phoenix-oss/llama-stack-mirror
0
0
Fork 1
mirror of https://github.com/meta-llama/llama-stack.git synced 2025-06-28 10:54:19 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

feat: support postgresql inference store (#2310)

Browse source 
# 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'
This commit is contained in:
ehhuang 2025-05-29 14:33:09 -07:00 • committed by Sumit Jaiswal
parent 6af13bbbf0
commit fbc8fc6eb5
No known key found for this signature in database
GPG key ID: A4604B39D64D6AEC
32 changed files with 516 additions and 53 deletions
Download patch file Download diff file Expand all files Collapse all files

21
tests/integration/inference/test_openai_completion.py
View file

@ -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: