fix: pull ollama embedding model if necessary (#1209)

Embedding models are tiny and can be pulled on-demand. Let's do that so
the user doesn't have to do "yet another thing" to get themselves set
up.

Thanks @hardikjshah for the suggestion.

Also fixed a build dependency miss (TODO: distro_codegen needs to
actually check that the build template contains all providers mentioned
for the run.yaml file)

## Test Plan 

First run `ollama rm all-minilm:latest`. 

Run `llama stack build --template ollama && llama stack run ollama --env
INFERENCE_MODEL=llama3.2:3b-instruct-fp16`. See that it outputs a
"Pulling embedding model `all-minilm:latest`" output and the stack
starts up correctly. Verify that `ollama list` shows the model is
correctly downloaded.

distributions/dependencies.json

@@ -356,6 +356,7 @@
     "scikit-learn",
     "scipy",
     "sentencepiece",
+    "sqlite-vec",
     "tqdm",
     "transformers",
     "uvicorn",

docs/source/distributions/self_hosted_distro/ollama.md

@@ -23,7 +23,7 @@ The `llamastack/distribution-ollama` distribution consists of the following prov
 | scoring | `inline::basic`, `inline::llm-as-judge`, `inline::braintrust` |
 | telemetry | `inline::meta-reference` |
 | tool_runtime | `remote::brave-search`, `remote::tavily-search`, `inline::code-interpreter`, `inline::rag-runtime` |
-| vector_io | `inline::faiss`, `remote::chromadb`, `remote::pgvector` |
+| vector_io | `inline::faiss`, `inline::sqlite_vec`, `remote::chromadb`, `remote::pgvector` |
 
 
 You should use this distribution if you have a regular desktop machine without very powerful GPUs. Of course, if you have powerful GPUs, you can still continue using this distribution since Ollama supports GPU acceleration.

llama_stack/providers/remote/inference/ollama/ollama.py

@@ -281,6 +281,8 @@ class OllamaInferenceAdapter(Inference, ModelsProtocolPrivate):
 
     async def register_model(self, model: Model) -> Model:
         if model.model_type == ModelType.embedding:
+            log.info(f"Pulling embedding model `{model.provider_resource_id}` if necessary...")
+            await self.client.pull(model.provider_resource_id)
             response = await self.client.list()
         else:
             response = await self.client.ps()

llama_stack/templates/ollama/build.yaml

@@ -6,6 +6,7 @@ distribution_spec:
     - remote::ollama
     vector_io:
     - inline::faiss
+    - inline::sqlite_vec
     - remote::chromadb
     - remote::pgvector
     safety:

llama_stack/templates/ollama/ollama.py

@@ -25,7 +25,7 @@ from llama_stack.templates.template import DistributionTemplate, RunConfigSettin
 def get_distribution_template() -> DistributionTemplate:
     providers = {
         "inference": ["remote::ollama"],
-        "vector_io": ["inline::faiss", "remote::chromadb", "remote::pgvector"],
+        "vector_io": ["inline::faiss", "inline::sqlite_vec", "remote::chromadb", "remote::pgvector"],
         "safety": ["inline::llama-guard"],
         "agents": ["inline::meta-reference"],
         "telemetry": ["inline::meta-reference"],