In markhocky/testthatMockr: Provides Mock Object Support to 'testthat'

knitr::opts_chunk$set(
  collapse = TRUE,
  comment = "#>",
  fig.path = "README-"
)

Purpose

The underlying purpose of testthatMockr is to facilitate better unit testing. It does this by allowing you to use Mock objects in place of real ones, which can make test setup faster and easier.

Secondly it allows you to confirm certain actions have been taken in interacting with the mock object. You can check the unit you are testing has made a certain function call on the mock with defined parameters.

It is mainly designed for use with object oriented development, which may not suit everyone's needs.

Example

The real benefit of using mocks (in my opinion at least) is when you want to test some part of your code and keep it separate in the test environment from the rest of the complexity of your package.

Here's an actual example of usage from something I was playing with. I wanted to check my Broker object was talking to my Market, without actually having to build a Market object with all the associated data behind it.

First building a mock Market:

market <- Mock("Market")

Next I note the function 'getBar' is something I want to track, and I specify that I want it to return a specific value (but this is optional).

ticker <- "AMP.AX"
AMP.AX <- loadStocks(ticker)[[1]]
timestamp <- "2010-01-04"
mockMethod(market, "getBar", return.value = AMP.AX[timestamp])

Then we build and exercise the Broker:

broker <- Broker()
broker <- addMarket(broker, market)
bar <- getBar(broker, ticker, timestamp)

And we can then check that 'getBar' was called with specific parameters:

expect_that(market, called_once_with("getBar", ticker, timestamp))

Check out the vignette for more examples.

Installation

From your R console type:

install.packages("testthatMockr")

Further details

Check out the vignette for more examples and discussion.

browseVignettes("testthatMockr")

I've based this package functionality on the Python Mock module. It's might be worth having a look at that to get an idea of the type of functionality too.

Following is a brief discussion of the main components to testthatMockr:

Mock objects

Mock objects allow you to mimick the behaviour of other objects, and have control over certain behaviour as well as record some behaviour. They are often used in test driven development.

Mock methods

Once you've created a mock you can assign certain methods to be a 'mock' method. This means you have full control over the output, and can record how the method was used on the mock object.

Mock assertions

The other main component of the package are several new assertions to be used in testthat unit testing. These typically are used to assert that some function call was made on the mock object.

markhocky/testthatMockr documentation built on May 21, 2019, 12:07 p.m.