TidierIteration.jl

What is TidierIteration.jl?

TidierIteration.jl is a 100% Julia implementation of the purrr R package. It aims to ease the different ways to map (or apply) functions to Julia objects: arrays, dicts and so on. It also provides some tools of functional programming: adverbs, composition, safe-functions and more.

Installation

For the stable version:

] add https://github.com/TidierOrg/TidierIteration

The ] character starts the Julia package manager. Press the backspace key to return to the Julia prompt.

or

using Pkg
Pkg.add(url="https://github.com/TidierOrg/TidierIteration")

Comparison to purrr

In R purrr package, there are many ways to apply a function to each element of a set: the many map_ variants (map_lgl returns a vector of logicals, map_chr returns a vector of strings/characters, etc) and so on exist because the map function returns a list in R, which is a wild beast (R lists allow any kind of elements, and few functions are optimized (i.e. vectorized) on them as opposed to vectors). In comparison, R vectors are always of just one type (integers, booleans, etc.)

In Julia, arrays are already the most general structure in the sense that they can contain any kind of element. Arrays of just one type (say, integers) are still arrays. We don’t need the _lgl and _chr variants.

Besides that, the name map is already taken in Julia and its arguments are fatally in the opposite order of the purrr map! Because of that, I decided to implement the apply function, which takes an iterable object x as the first argument and a function f as the second. So apply(x, f) means that for each element of x I want to apply the function f and return an array.

So, instead of defining a function like map_int in Julia, you can just type

apply(x, Int ∘ f)

or, even better, let f return an integer.

Base Julia mapping

In Julia we already have 3 ways to apply a function to each element of an array x, and they are very elegant ones:

map(f, x)

or

[f(xi) for xi ∈ x]

or the one I prefer:

f.(x)

Why use TidierIteration then?

• The object x is always the first element of the apply family, so it’s easy to pipe them;

• We extend the apply family to Julia objects which are not mapped by default, like dictionaries;

• We also provide the apply2, iapply and papply methods to map over 2 or n elements at the same time;

• We provide the flatten_* functions to tidy wild dictionaries (like JSON responses from APIs) and many adverbs.

Complete list of implemented functions

See the documentation for examples.

Apply

apply, apply_keys, apply_dfc, apply_dfr, walk, apply2, walk2, iapply, iwalk, papply, pwalk;

Modify

modify, modify!, modify_if, modify_if!, keep, keep!, keep_keys, discard, discard!, compact, compact!;

Predicates

is_empty, is_non_empty, every, some, none, detect_index, detect, has_element, has_key, get_value;

Adverbs

compose, compose_n, negate, possibly;

Flatten

flatten, flatten_n, flatten_dfr, flatten_json, flatten_dfr_json, json_string, to_json;