noflo-wrapper
NoFlo component/graph testing and embedding wrapper compatible with BDD/TDD/whatever
Last updated 3 years ago by trustmaster .
MIT · Repository · Bugs · Original npm · Tarball · package.json
$ cnpm install noflo-wrapper 
SYNC missed versions from official npm registry.

NoFlo component/graph testing and embedding wrapper

Build Status

Wraps a component to provide a convenient interface compatible with any testing paradigm: TDD/BDD/whatever. Also useful to embed NoFlo graphs into existing applications.

Benefits

  • Reduces boilerplate to set up a component testbed.
  • Provides common high-level methods.
  • Provides low-level access to the component, ports and events.
  • Compatible with different testing frameworks and complex test cases.

Getting started

Install noflo-wrapper and add it to your project's dev dependecies:

npm install --save-dev noflo-wrapper

Require it in your specs/tests:

Wrapper = require 'noflo-wrapper'

Use methods described below and run the tests just as you do it normally with your favorite testing framework.

API

Explanations below contain examples in CoffeeScript using Mocha and Chai in BDD style. You can also write your tests in JavaScript, using any other framework or style.

Loading a component

First you need to create a new Wrapper object to wrap your component or graph:

t = new Wrapper 'my-noflo-app/Multiplier'

The constructor accepts either a full component name (including namespace prefix), or an already instantiated component object, or a function returning such an object.

In general, components are loaded and wired up asynchronously, so you need to start the wrapper like this before running any tests:

before (done) ->
  t.start (err, instance) ->
    return done err if err # Error handling, optional
    # instance contains a ready to use component
    done()

Advanced options

If the component to be tested is a NoFlo graph, you can pass custom event handlers to the Wrapper constructor:

t = new Wrapper 'my-noflo-app/Multiplier',
  load: (err, instance) ->
    # This is call after loading the graph
  ready: (err, instance) ->
    # This is called when the network is ready to be attached

Sending inputs and expecting output

A high-level receive method listens on output ports for data and groups until a disconnect event.

A high-level send methods sends data followed by a disconnect to one or more input ports.

Here is an example that tests a simple multiplier component:

t.receive 'xy', (data) ->
  chai.expect(data).to.equal 30
  done()

t.send
  x: 5
  y: 6

Note that receive is called before send, because it binds event handlers asynchronously, while send is almost an instant operation.

Short syntax for send method to send data and disconnect to just one inport looks like this:

t.send 'x', 123

Direct access to component, ports and events

In more complex test cases you might want to send IPs and handle particular events manually:

t.outs.xy.on 'data', (data) ->
  chai.expect(data).to.equal 24
  done()

t.ins.x.send 8
t.ins.x.disconnect()
t.ins.y.send 3
t.ins.y.disconnect()

Wrapper object provides ins and outs hashmaps of sockets attached to the component.

You can also access the component directly via c property:

if t.c.outPorts.error.isAttached()
  # Do something

Receiving multiple data chunks and groups

As receive is triggered by a disconnect event, there might be multiple data packets in the transmission and also some group bracket IPs. In such case they are available as arrays and counts in the callback arguments:

t.receive 'xy', (data, groups, dataCount, groupCount) ->
  chai.expect(data).to.eql [4, 10, 18]
  chai.expect(dataCount).to.equal 3
  chai.expect(groups).to.eql ['foo', 'bar']
  chai.expect(groupCount).to.equal 2
  done()

Note that groupCount counts only closed groups via endGroup events, while groups contains unique groups sent to the output.

Receiving from multiple output ports

If a component sends output to multiple ports at the same time and you need to test results from all of them at once, that may require some syncrhonization spaghetti in your specs. But receive simplifies it by accepting a hashmap and returning a Promise that is resolved when results from all outputs in the map have been received:

div = null
mod = null

t.receive
  quotient: (data) ->
    div = data
  remainder: (data) ->
    mod = data
.then ->
  chai.expect(div).to.equal 3
  chai.expect(mod).to.equal 2
  done()

t.send
  dividend: 11
  divisor: 3

Using promises to chain subsequent receives

The receive method returns a Promise resolved when a transmission is received, so you can chain subsequent transmissions in a thenable way, e.g.:

t.receive 'quotient', (data) ->
  chai.expect(data).to.equal 5
.then ->
  t.receive 'quotient', (data) ->
    chai.expect(data).to.equal 8
    done()
  t.send
    dividend: 56
    divisor: 7
t.send
  dividend: 30
  divisor: 6

Examples

See complete BDD-style examples in spec folder.

Development

The first thing to start developing this package is:

npm install

Then run bundled Mocha specs:

npm test

Then feel free to hack on the lib and specs.

Current Tags

  • 0.2.1                                ...           latest (3 years ago)

3 Versions

  • 0.2.1                                ...           3 years ago
  • 0.2.0                                ...           3 years ago
  • 0.1.2                                ...           3 years ago
Maintainers (1)
Downloads
Today 0
This Week 0
This Month 1
Last Day 0
Last Week 1
Last Month 3
Dependencies (5)
Dev Dependencies (3)
Dependents (0)
None

Copyright 2014 - 2016 © taobao.org |