Regular readers will recall the “utility belt” post from back in April of this year. This is a follow-up to a request made asking for a list of all the % infix functions in those files.

We’re going to:

• collect up all of the sources
• parse them
• find all the definitions of % infix functions
• write them to a file

We’ll start by grabbing the data from the previous post and look at it as a refresher:

library(stringi)
library(tidyverse)

utils 

Note that we somewhat expected the file source to potentially come in handy at a later date and also expected the need to revisit that post, so the R data file [←direct link to RDS] included a file_src column.

Now, let’s find all the source files with at least one infix definition, collect them together and parse them so we can do more code spelunking:

filter(utils, stri_detect_fixed(file_src, "`%")) %>% # only find sources with infix definitions
  pull(file_src) %>%
  paste0(collapse="\n\n") %>%
  parse(text = ., keep.source=TRUE) -> infix_src

str(infix_src, 1)
## length 1364 expression(dplyr::`%>%`, `%||%` 

We can now take all of that lovely parsed source and tokenize it to work with the discrete elements in a very tidy manner:

infix_parsed %`     
##  7     5     1     5    49    51      0 expr           FALSE    ""        
##  8     5     1     5     6    16     18 SYMBOL         TRUE     `%||%`    
##  9     5     1     5     6    18     51 expr           FALSE    ""        
## 10     5     8     5     9    17     51 LEFT_ASSIGN    TRUE     

We just need to find a sequence of tokens that make up a function definition, then whittle those down to ones that look like our % infix names:

pat 

Now, write it out to a file so we can peruse the infix functions:

# nuke a file and fill it with the function definition
cat("", sep="", file="infix_functions.R")
walk2(
  getParseText(infix_parsed, infix_parsed$id[infix_defs]),     # extract the infix name
  getParseText(infix_parsed, infix_parsed$id[infix_defs + 3]), # extract the function definition body
  ~{
    cat(.x, " 

There are 106 of them so you can find the extracted ones in this gist.

Here’s an overview of what you can expect to find:

# A tibble: 39 x 2
   name                 n
 1 `%||%`              47
 2 `%+%`                7
 3 `%AND%`              4
 4 `%notin%`            4
 5 `%:::%`              3
 6 `%==%`               3
 7 `%!=%`               2
 8 `%*diag%`            2
 9 `%diag*%`            2
10 `%nin%`              2
11 `%OR%`               2
12 `%::%`               1
13 `%??%`               1
14 `%.%`                1
15 `%@%`                1
16 `%&&%`               1
17 `%&%`                1
18 `%+&%`               1
19 `%++%`               1
20 `%+|%`               1
21 `%%`               1
23 `%~~%`               1
24 `%assert_class%`     1
25 `%contains%`         1
26 `%din%`              1
27 `%fin%`              1
28 `%identical%`        1
29 `%In%`               1
30 `%inr%`              1
31 `%M%`                1
32 `%notchin%`          1
33 `%or%`               1
34 `%p%`                1
35 `%pin%`              1
36 `%R%`                1
37 `%s%`                1
38 `%sub_in%`           1
39 `%sub_nin%`          1

FIN

If any of those are useful, feel free to PR them in to https://github.com/hrbrmstr/freebase/blob/master/inst/templates/infix-helpers.R (and add yourself to the DESCRIPTION if you do).

Hopefully this provided some further inspiration to continue to use R not only as your language of choice but also as a fun data source.