Smritea search methods and rerankers: which to pick

A search method is the strategy Smritea uses to find memories. A reranker is a step that reorders the memories it found. You can choose both on each search call.

The three search methods

Pass method to choose one. If you leave it out, Smritea uses your app's storage engine.

If you omit method, Smritea charges the deep_search cost, even on a Quick Search app. Pass method when you want to control the cost.

The six rerankers

A reranker applies to deep_search only. Pass reranker_type to choose one. The default is cross_encoder.

Compare the methods on one question

import os

from smritea import MemoryScope, SmriteaClient, SmriteaError

client = SmriteaClient(
    api_key=os.environ["SMRITEA_API_KEY"],
    app_id=os.environ["SMRITEA_APP_ID"],
)
alice = MemoryScope(actor_id="alice", actor_type="user")

try:
    for method in ("quick_search", "deep_search"):
        results = client.search(
            "What does Alice eat?",
            scope=alice,
            method=method,
            limit=5,
        )
        print(method, [round(r.score, 2) for r in results])
except SmriteaError as e:
    print(e.http_status, e.code, e.request_id)

Compare scores only within one call. A cross_encoder score and an rrf score are different kinds of numbers.

Limits

The table gives the default and the cap for two search settings. It also says what happens above the cap.

The server rejects a value above the cap. It does not cut the value down.

Learn more

The docs page Search methods and rerankers has the full reference. A Quick Search app stores text without extracting facts or entities, so the deep methods suit a Deep Search app. Read How Smritea stores memory for the difference.