phoenix-oss/llama-stack-mirror
0
0
Fork 1
mirror of https://github.com/meta-llama/llama-stack.git synced 2025-10-07 04:45:44 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Merge branch 'main' into prompt-api

Browse source
This commit is contained in:
Francisco Arceo 2025-09-05 15:11:00 -06:00 • committed by GitHub
commit 5d610de5db
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
34 changed files with 679 additions and 240 deletions
Download patch file Download diff file Expand all files Collapse all files

26
llama_stack/core/routing_tables/vector_dbs.py
View file

@ -52,7 +52,6 @@ class VectorDBsRoutingTable(CommonRoutingTableImpl, VectorDBs):
provider_vector_db_id: str | None = None,
vector_db_name: str | None = None,
) -> VectorDB:
provider_vector_db_id = provider_vector_db_id or vector_db_id
if provider_id is None:
if len(self.impls_by_provider_id) > 0:
provider_id = list(self.impls_by_provider_id.keys())[0]
@ -69,14 +68,33 @@ class VectorDBsRoutingTable(CommonRoutingTableImpl, VectorDBs):
raise ModelTypeError(embedding_model, model.model_type, ModelType.embedding)
if "embedding_dimension" not in model.metadata:
raise ValueError(f"Model {embedding_model} does not have an embedding dimension")
provider = self.impls_by_provider_id[provider_id]
logger.warning(
"VectorDB is being deprecated in future releases in favor of VectorStore. Please migrate your usage accordingly."
)
vector_store = await provider.openai_create_vector_store(
name=vector_db_name or vector_db_id,
embedding_model=embedding_model,
embedding_dimension=model.metadata["embedding_dimension"],
provider_id=provider_id,
provider_vector_db_id=provider_vector_db_id,
)
vector_store_id = vector_store.id
actual_provider_vector_db_id = provider_vector_db_id or vector_store_id
logger.warning(
f"Ignoring vector_db_id {vector_db_id} and using vector_store_id {vector_store_id} instead. Setting VectorDB {vector_db_id} to VectorDB.vector_db_name"
)
vector_db_data = {
"identifier": vector_db_id,
"identifier": vector_store_id,
"type": ResourceType.vector_db.value,
"provider_id": provider_id,
"provider_resource_id": provider_vector_db_id,
"provider_resource_id": actual_provider_vector_db_id,
"embedding_model": embedding_model,
"embedding_dimension": model.metadata["embedding_dimension"],
"vector_db_name": vector_db_name,
"vector_db_name": vector_store.name,
}
vector_db = TypeAdapter(VectorDBWithOwner).validate_python(vector_db_data)
await self.register_object(vector_db)

69
llama_stack/providers/inline/batches/reference/batches.py
View file

@ -178,9 +178,9 @@ class ReferenceBatchesImpl(Batches):
# TODO: set expiration time for garbage collection
if endpoint not in ["/v1/chat/completions"]:
if endpoint not in ["/v1/chat/completions", "/v1/completions"]:
raise ValueError(
f"Invalid endpoint: {endpoint}. Supported values: /v1/chat/completions. Code: invalid_value. Param: endpoint",
f"Invalid endpoint: {endpoint}. Supported values: /v1/chat/completions, /v1/completions. Code: invalid_value. Param: endpoint",
)
if completion_window != "24h":
@ -424,13 +424,21 @@ class ReferenceBatchesImpl(Batches):
)
valid = False
for param, expected_type, type_string in [
("model", str, "a string"),
# messages is specific to /v1/chat/completions
# we could skip validating messages here and let inference fail. however,
# that would be a very expensive way to find out messages is wrong.
("messages", list, "an array"), # TODO: allow messages to be a string?
]:
if batch.endpoint == "/v1/chat/completions":
required_params = [
("model", str, "a string"),
# messages is specific to /v1/chat/completions
# we could skip validating messages here and let inference fail. however,
# that would be a very expensive way to find out messages is wrong.
("messages", list, "an array"), # TODO: allow messages to be a string?
]
else: # /v1/completions
required_params = [
("model", str, "a string"),
("prompt", str, "a string"), # TODO: allow prompt to be a list of strings??
]
for param, expected_type, type_string in required_params:
if param not in body:
errors.append(
BatchError(
@ -591,20 +599,37 @@ class ReferenceBatchesImpl(Batches):
try:
# TODO(SECURITY): review body for security issues
request.body["messages"] = [convert_to_openai_message_param(msg) for msg in request.body["messages"]]
chat_response = await self.inference_api.openai_chat_completion(**request.body)
if request.url == "/v1/chat/completions":
request.body["messages"] = [convert_to_openai_message_param(msg) for msg in request.body["messages"]]
chat_response = await self.inference_api.openai_chat_completion(**request.body)
# this is for mypy, we don't allow streaming so we'll get the right type
assert hasattr(chat_response, "model_dump_json"), "Chat response must have model_dump_json method"
return {
"id": request_id,
"custom_id": request.custom_id,
"response": {
"status_code": 200,
"request_id": request_id, # TODO: should this be different?
"body": chat_response.model_dump_json(),
},
}
# this is for mypy, we don't allow streaming so we'll get the right type
assert hasattr(chat_response, "model_dump_json"), "Chat response must have model_dump_json method"
return {
"id": request_id,
"custom_id": request.custom_id,
"response": {
"status_code": 200,
"request_id": request_id, # TODO: should this be different?
"body": chat_response.model_dump_json(),
},
}
else: # /v1/completions
completion_response = await self.inference_api.openai_completion(**request.body)
# this is for mypy, we don't allow streaming so we'll get the right type
assert hasattr(completion_response, "model_dump_json"), (
"Completion response must have model_dump_json method"
)
return {
"id": request_id,
"custom_id": request.custom_id,
"response": {
"status_code": 200,
"request_id": request_id,
"body": completion_response.model_dump_json(),
},
}
except Exception as e:
logger.info(f"Error processing request {request.custom_id} in batch {batch_id}: {e}")
return {