@tangle/reduce

Takes in a collection of tangle nodes which contain operational transforms, and reduces them into some final transform(s).

The reduce starts with root node(s) and traverses the tangle graphs, successively concatenating the transforms in each node until the tip(s) of the graph are reached, where we return the final state(s)

Example Usage

//     A       (root node)
//    / \                                       |
//   B   C     (two concurrent nodes)           |  causality
//    \ /                                       v
//     D       (leading tip)

const nodes = [
  {
    key: 'A',
    previous: null,
    data: {
      title: { set: 'spec gathering' },
      attendees: { mix: 1, luandro: 1 }
    }
  },
  {
    key: 'B',
    previous: ['A'],
    data: {
      attendees: { luandro: -1, cherese: 1 }
    }
  },
  {
    key: 'C',
    previous: ['A'],
    data: {
      attendees: { luandro: -1 }
    }
  },
  {
    key: 'D',
    previous: ['B', 'C'],
    data: {
      title: { set: 'Tangle Spec v1 meeting' }
    }
  },
]

const Reduce = require('@tangle/reduce')
const Strategy = require('@tangle/strategy')

const strategy = new Strategy({
  title: require('@tangle/simple-set')(),
  attendees: require('@tangle/overwrite')()
})

const result = new Reduce(strategy, { nodes })
// => {
//   D: {
//     title: { set: 'Tangle Spec v1 meeting' },
//     attendees: { mix: 1. luandro: -2, cherese: 1 }
//   }
// }

Here the only key in our result is D, as there's only one leading tip. If we reduced [A, B, C] there would be two tips [B, C] and an accumulated transform state for each.

Note you could use strategy.mapToOutput to convert this accumulated transformation into a 'real' state:

strategy.mapToOutput(result['D'])
// => {
//   title: 'Tangle Spec v1 meeting',
//   attendees: ['cherese', 'mix']
// }

API

new Reduce(strategy, opts) => reduce

Instantiates a new reduce helper

strategy Object defines how to reduce nodes. Produced by @tangle/strategy (version >=2)

opts Object (optional) is an object which lets you to customise the reduce:

• opts.nodes Array is a collection of tangle nodes, where each expected to include:

  • node.key - a unique identifier for the node
  • node.previous - an Array of keys for nodes which this node extended from
  • node.data - an object which contains the transformation which the strategy describes

• opts.isValidNextStep = fn(context, node): Boolean

  • determine whether the next node node should be included in the tangle
  • context is the context in the tangle immediately before node, and is an object { graph, tips }
    • graph is an instance of @tangle/graph, which has a bunch of helper functions for querying the tangle
    • tips which described the accumulated transform right before node as an object { key, T }
  • NOTE if a node is not valid, all nodes downstream of that node and considered invalid, and will not be included in the reduce

reduce.resolve() => resolvedState

A getter which accesses the current resolved state for the

alias: reduce.state (a getter)

reduce.addNodes(nodes)

Register more nodes to include in the reducing.

reduce.graph => Graph

The instance of [@tangle/graph](https://gitlab.com/tangle-js/tangle-graph for API) which may be useful for querying the tangle graph that the resolved state came from.

NOTES:

• ⚠️ ensure you call reduce.resolve() before lookin at this graph, as this does some work to validate/ invalidate nodes, and identify disconnected nodes.
• ⚠️ NEVER manipulate the underlying graph, only query (manipulating it will risk putting it out of sync with @tangle/reduce logic)