phoenix-oss/llama-stack-mirror
0
0
Fork 1
mirror of https://github.com/meta-llama/llama-stack.git synced 2025-12-03 18:00:36 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

Merge branch 'main' into add-mcp-authentication-param

Browse source
This commit is contained in:
Omar Abdelwahab 2025-11-04 16:20:38 -08:00 • committed by GitHub
commit 8632c705aa
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
1250 changed files with 2278 additions and 343484 deletions
Download patch file Download diff file Expand all files Collapse all files

756
src/llama_stack/apis/agents/agents.py
View file

@ -5,30 +5,13 @@
# the root directory of this source tree.
from collections.abc import AsyncIterator
from datetime import datetime
from enum import StrEnum
from typing import Annotated, Any, Literal, Protocol, runtime_checkable
from typing import Annotated, Protocol, runtime_checkable
from pydantic import BaseModel, ConfigDict, Field
from pydantic import BaseModel
from llama_stack.apis.common.content_types import URL, ContentDelta, InterleavedContent
from llama_stack.apis.common.responses import Order, PaginatedResponse
from llama_stack.apis.inference import (
CompletionMessage,
ResponseFormat,
SamplingParams,
ToolCall,
ToolChoice,
ToolConfig,
ToolPromptFormat,
ToolResponse,
ToolResponseMessage,
UserMessage,
)
from llama_stack.apis.safety import SafetyViolation
from llama_stack.apis.tools import ToolDef
from llama_stack.apis.version import LLAMA_STACK_API_V1, LLAMA_STACK_API_V1ALPHA
from llama_stack.schema_utils import ExtraBodyField, json_schema_type, register_schema, webmethod
from llama_stack.apis.common.responses import Order
from llama_stack.apis.version import LLAMA_STACK_API_V1
from llama_stack.schema_utils import ExtraBodyField, json_schema_type, webmethod
from .openai_responses import (
ListOpenAIResponseInputItem,
@ -57,729 +40,12 @@ class ResponseGuardrailSpec(BaseModel):
ResponseGuardrail = str | ResponseGuardrailSpec
class Attachment(BaseModel):
"""An attachment to an agent turn.
:param content: The content of the attachment.
:param mime_type: The MIME type of the attachment.
"""
content: InterleavedContent | URL
mime_type: str
class Document(BaseModel):
"""A document to be used by an agent.
:param content: The content of the document.
:param mime_type: The MIME type of the document.
"""
content: InterleavedContent | URL
mime_type: str
class StepCommon(BaseModel):
"""A common step in an agent turn.
:param turn_id: The ID of the turn.
:param step_id: The ID of the step.
:param started_at: The time the step started.
:param completed_at: The time the step completed.
"""
turn_id: str
step_id: str
started_at: datetime | None = None
completed_at: datetime | None = None
class StepType(StrEnum):
"""Type of the step in an agent turn.
:cvar inference: The step is an inference step that calls an LLM.
:cvar tool_execution: The step is a tool execution step that executes a tool call.
:cvar shield_call: The step is a shield call step that checks for safety violations.
:cvar memory_retrieval: The step is a memory retrieval step that retrieves context for vector dbs.
"""
inference = "inference"
tool_execution = "tool_execution"
shield_call = "shield_call"
memory_retrieval = "memory_retrieval"
@json_schema_type
class InferenceStep(StepCommon):
"""An inference step in an agent turn.
:param model_response: The response from the LLM.
"""
model_config = ConfigDict(protected_namespaces=())
step_type: Literal[StepType.inference] = StepType.inference
model_response: CompletionMessage
@json_schema_type
class ToolExecutionStep(StepCommon):
"""A tool execution step in an agent turn.
:param tool_calls: The tool calls to execute.
:param tool_responses: The tool responses from the tool calls.
"""
step_type: Literal[StepType.tool_execution] = StepType.tool_execution
tool_calls: list[ToolCall]
tool_responses: list[ToolResponse]
@json_schema_type
class ShieldCallStep(StepCommon):
"""A shield call step in an agent turn.
:param violation: The violation from the shield call.
"""
step_type: Literal[StepType.shield_call] = StepType.shield_call
violation: SafetyViolation | None
@json_schema_type
class MemoryRetrievalStep(StepCommon):
"""A memory retrieval step in an agent turn.
:param vector_store_ids: The IDs of the vector databases to retrieve context from.
:param inserted_context: The context retrieved from the vector databases.
"""
step_type: Literal[StepType.memory_retrieval] = StepType.memory_retrieval
# TODO: should this be List[str]?
vector_store_ids: str
inserted_context: InterleavedContent
Step = Annotated[
InferenceStep | ToolExecutionStep | ShieldCallStep | MemoryRetrievalStep,
Field(discriminator="step_type"),
]
@json_schema_type
class Turn(BaseModel):
"""A single turn in an interaction with an Agentic System.
:param turn_id: Unique identifier for the turn within a session
:param session_id: Unique identifier for the conversation session
:param input_messages: List of messages that initiated this turn
:param steps: Ordered list of processing steps executed during this turn
:param output_message: The model's generated response containing content and metadata
:param output_attachments: (Optional) Files or media attached to the agent's response
:param started_at: Timestamp when the turn began
:param completed_at: (Optional) Timestamp when the turn finished, if completed
"""
turn_id: str
session_id: str
input_messages: list[UserMessage | ToolResponseMessage]
steps: list[Step]
output_message: CompletionMessage
output_attachments: list[Attachment] | None = Field(default_factory=lambda: [])
started_at: datetime
completed_at: datetime | None = None
@json_schema_type
class Session(BaseModel):
"""A single session of an interaction with an Agentic System.
:param session_id: Unique identifier for the conversation session
:param session_name: Human-readable name for the session
:param turns: List of all turns that have occurred in this session
:param started_at: Timestamp when the session was created
"""
session_id: str
session_name: str
turns: list[Turn]
started_at: datetime
class AgentToolGroupWithArgs(BaseModel):
name: str
args: dict[str, Any]
AgentToolGroup = str | AgentToolGroupWithArgs
register_schema(AgentToolGroup, name="AgentTool")
class AgentConfigCommon(BaseModel):
sampling_params: SamplingParams | None = Field(default_factory=SamplingParams)
input_shields: list[str] | None = Field(default_factory=lambda: [])
output_shields: list[str] | None = Field(default_factory=lambda: [])
toolgroups: list[AgentToolGroup] | None = Field(default_factory=lambda: [])
client_tools: list[ToolDef] | None = Field(default_factory=lambda: [])
tool_choice: ToolChoice | None = Field(default=None, deprecated="use tool_config instead")
tool_prompt_format: ToolPromptFormat | None = Field(default=None, deprecated="use tool_config instead")
tool_config: ToolConfig | None = Field(default=None)
max_infer_iters: int | None = 10
def model_post_init(self, __context):
if self.tool_config:
if self.tool_choice and self.tool_config.tool_choice != self.tool_choice:
raise ValueError("tool_choice is deprecated. Use tool_choice in tool_config instead.")
if self.tool_prompt_format and self.tool_config.tool_prompt_format != self.tool_prompt_format:
raise ValueError("tool_prompt_format is deprecated. Use tool_prompt_format in tool_config instead.")
else:
params = {}
if self.tool_choice:
params["tool_choice"] = self.tool_choice
if self.tool_prompt_format:
params["tool_prompt_format"] = self.tool_prompt_format
self.tool_config = ToolConfig(**params)
@json_schema_type
class AgentConfig(AgentConfigCommon):
"""Configuration for an agent.
:param model: The model identifier to use for the agent
:param instructions: The system instructions for the agent
:param name: Optional name for the agent, used in telemetry and identification
:param enable_session_persistence: Optional flag indicating whether session data has to be persisted
:param response_format: Optional response format configuration
"""
model: str
instructions: str
name: str | None = None
enable_session_persistence: bool | None = False
response_format: ResponseFormat | None = None
@json_schema_type
class Agent(BaseModel):
"""An agent instance with configuration and metadata.
:param agent_id: Unique identifier for the agent
:param agent_config: Configuration settings for the agent
:param created_at: Timestamp when the agent was created
"""
agent_id: str
agent_config: AgentConfig
created_at: datetime
class AgentConfigOverridablePerTurn(AgentConfigCommon):
instructions: str | None = None
class AgentTurnResponseEventType(StrEnum):
step_start = "step_start"
step_complete = "step_complete"
step_progress = "step_progress"
turn_start = "turn_start"
turn_complete = "turn_complete"
turn_awaiting_input = "turn_awaiting_input"
@json_schema_type
class AgentTurnResponseStepStartPayload(BaseModel):
"""Payload for step start events in agent turn responses.
:param event_type: Type of event being reported
:param step_type: Type of step being executed
:param step_id: Unique identifier for the step within a turn
:param metadata: (Optional) Additional metadata for the step
"""
event_type: Literal[AgentTurnResponseEventType.step_start] = AgentTurnResponseEventType.step_start
step_type: StepType
step_id: str
metadata: dict[str, Any] | None = Field(default_factory=lambda: {})
@json_schema_type
class AgentTurnResponseStepCompletePayload(BaseModel):
"""Payload for step completion events in agent turn responses.
:param event_type: Type of event being reported
:param step_type: Type of step being executed
:param step_id: Unique identifier for the step within a turn
:param step_details: Complete details of the executed step
"""
event_type: Literal[AgentTurnResponseEventType.step_complete] = AgentTurnResponseEventType.step_complete
step_type: StepType
step_id: str
step_details: Step
@json_schema_type
class AgentTurnResponseStepProgressPayload(BaseModel):
"""Payload for step progress events in agent turn responses.
:param event_type: Type of event being reported
:param step_type: Type of step being executed
:param step_id: Unique identifier for the step within a turn
:param delta: Incremental content changes during step execution
"""
model_config = ConfigDict(protected_namespaces=())
event_type: Literal[AgentTurnResponseEventType.step_progress] = AgentTurnResponseEventType.step_progress
step_type: StepType
step_id: str
delta: ContentDelta
@json_schema_type
class AgentTurnResponseTurnStartPayload(BaseModel):
"""Payload for turn start events in agent turn responses.
:param event_type: Type of event being reported
:param turn_id: Unique identifier for the turn within a session
"""
event_type: Literal[AgentTurnResponseEventType.turn_start] = AgentTurnResponseEventType.turn_start
turn_id: str
@json_schema_type
class AgentTurnResponseTurnCompletePayload(BaseModel):
"""Payload for turn completion events in agent turn responses.
:param event_type: Type of event being reported
:param turn: Complete turn data including all steps and results
"""
event_type: Literal[AgentTurnResponseEventType.turn_complete] = AgentTurnResponseEventType.turn_complete
turn: Turn
@json_schema_type
class AgentTurnResponseTurnAwaitingInputPayload(BaseModel):
"""Payload for turn awaiting input events in agent turn responses.
:param event_type: Type of event being reported
:param turn: Turn data when waiting for external tool responses
"""
event_type: Literal[AgentTurnResponseEventType.turn_awaiting_input] = AgentTurnResponseEventType.turn_awaiting_input
turn: Turn
AgentTurnResponseEventPayload = Annotated[
AgentTurnResponseStepStartPayload
| AgentTurnResponseStepProgressPayload
| AgentTurnResponseStepCompletePayload
| AgentTurnResponseTurnStartPayload
| AgentTurnResponseTurnCompletePayload
| AgentTurnResponseTurnAwaitingInputPayload,
Field(discriminator="event_type"),
]
register_schema(AgentTurnResponseEventPayload, name="AgentTurnResponseEventPayload")
@json_schema_type
class AgentTurnResponseEvent(BaseModel):
"""An event in an agent turn response stream.
:param payload: Event-specific payload containing event data
"""
payload: AgentTurnResponseEventPayload
@json_schema_type
class AgentCreateResponse(BaseModel):
"""Response returned when creating a new agent.
:param agent_id: Unique identifier for the created agent
"""
agent_id: str
@json_schema_type
class AgentSessionCreateResponse(BaseModel):
"""Response returned when creating a new agent session.
:param session_id: Unique identifier for the created session
"""
session_id: str
@json_schema_type
class AgentTurnCreateRequest(AgentConfigOverridablePerTurn):
"""Request to create a new turn for an agent.
:param agent_id: Unique identifier for the agent
:param session_id: Unique identifier for the conversation session
:param messages: List of messages to start the turn with
:param documents: (Optional) List of documents to provide to the agent
:param toolgroups: (Optional) List of tool groups to make available for this turn
:param stream: (Optional) Whether to stream the response
:param tool_config: (Optional) Tool configuration to override agent defaults
"""
agent_id: str
session_id: str
# TODO: figure out how we can simplify this and make why
# ToolResponseMessage needs to be here (it is function call
# execution from outside the system)
messages: list[UserMessage | ToolResponseMessage]
documents: list[Document] | None = None
toolgroups: list[AgentToolGroup] | None = Field(default_factory=lambda: [])
stream: bool | None = False
tool_config: ToolConfig | None = None
@json_schema_type
class AgentTurnResumeRequest(BaseModel):
"""Request to resume an agent turn with tool responses.
:param agent_id: Unique identifier for the agent
:param session_id: Unique identifier for the conversation session
:param turn_id: Unique identifier for the turn within a session
:param tool_responses: List of tool responses to submit to continue the turn
:param stream: (Optional) Whether to stream the response
"""
agent_id: str
session_id: str
turn_id: str
tool_responses: list[ToolResponse]
stream: bool | None = False
@json_schema_type
class AgentTurnResponseStreamChunk(BaseModel):
"""Streamed agent turn completion response.
:param event: Individual event in the agent turn response stream
"""
event: AgentTurnResponseEvent
@json_schema_type
class AgentStepResponse(BaseModel):
"""Response containing details of a specific agent step.
:param step: The complete step data and execution details
"""
step: Step
@runtime_checkable
class Agents(Protocol):
"""Agents
APIs for creating and interacting with agentic systems."""
@webmethod(
route="/agents",
method="POST",
descriptive_name="create_agent",
deprecated=True,
level=LLAMA_STACK_API_V1,
)
@webmethod(
route="/agents",
method="POST",
descriptive_name="create_agent",
level=LLAMA_STACK_API_V1ALPHA,
)
async def create_agent(
self,
agent_config: AgentConfig,
) -> AgentCreateResponse:
"""Create an agent with the given configuration.
:param agent_config: The configuration for the agent.
:returns: An AgentCreateResponse with the agent ID.
"""
...
@webmethod(
route="/agents/{agent_id}/session/{session_id}/turn",
method="POST",
descriptive_name="create_agent_turn",
deprecated=True,
level=LLAMA_STACK_API_V1,
)
@webmethod(
route="/agents/{agent_id}/session/{session_id}/turn",
method="POST",
descriptive_name="create_agent_turn",
level=LLAMA_STACK_API_V1ALPHA,
)
async def create_agent_turn(
self,
agent_id: str,
session_id: str,
messages: list[UserMessage | ToolResponseMessage],
stream: bool | None = False,
documents: list[Document] | None = None,
toolgroups: list[AgentToolGroup] | None = None,
tool_config: ToolConfig | None = None,
) -> Turn | AsyncIterator[AgentTurnResponseStreamChunk]:
"""Create a new turn for an agent.
:param agent_id: The ID of the agent to create the turn for.
:param session_id: The ID of the session to create the turn for.
:param messages: List of messages to start the turn with.
:param stream: (Optional) If True, generate an SSE event stream of the response. Defaults to False.
:param documents: (Optional) List of documents to create the turn with.
:param toolgroups: (Optional) List of toolgroups to create the turn with, will be used in addition to the agent's config toolgroups for the request.
:param tool_config: (Optional) The tool configuration to create the turn with, will be used to override the agent's tool_config.
:returns: If stream=False, returns a Turn object.
If stream=True, returns an SSE event stream of AgentTurnResponseStreamChunk.
"""
...
@webmethod(
route="/agents/{agent_id}/session/{session_id}/turn/{turn_id}/resume",
method="POST",
descriptive_name="resume_agent_turn",
deprecated=True,
level=LLAMA_STACK_API_V1,
)
@webmethod(
route="/agents/{agent_id}/session/{session_id}/turn/{turn_id}/resume",
method="POST",
descriptive_name="resume_agent_turn",
level=LLAMA_STACK_API_V1ALPHA,
)
async def resume_agent_turn(
self,
agent_id: str,
session_id: str,
turn_id: str,
tool_responses: list[ToolResponse],
stream: bool | None = False,
) -> Turn | AsyncIterator[AgentTurnResponseStreamChunk]:
"""Resume an agent turn with executed tool call responses.
When a Turn has the status `awaiting_input` due to pending input from client side tool calls, this endpoint can be used to submit the outputs from the tool calls once they are ready.
:param agent_id: The ID of the agent to resume.
:param session_id: The ID of the session to resume.
:param turn_id: The ID of the turn to resume.
:param tool_responses: The tool call responses to resume the turn with.
:param stream: Whether to stream the response.
:returns: A Turn object if stream is False, otherwise an AsyncIterator of AgentTurnResponseStreamChunk objects.
"""
...
@webmethod(
route="/agents/{agent_id}/session/{session_id}/turn/{turn_id}",
method="GET",
deprecated=True,
level=LLAMA_STACK_API_V1,
)
@webmethod(
route="/agents/{agent_id}/session/{session_id}/turn/{turn_id}",
method="GET",
level=LLAMA_STACK_API_V1ALPHA,
)
async def get_agents_turn(
self,
agent_id: str,
session_id: str,
turn_id: str,
) -> Turn:
"""Retrieve an agent turn by its ID.
:param agent_id: The ID of the agent to get the turn for.
:param session_id: The ID of the session to get the turn for.
:param turn_id: The ID of the turn to get.
:returns: A Turn.
"""
...
@webmethod(
route="/agents/{agent_id}/session/{session_id}/turn/{turn_id}/step/{step_id}",
method="GET",
deprecated=True,
level=LLAMA_STACK_API_V1,
)
@webmethod(
route="/agents/{agent_id}/session/{session_id}/turn/{turn_id}/step/{step_id}",
method="GET",
level=LLAMA_STACK_API_V1ALPHA,
)
async def get_agents_step(
self,
agent_id: str,
session_id: str,
turn_id: str,
step_id: str,
) -> AgentStepResponse:
"""Retrieve an agent step by its ID.
:param agent_id: The ID of the agent to get the step for.
:param session_id: The ID of the session to get the step for.
:param turn_id: The ID of the turn to get the step for.
:param step_id: The ID of the step to get.
:returns: An AgentStepResponse.
"""
...
@webmethod(
route="/agents/{agent_id}/session",
method="POST",
descriptive_name="create_agent_session",
deprecated=True,
level=LLAMA_STACK_API_V1,
)
@webmethod(
route="/agents/{agent_id}/session",
method="POST",
descriptive_name="create_agent_session",
level=LLAMA_STACK_API_V1ALPHA,
)
async def create_agent_session(
self,
agent_id: str,
session_name: str,
) -> AgentSessionCreateResponse:
"""Create a new session for an agent.
:param agent_id: The ID of the agent to create the session for.
:param session_name: The name of the session to create.
:returns: An AgentSessionCreateResponse.
"""
...
@webmethod(
route="/agents/{agent_id}/session/{session_id}",
method="GET",
deprecated=True,
level=LLAMA_STACK_API_V1,
)
@webmethod(
route="/agents/{agent_id}/session/{session_id}",
method="GET",
level=LLAMA_STACK_API_V1ALPHA,
)
async def get_agents_session(
self,
session_id: str,
agent_id: str,
turn_ids: list[str] | None = None,
) -> Session:
"""Retrieve an agent session by its ID.
:param session_id: The ID of the session to get.
:param agent_id: The ID of the agent to get the session for.
:param turn_ids: (Optional) List of turn IDs to filter the session by.
:returns: A Session.
"""
...
@webmethod(
route="/agents/{agent_id}/session/{session_id}",
method="DELETE",
deprecated=True,
level=LLAMA_STACK_API_V1,
)
@webmethod(
route="/agents/{agent_id}/session/{session_id}",
method="DELETE",
level=LLAMA_STACK_API_V1ALPHA,
)
async def delete_agents_session(
self,
session_id: str,
agent_id: str,
) -> None:
"""Delete an agent session by its ID and its associated turns.
:param session_id: The ID of the session to delete.
:param agent_id: The ID of the agent to delete the session for.
"""
...
@webmethod(
route="/agents/{agent_id}",
method="DELETE",
deprecated=True,
level=LLAMA_STACK_API_V1,
)
@webmethod(route="/agents/{agent_id}", method="DELETE", level=LLAMA_STACK_API_V1ALPHA)
async def delete_agent(
self,
agent_id: str,
) -> None:
"""Delete an agent by its ID and its associated sessions and turns.
:param agent_id: The ID of the agent to delete.
"""
...
@webmethod(route="/agents", method="GET", deprecated=True, level=LLAMA_STACK_API_V1)
@webmethod(route="/agents", method="GET", level=LLAMA_STACK_API_V1ALPHA)
async def list_agents(self, start_index: int | None = None, limit: int | None = None) -> PaginatedResponse:
"""List all agents.
:param start_index: The index to start the pagination from.
:param limit: The number of agents to return.
:returns: A PaginatedResponse.
"""
...
@webmethod(
route="/agents/{agent_id}",
method="GET",
deprecated=True,
level=LLAMA_STACK_API_V1,
)
@webmethod(route="/agents/{agent_id}", method="GET", level=LLAMA_STACK_API_V1ALPHA)
async def get_agent(self, agent_id: str) -> Agent:
"""Describe an agent by its ID.
:param agent_id: ID of the agent.
:returns: An Agent of the agent.
"""
...
@webmethod(
route="/agents/{agent_id}/sessions",
method="GET",
deprecated=True,
level=LLAMA_STACK_API_V1,
)
@webmethod(route="/agents/{agent_id}/sessions", method="GET", level=LLAMA_STACK_API_V1ALPHA)
async def list_agent_sessions(
self,
agent_id: str,
start_index: int | None = None,
limit: int | None = None,
) -> PaginatedResponse:
"""List all session(s) of a given agent.
:param agent_id: The ID of the agent to list sessions for.
:param start_index: The index to start the pagination from.
:param limit: The number of sessions to return.
:returns: A PaginatedResponse.
"""
...
# We situate the OpenAI Responses API in the Agents API just like we did things
# for Inference. The Responses API, in its intent, serves the same purpose as
# the Agents API above -- it is essentially a lightweight "agentic loop" with
@ -787,12 +53,6 @@ class Agents(Protocol):
#
# Both of these APIs are inherently stateful.
@webmethod(
route="/openai/v1/responses/{response_id}",
method="GET",
level=LLAMA_STACK_API_V1,
deprecated=True,
)
@webmethod(route="/responses/{response_id}", method="GET", level=LLAMA_STACK_API_V1)
async def get_openai_response(
self,
@ -805,7 +65,6 @@ class Agents(Protocol):
"""
...
@webmethod(route="/openai/v1/responses", method="POST", level=LLAMA_STACK_API_V1, deprecated=True)
@webmethod(route="/responses", method="POST", level=LLAMA_STACK_API_V1)
async def create_openai_response(
self,
@ -842,7 +101,6 @@ class Agents(Protocol):
"""
...
@webmethod(route="/openai/v1/responses", method="GET", level=LLAMA_STACK_API_V1, deprecated=True)
@webmethod(route="/responses", method="GET", level=LLAMA_STACK_API_V1)
async def list_openai_responses(
self,
@ -861,9 +119,6 @@ class Agents(Protocol):
"""
...
@webmethod(
route="/openai/v1/responses/{response_id}/input_items", method="GET", level=LLAMA_STACK_API_V1, deprecated=True
)
@webmethod(route="/responses/{response_id}/input_items", method="GET", level=LLAMA_STACK_API_V1)
async def list_openai_response_input_items(
self,
@ -886,7 +141,6 @@ class Agents(Protocol):
"""
...
@webmethod(route="/openai/v1/responses/{response_id}", method="DELETE", level=LLAMA_STACK_API_V1, deprecated=True)
@webmethod(route="/responses/{response_id}", method="DELETE", level=LLAMA_STACK_API_V1)
async def delete_openai_response(self, response_id: str) -> OpenAIDeleteResponseObject:
"""Delete a response.

