Enable sane naming of registered objects with defaults (#429)

# What does this PR do? 

This is a follow-up to #425. That PR allows for specifying models in the
registry, but each entry needs to look like:

```yaml
- identifier: ...
  provider_id: ...
  provider_resource_identifier: ...
```

This is headache-inducing.

The current PR makes this situation better by adopting the shape of our
APIs. Namely, we need the user to only specify `model-id`. The rest
should be optional and figured out by the Stack. You can always override
it.

Here's what example `ollama` "full stack" registry looks like (we still
need to kill or simplify shield_type crap):
```yaml
models:
- model_id: Llama3.2-3B-Instruct
- model_id: Llama-Guard-3-1B
shields:
- shield_id: llama_guard
  shield_type: llama_guard
```

## Test Plan

See test plan for #425. Re-ran it.

llama_stack/apis/datasets/datasets.py

@@ -10,15 +10,13 @@ from llama_models.llama3.api.datatypes import URL
 
 from llama_models.schema_utils import json_schema_type, webmethod
 
-from pydantic import Field
+from pydantic import BaseModel, Field
 
 from llama_stack.apis.common.type_system import ParamType
-from llama_stack.apis.resource import Resource
+from llama_stack.apis.resource import Resource, ResourceType
 
 
-@json_schema_type
-class Dataset(Resource):
-    type: Literal["dataset"] = "dataset"
+class CommonDatasetFields(BaseModel):
     schema: Dict[str, ParamType]
     url: URL
     metadata: Dict[str, Any] = Field(
@@ -27,6 +25,26 @@ class Dataset(Resource):
     )
 
 
+@json_schema_type
+class Dataset(CommonDatasetFields, Resource):
+    type: Literal[ResourceType.dataset.value] = ResourceType.dataset.value
+
+    @property
+    def dataset_id(self) -> str:
+        return self.identifier
+
+    @property
+    def provider_dataset_id(self) -> str:
+        return self.provider_resource_id
+
+
+@json_schema_type
+class DatasetInput(CommonDatasetFields, BaseModel):
+    dataset_id: str
+    provider_id: Optional[str] = None
+    provider_dataset_id: Optional[str] = None
+
+
 class Datasets(Protocol):
     @webmethod(route="/datasets/register", method="POST")
     async def register_dataset(

llama_stack/apis/eval_tasks/eval_tasks.py

@@ -7,14 +7,12 @@ from typing import Any, Dict, List, Literal, Optional, Protocol, runtime_checkab
 
 from llama_models.schema_utils import json_schema_type, webmethod
 
-from pydantic import Field
+from pydantic import BaseModel, Field
 
-from llama_stack.apis.resource import Resource
+from llama_stack.apis.resource import Resource, ResourceType
 
 
-@json_schema_type
-class EvalTask(Resource):
-    type: Literal["eval_task"] = "eval_task"
+class CommonEvalTaskFields(BaseModel):
     dataset_id: str
     scoring_functions: List[str]
     metadata: Dict[str, Any] = Field(
@@ -23,6 +21,26 @@ class EvalTask(Resource):
     )
 
 
+@json_schema_type
+class EvalTask(CommonEvalTaskFields, Resource):
+    type: Literal[ResourceType.eval_task.value] = ResourceType.eval_task.value
+
+    @property
+    def eval_task_id(self) -> str:
+        return self.identifier
+
+    @property
+    def provider_eval_task_id(self) -> str:
+        return self.provider_resource_id
+
+
+@json_schema_type
+class EvalTaskInput(CommonEvalTaskFields, BaseModel):
+    eval_task_id: str
+    provider_id: Optional[str] = None
+    provider_eval_task_id: Optional[str] = None
+
+
 @runtime_checkable
 class EvalTasks(Protocol):
     @webmethod(route="/eval_tasks/list", method="GET")

llama_stack/apis/memory_banks/memory_banks.py

@@ -30,37 +30,8 @@ class MemoryBankType(Enum):
     graph = "graph"
 
