How the WL algorithm works

Wadler–Lindig pretty printing happens in two steps.

First, we represent the object as a nested collection of Wadler–Lindig documents, corresponding to unbroken text, places that newlines can be inserted, etc. This describes the pretty representation, but for which no maximum width has yet been set.

Second, once provided a width, this nested structure of documents is then passed to the Wadler–Lindig algorithm, which lays out the text within the space constraints.

There are five kinds of Wadler–Lindig document:

• wadler_lindig.BreakDoc: either inserts a newline or the specified text.
• wadler_lindig.ConcatDoc: concatenates multiple documents together.
• wadler_lindig.GroupDoc: all elements in the group will either be laid out horizontally together, or vertically together.
• wadler_lindig.NestDoc: if laying out vertically, then all child BreakDocs will have this much indent added after their newlines.
• wadler_lindig.TextDoc: an unbroken string of text.

As such, the wadler_lindig library actually does two distinct things:

1. It converts Python objects into Wadler–Lindig documents.
2. It runs the Wadler–Lindig algorithm for formatting Wadler–Lindig documents.

When we run wadler_lindig.pprint or wadler_lindig.pformat, then both of these happen back-to-back.

Now that we know this, we can customise how types are displayed by specifying how to turn custom Python objects into Wadler–Lindig documents. This is the subject of our next example.