Improve the Style/SelectByRegexp cop

[ydakuka]

Describe the solution you'd like

I have the following code:

array.select { |x| x.match?(/regexp/) }

If I run the command rubocop -A --only Style/SelectByRegexp I will get:

array.grep(/regexp/)

However, if I run the command rubocop -A --only Performance/StringInclude I will get:

array.select { |x| x.include?('regexp') }

So I conclude that these onelines do the same thing.

I propose (see benchmarks for more information):

# bad
array.select { |x| x.include?('regexp') }
array.reject { |x| x.include?('regexp') }

# good
array.grep(/regexp/)
array.grep_v(/regexp/)

[ydakuka]

I have the Dockerfile:

FROM ruby:3.3.0-preview3
COPY . /var/www/ruby  
WORKDIR /var/www/ruby  
RUN gem install benchmark-ips
RUN ruby -v
CMD ["ruby", "file.rb"]

[ydakuka]

1 case

require "benchmark/ips"

ARRAY = ("a".."zzz").to_a.shuffle.freeze

def initial
  ARRAY.select { |x| x.match?(/regexp/) }
end

def slow
  ARRAY.select { |x| x.include?('regexp') }
end

def fast
  ARRAY.grep(/regexp/)
end

Benchmark.ips do |x|
  x.report('initial') { initial }
  x.report('Array#select#include') { slow }
  x.report('Array#grep') { fast }
  x.compare!
end

Output:

ydakuka@yauhenid:~/ruby-docker-app$  docker run ruby-app  
Warming up --------------------------------------
             initial    60.000  i/100ms
Array#select#include    72.000  i/100ms
          Array#grep    95.000  i/100ms
Calculating -------------------------------------
             initial    582.673  (± 4.8%) i/s -      2.940k in   5.058520s
Array#select#include    777.983  (± 1.4%) i/s -      3.960k in   5.091249s
          Array#grep    966.763  (± 2.5%) i/s -      4.845k in   5.015105s

Comparison:
          Array#grep:      966.8 i/s
Array#select#include:      778.0 i/s - 1.24x  slower
             initial:      582.7 i/s - 1.66x  slower

[ydakuka]

2 case

require "benchmark/ips"

ARRAY = ("a".."zzz").to_a.shuffle.freeze

def initial
  ARRAY.reject { |x| x.match?(/regexp/) }
end

def slow
  ARRAY.reject { |x| x.include?('regexp') }
end

def fast
  ARRAY.grep_v(/regexp/)
end

Benchmark.ips do |x|
  x.report('initial') { initial }
  x.report('Array#reject#include') { slow }
  x.report('Array#grep_v') { fast }
  x.compare!
end

Output:

ydakuka@yauhenid:~/ruby-docker-app$  docker run ruby-app  
Warming up --------------------------------------
             initial    42.000  i/100ms
Array#reject#include    44.000  i/100ms
        Array#grep_v    76.000  i/100ms
Calculating -------------------------------------
             initial    504.497  (± 6.5%) i/s -      2.520k in   5.020388s
Array#reject#include    637.544  (± 8.0%) i/s -      3.168k in   5.025849s
        Array#grep_v    792.533  (± 2.4%) i/s -      4.028k in   5.085471s

Comparison:
        Array#grep_v:      792.5 i/s
Array#reject#include:      637.5 i/s - 1.24x  slower
             initial:      504.5 i/s - 1.57x  slower

[fatkodima]

I think this should be implemented as a separate cop. ARRAY.grep(/regexp/) should be really ARRAY.grep('regexp'), because grep/grep_v also supports strings, ranges, classes etc.