Local mode accepts relevance-feedback queries with an empty feedback list

[Harsh23Kashyap]

Problem

An empty relevance-feedback list is rejected by Qdrant server validation, but local mode accepts the same query and runs a normal target-vector search instead.

This makes a query behave differently when an application switches between QdrantClient(":memory:") and a server-backed client.

Reproduction

from qdrant_client import QdrantClient, models

client = QdrantClient(":memory:")
client.create_collection(
    "demo",
    vectors_config=models.VectorParams(size=2, distance=models.Distance.DOT),
)
client.upsert(
    "demo",
    [
        models.PointStruct(id=1, vector=[1.0, 0.0]),
        models.PointStruct(id=2, vector=[0.0, 1.0]),
    ],
)

query = models.RelevanceFeedbackQuery(
    relevance_feedback=models.RelevanceFeedbackInput(
        target=[1.0, 0.0],
        feedback=[],
        strategy=models.NaiveFeedbackStrategy(
            naive=models.NaiveFeedbackStrategyParams(a=0.5, b=1.0, c=0.7)
        ),
    )
)

print(client.query_points("demo", query=query).points)

On current master (cf747f4), local mode returns:

[ScoredPoint(id=1, score=0.5, ...), ScoredPoint(id=2, score=0.0, ...)]

The server rejects the request with feedback elements must be non-empty.

I reproduced the local result twice from a clean checkout. The mismatch comes from local query validation checking the naive strategy's b bound but not the server's non-empty feedback rule.

Expected behavior

Local mode should reject an empty feedback list before executing the query, matching the server.