Last active
August 29, 2015 14:17
-
-
Save inkel/51c27ade94a9f83d387c to your computer and use it in GitHub Desktop.
Elasticsearch::API::Utils#__pathify is slow and probably overly complex. This gist shows the result of my attempts at improving it's performance.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
$ PROTEST_REPORT=progress RUBYOPT=-rpry tmp/benchmark.rb | |
Calculating ------------------------------------- | |
original 11.231k i/100ms | |
__pathify 14.706k i/100ms | |
__pathify2 21.360k i/100ms | |
------------------------------------------------- | |
original 124.124k (±13.1%) i/s - 617.705k | |
__pathify 154.361k (±18.9%) i/s - 735.300k | |
__pathify2 209.829k (± 9.5%) i/s - 1.047M | |
Comparison: | |
__pathify2: 209829.4 i/s | |
__pathify: 154360.9 i/s - 1.36x slower | |
original: 124123.8 i/s - 1.69x slower | |
.. | |
2 tests, 221572 assertions (2 passed, 0 pending, 0 failed, 0 errored) | |
Ran in 6.345552 seconds |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
# This file contains several JSON lines with the following structure: | |
# | |
# [segments,expected] | |
# | |
# segments: an array of `*segments` parameter for __pathify | |
# expected: the result of running the original __pathify method | |
# for those segments. | |
# | |
# My current file has 110,786 lines, constructed by running my app's | |
# complete test suite and saving those values with the above format. |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
#! /usr/bin/env ruby | |
require "benchmark/ips" | |
require "elasticsearch" | |
require "oj" | |
def original(*segments) | |
Array(segments).flatten. | |
compact. | |
reject { |s| s.to_s =~ /^\s*$/ }. | |
join('/'). | |
squeeze('/') | |
end | |
def __pathify(*segments) | |
segments. | |
reject { |s| s.to_s =~ /^\s*$/ }. | |
join("/"). | |
squeeze("/") | |
end | |
def __pathify2(*segments) | |
segments.inject("") do |m, s| | |
next m if s.empty? | |
m << "/" << (s[0] == "/" ? s[1..-1] : s) | |
end[1..-1] | |
end | |
Benchmark.ips do |r| | |
r.report("original") { p = original("v1-datapoints-201501", "_mapping", "datapoint") } | |
r.report("__pathify") { p = __pathify("v1-datapoints-201501", "_mapping", "datapoint") } | |
r.report("__pathify2") { p = __pathify2("v1-datapoints-201501", "_mapping", "datapoint") } | |
r.compare! | |
end | |
require "protest" | |
Protest.describe "__pathify" do | |
LINES = open("tmp/__pathify.log").readlines.map(&:chomp!) | |
test "__pathify: produces same results" do | |
LINES.each do |line| | |
line.chomp! | |
segments, expected = Oj.load(line, mode: :compat) | |
actual = __pathify(*segments) | |
assert_equal expected, actual, "expected: #{expected}; actual: #{actual}; segments: #{segments.inspect}" | |
end | |
end | |
test "__pathify2: produces same results" do | |
LINES.each do |line| | |
line.chomp! | |
segments, expected = Oj.load(line, mode: :compat) | |
actual = __pathify2(*segments) | |
assert_equal expected, actual, "expected: #{expected}; actual: #{actual}; segments: #{segments.inspect}" | |
end | |
end | |
end |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment