feat: implement get chat completions APIs (#2200)

# What does this PR do?
* Provide sqlite implementation of the APIs introduced in
https://github.com/meta-llama/llama-stack/pull/2145.
* Introduced a SqlStore API: llama_stack/providers/utils/sqlstore/api.py
and the first Sqlite implementation
* Pagination support will be added in a future PR.

## Test Plan
Unit test on sql store:
<img width="1005" alt="image"
src="https://github.com/user-attachments/assets/9b8b7ec8-632b-4667-8127-5583426b2e29"
/>


Integration test:
```
INFERENCE_MODEL="llama3.2:3b-instruct-fp16" llama stack build --template ollama --image-type conda --run
```
```
LLAMA_STACK_CONFIG=http://localhost:5001 INFERENCE_MODEL="llama3.2:3b-instruct-fp16" python -m pytest -v tests/integration/inference/test_openai_completion.py --text-model "llama3.2:3b-instruct-fp16" -k 'inference_store and openai'
```

llama_stack/templates/bedrock/build.yaml

@@ -29,3 +29,5 @@ distribution_spec:
     - inline::rag-runtime
     - remote::model-context-protocol
 image_type: conda
+additional_pip_packages:
+- sqlalchemy[asyncio]

llama_stack/templates/bedrock/run.yaml

@@ -96,6 +96,9 @@ providers:
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/bedrock}/registry.db
+inference_store:
+  type: sqlite
+  db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/bedrock}/inference_store.db
 models:
 - metadata: {}
   model_id: meta.llama3-1-8b-instruct-v1:0

llama_stack/templates/cerebras/build.yaml

@@ -29,3 +29,5 @@ distribution_spec:
     - remote::tavily-search
     - inline::rag-runtime
 image_type: conda
+additional_pip_packages:
+- sqlalchemy[asyncio]

llama_stack/templates/cerebras/run.yaml

@@ -99,6 +99,9 @@ providers:
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/cerebras}/registry.db
+inference_store:
+  type: sqlite
+  db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/cerebras}/inference_store.db
 models:
 - metadata: {}
   model_id: llama3.1-8b

llama_stack/templates/ci-tests/build.yaml

@@ -30,3 +30,5 @@ distribution_spec:
     - inline::rag-runtime
     - remote::model-context-protocol
 image_type: conda
+additional_pip_packages:
+- sqlalchemy[asyncio]

llama_stack/templates/ci-tests/run.yaml

@@ -99,6 +99,9 @@ providers:
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/ci-tests}/registry.db
+inference_store:
+  type: sqlite
+  db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/ci-tests}/inference_store.db
 models:
 - metadata: {}
   model_id: accounts/fireworks/models/llama-v3p1-8b-instruct

llama_stack/templates/dell/build.yaml

@@ -30,3 +30,6 @@ distribution_spec:
     - remote::tavily-search
     - inline::rag-runtime
 image_type: conda
+additional_pip_packages:
+- sqlalchemy[asyncio]
+- sqlalchemy[asyncio]

llama_stack/templates/dell/run-with-safety.yaml

@@ -99,6 +99,9 @@ providers:
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/dell}/registry.db
+inference_store:
+  type: sqlite
+  db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/dell}/inference_store.db
 models:
 - metadata: {}
   model_id: ${env.INFERENCE_MODEL}

llama_stack/templates/dell/run.yaml

@@ -95,6 +95,9 @@ providers:
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/dell}/registry.db
+inference_store:
+  type: sqlite
+  db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/dell}/inference_store.db
 models:
 - metadata: {}
   model_id: ${env.INFERENCE_MODEL}

llama_stack/templates/dependencies.json

@@ -31,6 +31,7 @@
     "scikit-learn",
     "scipy",
     "sentencepiece",
+    "sqlalchemy[asyncio]",
     "tqdm",
     "transformers",
     "tree_sitter",
@@ -67,6 +68,7 @@
     "scikit-learn",
     "scipy",
     "sentencepiece",
+    "sqlalchemy[asyncio]",
     "tqdm",
     "transformers",
     "tree_sitter",
@@ -105,6 +107,7 @@
     "scikit-learn",
     "scipy",
     "sentencepiece",
+    "sqlalchemy[asyncio]",
     "sqlite-vec",
     "tqdm",
     "transformers",
@@ -145,6 +148,7 @@
     "scikit-learn",
     "scipy",
     "sentencepiece",
+    "sqlalchemy[asyncio]",
     "tqdm",
     "transformers",
     "tree_sitter",
@@ -184,6 +188,7 @@
     "scikit-learn",
     "scipy",
     "sentencepiece",
+    "sqlalchemy[asyncio]",
     "tqdm",
     "transformers",
     "tree_sitter",
@@ -221,6 +226,7 @@
     "scikit-learn",
     "scipy",
     "sentencepiece",
+    "sqlalchemy[asyncio]",
     "tqdm",
     "transformers",
     "tree_sitter",
@@ -259,6 +265,7 @@
     "scikit-learn",
     "scipy",
     "sentencepiece",
+    "sqlalchemy[asyncio]",
     "tqdm",
     "transformers",
     "tree_sitter",
@@ -297,6 +304,7 @@
     "scikit-learn",
     "scipy",
     "sentencepiece",
+    "sqlalchemy[asyncio]",
     "tqdm",
     "transformers",
     "tree_sitter",
@@ -335,6 +343,7 @@
     "scikit-learn",
     "scipy",
     "sentencepiece",
+    "sqlalchemy[asyncio]",
     "sqlite-vec",
     "tqdm",
     "transformers",
@@ -379,6 +388,7 @@
     "scipy",
     "sentence-transformers",
     "sentencepiece",
+    "sqlalchemy[asyncio]",
     "torch",
     "torchao==0.8.0",
     "torchvision",
@@ -414,6 +424,7 @@
     "scikit-learn",
     "scipy",
     "sentencepiece",
+    "sqlalchemy[asyncio]",
     "tqdm",
     "transformers",
     "uvicorn"
@@ -452,6 +463,7 @@
     "scikit-learn",
     "scipy",
     "sentencepiece",
+    "sqlalchemy[asyncio]",
     "torch",
     "tqdm",
     "transformers",
@@ -490,6 +502,7 @@
     "scikit-learn",
     "scipy",
     "sentencepiece",
+    "sqlalchemy[asyncio]",
     "sqlite-vec",
     "together",
     "tqdm",
@@ -528,6 +541,7 @@
     "scikit-learn",
     "scipy",
     "sentencepiece",
+    "sqlalchemy[asyncio]",
     "tqdm",
     "transformers",
     "tree_sitter",
@@ -566,6 +580,7 @@
     "scikit-learn",
     "scipy",
     "sentencepiece",
+    "sqlalchemy[asyncio]",
     "tqdm",
     "transformers",
     "tree_sitter",
@@ -599,6 +614,7 @@
     "scikit-learn",
     "scipy",
     "sentencepiece",
+    "sqlalchemy[asyncio]",
     "tqdm",
     "transformers",
     "uvicorn",
@@ -637,6 +653,7 @@
     "scikit-learn",
     "scipy",
     "sentencepiece",
+    "sqlalchemy[asyncio]",
     "sqlite-vec",
     "tqdm",
     "transformers",
@@ -678,6 +695,7 @@
     "scikit-learn",
     "scipy",
     "sentencepiece",
+    "sqlalchemy[asyncio]",
     "tqdm",
     "transformers",
     "tree_sitter",
@@ -716,6 +734,7 @@
     "scikit-learn",
     "scipy",
     "sentencepiece",
+    "sqlalchemy[asyncio]",
     "together",
     "tqdm",
     "transformers",
@@ -755,6 +774,7 @@
     "scikit-learn",
     "scipy",
     "sentencepiece",
+    "sqlalchemy[asyncio]",
     "sqlite-vec",
     "tqdm",
     "transformers",
@@ -794,6 +814,7 @@
     "scikit-learn",
     "scipy",
     "sentencepiece",
+    "sqlalchemy[asyncio]",
     "tqdm",
     "transformers",
     "tree_sitter",
@@ -833,6 +854,7 @@
     "scikit-learn",
     "scipy",
     "sentencepiece",
+    "sqlalchemy[asyncio]",
     "tqdm",
     "transformers",
     "tree_sitter",

