use nosql style map reduce functions to iterate over data in node.
1 function:
var mapR = require('map-reduce')
mapR({
on: collection to iterate over (can be an object or an Array)
map: iteration function (optional, defaults to mapR.identity
reduce: reduction function (optional, defaults to mapR.collect)
init: initial value for collection arg in reduce step, default to undefined
done: callback function
})
##map (emit, value, key) ##
calling emit(value[,key,etc...]) will send args to the reduce function.
map-reduce is async,
call emit.next() to go to the next item.
or, emit.done() to stop processing more items.
if the values to be emitted are passed into emit.next or emit.done that is like calling emit(args); emit.next(), or emit(args); emit.done()
##reduce (collection, value[, key,etc...]) ##
reduce will be called with a collection which is either,
- the return value of the previous reduce call
- the .init value you supplied to mapR
- undefined
the other arguments are the values you called emit with.
##TODO: traverse, heuristic, prune
expand this module to traverse graphs by adding a emit.traverse
calling emit.traverse(obj) will add obj to the queue of objects being mapped.
depending on some configuration this could be used to perform both depth-first,
and width-first traversals, and with a heuristic function, A* could be implemented.
##TODO: support async function as value for on:
for large mappings it will be necessary to partition and parallelize, will need to think up an effective way to partition the source.