4
src/llama_stack/apis/batches/batches.py
View file

@ -43,7 +43,6 @@ class Batches(Protocol):
Note: This API is currently under active development and may undergo changes.
"""
@webmethod(route="/openai/v1/batches", method="POST", level=LLAMA_STACK_API_V1, deprecated=True)
@webmethod(route="/batches", method="POST", level=LLAMA_STACK_API_V1)
async def create_batch(
self,
@ -64,7 +63,6 @@ class Batches(Protocol):
"""
...
@webmethod(route="/openai/v1/batches/{batch_id}", method="GET", level=LLAMA_STACK_API_V1, deprecated=True)
@webmethod(route="/batches/{batch_id}", method="GET", level=LLAMA_STACK_API_V1)
async def retrieve_batch(self, batch_id: str) -> BatchObject:
"""Retrieve information about a specific batch.
@ -74,7 +72,6 @@ class Batches(Protocol):
"""
...
@webmethod(route="/openai/v1/batches/{batch_id}/cancel", method="POST", level=LLAMA_STACK_API_V1, deprecated=True)
@webmethod(route="/batches/{batch_id}/cancel", method="POST", level=LLAMA_STACK_API_V1)
async def cancel_batch(self, batch_id: str) -> BatchObject:
"""Cancel a batch that is in progress.
@ -84,7 +81,6 @@ class Batches(Protocol):
"""
...
@webmethod(route="/openai/v1/batches", method="GET", level=LLAMA_STACK_API_V1, deprecated=True)
@webmethod(route="/batches", method="GET", level=LLAMA_STACK_API_V1)
async def list_batches(
self,

6
src/llama_stack/apis/benchmarks/benchmarks.py
View file

@ -8,7 +8,7 @@ from typing import Any, Literal, Protocol, runtime_checkable
from pydantic import BaseModel, Field
from llama_stack.apis.resource import Resource, ResourceType
from llama_stack.apis.version import LLAMA_STACK_API_V1, LLAMA_STACK_API_V1ALPHA
from llama_stack.apis.version import LLAMA_STACK_API_V1ALPHA
from llama_stack.schema_utils import json_schema_type, webmethod
@ -54,7 +54,6 @@ class ListBenchmarksResponse(BaseModel):
@runtime_checkable
class Benchmarks(Protocol):
@webmethod(route="/eval/benchmarks", method="GET", level=LLAMA_STACK_API_V1, deprecated=True)
@webmethod(route="/eval/benchmarks", method="GET", level=LLAMA_STACK_API_V1ALPHA)
async def list_benchmarks(self) -> ListBenchmarksResponse:
"""List all benchmarks.
@ -63,7 +62,6 @@ class Benchmarks(Protocol):
"""
...
@webmethod(route="/eval/benchmarks/{benchmark_id}", method="GET", level=LLAMA_STACK_API_V1, deprecated=True)
@webmethod(route="/eval/benchmarks/{benchmark_id}", method="GET", level=LLAMA_STACK_API_V1ALPHA)
async def get_benchmark(
self,
@ -76,7 +74,6 @@ class Benchmarks(Protocol):
"""
...
@webmethod(route="/eval/benchmarks", method="POST", level=LLAMA_STACK_API_V1, deprecated=True)
@webmethod(route="/eval/benchmarks", method="POST", level=LLAMA_STACK_API_V1ALPHA)
async def register_benchmark(
self,
@ -98,7 +95,6 @@ class Benchmarks(Protocol):
"""
...
@webmethod(route="/eval/benchmarks/{benchmark_id}", method="DELETE", level=LLAMA_STACK_API_V1, deprecated=True)
@webmethod(route="/eval/benchmarks/{benchmark_id}", method="DELETE", level=LLAMA_STACK_API_V1ALPHA)
async def unregister_benchmark(self, benchmark_id: str) -> None:
"""Unregister a benchmark.

