrit.models.identifiers.param_markers import Param
 from pyrit.models.parameter import ComponentType
 
 
@@ -46,5 +47,7 @@ class TargetIdentifier(ComponentIdentifier):
     top_p: Annotated[float | None, Evaluate.Include()] = None
     #: Maximum requests per minute.
     max_requests_per_minute: Annotated[int | None, Evaluate.Exclude()] = None
-    #: Inner targets of a multi-target (e.g., ``RoundRobinTarget``), typed recursively.
-    targets: Annotated[list[TargetIdentifier], Evaluate.Unwrap()] = Field(default_factory=list)
+    #: Inner targets of a multi-target (e.g., ``RoundRobinTarget``), typed
+    #: recursively. An included constructor parameter (the ctor arg is also
+    #: ``targets``, a list) resolved by name from the target registry.
+    targets: Annotated[list[TargetIdentifier], Evaluate.Unwrap(), Param.Include()] = Field(default_factory=list)