salmon.backend.sampler.Sampler

class salmon.backend.sampler.Sampler(ident: str = '')

Run a sampling algorithm. Provides hooks to connect with the database and the Dask cluster.

Parameters

identstr

The algorithm idenfifier. This value is used to identify the algorithm in the database.

Attributes

clear

Should the queries be cleared from the database?

Methods

clear_queries(rj)	Clear all queries that this sampler has posted from the database.
get_answers(rj[, clear])	Get all answers the frontend has received.
get_model()	Get the model underlying the algorithm.
post_queries(queries, scores[, rj, done])	Post scored queries to the database.
process_answers(answers)	Process answers.
redis_client([decode_responses])	Get the database (/Redis client)
reset(client, rj[, futures])	Stop the algorithm.
run(client)	Run the algorithm.
save()	Save the sampler's state and current embedding to the database.
serialize_query(q)	Serialize a query (so it can go in the database).

property clear

Should the queries be cleared from the database?

clear_queries(rj: RedisClient) → bool

Clear all queries that this sampler has posted from the database.

get_answers(rj: RedisClient, clear: bool = True) → List[Answer]

Get all answers the frontend has received.

get_model() → Dict[str, Any]

Get the model underlying the algorithm.

Returns

stateDict[str, Any]

The state of the algorithm. This can be used for display on the dashboard or with an HTTP get request.

post_queries(queries: List[Query], scores: List[float], rj: Optional[RedisClient] = None, done=None) → int

Post scored queries to the database.

Parameters

queriesList[Query]

Queries to post to the database

scoresList[float]

The scores for each query

rjRedisClient, optional

The databaase

Returns

n_queriesint

The number of queries posted to the database.

process_answers(answers: List[Answer])

Process answers.

Parameters

answersList[Answers]

Each answer is a dictionary. Each answer certainly has the keys “head”, “left”, “right” and “winner”, and may have the key “puid” for participant UID.

Returns

datadict

An update to self.__dict__.

redis_client(decode_responses=True) → RedisClient

Get the database (/Redis client)

reset(client, rj, futures=None)

Stop the algorithm. The algorithm will be deleted shortly after this function is called.

run(client: DaskClient)

Run the algorithm.

Parameters

clientDaskClient

A client to Dask.

rjRedisClient

A Redist Client, a rejson.Client

Notes

This function runs the adaptive algorithm. Because it’s asynchronous, this function should return if "reset" in rj.keys() and rj.jsonget("reset").

save() → bool

Save the sampler’s state and current embedding to the database.

serialize_query(q: Query) → str

Serialize a query (so it can go in the database).