8
src/llama_stack/apis/common/errors.py
View file

@ -56,14 +56,6 @@ class ToolGroupNotFoundError(ResourceNotFoundError):
super().__init__(toolgroup_name, "Tool Group", "client.toolgroups.list()")
class SessionNotFoundError(ValueError):
"""raised when Llama Stack cannot find a referenced session or access is denied"""
def __init__(self, session_name: str) -> None:
message = f"Session '{session_name}' not found or access denied."
super().__init__(message)
class ModelTypeError(TypeError):
"""raised when a model is present but not the correct type"""

14
src/llama_stack/apis/common/type_system.py
View file

@ -103,17 +103,6 @@ class CompletionInputType(BaseModel):
type: Literal["completion_input"] = "completion_input"
@json_schema_type
class AgentTurnInputType(BaseModel):
"""Parameter type for agent turn input.
:param type: Discriminator type. Always "agent_turn_input"
"""
# expects List[Message] for messages (may also include attachments?)
type: Literal["agent_turn_input"] = "agent_turn_input"
@json_schema_type
class DialogType(BaseModel):
"""Parameter type for dialog data with semantic output labels.
@ -135,8 +124,7 @@ ParamType = Annotated[
| JsonType
| UnionType
| ChatCompletionInputType
| CompletionInputType
| AgentTurnInputType,
| CompletionInputType,
Field(discriminator="type"),
]
register_schema(ParamType, name="ParamType")

6
src/llama_stack/apis/datasetio/datasetio.py
View file

@ -8,7 +8,7 @@ from typing import Any, Protocol, runtime_checkable
from llama_stack.apis.common.responses import PaginatedResponse
from llama_stack.apis.datasets import Dataset
from llama_stack.apis.version import LLAMA_STACK_API_V1, LLAMA_STACK_API_V1BETA
from llama_stack.apis.version import LLAMA_STACK_API_V1BETA
from llama_stack.schema_utils import webmethod
@ -21,7 +21,6 @@ class DatasetIO(Protocol):
# keeping for aligning with inference/safety, but this is not used
dataset_store: DatasetStore
@webmethod(route="/datasetio/iterrows/{dataset_id:path}", method="GET", deprecated=True, level=LLAMA_STACK_API_V1)
@webmethod(route="/datasetio/iterrows/{dataset_id:path}", method="GET", level=LLAMA_STACK_API_V1BETA)
async def iterrows(
self,
@ -46,9 +45,6 @@ class DatasetIO(Protocol):
"""
...
@webmethod(
route="/datasetio/append-rows/{dataset_id:path}", method="POST", deprecated=True, level=LLAMA_STACK_API_V1
)
@webmethod(route="/datasetio/append-rows/{dataset_id:path}", method="POST", level=LLAMA_STACK_API_V1BETA)
async def append_rows(self, dataset_id: str, rows: list[dict[str, Any]]) -> None:
"""Append rows to a dataset.

6
src/llama_stack/apis/datasets/datasets.py
View file

@ -10,7 +10,7 @@ from typing import Annotated, Any, Literal, Protocol
from pydantic import BaseModel, Field
from llama_stack.apis.resource import Resource, ResourceType
from llama_stack.apis.version import LLAMA_STACK_API_V1, LLAMA_STACK_API_V1BETA
from llama_stack.apis.version import LLAMA_STACK_API_V1BETA
from llama_stack.schema_utils import json_schema_type, register_schema, webmethod
@ -146,7 +146,6 @@ class ListDatasetsResponse(BaseModel):
class Datasets(Protocol):
@webmethod(route="/datasets", method="POST", deprecated=True, level=LLAMA_STACK_API_V1)
@webmethod(route="/datasets", method="POST", level=LLAMA_STACK_API_V1BETA)
async def register_dataset(
self,
@ -216,7 +215,6 @@ class Datasets(Protocol):
"""
...
@webmethod(route="/datasets/{dataset_id:path}", method="GET", deprecated=True, level=LLAMA_STACK_API_V1)
@webmethod(route="/datasets/{dataset_id:path}", method="GET", level=LLAMA_STACK_API_V1BETA)
async def get_dataset(
self,
@ -229,7 +227,6 @@ class Datasets(Protocol):
"""
...
@webmethod(route="/datasets", method="GET", deprecated=True, level=LLAMA_STACK_API_V1)
@webmethod(route="/datasets", method="GET", level=LLAMA_STACK_API_V1BETA)
async def list_datasets(self) -> ListDatasetsResponse:
"""List all datasets.
@ -238,7 +235,6 @@ class Datasets(Protocol):
"""
...
@webmethod(route="/datasets/{dataset_id:path}", method="DELETE", deprecated=True, level=LLAMA_STACK_API_V1)
@webmethod(route="/datasets/{dataset_id:path}", method="DELETE", level=LLAMA_STACK_API_V1BETA)
async def unregister_dataset(
self,

40
src/llama_stack/apis/eval/eval.py
View file

@ -4,17 +4,16 @@
# This source code is licensed under the terms described in the LICENSE file in
# the root directory of this source tree.
from typing import Annotated, Any, Literal, Protocol
from typing import Any, Literal, Protocol
from pydantic import BaseModel, Field
from llama_stack.apis.agents import AgentConfig
from llama_stack.apis.common.job_types import Job
from llama_stack.apis.inference import SamplingParams, SystemMessage
from llama_stack.apis.scoring import ScoringResult
from llama_stack.apis.scoring_functions import ScoringFnParams
from llama_stack.apis.version import LLAMA_STACK_API_V1, LLAMA_STACK_API_V1ALPHA
from llama_stack.schema_utils import json_schema_type, register_schema, webmethod
from llama_stack.apis.version import LLAMA_STACK_API_V1ALPHA
from llama_stack.schema_utils import json_schema_type, webmethod
@json_schema_type
@ -32,19 +31,7 @@ class ModelCandidate(BaseModel):
system_message: SystemMessage | None = None
@json_schema_type
class AgentCandidate(BaseModel):
"""An agent candidate for evaluation.
:param config: The configuration for the agent candidate.
"""
type: Literal["agent"] = "agent"
config: AgentConfig
EvalCandidate = Annotated[ModelCandidate | AgentCandidate, Field(discriminator="type")]
register_schema(EvalCandidate, name="EvalCandidate")
EvalCandidate = ModelCandidate
@json_schema_type
@ -86,7 +73,6 @@ class Eval(Protocol):
Llama Stack Evaluation API for running evaluations on model and agent candidates."""
@webmethod(route="/eval/benchmarks/{benchmark_id}/jobs", method="POST", level=LLAMA_STACK_API_V1, deprecated=True)
@webmethod(route="/eval/benchmarks/{benchmark_id}/jobs", method="POST", level=LLAMA_STACK_API_V1ALPHA)
async def run_eval(
self,
@ -101,9 +87,6 @@ class Eval(Protocol):
"""
...
@webmethod(
route="/eval/benchmarks/{benchmark_id}/evaluations", method="POST", level=LLAMA_STACK_API_V1, deprecated=True
)
@webmethod(route="/eval/benchmarks/{benchmark_id}/evaluations", method="POST", level=LLAMA_STACK_API_V1ALPHA)
async def evaluate_rows(
self,
@ -122,9 +105,6 @@ class Eval(Protocol):
"""
...
@webmethod(
route="/eval/benchmarks/{benchmark_id}/jobs/{job_id}", method="GET", level=LLAMA_STACK_API_V1, deprecated=True
)
@webmethod(route="/eval/benchmarks/{benchmark_id}/jobs/{job_id}", method="GET", level=LLAMA_STACK_API_V1ALPHA)
async def job_status(self, benchmark_id: str, job_id: str) -> Job:
"""Get the status of a job.
@ -135,12 +115,6 @@ class Eval(Protocol):
"""
...
@webmethod(
route="/eval/benchmarks/{benchmark_id}/jobs/{job_id}",
method="DELETE",
level=LLAMA_STACK_API_V1,
deprecated=True,
)
@webmethod(route="/eval/benchmarks/{benchmark_id}/jobs/{job_id}", method="DELETE", level=LLAMA_STACK_API_V1ALPHA)
async def job_cancel(self, benchmark_id: str, job_id: str) -> None:
"""Cancel a job.
@ -150,12 +124,6 @@ class Eval(Protocol):
"""
...
@webmethod(
route="/eval/benchmarks/{benchmark_id}/jobs/{job_id}/result",
method="GET",
level=LLAMA_STACK_API_V1,
deprecated=True,
)
@webmethod(
route="/eval/benchmarks/{benchmark_id}/jobs/{job_id}/result", method="GET", level=LLAMA_STACK_API_V1ALPHA
)

5
src/llama_stack/apis/files/files.py
View file

@ -110,7 +110,6 @@ class Files(Protocol):
"""
# OpenAI Files API Endpoints
@webmethod(route="/openai/v1/files", method="POST", level=LLAMA_STACK_API_V1, deprecated=True)
@webmethod(route="/files", method="POST", level=LLAMA_STACK_API_V1)
async def openai_upload_file(
self,
@ -134,7 +133,6 @@ class Files(Protocol):
"""
...
@webmethod(route="/openai/v1/files", method="GET", level=LLAMA_STACK_API_V1, deprecated=True)
@webmethod(route="/files", method="GET", level=LLAMA_STACK_API_V1)
async def openai_list_files(
self,
@ -155,7 +153,6 @@ class Files(Protocol):
"""
...
@webmethod(route="/openai/v1/files/{file_id}", method="GET", level=LLAMA_STACK_API_V1, deprecated=True)
@webmethod(route="/files/{file_id}", method="GET", level=LLAMA_STACK_API_V1)
async def openai_retrieve_file(
self,
@ -170,7 +167,6 @@ class Files(Protocol):
"""
...
@webmethod(route="/openai/v1/files/{file_id}", method="DELETE", level=LLAMA_STACK_API_V1, deprecated=True)
@webmethod(route="/files/{file_id}", method="DELETE", level=LLAMA_STACK_API_V1)
async def openai_delete_file(
self,
@ -183,7 +179,6 @@ class Files(Protocol):
"""
...
@webmethod(route="/openai/v1/files/{file_id}/content", method="GET", level=LLAMA_STACK_API_V1, deprecated=True)
@webmethod(route="/files/{file_id}/content", method="GET", level=LLAMA_STACK_API_V1)
async def openai_retrieve_file_content(
self,

7
src/llama_stack/apis/inference/inference.py
View file

@ -1189,7 +1189,6 @@ class InferenceProvider(Protocol):
raise NotImplementedError("Reranking is not implemented")
return # this is so mypy's safe-super rule will consider the method concrete
@webmethod(route="/openai/v1/completions", method="POST", level=LLAMA_STACK_API_V1, deprecated=True)
@webmethod(route="/completions", method="POST", level=LLAMA_STACK_API_V1)
async def openai_completion(
self,
@ -1202,7 +1201,6 @@ class InferenceProvider(Protocol):
"""
...
@webmethod(route="/openai/v1/chat/completions", method="POST", level=LLAMA_STACK_API_V1, deprecated=True)
@webmethod(route="/chat/completions", method="POST", level=LLAMA_STACK_API_V1)
async def openai_chat_completion(
self,
@ -1215,7 +1213,6 @@ class InferenceProvider(Protocol):
"""
...
@webmethod(route="/openai/v1/embeddings", method="POST", level=LLAMA_STACK_API_V1, deprecated=True)
@webmethod(route="/embeddings", method="POST", level=LLAMA_STACK_API_V1)
async def openai_embeddings(
self,
@ -1240,7 +1237,6 @@ class Inference(InferenceProvider):
- Rerank models: these models reorder the documents based on their relevance to a query.
"""
@webmethod(route="/openai/v1/chat/completions", method="GET", level=LLAMA_STACK_API_V1, deprecated=True)
@webmethod(route="/chat/completions", method="GET", level=LLAMA_STACK_API_V1)
async def list_chat_completions(
self,
@ -1259,9 +1255,6 @@ class Inference(InferenceProvider):
"""
raise NotImplementedError("List chat completions is not implemented")
@webmethod(
route="/openai/v1/chat/completions/{completion_id}", method="GET", level=LLAMA_STACK_API_V1, deprecated=True
)
@webmethod(route="/chat/completions/{completion_id}", method="GET", level=LLAMA_STACK_API_V1)
async def get_chat_completion(self, completion_id: str) -> OpenAICompletionWithInputMessages:
"""Get chat completion.

