Pipelines

Mix.install([
  {:kino, "~> 0.12.3"}
])

Introduction

Pipelining is a pattern whereby a series of functions are called, and the result of any one function is used as the first input to the following function.

Example:

name      = input("your name?")
resultset = lookup(name, database)
processed = process(resultset)
outcome   = report(processed)

Pipe Operator

This pattern is so common that Elixir – like many other functional languages – has introducted a pipe operator |> that allows the above snippet to be rewritten as:

outcome =
  input("your name?")
  |> lookup(database)
  |> process()
  |> report()

Note: If your editor is using a font with ligatures, the pipe operator will be rendered as ▷.

Example

Sample data:

data = """
evening Monday
morning Friday
lunch tUesday
night monday

morning thusday


"""

Mapping from day abbreviation to day index:

days = %{"mon" => 0, "tue" => 1, "wed" => 2, "thu" => 3, "fri" => 4, "sat" => 5, "sun" => 6}

Lets extract the set of date indices that are mentioned in the sample data:

mentioned_days =
  data
  |> String.trim_trailing("\n")
  |> String.downcase()
  |> String.split("\n")
  |> Enum.filter(fn entry -> entry != "" end)
  |> Enum.map(fn entry ->
    entry
    |> String.split(" ")
    |> Enum.at(1)
    |> String.slice(0, 3)
    |> (fn abbrev -> Map.get(days, abbrev) end).()
  end)
  |> MapSet.new()
  |> dbg()

Here, the last step of the pipeline (the call to dbg()) will – when the Kino module is loaded – provide a number of extra features:

• Firstly, it does not affect what the expression is evaluating to.
• Each step of the pipeline is listed between the code cell and its result.
• Visual controls are added to the right of each pipeline step. These can be used to toggle the step.
• When clicking on a pipeline step, the output of that step can be inspected.