llama_stack/templates/fireworks/build.yaml

@@ -31,3 +31,6 @@ distribution_spec:
     - inline::rag-runtime
     - remote::model-context-protocol
 image_type: conda
+additional_pip_packages:
+- sqlalchemy[asyncio]
+- sqlalchemy[asyncio]

llama_stack/templates/fireworks/run-with-safety.yaml

@@ -111,6 +111,9 @@ providers:
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/fireworks}/registry.db
+inference_store:
+  type: sqlite
+  db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/fireworks}/inference_store.db
 models:
 - metadata: {}
   model_id: accounts/fireworks/models/llama-v3p1-8b-instruct

llama_stack/templates/fireworks/run.yaml

@@ -106,6 +106,9 @@ providers:
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/fireworks}/registry.db
+inference_store:
+  type: sqlite
+  db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/fireworks}/inference_store.db
 models:
 - metadata: {}
   model_id: accounts/fireworks/models/llama-v3p1-8b-instruct

llama_stack/templates/groq/build.yaml

@@ -26,3 +26,5 @@ distribution_spec:
     - remote::tavily-search
     - inline::rag-runtime
 image_type: conda
+additional_pip_packages:
+- sqlalchemy[asyncio]

llama_stack/templates/groq/run.yaml

@@ -99,6 +99,9 @@ providers:
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/groq}/registry.db
+inference_store:
+  type: sqlite
+  db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/groq}/inference_store.db
 models:
 - metadata: {}
   model_id: groq/llama3-8b-8192

llama_stack/templates/hf-endpoint/build.yaml

@@ -29,3 +29,6 @@ distribution_spec:
     - inline::rag-runtime
     - remote::model-context-protocol
 image_type: conda
+additional_pip_packages:
+- sqlalchemy[asyncio]
+- sqlalchemy[asyncio]

llama_stack/templates/hf-endpoint/run-with-safety.yaml

@@ -107,6 +107,9 @@ providers:
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/hf-endpoint}/registry.db
+inference_store:
+  type: sqlite
+  db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/hf-endpoint}/inference_store.db
 models:
 - metadata: {}
   model_id: ${env.INFERENCE_MODEL}

llama_stack/templates/hf-endpoint/run.yaml

@@ -102,6 +102,9 @@ providers:
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/hf-endpoint}/registry.db
+inference_store:
+  type: sqlite
+  db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/hf-endpoint}/inference_store.db
 models:
 - metadata: {}
   model_id: ${env.INFERENCE_MODEL}

llama_stack/templates/hf-serverless/build.yaml

@@ -30,3 +30,6 @@ distribution_spec:
     - inline::rag-runtime
     - remote::model-context-protocol
 image_type: conda
+additional_pip_packages:
+- sqlalchemy[asyncio]
+- sqlalchemy[asyncio]

llama_stack/templates/hf-serverless/run-with-safety.yaml

@@ -107,6 +107,9 @@ providers:
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/hf-serverless}/registry.db
+inference_store:
+  type: sqlite
+  db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/hf-serverless}/inference_store.db
 models:
 - metadata: {}
   model_id: ${env.INFERENCE_MODEL}

llama_stack/templates/hf-serverless/run.yaml

@@ -102,6 +102,9 @@ providers:
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/hf-serverless}/registry.db
+inference_store:
+  type: sqlite
+  db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/hf-serverless}/inference_store.db
 models:
 - metadata: {}
   model_id: ${env.INFERENCE_MODEL}

llama_stack/templates/llama_api/build.yaml

@@ -30,3 +30,5 @@ distribution_spec:
     - inline::rag-runtime
     - remote::model-context-protocol
 image_type: conda
+additional_pip_packages:
+- sqlalchemy[asyncio]

