DistributionRange

class DistributionRange(n, dist, ...params)

Wraps a distributing function into a RangeObject-like API, so a statistical distribution (e.g. normal, poisson, binomial) can be dropped directly into a job’s input set, or used as one of the ranges inside a MultiRangeObject.

DistributionRange objects are normally created via the factory functions in the set export of dcp/stats-ranges, e.g. stats.set.normal(10, 0, 1), rather than by calling this constructor directly.

Arguments:

• n (number) – Size of the generated set.

• dist (string) – Name of a distributing function exported by dcp/stats-ranges’s distributor object (e.g. 'normal', 'poisson', 'binomial').

• params – Additional parameters passed through to the named distributing function, following its k and n parameters.

Properties

length

DistributionRange.length

Type:

number

Size of the generated set, as passed to the constructor.

dist

DistributionRange.dist

Type:

string

Name of the distributing function backing this object.

params

DistributionRange.params

Type:

Array

Additional parameters passed to the distributing function.

Methods

filter

See SuperRangeObject.filter()

slice

See SuperRangeObject.slice()

nthValue

nthValue(n)

Returns the nth value of the distribution, by calling the underlying distributing function with (n, length, …params). If the distribution has already been resolved (see DistributionRange.materialize()), returns the nth value of the resolved array instead.

Arguments:

• n (number) – Index to look up.

Return type:

number

materialize

DistributionRange.materialize(now=true)

Resolve the distribution to a static array

Arguments:

• now – If false, then set a flag to materialize on the scheduler. Default: materialize now

toObject

toObject()

Resolves the distribution (see DistributionRange.materialize()) and returns a plain object with length, dist, params, and the resolved values, suitable for reconstructing an equivalent DistributionRange later.

Return type:

object

toString

toString()

Returns a descriptor string, e.g. [object DistributionRange normal(10,0,1)], which can be passed back into the DistributionRange constructor to reconstruct an equivalent object.

Return type:

string

Example

const stats = require('dcp/stats-ranges');

const job = compute.for(stats.set.normal(10, 0, 1), function (x) {
  progress(1);
  return x;
});