5
src/llama_stack/apis/models/models.py
View file

@ -90,12 +90,14 @@ class OpenAIModel(BaseModel):
:object: The object type, which will be "model"
:created: The Unix timestamp in seconds when the model was created
:owned_by: The owner of the model
:custom_metadata: Llama Stack-specific metadata including model_type, provider info, and additional metadata
"""
id: str
object: Literal["model"] = "model"
created: int
owned_by: str
custom_metadata: dict[str, Any] | None = None
class OpenAIListModelsResponse(BaseModel):
@ -105,7 +107,6 @@ class OpenAIListModelsResponse(BaseModel):
@runtime_checkable
@trace_protocol
class Models(Protocol):
@webmethod(route="/models", method="GET", level=LLAMA_STACK_API_V1)
async def list_models(self) -> ListModelsResponse:
"""List all models.
@ -113,7 +114,7 @@ class Models(Protocol):
"""
...
@webmethod(route="/openai/v1/models", method="GET", level=LLAMA_STACK_API_V1, deprecated=True)
@webmethod(route="/models", method="GET", level=LLAMA_STACK_API_V1)
async def openai_list_models(self) -> OpenAIListModelsResponse:
"""List models using the OpenAI API.

8
src/llama_stack/apis/post_training/post_training.py
View file

@ -13,7 +13,7 @@ from pydantic import BaseModel, Field
from llama_stack.apis.common.content_types import URL
from llama_stack.apis.common.job_types import JobStatus
from llama_stack.apis.common.training_types import Checkpoint
from llama_stack.apis.version import LLAMA_STACK_API_V1, LLAMA_STACK_API_V1ALPHA
from llama_stack.apis.version import LLAMA_STACK_API_V1ALPHA
from llama_stack.schema_utils import json_schema_type, register_schema, webmethod
@ -284,7 +284,6 @@ class PostTrainingJobArtifactsResponse(BaseModel):
class PostTraining(Protocol):
@webmethod(route="/post-training/supervised-fine-tune", method="POST", level=LLAMA_STACK_API_V1, deprecated=True)
@webmethod(route="/post-training/supervised-fine-tune", method="POST", level=LLAMA_STACK_API_V1ALPHA)
async def supervised_fine_tune(
self,
@ -312,7 +311,6 @@ class PostTraining(Protocol):
"""
...
@webmethod(route="/post-training/preference-optimize", method="POST", level=LLAMA_STACK_API_V1, deprecated=True)
@webmethod(route="/post-training/preference-optimize", method="POST", level=LLAMA_STACK_API_V1ALPHA)
async def preference_optimize(
self,
@ -335,7 +333,6 @@ class PostTraining(Protocol):
"""
...
@webmethod(route="/post-training/jobs", method="GET", level=LLAMA_STACK_API_V1, deprecated=True)
@webmethod(route="/post-training/jobs", method="GET", level=LLAMA_STACK_API_V1ALPHA)
async def get_training_jobs(self) -> ListPostTrainingJobsResponse:
"""Get all training jobs.
@ -344,7 +341,6 @@ class PostTraining(Protocol):
"""
...
@webmethod(route="/post-training/job/status", method="GET", level=LLAMA_STACK_API_V1, deprecated=True)
@webmethod(route="/post-training/job/status", method="GET", level=LLAMA_STACK_API_V1ALPHA)
async def get_training_job_status(self, job_uuid: str) -> PostTrainingJobStatusResponse:
"""Get the status of a training job.
@ -354,7 +350,6 @@ class PostTraining(Protocol):
"""
...
@webmethod(route="/post-training/job/cancel", method="POST", level=LLAMA_STACK_API_V1, deprecated=True)
@webmethod(route="/post-training/job/cancel", method="POST", level=LLAMA_STACK_API_V1ALPHA)
async def cancel_training_job(self, job_uuid: str) -> None:
"""Cancel a training job.
@ -363,7 +358,6 @@ class PostTraining(Protocol):
"""
...
@webmethod(route="/post-training/job/artifacts", method="GET", level=LLAMA_STACK_API_V1, deprecated=True)
@webmethod(route="/post-training/job/artifacts", method="GET", level=LLAMA_STACK_API_V1ALPHA)
async def get_training_job_artifacts(self, job_uuid: str) -> PostTrainingJobArtifactsResponse:
"""Get the artifacts of a training job.

1
src/llama_stack/apis/safety/safety.py
View file

@ -121,7 +121,6 @@ class Safety(Protocol):
"""
...
@webmethod(route="/openai/v1/moderations", method="POST", level=LLAMA_STACK_API_V1, deprecated=True)
@webmethod(route="/moderations", method="POST", level=LLAMA_STACK_API_V1)
async def run_moderation(self, input: str | list[str], model: str | None = None) -> ModerationObject:
"""Create moderation.

7
src/llama_stack/apis/synthetic_data_generation/__init__.py
View file

@ -1,7 +0,0 @@
# Copyright (c) Meta Platforms, Inc. and affiliates.
# All rights reserved.
#
# This source code is licensed under the terms described in the LICENSE file in
# the root directory of this source tree.
from .synthetic_data_generation import *

77
src/llama_stack/apis/synthetic_data_generation/synthetic_data_generation.py
View file

@ -1,77 +0,0 @@
# Copyright (c) Meta Platforms, Inc. and affiliates.
# All rights reserved.
#
# This source code is licensed under the terms described in the LICENSE file in
# the root directory of this source tree.
from enum import Enum
from typing import Any, Protocol
from pydantic import BaseModel
from llama_stack.apis.inference import Message
from llama_stack.apis.version import LLAMA_STACK_API_V1
from llama_stack.schema_utils import json_schema_type, webmethod
class FilteringFunction(Enum):
"""The type of filtering function.
:cvar none: No filtering applied, accept all generated synthetic data
:cvar random: Random sampling of generated data points
:cvar top_k: Keep only the top-k highest scoring synthetic data samples
:cvar top_p: Nucleus-style filtering, keep samples exceeding cumulative score threshold
:cvar top_k_top_p: Combined top-k and top-p filtering strategy
:cvar sigmoid: Apply sigmoid function for probability-based filtering
"""
none = "none"
random = "random"
top_k = "top_k"
top_p = "top_p"
top_k_top_p = "top_k_top_p"
sigmoid = "sigmoid"
@json_schema_type
class SyntheticDataGenerationRequest(BaseModel):
"""Request to generate synthetic data. A small batch of prompts and a filtering function
:param dialogs: List of conversation messages to use as input for synthetic data generation
:param filtering_function: Type of filtering to apply to generated synthetic data samples
:param model: (Optional) The identifier of the model to use. The model must be registered with Llama Stack and available via the /models endpoint
"""
dialogs: list[Message]
filtering_function: FilteringFunction = FilteringFunction.none
model: str | None = None
@json_schema_type
class SyntheticDataGenerationResponse(BaseModel):
"""Response from the synthetic data generation. Batch of (prompt, response, score) tuples that pass the threshold.
:param synthetic_data: List of generated synthetic data samples that passed the filtering criteria
:param statistics: (Optional) Statistical information about the generation process and filtering results
"""
synthetic_data: list[dict[str, Any]]
statistics: dict[str, Any] | None = None
class SyntheticDataGeneration(Protocol):
@webmethod(route="/synthetic-data-generation/generate", level=LLAMA_STACK_API_V1)
def synthetic_data_generate(
self,
dialogs: list[Message],
filtering_function: FilteringFunction = FilteringFunction.none,
model: str | None = None,
) -> SyntheticDataGenerationResponse:
"""Generate synthetic data based on input dialogs and apply filtering.
:param dialogs: List of conversation messages to use as input for synthetic data generation
:param filtering_function: Type of filtering to apply to generated synthetic data samples
:param model: (Optional) The identifier of the model to use. The model must be registered with Llama Stack and available via the /models endpoint
:returns: Response containing filtered synthetic data samples and optional statistics
"""
...

52
src/llama_stack/apis/tools/rag_tool.py
View file

@ -5,18 +5,13 @@
# the root directory of this source tree.
from enum import Enum, StrEnum
from typing import Annotated, Any, Literal, Protocol
from typing import Annotated, Any, Literal
from pydantic import BaseModel, Field, field_validator
from typing_extensions import runtime_checkable
from llama_stack.apis.common.content_types import URL, InterleavedContent
from llama_stack.apis.version import LLAMA_STACK_API_V1
from llama_stack.core.telemetry.trace_protocol import trace_protocol
from llama_stack.schema_utils import json_schema_type, register_schema, webmethod
@json_schema_type
class RRFRanker(BaseModel):
"""
Reciprocal Rank Fusion (RRF) ranker configuration.
@ -30,7 +25,6 @@ class RRFRanker(BaseModel):
impact_factor: float = Field(default=60.0, gt=0.0) # default of 60 for optimal performance
@json_schema_type
class WeightedRanker(BaseModel):
"""
Weighted ranker configuration that combines vector and keyword scores.
@ -55,10 +49,8 @@ Ranker = Annotated[
RRFRanker | WeightedRanker,
Field(discriminator="type"),
]
register_schema(Ranker, name="Ranker")
@json_schema_type
class RAGDocument(BaseModel):
"""
A document to be used for document ingestion in the RAG Tool.
@ -75,7 +67,6 @@ class RAGDocument(BaseModel):
metadata: dict[str, Any] = Field(default_factory=dict)
@json_schema_type
class RAGQueryResult(BaseModel):
"""Result of a RAG query containing retrieved content and metadata.
@ -87,7 +78,6 @@ class RAGQueryResult(BaseModel):
metadata: dict[str, Any] = Field(default_factory=dict)
@json_schema_type
class RAGQueryGenerator(Enum):
"""Types of query generators for RAG systems.
@ -101,7 +91,6 @@ class RAGQueryGenerator(Enum):
custom = "custom"
@json_schema_type
class RAGSearchMode(StrEnum):
"""
Search modes for RAG query retrieval:
@ -115,7 +104,6 @@ class RAGSearchMode(StrEnum):
HYBRID = "hybrid"
@json_schema_type
class DefaultRAGQueryGeneratorConfig(BaseModel):
"""Configuration for the default RAG query generator.
@ -127,7 +115,6 @@ class DefaultRAGQueryGeneratorConfig(BaseModel):
separator: str = " "
@json_schema_type
class LLMRAGQueryGeneratorConfig(BaseModel):
"""Configuration for the LLM-based RAG query generator.
@ -145,10 +132,8 @@ RAGQueryGeneratorConfig = Annotated[
DefaultRAGQueryGeneratorConfig | LLMRAGQueryGeneratorConfig,
Field(discriminator="type"),
]
register_schema(RAGQueryGeneratorConfig, name="RAGQueryGeneratorConfig")
@json_schema_type
class RAGQueryConfig(BaseModel):
"""
Configuration for the RAG query generation.
@ -181,38 +166,3 @@ class RAGQueryConfig(BaseModel):
if len(v) == 0:
raise ValueError("chunk_template must not be empty")
return v
@runtime_checkable
@trace_protocol
class RAGToolRuntime(Protocol):
@webmethod(route="/tool-runtime/rag-tool/insert", method="POST", level=LLAMA_STACK_API_V1)
async def insert(
self,
documents: list[RAGDocument],
vector_store_id: str,
chunk_size_in_tokens: int = 512,
) -> None:
"""Index documents so they can be used by the RAG system.
:param documents: List of documents to index in the RAG system
:param vector_store_id: ID of the vector database to store the document embeddings
:param chunk_size_in_tokens: (Optional) Size in tokens for document chunking during indexing
"""
...
@webmethod(route="/tool-runtime/rag-tool/query", method="POST", level=LLAMA_STACK_API_V1)
async def query(
self,
content: InterleavedContent,
vector_store_ids: list[str],
query_config: RAGQueryConfig | None = None,
) -> RAGQueryResult:
"""Query the RAG system for context; typically invoked by the agent.
:param content: The query content to search for in the indexed documents
:param vector_store_ids: List of vector database IDs to search within
:param query_config: (Optional) Configuration parameters for the query operation
:returns: RAGQueryResult containing the retrieved content and metadata
"""
...