llama_stack/templates/llama_api/run.yaml

@@ -111,6 +111,9 @@ providers:
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/llama_api}/registry.db
+inference_store:
+  type: sqlite
+  db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/llama_api}/inference_store.db
 models:
 - metadata: {}
   model_id: Llama-3.3-70B-Instruct

llama_stack/templates/meta-reference-gpu/build.yaml

@@ -29,3 +29,6 @@ distribution_spec:
     - inline::rag-runtime
     - remote::model-context-protocol
 image_type: conda
+additional_pip_packages:
+- sqlalchemy[asyncio]
+- sqlalchemy[asyncio]

llama_stack/templates/meta-reference-gpu/run-with-safety.yaml

@@ -117,6 +117,9 @@ providers:
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/meta-reference-gpu}/registry.db
+inference_store:
+  type: sqlite
+  db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/meta-reference-gpu}/inference_store.db
 models:
 - metadata: {}
   model_id: ${env.INFERENCE_MODEL}

llama_stack/templates/meta-reference-gpu/run.yaml

@@ -107,6 +107,9 @@ providers:
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/meta-reference-gpu}/registry.db
+inference_store:
+  type: sqlite
+  db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/meta-reference-gpu}/inference_store.db
 models:
 - metadata: {}
   model_id: ${env.INFERENCE_MODEL}

llama_stack/templates/nvidia/build.yaml

@@ -24,3 +24,6 @@ distribution_spec:
     tool_runtime:
     - inline::rag-runtime
 image_type: conda
+additional_pip_packages:
+- sqlalchemy[asyncio]
+- sqlalchemy[asyncio]

llama_stack/templates/nvidia/run-with-safety.yaml

@@ -92,6 +92,9 @@ providers:
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/nvidia}/registry.db
+inference_store:
+  type: sqlite
+  db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/nvidia}/inference_store.db
 models:
 - metadata: {}
   model_id: ${env.INFERENCE_MODEL}

llama_stack/templates/nvidia/run.yaml

@@ -80,6 +80,9 @@ providers:
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/nvidia}/registry.db
+inference_store:
+  type: sqlite
+  db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/nvidia}/inference_store.db
 models:
 - metadata: {}
   model_id: meta/llama3-8b-instruct

llama_stack/templates/ollama/build.yaml

@@ -32,3 +32,6 @@ distribution_spec:
     - remote::model-context-protocol
     - remote::wolfram-alpha
 image_type: conda
+additional_pip_packages:
+- sqlalchemy[asyncio]
+- sqlalchemy[asyncio]

llama_stack/templates/ollama/run-with-safety.yaml

@@ -112,6 +112,9 @@ providers:
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/ollama}/registry.db
+inference_store:
+  type: sqlite
+  db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/ollama}/inference_store.db
 models:
 - metadata: {}
   model_id: ${env.INFERENCE_MODEL}

llama_stack/templates/ollama/run.yaml

@@ -110,6 +110,9 @@ providers:
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/ollama}/registry.db
+inference_store:
+  type: sqlite
+  db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/ollama}/inference_store.db
 models:
 - metadata: {}
   model_id: ${env.INFERENCE_MODEL}

llama_stack/templates/open-benchmark/build.yaml

@@ -33,3 +33,5 @@ distribution_spec:
     - inline::rag-runtime
     - remote::model-context-protocol
 image_type: conda
+additional_pip_packages:
+- sqlalchemy[asyncio]

llama_stack/templates/open-benchmark/run.yaml

@@ -125,6 +125,9 @@ providers:
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/open-benchmark}/registry.db
+inference_store:
+  type: sqlite
+  db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/open-benchmark}/inference_store.db
 models:
 - metadata: {}
   model_id: openai/gpt-4o

llama_stack/templates/passthrough/build.yaml

@@ -31,3 +31,6 @@ distribution_spec:
     - inline::rag-runtime
     - remote::model-context-protocol
 image_type: conda
+additional_pip_packages:
+- sqlalchemy[asyncio]
+- sqlalchemy[asyncio]

