tff.analytics.IbltFactory

Factory for string and values aggregation by IBLT.

Inherits From: UnweightedAggregationFactory

tff.analytics.IbltFactory(
    *,
    capacity: int,
    string_max_bytes: int,
    encoding: tff.analytics.heavy_hitters.iblt.CharacterEncoding = tff.analytics.heavy_hitters.iblt.CharacterEncoding.UTF8,
    repetitions: int,
    seed: int = 0,
    sketch_agg_factory: Optional[tff.aggregators.UnweightedAggregationFactory] = None,
    value_tensor_agg_factory: Optional[tff.aggregators.UnweightedAggregationFactory] = None
) -> None

capacity The capacity of the IBLT sketch. Must be positive.
string_max_bytes The maximum length in bytes of a string in the IBLT. Must be positive.
encoding The character encoding of the string data to encode. For non-character binary data or strings with unknown encoding, specify CharacterEncoding.UNKNOWN.
repetitions The number of repetitions in IBLT data structure (must be >= 3). Must be at least 3.
seed An integer seed for hash functions. Defaults to 0.
sketch_agg_factory (Optional) A UnweightedAggregationFactory specifying the value aggregation to sum IBLT sketches. Defaults to tff.aggregators.SumFactory. If sketch_agg_factory is set to a tff.aggregators.SecureSumFactory, then the upper_bound_threshold should be at least 2 ** 32 - 1.
value_tensor_agg_factory (Optional) A UnweightedAggregationFactory specifying the value aggregation to sum value tensors. Defaults to tff.aggregators.SumFactory. Note that when using sketch_agg_factory is set to a tff.aggregators.SecureSumFactory, the value to be summed might be clipped depends on the choices of upper_bound_threshold and lower_bound_threshold parameters in SecureSumFactory.

ValueError if parameters don't meet expectations.

Methods

create

View source

create(
    value_type: tff.types.SequenceType
) -> tff.templates.AggregationProcess

Creates an AggregationProcess using IBLT to aggregate strings.

Args
value_type A tff.SequenceType representing the type of the input dataset, must be compatible with the following tff.Type: tff.SequenceType(collections.OrderedDict([ (DATASETKEY, np.str), (DATASET_VALUE, tff.TensorType(shape=[None], dtype=np.int64)), ]))

Raises
ValueError If value_type is not as expected.

Returns
A tff.templates.AggregationProcess to aggregate strings and values associate with the strings.