@tangle/complex-set

A strategy for tracking a set over time, when it matters when a certain thing was added to a Set, and when it was removed

Example Usage

const ComplexSet = require('@tangle/complex-set')

const input = {
  add: [
    { id: '@ben', seq: 25 },
    { id: '@cherese', seq: 2 }
  ]
}

strategy.mapFromInput(input, [strategy.identity()])
// => {
//   '@ben': { 25: 1 },
//   'cherese: { 2: 1 }
// }

The raw transformations look like this:

{
  [id]: {
    [seq]: state
  }
}

where

• id String - the thing being added / removed from the set
• seq Integer - some Integer which is a unique representation of when. (e.g. logical clock or clock wall time). Must be >= 0
• state Integer - an Integer representing whether it was an add or remove.
  • state > 0 - add
  • state <= 0 - remove

However there are convenience methods for making it easier to work with these.

API

ComplexSet(idPattern) => complexSet

Instantiates a strategy, complexSet.

• idPattern String (optional)
  • a pattern which is converted into a regexp validating the id field
  • e.g. '^@\w+$' would make only ids like @mixmix valid
  • default: '^.+$'

complexSet.isValid(T) => Boolean

complexSet.schema

Access the JSON shcema that isValid was built with.

complexSet.concat(A, B) => C

complexSet.identity() => I

returns "identity transformation"

complexSet.mapFromInput(input, currentTips) => T

Takes a current transformation state, currentTips, an array of Ts that are the tips of the graph, along with a human friendly (descriptive) input and returns a transformation T which satisfies the change requested in `input.

Format of input:

{
  add: [{ id: Id, seq: Integer }, ... ]
  remove: [{ id: Id, seq: Integer }, ... ]
}

where:

• id is a unique identifier for something being added.
  • Id must pass the idPattern validator
• seq is a "sequence", some whole number which represents a position in time
  • this could be a vector clock, or
  • a wall clock (UTC time)

You can provide EITHER add / remove OR both.

complexSet.mapToOutput(T) => t

Takes a transformation T and returns an output state t, which is more "human readable"

Format of output t:

{
  Id: [Interval],
}

A series of Intervals is computed for each Id from all the add and remove events, and these are ordered lowest to highest in terms of sequence.

Intervals will be of form:

• { start: Integer, end: Integer } - a "closed" interval
• { start: Integer, end: null }
  • an "open" interval
  • this will only ever occur as the last interval in a series

An Example transformation

When users were added / removed as permitted authors for an ssb-record.

// an example of a transformation
T = {
  // an example set where each @user was permitted at different intervals
  '@cherese': {
    200: 1, // seq=200, state=1=add
    1000: -1 // seq=1000, state=-1=remove
  },
  '@ben': {
    1000: 1, // seq=1000, state=1=add
    1100: 2, // seq=1100, state=2=add
    2000: -1 // seq=2000, state=-1=remove
  },
  '@mix': {
    300: -1 // seq=300, state=-1=remove
  }
}

// the reified state of that transformation
mapToOutput(T) = {
  '@cherese': [{ start: 200, end: 1000 }],
  '@ben': [{ start: 1000, end: 2000 }], // note how 1100 wasnt added
  // note @mix is not here
  // ids with no "add" states are ignored
}

simpleSet.isConflict() => False

simpleSet.isValidMerge() => True

simpleSet.merge(graph, mergeNode, field) => T

where:

• graph is a @tangle/graph instance
• mergeNode is the proposed merge-node
• field String contains the the data fields node.data[field]