4
src/llama_stack/apis/tools/tools.py
View file

@ -16,8 +16,6 @@ from llama_stack.apis.version import LLAMA_STACK_API_V1
from llama_stack.core.telemetry.trace_protocol import trace_protocol
from llama_stack.schema_utils import json_schema_type, webmethod
from .rag_tool import RAGToolRuntime
@json_schema_type
class ToolDef(BaseModel):
@ -195,8 +193,6 @@ class SpecialToolGroup(Enum):
class ToolRuntime(Protocol):
tool_store: ToolStore | None = None
rag_tool: RAGToolRuntime | None = None
# TODO: This needs to be renamed once OPEN API generator name conflict issue is fixed.
@webmethod(route="/tool-runtime/list-tools", method="GET", level=LLAMA_STACK_API_V1)
async def list_runtime_tools(

77
src/llama_stack/apis/vector_io/vector_io.py
View file

@ -545,7 +545,6 @@ class VectorIO(Protocol):
...
# OpenAI Vector Stores API endpoints
@webmethod(route="/openai/v1/vector_stores", method="POST", level=LLAMA_STACK_API_V1, deprecated=True)
@webmethod(route="/vector_stores", method="POST", level=LLAMA_STACK_API_V1)
async def openai_create_vector_store(
self,
@ -558,7 +557,6 @@ class VectorIO(Protocol):
"""
...
@webmethod(route="/openai/v1/vector_stores", method="GET", level=LLAMA_STACK_API_V1, deprecated=True)
@webmethod(route="/vector_stores", method="GET", level=LLAMA_STACK_API_V1)
async def openai_list_vector_stores(
self,
@ -577,9 +575,6 @@ class VectorIO(Protocol):
"""
...
@webmethod(
route="/openai/v1/vector_stores/{vector_store_id}", method="GET", level=LLAMA_STACK_API_V1, deprecated=True
)
@webmethod(route="/vector_stores/{vector_store_id}", method="GET", level=LLAMA_STACK_API_V1)
async def openai_retrieve_vector_store(
self,
@ -592,9 +587,6 @@ class VectorIO(Protocol):
"""
...
@webmethod(
route="/openai/v1/vector_stores/{vector_store_id}", method="POST", level=LLAMA_STACK_API_V1, deprecated=True
)
@webmethod(
route="/vector_stores/{vector_store_id}",
method="POST",
@ -617,9 +609,6 @@ class VectorIO(Protocol):
"""
...
@webmethod(
route="/openai/v1/vector_stores/{vector_store_id}", method="DELETE", level=LLAMA_STACK_API_V1, deprecated=True
)
@webmethod(
route="/vector_stores/{vector_store_id}",
method="DELETE",
@ -636,12 +625,6 @@ class VectorIO(Protocol):
"""
...
@webmethod(
route="/openai/v1/vector_stores/{vector_store_id}/search",
method="POST",
level=LLAMA_STACK_API_V1,
deprecated=True,
)
@webmethod(
route="/vector_stores/{vector_store_id}/search",
method="POST",
@ -674,12 +657,6 @@ class VectorIO(Protocol):
"""
...
@webmethod(
route="/openai/v1/vector_stores/{vector_store_id}/files",
method="POST",
level=LLAMA_STACK_API_V1,
deprecated=True,
)
@webmethod(
route="/vector_stores/{vector_store_id}/files",
method="POST",
@ -702,12 +679,6 @@ class VectorIO(Protocol):
"""
...
@webmethod(
route="/openai/v1/vector_stores/{vector_store_id}/files",
method="GET",
level=LLAMA_STACK_API_V1,
deprecated=True,
)
@webmethod(
route="/vector_stores/{vector_store_id}/files",
method="GET",
@ -734,12 +705,6 @@ class VectorIO(Protocol):
"""
...
@webmethod(
route="/openai/v1/vector_stores/{vector_store_id}/files/{file_id}",
method="GET",
level=LLAMA_STACK_API_V1,
deprecated=True,
)
@webmethod(
route="/vector_stores/{vector_store_id}/files/{file_id}",
method="GET",
@ -758,12 +723,6 @@ class VectorIO(Protocol):
"""
...
@webmethod(
route="/openai/v1/vector_stores/{vector_store_id}/files/{file_id}/content",
method="GET",
level=LLAMA_STACK_API_V1,
deprecated=True,
)
@webmethod(
route="/vector_stores/{vector_store_id}/files/{file_id}/content",
method="GET",
@ -782,12 +741,6 @@ class VectorIO(Protocol):
"""
...
@webmethod(
route="/openai/v1/vector_stores/{vector_store_id}/files/{file_id}",
method="POST",
level=LLAMA_STACK_API_V1,
deprecated=True,
)
@webmethod(
route="/vector_stores/{vector_store_id}/files/{file_id}",
method="POST",
@ -808,12 +761,6 @@ class VectorIO(Protocol):
"""
...
@webmethod(
route="/openai/v1/vector_stores/{vector_store_id}/files/{file_id}",
method="DELETE",
level=LLAMA_STACK_API_V1,
deprecated=True,
)
@webmethod(
route="/vector_stores/{vector_store_id}/files/{file_id}",
method="DELETE",
@ -837,12 +784,6 @@ class VectorIO(Protocol):
method="POST",
level=LLAMA_STACK_API_V1,
)
@webmethod(
route="/openai/v1/vector_stores/{vector_store_id}/file_batches",
method="POST",
level=LLAMA_STACK_API_V1,
deprecated=True,
)
async def openai_create_vector_store_file_batch(
self,
vector_store_id: str,
@ -861,12 +802,6 @@ class VectorIO(Protocol):
method="GET",
level=LLAMA_STACK_API_V1,
)
@webmethod(
route="/openai/v1/vector_stores/{vector_store_id}/file_batches/{batch_id}",
method="GET",
level=LLAMA_STACK_API_V1,
deprecated=True,
)
async def openai_retrieve_vector_store_file_batch(
self,
batch_id: str,
@ -880,12 +815,6 @@ class VectorIO(Protocol):
"""
...
@webmethod(
route="/openai/v1/vector_stores/{vector_store_id}/file_batches/{batch_id}/files",
method="GET",
level=LLAMA_STACK_API_V1,
deprecated=True,
)
@webmethod(
route="/vector_stores/{vector_store_id}/file_batches/{batch_id}/files",
method="GET",
@ -914,12 +843,6 @@ class VectorIO(Protocol):
"""
...
@webmethod(
route="/openai/v1/vector_stores/{vector_store_id}/file_batches/{batch_id}/cancel",
method="POST",
level=LLAMA_STACK_API_V1,
deprecated=True,
)
@webmethod(
route="/vector_stores/{vector_store_id}/file_batches/{batch_id}/cancel",
method="POST",

2
src/llama_stack/cli/stack/run.py
View file

@ -253,7 +253,7 @@ class StackRun(Subcommand):
)
return
ui_dir = REPO_ROOT / "llama_stack" / "ui"
ui_dir = REPO_ROOT / "llama_stack_ui"
logs_dir = Path("~/.llama/ui/logs").expanduser()
try:
# Create logs directory if it doesn't exist

40
src/llama_stack/core/routers/tool_runtime.py
View file

@ -8,14 +8,9 @@ from typing import Any
from llama_stack.apis.common.content_types import (
URL,
InterleavedContent,
)
from llama_stack.apis.tools import (
ListToolDefsResponse,
RAGDocument,
RAGQueryConfig,
RAGQueryResult,
RAGToolRuntime,
ToolRuntime,
)
from llama_stack.log import get_logger
@ -26,36 +21,6 @@ logger = get_logger(name=__name__, category="core::routers")
class ToolRuntimeRouter(ToolRuntime):
class RagToolImpl(RAGToolRuntime):
def __init__(
self,
routing_table: ToolGroupsRoutingTable,
) -> None:
logger.debug("Initializing ToolRuntimeRouter.RagToolImpl")
self.routing_table = routing_table
async def query(
self,
content: InterleavedContent,
vector_store_ids: list[str],
query_config: RAGQueryConfig | None = None,
) -> RAGQueryResult:
logger.debug(f"ToolRuntimeRouter.RagToolImpl.query: {vector_store_ids}")
provider = await self.routing_table.get_provider_impl("knowledge_search")
return await provider.query(content, vector_store_ids, query_config)
async def insert(
self,
documents: list[RAGDocument],
vector_store_id: str,
chunk_size_in_tokens: int = 512,
) -> None:
logger.debug(
f"ToolRuntimeRouter.RagToolImpl.insert: {vector_store_id}, {len(documents)} documents, chunk_size={chunk_size_in_tokens}"
)
provider = await self.routing_table.get_provider_impl("insert_into_memory")
return await provider.insert(documents, vector_store_id, chunk_size_in_tokens)
def __init__(
self,
routing_table: ToolGroupsRoutingTable,
@ -63,11 +28,6 @@ class ToolRuntimeRouter(ToolRuntime):
logger.debug("Initializing ToolRuntimeRouter")
self.routing_table = routing_table
# HACK ALERT this should be in sync with "get_all_api_endpoints()"
self.rag_tool = self.RagToolImpl(routing_table)
for method in ("query", "insert"):
setattr(self, f"rag_tool.{method}", getattr(self.rag_tool, method))
async def initialize(self) -> None:
logger.debug("ToolRuntimeRouter.initialize")
pass

6
src/llama_stack/core/routing_tables/models.py
View file

@ -134,6 +134,12 @@ class ModelsRoutingTable(CommonRoutingTableImpl, Models):
object="model",
created=int(time.time()),
owned_by="llama_stack",
custom_metadata={
"model_type": model.model_type,
"provider_id": model.provider_id,
"provider_resource_id": model.provider_resource_id,
**model.metadata,
},
)
for model in all_models
]

18
src/llama_stack/core/server/routes.py
View file

@ -13,7 +13,6 @@ from aiohttp import hdrs
from starlette.routing import Route
from llama_stack.apis.datatypes import Api, ExternalApiSpec
from llama_stack.apis.tools import RAGToolRuntime, SpecialToolGroup
from llama_stack.core.resolver import api_protocol_map
from llama_stack.schema_utils import WebMethod
@ -25,33 +24,16 @@ RouteImpls = dict[str, PathImpl]
RouteMatch = tuple[EndpointFunc, PathParams, str, WebMethod]
def toolgroup_protocol_map():
return {
SpecialToolGroup.rag_tool: RAGToolRuntime,
}
def get_all_api_routes(
external_apis: dict[Api, ExternalApiSpec] | None = None,
) -> dict[Api, list[tuple[Route, WebMethod]]]:
apis = {}
protocols = api_protocol_map(external_apis)
toolgroup_protocols = toolgroup_protocol_map()
for api, protocol in protocols.items():
routes = []
protocol_methods = inspect.getmembers(protocol, predicate=inspect.isfunction)
# HACK ALERT
if api == Api.tool_runtime:
for tool_group in SpecialToolGroup:
sub_protocol = toolgroup_protocols[tool_group]
sub_protocol_methods = inspect.getmembers(sub_protocol, predicate=inspect.isfunction)
for name, method in sub_protocol_methods:
if not hasattr(method, "__webmethod__"):
continue
protocol_methods.append((f"{tool_group.value}.{name}", method))
for name, method in protocol_methods:
# Get all webmethods for this method (supports multiple decorators)
webmethods = getattr(method, "__webmethods__", [])

5
src/llama_stack/core/stack.py
View file

@ -31,8 +31,7 @@ from llama_stack.apis.safety import Safety
from llama_stack.apis.scoring import Scoring
from llama_stack.apis.scoring_functions import ScoringFunctions
from llama_stack.apis.shields import Shields
from llama_stack.apis.synthetic_data_generation import SyntheticDataGeneration
from llama_stack.apis.tools import RAGToolRuntime, ToolGroups, ToolRuntime
from llama_stack.apis.tools import ToolGroups, ToolRuntime
from llama_stack.apis.vector_io import VectorIO
from llama_stack.core.conversations.conversations import ConversationServiceConfig, ConversationServiceImpl
from llama_stack.core.datatypes import Provider, SafetyConfig, StackRunConfig, VectorStoresConfig
@ -66,7 +65,6 @@ class LlamaStack(
Agents,
Batches,
Safety,
SyntheticDataGeneration,
Datasets,
PostTraining,
VectorIO,
@ -80,7 +78,6 @@ class LlamaStack(
Inspect,
ToolGroups,
ToolRuntime,
RAGToolRuntime,
Files,
Prompts,
Conversations,

2
src/llama_stack/core/ui/page/distribution/models.py
View file

@ -12,7 +12,7 @@ from llama_stack.core.ui.modules.api import llama_stack_api
def models():
# Models Section
st.header("Models")
models_info = {m.identifier: m.to_dict() for m in llama_stack_api.client.models.list()}
models_info = {m.id: m.model_dump() for m in llama_stack_api.client.models.list()}
selected_model = st.selectbox("Select a model", list(models_info.keys()))
st.json(models_info[selected_model])

6
src/llama_stack/core/ui/page/playground/chat.py
View file

@ -12,7 +12,11 @@ from llama_stack.core.ui.modules.api import llama_stack_api
with st.sidebar:
st.header("Configuration")
available_models = llama_stack_api.client.models.list()
available_models = [model.identifier for model in available_models if model.model_type == "llm"]
available_models = [
model.id
for model in available_models
if model.custom_metadata and model.custom_metadata.get("model_type") == "llm"
]
selected_model = st.selectbox(
"Choose a model",
available_models,

1056
src/llama_stack/providers/inline/agents/meta_reference/agent_instance.py
View file

