mirrors/valkey
1
0
Fork 0
mirror of http://github.com/valkey-io/valkey synced 2024-11-22 00:52:38 +00:00
Code Issues Packages Projects Releases Wiki Activity

hll-gnuplot-graph.rb improved with new filter.

Browse Source 
The function to generate graphs is also more flexible as now includes
step and max value. The step of the samples generation function is no
longer limited to min step of 1000.
This commit is contained in:
antirez 2014-04-02 12:30:11 +02:00
parent 096b5e921e
commit cf34507b87
1 changed files with 22 additions and 13 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

35
utils/hyperloglog/hll-gnuplot-graph.rb
View File

@ -10,7 +10,7 @@ require 'redis'
require 'digest/sha1' require 'digest/sha1'
# Generate an array of [cardinality,relative_error] pairs # Generate an array of [cardinality,relative_error] pairs
# in the 0 - max range with step of 1000*step. # in the 0 - max range, with the specified step.
# #
# 'r' is the Redis object used to perform the queries. # 'r' is the Redis object used to perform the queries.
# 'seed' must be different every time you want a test performed # 'seed' must be different every time you want a test performed
@ -22,16 +22,15 @@ def run_experiment(r,seed,max,step)
r.del('hll') r.del('hll')
i = 0 i = 0
samples = [] samples = []
step = 1000 if step > 1000
while i < max do while i < max do
elements = []
step.times { step.times {
elements = [] ele = Digest::SHA1.hexdigest(i.to_s+seed.to_s)
1000.times { elements << ele
ele = Digest::SHA1.hexdigest(i.to_s+seed.to_s) i += 1
elements << ele
i += 1
}
r.pfadd('hll',*elements)
} }
r.pfadd('hll',*elements)
approx = r.pfcount('hll') approx = r.pfcount('hll')
err = approx-i err = approx-i
rel_err = 100.to_f*err/i rel_err = 100.to_f*err/i
@ -40,11 +39,12 @@ def run_experiment(r,seed,max,step)
samples samples
end end
def filter_samples(numsets,filter) def filter_samples(numsets,max,step,filter)
r = Redis.new r = Redis.new
dataset = {} dataset = {}
(0...numsets).each{|i| (0...numsets).each{|i|
dataset[i] = run_experiment(r,i,100000,1) dataset[i] = run_experiment(r,i,max,step)
STDERR.puts "Set #{i}"
} }
dataset[0].each_with_index{|ele,index| dataset[0].each_with_index{|ele,index|
if filter == :max if filter == :max
@ -62,6 +62,14 @@ def filter_samples(numsets,filter)
} }
err /= numsets err /= numsets
puts "#{card} #{err}" puts "#{card} #{err}"
elsif filter == :absavg
card=ele[0]
err = 0
(0...numsets).each{|i|
err += dataset[i][index][1].abs
}
err /= numsets
puts "#{card} #{err}"
elsif filter == :all elsif filter == :all
(0...numsets).each{|i| (0...numsets).each{|i|
card,err = dataset[i][index] card,err = dataset[i][index]
@ -73,6 +81,7 @@ def filter_samples(numsets,filter)
} }
end end
filter_samples(100,:all) filter_samples(100,100000,1000,:absavg)
#filter_samples(100,:max) #filter_samples(100,1000,10,:all)
#filter_samples(100,:avg) #filter_samples(100,10000,1000,:max)
#filter_samples(100,10000,1000,:avg)