Make embedding generation go through inference (#606)

This PR does the following:
1) adds the ability to generate embeddings in all supported inference
providers.
2) Moves all the memory providers to use the inference API and improved
the memory tests to setup the inference stack correctly and use the
embedding models

This is a merge from #589 and #598

llama_stack/providers/tests/memory/fixtures.py

@@ -10,6 +10,8 @@ import tempfile
 import pytest
 import pytest_asyncio
 
+from llama_stack.apis.inference import ModelInput, ModelType
+
 from llama_stack.distribution.datatypes import Api, Provider
 from llama_stack.providers.inline.memory.chroma import ChromaInlineImplConfig
 from llama_stack.providers.inline.memory.faiss import FaissImplConfig
@@ -105,14 +107,30 @@ MEMORY_FIXTURES = ["faiss", "pgvector", "weaviate", "remote", "chroma"]
 
 
 @pytest_asyncio.fixture(scope="session")
-async def memory_stack(request):
-    fixture_name = request.param
-    fixture = request.getfixturevalue(f"memory_{fixture_name}")
+async def memory_stack(inference_model, request):
+    fixture_dict = request.param
+
+    providers = {}
+    provider_data = {}
+    for key in ["inference", "memory"]:
+        fixture = request.getfixturevalue(f"{key}_{fixture_dict[key]}")
+        providers[key] = fixture.providers
+        if fixture.provider_data:
+            provider_data.update(fixture.provider_data)
 
     test_stack = await construct_stack_for_test(
-        [Api.memory],
-        {"memory": fixture.providers},
-        fixture.provider_data,
+        [Api.memory, Api.inference],
+        providers,
+        provider_data,
+        models=[
+            ModelInput(
+                model_id=inference_model,
+                model_type=ModelType.embedding_model,
+                metadata={
+                    "embedding_dimension": get_env_or_fail("EMBEDDING_DIMENSION"),
+                },
+            )
+        ],
     )
 
     return test_stack.impls[Api.memory], test_stack.impls[Api.memory_banks]