/afpj

Applied Functional Programming in Julia - Tools to simplify iterative processes. Based on afp (https://github.com/robertschnitman/afp).

Primary LanguageJuliaGNU General Public License v3.0GPL-3.0

Robert Schnitman
2018-07-16
Recommended Citation:
       Schnitman, Robert (2018). afpj v0.0.0.1. https://github.com/robertschnitman/afpj

Outline

  1. Introduction
  2. mapcat()
  3. mrchop()
  4. mop()
  5. Conclusion
  6. References
  7. See also

0. Installation

    # Julia >= 0.6.2
    Pkg.clone("https://github.com/robertschnitman/afpj.git")

1. Introduction

The afpj package--Applied Functional Programming in Julia--is based on the original R library, afp. Some functions in this library are direct translations, while others cover gaps in Julia functionality. For example, mapcat() is the equivalent of do.bind and mrchop() replicates the process of its R counterpart, whereas mop() attempts to simulate its R equivalent that simplifies sweep() from the same programming language (an attempted replication of the latter function is also in this package).

Thus, the purpose of afpj is to supplement base Julia and its libraries to support efficient and concise programming.

The following sections provide examples for the primary functions: mapcat(), mrchop(), and mop(). Tables were made with the CSV to Markdown Table Generator tool (Donat Studios).

2. mapcat()

After mapping a function to an array of arrays, one may wish to concatenate the results with [v/h]cat(). To streamline this procedure, mapcat() takes the function, array of arrays, and dimension into a single call--the associated parameters are f, a, and d (1 for row-wise concatenation, 2 for column-wise) respectively.

EXAMPLE

 A = [[1:5;] [6:10;] [11:15;]] # Create 3 arrays (A, B, C).
 B = [[16:20;] [21:25;] [26:30;]]
 C = [[31:35;] [36:40;] [41:45;]]
 ABC = [A, B, C] # 3-element Array of Arrays
 mapcat(x -> x.^2, ABC, 1) # Row-wise concatenation (default).
1 36 121
4 49 144
9 64 169
16 81 196
25 100 225
256 441 676
289 484 729
324 529 784
361 576 841
400 625 900
961 1296 1681
1024 1369 1764
1089 1444 1849
1156 1521 1936
1225 1600 2025 
 mapcat(x -> x.^2, ABC, 2) # Column-wise concatenation of the same.
1 36 121 256 441 676 961 1296 1681
4 49 144 289 484 729 1024 1369 1764
9 64 169 324 529 784 1089 1444 1849
16 81 196 361 576 841 1156 1521 1936
25 100 225 400 625 900 1225 1600 2025

3. mrchop()

As inspired by mapslices(), mrchop() acts as a dimension-specific version of mapreduce().

The parameters are f, o, x, and d--the function, (binary) operator, collection, and dimension respectively.

EXAMPLE

A = [[1:5;] [6:10;] [11:15;]]
mrchop(x -> x.^2, /, A, 1) # Column-wise reduction of squared elements.
6.94444e-5 1.41723e-6 1.12745e-7 
mrchop(x -> x.^2, /, A, 2) # Row-wise of the same.
0.000229568
0.000566893
0.000832101
0.00100781
0.00111111

4. mop()

The function mop() attempts to simulate its R equivalent from the afp library: it operates on an array by a summary statistic function according to a given dimension. In other words, it indexes values by the summary statistic function column-wise or row-wise. The advantages of this function includes (1) the ability to compare individual values to the median of a variable, for example, and (2) the convenience to users of having them avoid writing anonymous functions within mapslices() to accomplish the same task (e.g. mapslices(x -> x./median(x), A, 1)).

The parameters are the same as in mrchop().

EXAMPLE

# Divide each element by the median of the associated column.
A = [1:10 11:20 21:30;]
mop(median, /, A, 1) # == mapslices(x -> x./median(x), A, 1)
0.181818 0.709677 0.823529
0.363636 0.774194 0.862745
0.545455 0.83871 0.901961
0.727273 0.903226 0.941176
0.909091 0.967742 0.980392
1.09091 1.03226 1.01961
1.27273 1.09677 1.05882
1.45455 1.16129 1.09804
1.63636 1.22581 1.13725
1.81818 1.29032 1.17647

5. Conclusion

The discussed functions will be improved on a continuous basis to (1) minimize repetitive iterative computations and (2) emphasize code efficiency and brevity. New functions to be added based on feasibility and future needs as appropriate.

6. References

  1. Julia - map()
  2. Julia - mapreduce()
  3. Julia - mapslices()
  4. R programming language
  5. R - afp, the original library
  6. R - afp - mop()
  7. R - sweep()
  8. Donat Studios (2017). CSV To Markdown Table Generator.

7. See also

Advanced R by Hadley Wickham - Functionals chapter

End of Document