Milvus Retriever¶

Use MilvusRetriever to evaluate Milvus search results.

Install¶

pip install "evret[milvus]"

Basic Usage¶

from evret.retrievers import MilvusRetriever

def encode_query(query: str) -> list[float]:
    return embedding_model.embed_query(query)

retriever = MilvusRetriever(
    collection_name="docs",
    query_encoder=encode_query,
    uri="http://localhost:19530",
    id_field="doc_id",
    anns_field="embedding",
    output_fields=["doc_id", "text", "source"],
    search_params={"metric_type": "COSINE", "params": {"nprobe": 10}},
)

results = retriever.retrieve("difference between mrr and ndcg", k=5)
for item in results:
    print(item.doc_id, item.score, item.metadata)

With Auth¶

retriever = MilvusRetriever(
    collection_name="docs",
    query_encoder=encode_query,
    uri="http://localhost:19530",
    token="username:password",
)

Notes¶

query_encoder must return a non-empty vector
You can pass search_filter for server-side filtering
Milvus distance is mapped to score using distance_to_score (identity by default)