phoenix-oss/llama-stack-mirror
0
0
Fork 1
mirror of https://github.com/meta-llama/llama-stack.git synced 2025-06-27 18:50:41 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

feat: support filters in file search (#2472)

Browse source 
# What does this PR do?
Move to use vector_stores.search for file search tool in Responses,
which supports filters.

closes #2435 

## Test Plan
Added e2e test with fitlers.
myenv ❯ llama stack run llama_stack/templates/fireworks/run.yaml

pytest -sv tests/verifications/openai_api/test_responses.py \
  -k 'file_search and filters' \
  --base-url=http://localhost:8321/v1/openai/v1 \
  --model=meta-llama/Llama-3.3-70B-Instruct
This commit is contained in:
ehhuang 2025-06-18 21:50:55 -07:00 • committed by GitHub
parent fd37a50e6a
commit db2cd9e8f3
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
13 changed files with 449 additions and 63 deletions
Download patch file Download diff file Expand all files Collapse all files

6
llama_stack/apis/agents/openai_responses.py
View file

@ -9,6 +9,7 @@ from typing import Annotated, Any, Literal
from pydantic import BaseModel, Field
from typing_extensions import TypedDict
from llama_stack.apis.vector_io import SearchRankingOptions as FileSearchRankingOptions
from llama_stack.schema_utils import json_schema_type, register_schema
# NOTE(ashwin): this file is literally a copy of the OpenAI responses API schema. We should probably
@ -400,11 +401,6 @@ class OpenAIResponseInputToolFunction(BaseModel):
strict: bool | None = None
class FileSearchRankingOptions(BaseModel):
ranker: str | None = None
score_threshold: float | None = Field(default=0.0, ge=0.0, le=1.0)
@json_schema_type
class OpenAIResponseInputToolFileSearch(BaseModel):
type: Literal["file_search"] = "file_search"

7
llama_stack/apis/vector_io/vector_io.py
View file

@ -157,6 +157,11 @@ VectorStoreChunkingStrategy = Annotated[
register_schema(VectorStoreChunkingStrategy, name="VectorStoreChunkingStrategy")
class SearchRankingOptions(BaseModel):
ranker: str | None = None
score_threshold: float | None = Field(default=0.0, ge=0.0, le=1.0)
@json_schema_type
class VectorStoreFileLastError(BaseModel):
code: Literal["server_error"] | Literal["rate_limit_exceeded"]
@ -319,7 +324,7 @@ class VectorIO(Protocol):
query: str | list[str],
filters: dict[str, Any] | None = None,
max_num_results: int | None = 10,
ranking_options: dict[str, Any] | None = None,
ranking_options: SearchRankingOptions | None = None,
rewrite_query: bool | None = False,
) -> VectorStoreSearchResponsePage:
"""Search for chunks in a vector store.