 
-@json_schema_type
-class VectorMemoryBank(Resource):
-    type: Literal[ResourceType.memory_bank.value] = ResourceType.memory_bank.value
-    memory_bank_type: Literal[MemoryBankType.vector.value] = MemoryBankType.vector.value
-    embedding_model: str
-    chunk_size_in_tokens: int
-    overlap_size_in_tokens: Optional[int] = None
-
-
-@json_schema_type
-class KeyValueMemoryBank(Resource):
-    type: Literal[ResourceType.memory_bank.value] = ResourceType.memory_bank.value
-    memory_bank_type: Literal[MemoryBankType.keyvalue.value] = (
-        MemoryBankType.keyvalue.value
-    )
-
-
-@json_schema_type
-class KeywordMemoryBank(Resource):
-    type: Literal[ResourceType.memory_bank.value] = ResourceType.memory_bank.value
-    memory_bank_type: Literal[MemoryBankType.keyword.value] = (
-        MemoryBankType.keyword.value
-    )
-
-
-@json_schema_type
-class GraphMemoryBank(Resource):
-    type: Literal[ResourceType.memory_bank.value] = ResourceType.memory_bank.value
-    memory_bank_type: Literal[MemoryBankType.graph.value] = MemoryBankType.graph.value
-
-
+# define params for each type of memory bank, this leads to a tagged union
+# accepted as input from the API or from the config.
 @json_schema_type
 class VectorMemoryBankParams(BaseModel):
     memory_bank_type: Literal[MemoryBankType.vector.value] = MemoryBankType.vector.value
@@ -88,6 +59,58 @@ class GraphMemoryBankParams(BaseModel):
     memory_bank_type: Literal[MemoryBankType.graph.value] = MemoryBankType.graph.value
 
 
+BankParams = Annotated[
+    Union[
+        VectorMemoryBankParams,
+        KeyValueMemoryBankParams,
+        KeywordMemoryBankParams,
+        GraphMemoryBankParams,
+    ],
+    Field(discriminator="memory_bank_type"),
+]
+
+
+# Some common functionality for memory banks.
+class MemoryBankResourceMixin(Resource):
+    type: Literal[ResourceType.memory_bank.value] = ResourceType.memory_bank.value
+
+    @property
+    def memory_bank_id(self) -> str:
+        return self.identifier
+
+    @property
+    def provider_memory_bank_id(self) -> str:
+        return self.provider_resource_id
+
+
+@json_schema_type
+class VectorMemoryBank(MemoryBankResourceMixin):
+    memory_bank_type: Literal[MemoryBankType.vector.value] = MemoryBankType.vector.value
+    embedding_model: str
+    chunk_size_in_tokens: int
+    overlap_size_in_tokens: Optional[int] = None
+
+
+@json_schema_type
+class KeyValueMemoryBank(MemoryBankResourceMixin):
+    memory_bank_type: Literal[MemoryBankType.keyvalue.value] = (
+        MemoryBankType.keyvalue.value
+    )
+
+
+# TODO: KeyValue and Keyword are so similar in name, oof. Get a better naming convention.
+@json_schema_type
+class KeywordMemoryBank(MemoryBankResourceMixin):
+    memory_bank_type: Literal[MemoryBankType.keyword.value] = (
+        MemoryBankType.keyword.value
+    )
+
+
+@json_schema_type
+class GraphMemoryBank(MemoryBankResourceMixin):
+    memory_bank_type: Literal[MemoryBankType.graph.value] = MemoryBankType.graph.value
+
+
 MemoryBank = Annotated[
     Union[
         VectorMemoryBank,
@@ -98,15 +121,12 @@ MemoryBank = Annotated[
     Field(discriminator="memory_bank_type"),
 ]
 
-BankParams = Annotated[
-    Union[
-        VectorMemoryBankParams,
-        KeyValueMemoryBankParams,
-        KeywordMemoryBankParams,
-        GraphMemoryBankParams,
-    ],
-    Field(discriminator="memory_bank_type"),
-]
+
+@json_schema_type
+class MemoryBankInput(BaseModel):
+    memory_bank_id: str
+    params: BankParams
+    provider_memory_bank_id: Optional[str] = None
 
 
 @runtime_checkable
@@ -123,5 +143,5 @@ class MemoryBanks(Protocol):
         memory_bank_id: str,
         params: BankParams,
         provider_id: Optional[str] = None,
-        provider_memorybank_id: Optional[str] = None,
+        provider_memory_bank_id: Optional[str] = None,
     ) -> MemoryBank: ...

