afs - array-functional-streams

Allows arrays to be turned into a stream of objects and then run asynchronous map, filter, and reduce operations against them.

The module exposes methods to create three types of stream:

• Reader: read stream that turns an array into a stream of objects;
• Transformer: transform stream that can perform map or filter like operations against a stream of objects;
• Reducer: transform stream the turns a stream of objects into one object.

API

createReader(array)

This is the function used to transform an array into a stream. The function returns a read stream that operates in object mode. Each read of the stream will return one object from the original array.

createTransformer(cb[, stringify])

This returns a transform stream that calls cb(chunk) for each object in the stream. You can simulate map operations by having the callback function transform the object in some way before returning it and you can simulate filter operations by returning null or undefined for any object that you don't want to be included in the output. stringify is an optional boolean value which, if set to true, causes each object to be parsed with JSON.stringify before being added to the output. This allows the stream to be piped directly to other streams that only support strings. This parameter defaults to false.

Differences from Array.prototype.map / Array.prototype.filter:

• You cannot add null elements to a stream without aborting. If your callback returns null or undefined no output is produced for that value.
• The callback only receives an object. No array index is supplied, and obviously the original array cannot be supplied.
• Map and filter do not alter the source array, which obviously makes no sense when applied to a stream. (Although, do note that the source array used to create the reader remains unaffected.)

createReducer(cb[, initValue, stringify])

This returns a transform stream that performs the operation acc = cb(chunk, acc) for each object in the input. Once all objects have been processed, acc is written to the output. If initValue is not null or undefined then acc is initialised to initValue, otherwise it is initialised to null. When an object is read from the input, if acc is null then it is initialised to be the object, otherwise the above operation is performed.

Example

There are a few examples in the samples folder. The following is the first.

var list = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10],
	filterFunc = function(val) { if (val < 6) return val; },
	reducerFunc = function(val, acc) { acc.push(val); return acc; },
	ArrayStream = require("../index.js"),
	reader = ArrayStream.createReader(list),
	filter = ArrayStream.createTransformer(filterFunc),
	// Note that initialValue is an empty array. This array is repeatedly passed to the 
	// reducerFunc which pushes the object to it and then returns it.
	reducer = ArrayStream.createReducer(reducerFunc, [], true);

reader.pipe(filter).pipe(reducer).pipe(process.stdout);	

Testing

Testing requires mocha to be installed globally and chai to be installed locally. Use the following commands to set up for testing:

• npm install -g mocha
• npm install chai

Testing can be run using the standard command npm test.