File diff suppressed because it is too large Load diff

245
src/llama_stack/providers/inline/agents/meta_reference/agents.py
View file

@ -4,21 +4,9 @@
# This source code is licensed under the terms described in the LICENSE file in
# the root directory of this source tree.
import uuid
from collections.abc import AsyncGenerator
from datetime import UTC, datetime
from llama_stack.apis.agents import (
Agent,
AgentConfig,
AgentCreateResponse,
Agents,
AgentSessionCreateResponse,
AgentStepResponse,
AgentToolGroup,
AgentTurnCreateRequest,
AgentTurnResumeRequest,
Document,
ListOpenAIResponseInputItem,
ListOpenAIResponseObject,
OpenAIDeleteResponseObject,
@ -26,19 +14,12 @@ from llama_stack.apis.agents import (
OpenAIResponseInputTool,
OpenAIResponseObject,
Order,
Session,
Turn,
)
from llama_stack.apis.agents.agents import ResponseGuardrail
from llama_stack.apis.agents.openai_responses import OpenAIResponsePrompt, OpenAIResponseText
from llama_stack.apis.common.responses import PaginatedResponse
from llama_stack.apis.conversations import Conversations
from llama_stack.apis.inference import (
Inference,
ToolConfig,
ToolResponse,
ToolResponseMessage,
UserMessage,
)
from llama_stack.apis.safety import Safety
from llama_stack.apis.tools import ToolGroups, ToolRuntime
@ -46,12 +27,9 @@ from llama_stack.apis.vector_io import VectorIO
from llama_stack.core.datatypes import AccessRule
from llama_stack.log import get_logger
from llama_stack.providers.utils.kvstore import InmemoryKVStoreImpl, kvstore_impl
from llama_stack.providers.utils.pagination import paginate_records
from llama_stack.providers.utils.responses.responses_store import ResponsesStore
from .agent_instance import ChatAgent
from .config import MetaReferenceAgentsImplConfig
from .persistence import AgentInfo
from .responses.openai_responses import OpenAIResponsesImpl
logger = get_logger(name=__name__, category="agents::meta_reference")
@ -97,229 +75,6 @@ class MetaReferenceAgentsImpl(Agents):
conversations_api=self.conversations_api,
)
async def create_agent(
self,
agent_config: AgentConfig,
) -> AgentCreateResponse:
agent_id = str(uuid.uuid4())
created_at = datetime.now(UTC)
agent_info = AgentInfo(
**agent_config.model_dump(),
created_at=created_at,
)
# Store the agent info
await self.persistence_store.set(
key=f"agent:{agent_id}",
value=agent_info.model_dump_json(),
)
return AgentCreateResponse(
agent_id=agent_id,
)
async def _get_agent_impl(self, agent_id: str) -> ChatAgent:
agent_info_json = await self.persistence_store.get(
key=f"agent:{agent_id}",
)
if not agent_info_json:
raise ValueError(f"Could not find agent info for {agent_id}")
try:
agent_info = AgentInfo.model_validate_json(agent_info_json)
except Exception as e:
raise ValueError(f"Could not validate agent info for {agent_id}") from e
return ChatAgent(
agent_id=agent_id,
agent_config=agent_info,
inference_api=self.inference_api,
safety_api=self.safety_api,
vector_io_api=self.vector_io_api,
tool_runtime_api=self.tool_runtime_api,
tool_groups_api=self.tool_groups_api,
persistence_store=(
self.persistence_store if agent_info.enable_session_persistence else self.in_memory_store
),
created_at=agent_info.created_at.isoformat(),
policy=self.policy,
telemetry_enabled=self.telemetry_enabled,
)
async def create_agent_session(
self,
agent_id: str,
session_name: str,
) -> AgentSessionCreateResponse:
agent = await self._get_agent_impl(agent_id)
session_id = await agent.create_session(session_name)
return AgentSessionCreateResponse(
session_id=session_id,
)
async def create_agent_turn(
self,
agent_id: str,
session_id: str,
messages: list[UserMessage | ToolResponseMessage],
stream: bool | None = False,
documents: list[Document] | None = None,
toolgroups: list[AgentToolGroup] | None = None,
tool_config: ToolConfig | None = None,
) -> AsyncGenerator:
request = AgentTurnCreateRequest(
agent_id=agent_id,
session_id=session_id,
messages=messages,
stream=True,
toolgroups=toolgroups,
documents=documents,
tool_config=tool_config,
)
if stream:
return self._create_agent_turn_streaming(request)
else:
raise NotImplementedError("Non-streaming agent turns not yet implemented")
async def _create_agent_turn_streaming(
self,
request: AgentTurnCreateRequest,
) -> AsyncGenerator:
agent = await self._get_agent_impl(request.agent_id)
async for event in agent.create_and_execute_turn(request):
yield event
async def resume_agent_turn(
self,
agent_id: str,
session_id: str,
turn_id: str,
tool_responses: list[ToolResponse],
stream: bool | None = False,
) -> AsyncGenerator:
request = AgentTurnResumeRequest(
agent_id=agent_id,
session_id=session_id,
turn_id=turn_id,
tool_responses=tool_responses,
stream=stream,
)
if stream:
return self._continue_agent_turn_streaming(request)
else:
raise NotImplementedError("Non-streaming agent turns not yet implemented")
async def _continue_agent_turn_streaming(
self,
request: AgentTurnResumeRequest,
) -> AsyncGenerator:
agent = await self._get_agent_impl(request.agent_id)
async for event in agent.resume_turn(request):
yield event
async def get_agents_turn(self, agent_id: str, session_id: str, turn_id: str) -> Turn:
agent = await self._get_agent_impl(agent_id)
turn = await agent.storage.get_session_turn(session_id, turn_id)
if turn is None:
raise ValueError(f"Turn {turn_id} not found in session {session_id}")
return turn
async def get_agents_step(self, agent_id: str, session_id: str, turn_id: str, step_id: str) -> AgentStepResponse:
turn = await self.get_agents_turn(agent_id, session_id, turn_id)
for step in turn.steps:
if step.step_id == step_id:
return AgentStepResponse(step=step)
raise ValueError(f"Provided step_id {step_id} could not be found")
async def get_agents_session(
self,
session_id: str,
agent_id: str,
turn_ids: list[str] | None = None,
) -> Session:
agent = await self._get_agent_impl(agent_id)
session_info = await agent.storage.get_session_info(session_id)
if session_info is None:
raise ValueError(f"Session {session_id} not found")
turns = await agent.storage.get_session_turns(session_id)
if turn_ids:
turns = [turn for turn in turns if turn.turn_id in turn_ids]
return Session(
session_name=session_info.session_name,
session_id=session_id,
turns=turns,
started_at=session_info.started_at,
)
async def delete_agents_session(self, session_id: str, agent_id: str) -> None:
agent = await self._get_agent_impl(agent_id)
# Delete turns first, then the session
await agent.storage.delete_session_turns(session_id)
await agent.storage.delete_session(session_id)
async def delete_agent(self, agent_id: str) -> None:
# First get all sessions for this agent
agent = await self._get_agent_impl(agent_id)
sessions = await agent.storage.list_sessions()
# Delete all sessions
for session in sessions:
await self.delete_agents_session(agent_id, session.session_id)
# Finally delete the agent itself
await self.persistence_store.delete(f"agent:{agent_id}")
async def list_agents(self, start_index: int | None = None, limit: int | None = None) -> PaginatedResponse:
agent_keys = await self.persistence_store.keys_in_range("agent:", "agent:\xff")
agent_list: list[Agent] = []
for agent_key in agent_keys:
agent_id = agent_key.split(":")[1]
# Get the agent info using the key
agent_info_json = await self.persistence_store.get(agent_key)
if not agent_info_json:
logger.error(f"Could not find agent info for key {agent_key}")
continue
try:
agent_info = AgentInfo.model_validate_json(agent_info_json)
agent_list.append(
Agent(
agent_id=agent_id,
agent_config=agent_info,
created_at=agent_info.created_at,
)
)
except Exception as e:
logger.error(f"Error parsing agent info for {agent_id}: {e}")
continue
# Convert Agent objects to dictionaries
agent_dicts = [agent.model_dump() for agent in agent_list]
return paginate_records(agent_dicts, start_index, limit)
async def get_agent(self, agent_id: str) -> Agent:
chat_agent = await self._get_agent_impl(agent_id)
agent = Agent(
agent_id=agent_id,
agent_config=chat_agent.agent_config,
created_at=datetime.fromisoformat(chat_agent.created_at),
)
return agent
async def list_agent_sessions(
self, agent_id: str, start_index: int | None = None, limit: int | None = None
) -> PaginatedResponse:
agent = await self._get_agent_impl(agent_id)
sessions = await agent.storage.list_sessions()
# Convert Session objects to dictionaries
session_dicts = [session.model_dump() for session in sessions]
return paginate_records(session_dicts, start_index, limit)
async def shutdown(self) -> None:
pass

261
src/llama_stack/providers/inline/agents/meta_reference/persistence.py
View file

@ -1,261 +0,0 @@
# Copyright (c) Meta Platforms, Inc. and affiliates.
# All rights reserved.
#
# This source code is licensed under the terms described in the LICENSE file in
# the root directory of this source tree.
import json
import uuid
from dataclasses import dataclass
from datetime import UTC, datetime
from llama_stack.apis.agents import AgentConfig, Session, ToolExecutionStep, Turn
from llama_stack.apis.common.errors import SessionNotFoundError
from llama_stack.core.access_control.access_control import AccessDeniedError, is_action_allowed
from llama_stack.core.access_control.conditions import User as ProtocolUser
from llama_stack.core.access_control.datatypes import AccessRule, Action
from llama_stack.core.datatypes import User
from llama_stack.core.request_headers import get_authenticated_user
from llama_stack.log import get_logger
from llama_stack.providers.utils.kvstore import KVStore
log = get_logger(name=__name__, category="agents::meta_reference")
class AgentSessionInfo(Session):
# TODO: is this used anywhere?
vector_store_id: str | None = None
started_at: datetime
owner: User | None = None
identifier: str | None = None
type: str = "session"
class AgentInfo(AgentConfig):
created_at: datetime
@dataclass
class SessionResource:
"""Concrete implementation of ProtectedResource for session access control."""
type: str
identifier: str
owner: ProtocolUser # Use the protocol type for structural compatibility
class AgentPersistence:
def __init__(self, agent_id: str, kvstore: KVStore, policy: list[AccessRule]):
self.agent_id = agent_id
self.kvstore = kvstore
self.policy = policy
async def create_session(self, name: str) -> str:
session_id = str(uuid.uuid4())
# Get current user's auth attributes for new sessions
user = get_authenticated_user()
session_info = AgentSessionInfo(
session_id=session_id,
session_name=name,
started_at=datetime.now(UTC),
owner=user,
turns=[],
identifier=name, # should this be qualified in any way?
)
# Only perform access control if we have an authenticated user
if user is not None and session_info.identifier is not None:
resource = SessionResource(
type=session_info.type,
identifier=session_info.identifier,
owner=user,
)
if not is_action_allowed(self.policy, Action.CREATE, resource, user):
raise AccessDeniedError(Action.CREATE, resource, user)
await self.kvstore.set(
key=f"session:{self.agent_id}:{session_id}",
value=session_info.model_dump_json(),
)
return session_id
async def get_session_info(self, session_id: str) -> AgentSessionInfo | None:
value = await self.kvstore.get(
key=f"session:{self.agent_id}:{session_id}",
)
if not value:
raise SessionNotFoundError(session_id)
session_info = AgentSessionInfo(**json.loads(value))
# Check access to session
if not self._check_session_access(session_info):
return None
return session_info
def _check_session_access(self, session_info: AgentSessionInfo) -> bool:
"""Check if current user has access to the session."""
# Handle backward compatibility for old sessions without access control
if not hasattr(session_info, "access_attributes") and not hasattr(session_info, "owner"):
return True
# Get current user - if None, skip access control (e.g., in tests)
user = get_authenticated_user()
if user is None:
return True
# Access control requires identifier and owner to be set
if session_info.identifier is None or session_info.owner is None:
return True
# At this point, both identifier and owner are guaranteed to be non-None
resource = SessionResource(
type=session_info.type,
identifier=session_info.identifier,
owner=session_info.owner,
)
return is_action_allowed(self.policy, Action.READ, resource, user)
async def get_session_if_accessible(self, session_id: str) -> AgentSessionInfo | None:
"""Get session info if the user has access to it. For internal use by sub-session methods."""
session_info = await self.get_session_info(session_id)
if not session_info:
return None
return session_info
async def add_vector_db_to_session(self, session_id: str, vector_store_id: str):
session_info = await self.get_session_if_accessible(session_id)
if session_info is None:
raise SessionNotFoundError(session_id)
session_info.vector_store_id = vector_store_id
await self.kvstore.set(
key=f"session:{self.agent_id}:{session_id}",
value=session_info.model_dump_json(),
)
async def add_turn_to_session(self, session_id: str, turn: Turn):
if not await self.get_session_if_accessible(session_id):
raise SessionNotFoundError(session_id)
await self.kvstore.set(
key=f"session:{self.agent_id}:{session_id}:{turn.turn_id}",
value=turn.model_dump_json(),
)
async def get_session_turns(self, session_id: str) -> list[Turn]:
if not await self.get_session_if_accessible(session_id):
raise SessionNotFoundError(session_id)
values = await self.kvstore.values_in_range(
start_key=f"session:{self.agent_id}:{session_id}:",
end_key=f"session:{self.agent_id}:{session_id}:\xff\xff\xff\xff",
)
turns = []
for value in values:
try:
turn = Turn(**json.loads(value))
turns.append(turn)
except Exception as e:
log.error(f"Error parsing turn: {e}")
continue
# The kvstore does not guarantee order, so we sort by started_at
# to ensure consistent ordering of turns.
turns.sort(key=lambda t: t.started_at)
return turns
async def get_session_turn(self, session_id: str, turn_id: str) -> Turn | None:
if not await self.get_session_if_accessible(session_id):
raise SessionNotFoundError(session_id)
value = await self.kvstore.get(
key=f"session:{self.agent_id}:{session_id}:{turn_id}",
)
if not value:
return None
return Turn(**json.loads(value))
async def set_in_progress_tool_call_step(self, session_id: str, turn_id: str, step: ToolExecutionStep):
if not await self.get_session_if_accessible(session_id):
raise SessionNotFoundError(session_id)
await self.kvstore.set(
key=f"in_progress_tool_call_step:{self.agent_id}:{session_id}:{turn_id}",
value=step.model_dump_json(),
)
async def get_in_progress_tool_call_step(self, session_id: str, turn_id: str) -> ToolExecutionStep | None:
if not await self.get_session_if_accessible(session_id):
return None
value = await self.kvstore.get(
key=f"in_progress_tool_call_step:{self.agent_id}:{session_id}:{turn_id}",
)
return ToolExecutionStep(**json.loads(value)) if value else None
async def set_num_infer_iters_in_turn(self, session_id: str, turn_id: str, num_infer_iters: int):
if not await self.get_session_if_accessible(session_id):
raise SessionNotFoundError(session_id)
await self.kvstore.set(
key=f"num_infer_iters_in_turn:{self.agent_id}:{session_id}:{turn_id}",
value=str(num_infer_iters),
)
async def get_num_infer_iters_in_turn(self, session_id: str, turn_id: str) -> int | None:
if not await self.get_session_if_accessible(session_id):
return None
value = await self.kvstore.get(
key=f"num_infer_iters_in_turn:{self.agent_id}:{session_id}:{turn_id}",
)
return int(value) if value else None
async def list_sessions(self) -> list[Session]:
values = await self.kvstore.values_in_range(
start_key=f"session:{self.agent_id}:",
end_key=f"session:{self.agent_id}:\xff\xff\xff\xff",
)
sessions = []
for value in values:
try:
data = json.loads(value)
if "turn_id" in data:
continue
session_info = Session(**data)
sessions.append(session_info)
except Exception as e:
log.error(f"Error parsing session info: {e}")
continue
return sessions
async def delete_session_turns(self, session_id: str) -> None:
"""Delete all turns and their associated data for a session.
Args:
session_id: The ID of the session whose turns should be deleted.
"""
turns = await self.get_session_turns(session_id)
for turn in turns:
await self.kvstore.delete(key=f"session:{self.agent_id}:{session_id}:{turn.turn_id}")
async def delete_session(self, session_id: str) -> None:
"""Delete a session and all its associated turns.
Args:
session_id: The ID of the session to delete.
Raises:
ValueError: If the session does not exist.
"""
session_info = await self.get_session_info(session_id)
if session_info is None:
raise SessionNotFoundError(session_id)
await self.kvstore.delete(key=f"session:{self.agent_id}:{session_id}")