llama_stack/apis/models/models.py

@@ -7,20 +7,38 @@
 from typing import Any, Dict, List, Literal, Optional, Protocol, runtime_checkable
 
 from llama_models.schema_utils import json_schema_type, webmethod
-from pydantic import Field
+from pydantic import BaseModel, Field
 
 from llama_stack.apis.resource import Resource, ResourceType
 
 
-@json_schema_type
-class Model(Resource):
-    type: Literal[ResourceType.model.value] = ResourceType.model.value
+class CommonModelFields(BaseModel):
     metadata: Dict[str, Any] = Field(
         default_factory=dict,
         description="Any additional metadata for this model",
     )
 
 
+@json_schema_type
+class Model(CommonModelFields, Resource):
+    type: Literal[ResourceType.model.value] = ResourceType.model.value
+
+    @property
+    def model_id(self) -> str:
+        return self.identifier
+
+    @property
+    def provider_model_id(self) -> str:
+        return self.provider_resource_id
+
+
+@json_schema_type
+class ModelInput(CommonModelFields):
+    model_id: str
+    provider_id: Optional[str] = None
+    provider_model_id: Optional[str] = None
+
+
 @runtime_checkable
 class Models(Protocol):
     @webmethod(route="/models/list", method="GET")

llama_stack/apis/resource.py

@@ -17,14 +17,12 @@ class ResourceType(Enum):
     memory_bank = "memory_bank"
     dataset = "dataset"
     scoring_function = "scoring_function"
+    eval_task = "eval_task"
 
 
 class Resource(BaseModel):
     """Base class for all Llama Stack resources"""
 
-    # TODO: I think we need to move these into the child classes
-    # and make them `model_id`, `shield_id`, etc. because otherwise
-    # the config file has these confusing generic names in there
     identifier: str = Field(
         description="Unique identifier for this resource in llama stack"
     )

llama_stack/apis/scoring_functions/scoring_functions.py

@@ -66,11 +66,7 @@ ScoringFnParams = Annotated[
 ]
 
 
-@json_schema_type
-class ScoringFn(Resource):
-    type: Literal[ResourceType.scoring_function.value] = (
-        ResourceType.scoring_function.value
-    )
+class CommonScoringFnFields(BaseModel):
     description: Optional[str] = None
     metadata: Dict[str, Any] = Field(
         default_factory=dict,
@@ -85,6 +81,28 @@ class ScoringFn(Resource):
     )
 
 
+@json_schema_type
+class ScoringFn(CommonScoringFnFields, Resource):
+    type: Literal[ResourceType.scoring_function.value] = (
+        ResourceType.scoring_function.value
+    )
+
+    @property
+    def scoring_fn_id(self) -> str:
+        return self.identifier
+
+    @property
+    def provider_scoring_fn_id(self) -> str:
+        return self.provider_resource_id
+
+
+@json_schema_type
+class ScoringFnInput(CommonScoringFnFields, BaseModel):
+    scoring_fn_id: str
+    provider_id: Optional[str] = None
+    provider_scoring_fn_id: Optional[str] = None
+
+
 @runtime_checkable
 class ScoringFunctions(Protocol):
     @webmethod(route="/scoring_functions/list", method="GET")

llama_stack/apis/shields/shields.py

@@ -8,6 +8,7 @@ from enum import Enum
 from typing import Any, Dict, List, Literal, Optional, Protocol, runtime_checkable
 
 from llama_models.schema_utils import json_schema_type, webmethod
+from pydantic import BaseModel
 
 from llama_stack.apis.resource import Resource, ResourceType
 
@@ -20,13 +21,30 @@ class ShieldType(Enum):
     prompt_guard = "prompt_guard"
 
 
+class CommonShieldFields(BaseModel):
+    shield_type: ShieldType
+    params: Optional[Dict[str, Any]] = None
+
+
 @json_schema_type
-class Shield(Resource):
+class Shield(CommonShieldFields, Resource):
     """A safety shield resource that can be used to check content"""
 
     type: Literal[ResourceType.shield.value] = ResourceType.shield.value
-    shield_type: ShieldType
-    params: Dict[str, Any] = {}
+
+    @property
+    def shield_id(self) -> str:
+        return self.identifier
+
+    @property
+    def provider_shield_id(self) -> str:
+        return self.provider_resource_id
+
+
+class ShieldInput(CommonShieldFields):
+    shield_id: str
+    provider_id: Optional[str] = None
+    provider_shield_id: Optional[str] = None
 
 
 @runtime_checkable