72ef18e39d
Every section inherited plain BaseModel, so unknown keys were dropped silently: providers.docling_serve.timeout was documented for months while being ignored, and a typo in any setting took the default. Sections now derive from ConfigModel, which forbids extras, so a stale or misspelled key fails with its path. This already found search.context_radius in a live app config and providers.vllm in soliplex's example. converter, chunker and chunker_type are Literals. Sizes, limits, dimensions, token budgets, attempt counts and breaker thresholds must be positive; retention, delays, intervals and cooldowns non-negative; similarity_threshold within 0-1; port within 0-65535. port 0 keeps its OS-assigned meaning and worker_count allows 0 for an API-and-reaper-only process. get_reranker caught ImportError and returned None, so a configured reranker whose extra was missing silently disappeared. It now propagates. raise_missing_extra names the install command and re-raises when the failure came from inside an installed package, so a broken transitive import is not reported as a missing one. zeroentropy imported bare and now guards like the others. The haiku.rag package declares the jina extra. jina-local already worked there through cross-encoder's transitive transformers and torch; the resolved package set is unchanged, but the support is now promised rather than inherited. Provider fields stay unconstrained: get_model ends in a pass-through to pydantic-ai for any provider it supports, so a Literal there would reject valid configurations.
37 lines
1.2 KiB
Python
37 lines
1.2 KiB
Python
import asyncio
|
|
|
|
from haiku.rag.utils import raise_missing_extra
|
|
|
|
try:
|
|
from transformers import AutoModel
|
|
except ModuleNotFoundError as e: # pragma: no cover
|
|
if e.name not in ("torch", "transformers"):
|
|
raise
|
|
raise_missing_extra(e.name, "jina", e)
|
|
|
|
from haiku.rag.reranking.base import RerankerBase
|
|
from haiku.rag.store.models.chunk import Chunk
|
|
|
|
|
|
class JinaLocalReranker(RerankerBase): # pragma: no cover
|
|
"""Jina reranker using local model inference via transformers.
|
|
|
|
Note: The Jina Reranker v3 model is licensed under CC BY-NC 4.0,
|
|
which restricts commercial use.
|
|
"""
|
|
|
|
def __init__(self, model: str = "jinaai/jina-reranker-v3"):
|
|
self._model = model
|
|
self._reranker = AutoModel.from_pretrained(model, trust_remote_code=True)
|
|
self._reranker.eval()
|
|
|
|
async def _rerank(
|
|
self, query: str, chunks: list[Chunk], top_n: int = 10
|
|
) -> list[tuple[Chunk, float]]:
|
|
documents = [chunk.content for chunk in chunks]
|
|
|
|
results = await asyncio.to_thread(
|
|
lambda: self._reranker.rerank(query, documents, top_n=top_n)
|
|
)
|
|
|
|
return [(chunks[r["index"]], float(r["relevance_score"])) for r in results]
|