54
src/llama_stack/providers/inline/eval/meta_reference/eval.py
View file

@ -8,7 +8,7 @@ from typing import Any
from tqdm import tqdm
from llama_stack.apis.agents import Agents, StepType
from llama_stack.apis.agents import Agents
from llama_stack.apis.benchmarks import Benchmark
from llama_stack.apis.datasetio import DatasetIO
from llama_stack.apis.datasets import Datasets
@ -18,13 +18,9 @@ from llama_stack.apis.inference import (
OpenAICompletionRequestWithExtraBody,
OpenAISystemMessageParam,
OpenAIUserMessageParam,
UserMessage,
)
from llama_stack.apis.scoring import Scoring
from llama_stack.providers.datatypes import BenchmarksProtocolPrivate
from llama_stack.providers.inline.agents.meta_reference.agent_instance import (
MEMORY_QUERY_TOOL,
)
from llama_stack.providers.utils.common.data_schema_validator import ColumnName
from llama_stack.providers.utils.kvstore import kvstore_impl
@ -118,49 +114,6 @@ class MetaReferenceEvalImpl(
self.jobs[job_id] = res
return Job(job_id=job_id, status=JobStatus.completed)
async def _run_agent_generation(
self, input_rows: list[dict[str, Any]], benchmark_config: BenchmarkConfig
) -> list[dict[str, Any]]:
candidate = benchmark_config.eval_candidate
create_response = await self.agents_api.create_agent(candidate.config)
agent_id = create_response.agent_id
generations = []
for i, x in tqdm(enumerate(input_rows)):
assert ColumnName.chat_completion_input.value in x, "Invalid input row"
input_messages = json.loads(x[ColumnName.chat_completion_input.value])
input_messages = [UserMessage(**x) for x in input_messages if x["role"] == "user"]
# NOTE: only single-turn agent generation is supported. Create a new session for each input row
session_create_response = await self.agents_api.create_agent_session(agent_id, f"session-{i}")
session_id = session_create_response.session_id
turn_request = dict(
agent_id=agent_id,
session_id=session_id,
messages=input_messages,
stream=True,
)
turn_response = [chunk async for chunk in await self.agents_api.create_agent_turn(**turn_request)]
final_event = turn_response[-1].event.payload
# check if there's a memory retrieval step and extract the context
memory_rag_context = None
for step in final_event.turn.steps:
if step.step_type == StepType.tool_execution.value:
for tool_response in step.tool_responses:
if tool_response.tool_name == MEMORY_QUERY_TOOL:
memory_rag_context = " ".join(x.text for x in tool_response.content)
agent_generation = {}
agent_generation[ColumnName.generated_answer.value] = final_event.turn.output_message.content
if memory_rag_context:
agent_generation[ColumnName.context.value] = memory_rag_context
generations.append(agent_generation)
return generations
async def _run_model_generation(
self, input_rows: list[dict[str, Any]], benchmark_config: BenchmarkConfig
) -> list[dict[str, Any]]:
@ -215,9 +168,8 @@ class MetaReferenceEvalImpl(
benchmark_config: BenchmarkConfig,
) -> EvaluateResponse:
candidate = benchmark_config.eval_candidate
if candidate.type == "agent":
generations = await self._run_agent_generation(input_rows, benchmark_config)
elif candidate.type == "model":
# Agent evaluation removed
if candidate.type == "model":
generations = await self._run_model_generation(input_rows, benchmark_config)
else:
raise ValueError(f"Invalid candidate type: {candidate.type}")

3
src/llama_stack/providers/inline/tool_runtime/rag/memory.py
View file

@ -27,7 +27,6 @@ from llama_stack.apis.tools import (
RAGDocument,
RAGQueryConfig,
RAGQueryResult,
RAGToolRuntime,
ToolDef,
ToolGroup,
ToolInvocationResult,
@ -91,7 +90,7 @@ async def raw_data_from_doc(doc: RAGDocument) -> tuple[bytes, str]:
return content_str.encode("utf-8"), "text/plain"
class MemoryToolRuntimeImpl(ToolGroupsProtocolPrivate, ToolRuntime, RAGToolRuntime):
class MemoryToolRuntimeImpl(ToolGroupsProtocolPrivate, ToolRuntime):
def __init__(
self,
config: RagToolRuntimeConfig,

10
src/llama_stack/providers/remote/inference/vertexai/vertexai.py
View file

@ -4,6 +4,7 @@
# This source code is licensed under the terms described in the LICENSE file in
# the root directory of this source tree.
from collections.abc import Iterable
import google.auth.transport.requests
from google.auth import default
@ -42,3 +43,12 @@ class VertexAIInferenceAdapter(OpenAIMixin):
Source: https://cloud.google.com/vertex-ai/generative-ai/docs/start/openai
"""
return f"https://{self.config.location}-aiplatform.googleapis.com/v1/projects/{self.config.project}/locations/{self.config.location}/endpoints/openapi"
async def list_provider_model_ids(self) -> Iterable[str]:
"""
VertexAI doesn't currently offer a way to query a list of available models from Google's Model Garden
For now we return a hardcoded version of the available models
:return: An iterable of model IDs
"""
return ["vertexai/gemini-2.0-flash", "vertexai/gemini-2.5-flash", "vertexai/gemini-2.5-pro"]

21
src/llama_stack/providers/utils/inference/inference_store.py
View file

@ -35,6 +35,7 @@ class InferenceStore:
self.reference = reference
self.sql_store = None
self.policy = policy
self.enable_write_queue = True
# Async write queue and worker control
self._queue: asyncio.Queue[tuple[OpenAIChatCompletion, list[OpenAIMessageParam]]] | None = None
@ -47,14 +48,13 @@ class InferenceStore:
base_store = sqlstore_impl(self.reference)
self.sql_store = AuthorizedSqlStore(base_store, self.policy)
# Disable write queue for SQLite to avoid concurrency issues
backend_name = self.reference.backend
backend_config = _SQLSTORE_BACKENDS.get(backend_name)
if backend_config is None:
raise ValueError(
f"Unregistered SQL backend '{backend_name}'. Registered backends: {sorted(_SQLSTORE_BACKENDS)}"
)
self.enable_write_queue = backend_config.type != StorageBackendType.SQL_SQLITE
# Disable write queue for SQLite since WAL mode handles concurrency
# Keep it enabled for other backends (like Postgres) for performance
backend_config = _SQLSTORE_BACKENDS.get(self.reference.backend)
if backend_config and backend_config.type == StorageBackendType.SQL_SQLITE:
self.enable_write_queue = False
logger.debug("Write queue disabled for SQLite (WAL mode handles concurrency)")
await self.sql_store.create_table(
"chat_completions",
{
@ -70,8 +70,9 @@ class InferenceStore:
self._queue = asyncio.Queue(maxsize=self._max_write_queue_size)
for _ in range(self._num_writers):
self._worker_tasks.append(asyncio.create_task(self._worker_loop()))
else:
logger.info("Write queue disabled for SQLite to avoid concurrency issues")
logger.debug(
f"Inference store write queue enabled with {self._num_writers} writers, max queue size {self._max_write_queue_size}"
)
async def shutdown(self) -> None:
if not self._worker_tasks:

15
src/llama_stack/providers/utils/responses/responses_store.py
View file

@ -70,13 +70,13 @@ class ResponsesStore:
base_store = sqlstore_impl(self.reference)
self.sql_store = AuthorizedSqlStore(base_store, self.policy)
# Disable write queue for SQLite since WAL mode handles concurrency
# Keep it enabled for other backends (like Postgres) for performance
backend_config = _SQLSTORE_BACKENDS.get(self.reference.backend)
if backend_config is None:
raise ValueError(
f"Unregistered SQL backend '{self.reference.backend}'. Registered backends: {sorted(_SQLSTORE_BACKENDS)}"
)
if backend_config.type == StorageBackendType.SQL_SQLITE:
if backend_config and backend_config.type == StorageBackendType.SQL_SQLITE:
self.enable_write_queue = False
logger.debug("Write queue disabled for SQLite (WAL mode handles concurrency)")
await self.sql_store.create_table(
"openai_responses",
{
@ -99,8 +99,9 @@ class ResponsesStore:
self._queue = asyncio.Queue(maxsize=self._max_write_queue_size)
for _ in range(self._num_writers):
self._worker_tasks.append(asyncio.create_task(self._worker_loop()))
else:
logger.debug("Write queue disabled for SQLite to avoid concurrency issues")
logger.debug(
f"Responses store write queue enabled with {self._num_writers} writers, max queue size {self._max_write_queue_size}"
)
async def shutdown(self) -> None:
if not self._worker_tasks:

32
src/llama_stack/providers/utils/sqlstore/sqlalchemy_sqlstore.py
View file

@ -17,6 +17,7 @@ from sqlalchemy import (
String,
Table,
Text,
event,
inspect,
select,
text,
@ -75,7 +76,36 @@ class SqlAlchemySqlStoreImpl(SqlStore):
self.metadata = MetaData()
def create_engine(self) -> AsyncEngine:
return create_async_engine(self.config.engine_str, pool_pre_ping=True)
# Configure connection args for better concurrency support
connect_args = {}
if "sqlite" in self.config.engine_str:
# SQLite-specific optimizations for concurrent access
# With WAL mode, most locks resolve in milliseconds, but allow up to 5s for edge cases
connect_args["timeout"] = 5.0
connect_args["check_same_thread"] = False # Allow usage across asyncio tasks
engine = create_async_engine(
self.config.engine_str,
pool_pre_ping=True,
connect_args=connect_args,
)
# Enable WAL mode for SQLite to support concurrent readers and writers
if "sqlite" in self.config.engine_str:
@event.listens_for(engine.sync_engine, "connect")
def set_sqlite_pragma(dbapi_conn, connection_record):
cursor = dbapi_conn.cursor()
# Enable Write-Ahead Logging for better concurrency
cursor.execute("PRAGMA journal_mode=WAL")
# Set busy timeout to 5 seconds (retry instead of immediate failure)
# With WAL mode, locks should be brief; if we hit 5s there's a bigger issue
cursor.execute("PRAGMA busy_timeout=5000")
# Use NORMAL synchronous mode for better performance (still safe with WAL)
cursor.execute("PRAGMA synchronous=NORMAL")
cursor.close()
return engine
async def create_table(
self,