llama_stack/templates/passthrough/run-with-safety.yaml

@@ -111,6 +111,9 @@ providers:
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/passthrough}/registry.db
+inference_store:
+  type: sqlite
+  db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/passthrough}/inference_store.db
 models:
 - metadata: {}
   model_id: meta-llama/Llama-3.1-8B-Instruct

llama_stack/templates/passthrough/run.yaml

@@ -106,6 +106,9 @@ providers:
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/passthrough}/registry.db
+inference_store:
+  type: sqlite
+  db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/passthrough}/inference_store.db
 models:
 - metadata: {}
   model_id: meta-llama/Llama-3.1-8B-Instruct

llama_stack/templates/remote-vllm/build.yaml

@@ -31,3 +31,6 @@ distribution_spec:
     - remote::model-context-protocol
     - remote::wolfram-alpha
 image_type: conda
+additional_pip_packages:
+- sqlalchemy[asyncio]
+- sqlalchemy[asyncio]

llama_stack/templates/remote-vllm/run-with-safety.yaml

@@ -115,6 +115,9 @@ providers:
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/remote-vllm}/registry.db
+inference_store:
+  type: sqlite
+  db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/remote-vllm}/inference_store.db
 models:
 - metadata: {}
   model_id: ${env.INFERENCE_MODEL}

llama_stack/templates/remote-vllm/run.yaml

@@ -108,6 +108,9 @@ providers:
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/remote-vllm}/registry.db
+inference_store:
+  type: sqlite
+  db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/remote-vllm}/inference_store.db
 models:
 - metadata: {}
   model_id: ${env.INFERENCE_MODEL}

llama_stack/templates/sambanova/build.yaml

@@ -22,3 +22,5 @@ distribution_spec:
     - remote::model-context-protocol
     - remote::wolfram-alpha
 image_type: conda
+additional_pip_packages:
+- sqlalchemy[asyncio]

llama_stack/templates/sambanova/run.yaml

@@ -82,6 +82,9 @@ providers:
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/sambanova}/registry.db
+inference_store:
+  type: sqlite
+  db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/sambanova}/inference_store.db
 models:
 - metadata: {}
   model_id: sambanova/Meta-Llama-3.1-8B-Instruct

llama_stack/templates/starter/build.yaml

@@ -35,3 +35,5 @@ distribution_spec:
     - inline::rag-runtime
     - remote::model-context-protocol
 image_type: conda
+additional_pip_packages:
+- sqlalchemy[asyncio]

llama_stack/templates/starter/run.yaml

@@ -133,6 +133,9 @@ providers:
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/starter}/registry.db
+inference_store:
+  type: sqlite
+  db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/starter}/inference_store.db
 models:
 - metadata: {}
   model_id: openai/gpt-4o

llama_stack/templates/template.py

@@ -29,6 +29,7 @@ from llama_stack.distribution.distribution import get_provider_registry
 from llama_stack.distribution.utils.dynamic import instantiate_class_type
 from llama_stack.providers.utils.inference.model_registry import ProviderModelEntry
 from llama_stack.providers.utils.kvstore.config import SqliteKVStoreConfig
+from llama_stack.providers.utils.sqlstore.sqlstore import SqliteSqlStoreConfig
 
 
 def get_model_registry(
@@ -117,6 +118,10 @@ class RunConfigSettings(BaseModel):
                 __distro_dir__=f"~/.llama/distributions/{name}",
                 db_name="registry.db",
             ),
+            inference_store=SqliteSqlStoreConfig.sample_run_config(
+                __distro_dir__=f"~/.llama/distributions/{name}",
+                db_name="inference_store.db",
+            ),
             models=self.default_models or [],
             shields=self.default_shields or [],
             tool_groups=self.default_tool_groups or [],
@@ -146,14 +151,20 @@ class DistributionTemplate(BaseModel):
     available_models_by_provider: dict[str, list[ProviderModelEntry]] | None = None
 
