feat: support postgresql inference store (#2310)

# What does this PR do?
* Added support postgresql inference store
* Added 'oracle' template that demos how to config postgresql stores
(except for telemetry, which is not supported currently)


## Test Plan

llama stack build --template oracle --image-type conda --run
LLAMA_STACK_CONFIG=http://localhost:8321 pytest -s -v tests/integration/
--text-model accounts/fireworks/models/llama-v3p3-70b-instruct -k
'inference_store'

tests/integration/inference/test_openai_completion.py

@@ -268,9 +268,9 @@ def test_openai_chat_completion_streaming_with_n(compat_client, client_with_mode
         False,
     ],
 )
-def test_inference_store(openai_client, client_with_models, text_model_id, stream):
+def test_inference_store(compat_client, client_with_models, text_model_id, stream):
     skip_if_model_doesnt_support_openai_chat_completion(client_with_models, text_model_id)
-    client = openai_client
+    client = compat_client
     # make a chat completion
     message = "Hello, world!"
     response = client.chat.completions.create(
@@ -301,9 +301,14 @@ def test_inference_store(openai_client, client_with_models, text_model_id, strea
 
     retrieved_response = client.chat.completions.retrieve(response_id)
     assert retrieved_response.id == response_id
-    assert retrieved_response.input_messages[0]["content"] == message, retrieved_response
     assert retrieved_response.choices[0].message.content == content, retrieved_response
 
+    input_content = (
+        getattr(retrieved_response.input_messages[0], "content", None)
+        or retrieved_response.input_messages[0]["content"]
+    )
+    assert input_content == message, retrieved_response
+
 
 @pytest.mark.parametrize(
     "stream",
@@ -312,9 +317,9 @@ def test_inference_store(openai_client, client_with_models, text_model_id, strea
         False,
     ],
 )
-def test_inference_store_tool_calls(openai_client, client_with_models, text_model_id, stream):
+def test_inference_store_tool_calls(compat_client, client_with_models, text_model_id, stream):
     skip_if_model_doesnt_support_openai_chat_completion(client_with_models, text_model_id)
-    client = openai_client
+    client = compat_client
     # make a chat completion
     message = "What's the weather in Tokyo? Use the get_weather function to get the weather."
     response = client.chat.completions.create(
@@ -361,7 +366,11 @@ def test_inference_store_tool_calls(openai_client, client_with_models, text_mode
 
     retrieved_response = client.chat.completions.retrieve(response_id)
     assert retrieved_response.id == response_id
-    assert retrieved_response.input_messages[0]["content"] == message
+    input_content = (
+        getattr(retrieved_response.input_messages[0], "content", None)
+        or retrieved_response.input_messages[0]["content"]
+    )
+    assert input_content == message, retrieved_response
     tool_calls = retrieved_response.choices[0].message.tool_calls
     # sometimes model doesn't ouptut tool calls, but we still want to test that the tool was called
     if tool_calls:

tests/unit/utils/test_sqlstore.py

@@ -9,7 +9,7 @@ from tempfile import TemporaryDirectory
 import pytest
 
 from llama_stack.providers.utils.sqlstore.api import ColumnType
-from llama_stack.providers.utils.sqlstore.sqlite.sqlite import SqliteSqlStoreImpl
+from llama_stack.providers.utils.sqlstore.sqlalchemy_sqlstore import SqlAlchemySqlStoreImpl
 from llama_stack.providers.utils.sqlstore.sqlstore import SqliteSqlStoreConfig
 
 
@@ -17,7 +17,7 @@ from llama_stack.providers.utils.sqlstore.sqlstore import SqliteSqlStoreConfig
 async def test_sqlite_sqlstore():
     with TemporaryDirectory() as tmp_dir:
         db_name = "test.db"
-        sqlstore = SqliteSqlStoreImpl(
+        sqlstore = SqlAlchemySqlStoreImpl(
             SqliteSqlStoreConfig(
                 db_path=tmp_dir + "/" + db_name,
             )