17
src/llama_stack/testing/api_recorder.py
View file

@ -156,7 +156,7 @@ def normalize_inference_request(method: str, url: str, headers: dict[str, Any],
}
# Include test_id for isolation, except for shared infrastructure endpoints
if parsed.path not in ("/api/tags", "/v1/models"):
if parsed.path not in ("/api/tags", "/v1/models", "/v1/openai/v1/models"):
normalized["test_id"] = test_id
normalized_json = json.dumps(normalized, sort_keys=True)
@ -430,7 +430,7 @@ class ResponseStorage:
# For model-list endpoints, include digest in filename to distinguish different model sets
endpoint = request.get("endpoint")
if endpoint in ("/api/tags", "/v1/models"):
if endpoint in ("/api/tags", "/v1/models", "/v1/openai/v1/models"):
digest = _model_identifiers_digest(endpoint, response)
response_file = f"models-{request_hash}-{digest}.json"
@ -554,13 +554,14 @@ def _model_identifiers_digest(endpoint: str, response: dict[str, Any]) -> str:
Supported endpoints:
- '/api/tags' (Ollama): response body has 'models': [ { name/model/digest/id/... }, ... ]
- '/v1/models' (OpenAI): response body is: [ { id: ... }, ... ]
- '/v1/openai/v1/models' (OpenAI): response body is: [ { id: ... }, ... ]
Returns a list of unique identifiers or None if structure doesn't match.
"""
if "models" in response["body"]:
# ollama
items = response["body"]["models"]
else:
# openai
# openai or openai-style endpoints
items = response["body"]
idents = [m.model if endpoint == "/api/tags" else m.id for m in items]
return sorted(set(idents))
@ -581,7 +582,7 @@ def _combine_model_list_responses(endpoint: str, records: list[dict[str, Any]])
seen: dict[str, dict[str, Any]] = {}
for rec in records:
body = rec["response"]["body"]
if endpoint == "/v1/models":
if endpoint in ("/v1/models", "/v1/openai/v1/models"):
for m in body:
key = m.id
seen[key] = m
@ -665,7 +666,7 @@ async def _patched_inference_method(original_method, self, client_type, endpoint
logger.info(f" Test context: {get_test_context()}")
if mode == APIRecordingMode.LIVE or storage is None:
if endpoint == "/v1/models":
if endpoint in ("/v1/models", "/v1/openai/v1/models"):
return original_method(self, *args, **kwargs)
else:
return await original_method(self, *args, **kwargs)
@ -699,7 +700,7 @@ async def _patched_inference_method(original_method, self, client_type, endpoint
recording = None
if mode == APIRecordingMode.REPLAY or mode == APIRecordingMode.RECORD_IF_MISSING:
# Special handling for model-list endpoints: merge all recordings with this hash
if endpoint in ("/api/tags", "/v1/models"):
if endpoint in ("/api/tags", "/v1/models", "/v1/openai/v1/models"):
records = storage._model_list_responses(request_hash)
recording = _combine_model_list_responses(endpoint, records)
else:
@ -739,13 +740,13 @@ async def _patched_inference_method(original_method, self, client_type, endpoint
)
if mode == APIRecordingMode.RECORD or (mode == APIRecordingMode.RECORD_IF_MISSING and not recording):
if endpoint == "/v1/models":
if endpoint in ("/v1/models", "/v1/openai/v1/models"):
response = original_method(self, *args, **kwargs)
else:
response = await original_method(self, *args, **kwargs)
# we want to store the result of the iterator, not the iterator itself
if endpoint == "/v1/models":
if endpoint in ("/v1/models", "/v1/openai/v1/models"):
response = [m async for m in response]
request_data = {

0
src/llama_stack/ui/.gitignore → src/llama_stack_ui/.gitignore vendored
View file

0
src/llama_stack/ui/.nvmrc → src/llama_stack_ui/.nvmrc
View file

0
src/llama_stack/ui/.prettierignore → src/llama_stack_ui/.prettierignore
View file

0
src/llama_stack/ui/.prettierrc → src/llama_stack_ui/.prettierrc
View file

0
src/llama_stack/ui/README.md → src/llama_stack_ui/README.md
View file

0
src/llama_stack/ui/app/api/auth/[...nextauth]/route.ts → src/llama_stack_ui/app/api/auth/[...nextauth]/route.ts
View file

0
src/llama_stack/ui/app/api/v1/[...path]/route.ts → src/llama_stack_ui/app/api/v1/[...path]/route.ts
View file

0
src/llama_stack/ui/app/auth/signin/page.tsx → src/llama_stack_ui/app/auth/signin/page.tsx
View file

0
src/llama_stack/ui/app/chat-playground/chunk-processor.test.tsx → src/llama_stack_ui/app/chat-playground/chunk-processor.test.tsx
View file

0
src/llama_stack/ui/app/chat-playground/page.test.tsx → src/llama_stack_ui/app/chat-playground/page.test.tsx
View file

0
src/llama_stack/ui/app/chat-playground/page.tsx → src/llama_stack_ui/app/chat-playground/page.tsx
View file

0
src/llama_stack/ui/app/globals.css → src/llama_stack_ui/app/globals.css
View file

0
src/llama_stack/ui/app/layout.tsx → src/llama_stack_ui/app/layout.tsx
View file

0
src/llama_stack/ui/app/logs/chat-completions/[id]/page.tsx → src/llama_stack_ui/app/logs/chat-completions/[id]/page.tsx
View file

0
src/llama_stack/ui/app/logs/chat-completions/layout.tsx → src/llama_stack_ui/app/logs/chat-completions/layout.tsx
View file

0
src/llama_stack/ui/app/logs/chat-completions/page.tsx → src/llama_stack_ui/app/logs/chat-completions/page.tsx
View file

0
src/llama_stack/ui/app/logs/responses/[id]/page.tsx → src/llama_stack_ui/app/logs/responses/[id]/page.tsx
View file

0
src/llama_stack/ui/app/logs/responses/layout.tsx → src/llama_stack_ui/app/logs/responses/layout.tsx
View file

0
src/llama_stack/ui/app/logs/responses/page.tsx → src/llama_stack_ui/app/logs/responses/page.tsx
View file

0
src/llama_stack/ui/app/logs/vector-stores/[id]/files/[fileId]/contents/[contentId]/page.test.tsx → src/llama_stack_ui/app/logs/vector-stores/[id]/files/[fileId]/contents/[contentId]/page.test.tsx
View file

0
src/llama_stack/ui/app/logs/vector-stores/[id]/files/[fileId]/contents/[contentId]/page.tsx → src/llama_stack_ui/app/logs/vector-stores/[id]/files/[fileId]/contents/[contentId]/page.tsx
View file

0
src/llama_stack/ui/app/logs/vector-stores/[id]/files/[fileId]/contents/page.test.tsx → src/llama_stack_ui/app/logs/vector-stores/[id]/files/[fileId]/contents/page.test.tsx
View file

0
src/llama_stack/ui/app/logs/vector-stores/[id]/files/[fileId]/contents/page.tsx → src/llama_stack_ui/app/logs/vector-stores/[id]/files/[fileId]/contents/page.tsx
View file

0
src/llama_stack/ui/app/logs/vector-stores/[id]/files/[fileId]/page.test.tsx → src/llama_stack_ui/app/logs/vector-stores/[id]/files/[fileId]/page.test.tsx
View file

0
src/llama_stack/ui/app/logs/vector-stores/[id]/files/[fileId]/page.tsx → src/llama_stack_ui/app/logs/vector-stores/[id]/files/[fileId]/page.tsx
View file

0
src/llama_stack/ui/app/logs/vector-stores/[id]/page.tsx → src/llama_stack_ui/app/logs/vector-stores/[id]/page.tsx
View file

0
src/llama_stack/ui/app/logs/vector-stores/layout.tsx → src/llama_stack_ui/app/logs/vector-stores/layout.tsx
View file

0
src/llama_stack/ui/app/logs/vector-stores/page.tsx → src/llama_stack_ui/app/logs/vector-stores/page.tsx
View file

0
src/llama_stack/ui/app/page.tsx → src/llama_stack_ui/app/page.tsx
View file

0
src/llama_stack/ui/app/prompts/page.tsx → src/llama_stack_ui/app/prompts/page.tsx
View file

0
src/llama_stack/ui/components.json → src/llama_stack_ui/components.json
View file

0
src/llama_stack/ui/components/chat-completions/chat-completion-detail.test.tsx → src/llama_stack_ui/components/chat-completions/chat-completion-detail.test.tsx
View file

0
src/llama_stack/ui/components/chat-completions/chat-completion-detail.tsx → src/llama_stack_ui/components/chat-completions/chat-completion-detail.tsx
View file

0
src/llama_stack/ui/components/chat-completions/chat-completion-table.test.tsx → src/llama_stack_ui/components/chat-completions/chat-completion-table.test.tsx
View file

0
src/llama_stack/ui/components/chat-completions/chat-completions-table.tsx → src/llama_stack_ui/components/chat-completions/chat-completions-table.tsx
View file

0
src/llama_stack/ui/components/chat-completions/chat-messasge-item.tsx → src/llama_stack_ui/components/chat-completions/chat-messasge-item.tsx
View file

0
src/llama_stack/ui/components/chat-playground/chat-message.tsx → src/llama_stack_ui/components/chat-playground/chat-message.tsx
View file

0
src/llama_stack/ui/components/chat-playground/chat.tsx → src/llama_stack_ui/components/chat-playground/chat.tsx
View file

0
src/llama_stack/ui/components/chat-playground/conversations.test.tsx → src/llama_stack_ui/components/chat-playground/conversations.test.tsx
View file

0
src/llama_stack/ui/components/chat-playground/conversations.tsx → src/llama_stack_ui/components/chat-playground/conversations.tsx
View file

0
src/llama_stack/ui/components/chat-playground/interrupt-prompt.tsx → src/llama_stack_ui/components/chat-playground/interrupt-prompt.tsx
View file

0
src/llama_stack/ui/components/chat-playground/markdown-renderer.tsx → src/llama_stack_ui/components/chat-playground/markdown-renderer.tsx
View file

0
src/llama_stack/ui/components/chat-playground/message-components.tsx → src/llama_stack_ui/components/chat-playground/message-components.tsx
View file

0
src/llama_stack/ui/components/chat-playground/message-input.tsx → src/llama_stack_ui/components/chat-playground/message-input.tsx
View file

0
src/llama_stack/ui/components/chat-playground/message-list.tsx → src/llama_stack_ui/components/chat-playground/message-list.tsx
View file

0
src/llama_stack/ui/components/chat-playground/prompt-suggestions.tsx → src/llama_stack_ui/components/chat-playground/prompt-suggestions.tsx
View file

0
src/llama_stack/ui/components/chat-playground/typing-indicator.tsx → src/llama_stack_ui/components/chat-playground/typing-indicator.tsx
View file

0
src/llama_stack/ui/components/chat-playground/vector-db-creator.tsx → src/llama_stack_ui/components/chat-playground/vector-db-creator.tsx
View file

0
src/llama_stack/ui/components/layout/app-sidebar.tsx → src/llama_stack_ui/components/layout/app-sidebar.tsx
View file

0
src/llama_stack/ui/components/layout/detail-layout.tsx → src/llama_stack_ui/components/layout/detail-layout.tsx
View file

0
src/llama_stack/ui/components/layout/logs-layout.tsx → src/llama_stack_ui/components/layout/logs-layout.tsx
View file

0
src/llama_stack/ui/components/layout/page-breadcrumb.tsx → src/llama_stack_ui/components/layout/page-breadcrumb.tsx
View file

0
src/llama_stack/ui/components/logs/logs-table-scroll.test.tsx → src/llama_stack_ui/components/logs/logs-table-scroll.test.tsx
View file

0
src/llama_stack/ui/components/logs/logs-table.test.tsx → src/llama_stack_ui/components/logs/logs-table.test.tsx
View file

0
src/llama_stack/ui/components/logs/logs-table.tsx → src/llama_stack_ui/components/logs/logs-table.tsx
View file

0
src/llama_stack/ui/components/prompts/index.ts → src/llama_stack_ui/components/prompts/index.ts
View file

0
src/llama_stack/ui/components/prompts/prompt-editor.test.tsx → src/llama_stack_ui/components/prompts/prompt-editor.test.tsx
View file

0
src/llama_stack/ui/components/prompts/prompt-editor.tsx → src/llama_stack_ui/components/prompts/prompt-editor.tsx
View file

0
src/llama_stack/ui/components/prompts/prompt-list.test.tsx → src/llama_stack_ui/components/prompts/prompt-list.test.tsx
View file

0
src/llama_stack/ui/components/prompts/prompt-list.tsx → src/llama_stack_ui/components/prompts/prompt-list.tsx
View file

0
src/llama_stack/ui/components/prompts/prompt-management.test.tsx → src/llama_stack_ui/components/prompts/prompt-management.test.tsx
View file

0
src/llama_stack/ui/components/prompts/prompt-management.tsx → src/llama_stack_ui/components/prompts/prompt-management.tsx
View file

0
src/llama_stack/ui/components/prompts/types.ts → src/llama_stack_ui/components/prompts/types.ts
View file

0
src/llama_stack/ui/components/providers/session-provider.tsx → src/llama_stack_ui/components/providers/session-provider.tsx
View file

0
src/llama_stack/ui/components/responses/grouping/grouped-items-display.tsx → src/llama_stack_ui/components/responses/grouping/grouped-items-display.tsx
View file

Some files were not shown because too many files have changed in this diff Show more