     def build_config(self) -> BuildConfig:
+        additional_pip_packages: list[str] = []
+        for run_config in self.run_configs.values():
+            run_config_ = run_config.run_config(self.name, self.providers, self.container_image)
+            if run_config_.inference_store:
+                additional_pip_packages.extend(run_config_.inference_store.pip_packages)
+
         return BuildConfig(
-            name=self.name,
             distribution_spec=DistributionSpec(
                 description=self.description,
                 container_image=self.container_image,
                 providers=self.providers,
             ),
             image_type="conda",  # default to conda, can be overridden
+            additional_pip_packages=additional_pip_packages,
         )
 
     def generate_markdown_docs(self) -> str:

llama_stack/templates/tgi/build.yaml

@@ -30,3 +30,6 @@ distribution_spec:
     - inline::rag-runtime
     - remote::model-context-protocol
 image_type: conda
+additional_pip_packages:
+- sqlalchemy[asyncio]
+- sqlalchemy[asyncio]

llama_stack/templates/tgi/run-with-safety.yaml

@@ -102,6 +102,9 @@ providers:
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/tgi}/registry.db
+inference_store:
+  type: sqlite
+  db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/tgi}/inference_store.db
 models:
 - metadata: {}
   model_id: ${env.INFERENCE_MODEL}

llama_stack/templates/tgi/run.yaml

@@ -101,6 +101,9 @@ providers:
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/tgi}/registry.db
+inference_store:
+  type: sqlite
+  db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/tgi}/inference_store.db
 models:
 - metadata: {}
   model_id: ${env.INFERENCE_MODEL}

llama_stack/templates/together/build.yaml

@@ -31,3 +31,6 @@ distribution_spec:
     - remote::model-context-protocol
     - remote::wolfram-alpha
 image_type: conda
+additional_pip_packages:
+- sqlalchemy[asyncio]
+- sqlalchemy[asyncio]

llama_stack/templates/together/run-with-safety.yaml

@@ -111,6 +111,9 @@ providers:
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/together}/registry.db
+inference_store:
+  type: sqlite
+  db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/together}/inference_store.db
 models:
 - metadata: {}
   model_id: meta-llama/Meta-Llama-3.1-8B-Instruct-Turbo

llama_stack/templates/together/run.yaml

@@ -106,6 +106,9 @@ providers:
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/together}/registry.db
+inference_store:
+  type: sqlite
+  db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/together}/inference_store.db
 models:
 - metadata: {}
   model_id: meta-llama/Meta-Llama-3.1-8B-Instruct-Turbo

llama_stack/templates/verification/build.yaml

@@ -35,3 +35,5 @@ distribution_spec:
     - inline::rag-runtime
     - remote::model-context-protocol
 image_type: conda
+additional_pip_packages:
+- sqlalchemy[asyncio]

llama_stack/templates/verification/run.yaml

@@ -135,6 +135,9 @@ providers:
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/verification}/registry.db
+inference_store:
+  type: sqlite
+  db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/verification}/inference_store.db
 models:
 - metadata: {}
   model_id: openai/gpt-4o

llama_stack/templates/vllm-gpu/build.yaml

@@ -30,3 +30,5 @@ distribution_spec:
     - inline::rag-runtime
     - remote::model-context-protocol
 image_type: conda
+additional_pip_packages:
+- sqlalchemy[asyncio]

llama_stack/templates/vllm-gpu/run.yaml

@@ -106,6 +106,9 @@ providers:
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/vllm-gpu}/registry.db
+inference_store:
+  type: sqlite
+  db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/vllm-gpu}/inference_store.db
 models:
 - metadata: {}
   model_id: ${env.INFERENCE_MODEL}

llama_stack/templates/watsonx/build.yaml

@@ -28,3 +28,5 @@ distribution_spec:
     - inline::rag-runtime
     - remote::model-context-protocol
 image_type: conda
+additional_pip_packages:
+- sqlalchemy[asyncio]

llama_stack/templates/watsonx/run.yaml

@@ -103,6 +103,9 @@ providers:
 metadata_store:
   type: sqlite
   db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/watsonx}/registry.db
+inference_store:
+  type: sqlite
+  db_path: ${env.SQLITE_STORE_DIR:~/.llama/distributions/watsonx}/inference_store.db
 models:
 - metadata: {}
   model_id: meta-llama/llama-3-3-70b-instruct