running_stat 0.0.1 → 0.1.0
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/CHANGELOG.md +8 -0
- data/README.md +2 -0
- data/lib/running_stat.rb +8 -15
- data/lib/running_stat/lua/push_datum.rb +7 -10
- data/lib/running_stat/lua/variance.rb +6 -6
- data/lib/running_stat/redis_backend.rb +10 -0
- data/lib/running_stat/version.rb +1 -1
- metadata +3 -2
checksums.yaml
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
---
|
2
2
|
SHA1:
|
3
|
-
metadata.gz:
|
4
|
-
data.tar.gz:
|
3
|
+
metadata.gz: 692e627aba71ce44c80b57d406c85dd69f38edfc
|
4
|
+
data.tar.gz: 0caada43394fd224b1e8932958bbe18ed1fd4815
|
5
5
|
SHA512:
|
6
|
-
metadata.gz:
|
7
|
-
data.tar.gz:
|
6
|
+
metadata.gz: b173a92b2b32bd8decf100e7c4f4f79b179ff4681839abbb4fcb390db1553d411d9ac552779279ee9d4ac8cef767d133f6780dbce8e8927c8ecb8a8560e8f965
|
7
|
+
data.tar.gz: fe7e30605e41b8e2980fa9760b7f722e33f8bf6f1f7b2dec5a7c56ced7c8dd648e345d4a90ac5b676e624d605582e54b8fdd602c67e10f21e0b0a8d6cd703aee
|
data/CHANGELOG.md
CHANGED
@@ -2,6 +2,14 @@
|
|
2
2
|
All notable changes to this project will be documented in this file.
|
3
3
|
This project adheres to [Semantic Versioning](http://semver.org/).
|
4
4
|
|
5
|
+
## [0.1.0] - 2016-01-12
|
6
|
+
### Added
|
7
|
+
- n/a
|
8
|
+
|
9
|
+
### Changed
|
10
|
+
- All keys for each metric combined into a single redis hash
|
11
|
+
(this will null out old metrics)
|
12
|
+
|
5
13
|
## [0.0.1] - 2016-01-11
|
6
14
|
### Added
|
7
15
|
- Initial release
|
data/README.md
CHANGED
@@ -8,6 +8,8 @@ RunningStat provides distributed redis-backed data buckets on which various stat
|
|
8
8
|
|
9
9
|
The algorithm used is based on Knuth's TAOCP and is numerically stable; a brief writeup is available on Wikipedia under [Algorithms for calculating online variances](https://en.wikipedia.org/wiki/Algorithms_for_calculating_variance#Online_algorithm).
|
10
10
|
|
11
|
+
RunningStat supports pretty much all versions of ruby, including MRI 1.8.7, 1.9.3, 2.2, and JRuby 1.7/9k.
|
12
|
+
|
11
13
|
Disclaimer: Because it uses atomic Lua scripting, RunningStat requires redis 2.6+.
|
12
14
|
|
13
15
|
|
data/lib/running_stat.rb
CHANGED
@@ -2,6 +2,7 @@ require 'running_stat/version'
|
|
2
2
|
|
3
3
|
require 'redis'
|
4
4
|
|
5
|
+
require 'running_stat/redis_backend'
|
5
6
|
require 'running_stat/lua/push_datum'
|
6
7
|
require 'running_stat/lua/variance'
|
7
8
|
require 'running_stat/insufficient_data_error'
|
@@ -22,26 +23,26 @@ class RunningStat
|
|
22
23
|
|
23
24
|
# Adds a piece of numerical data to the dataset's stats
|
24
25
|
def push(datum)
|
25
|
-
redis.eval(Lua::PUSH_DATUM, [
|
26
|
+
redis.eval(Lua::PUSH_DATUM, [bucket_key], [Float(datum)])
|
26
27
|
rescue ArgumentError => e
|
27
28
|
raise InvalidDataError.new(e) # datum was non-numerical
|
28
29
|
end
|
29
30
|
|
30
31
|
# Returns the number of data points seen, or 0 if the stat does not exist
|
31
32
|
def cardinality
|
32
|
-
redis.
|
33
|
+
redis.hget(bucket_key, RedisBackend::COUNT_FIELD).to_i
|
33
34
|
end
|
34
35
|
|
35
36
|
# Returns the arithmetic mean of data points seen, or 0.0 if non-existent
|
36
37
|
def mean
|
37
|
-
redis.
|
38
|
+
redis.hget(bucket_key, RedisBackend::MEAN_FIELD).to_f
|
38
39
|
end
|
39
40
|
|
40
41
|
# Returns the sample variance of the dataset so far, or raises
|
41
42
|
# an InsufficientDataError if insufficient data (< 2 datapoints)
|
42
43
|
# has been pushed
|
43
44
|
def variance
|
44
|
-
redis.eval(Lua::VARIANCE, [
|
45
|
+
redis.eval(Lua::VARIANCE, [bucket_key], []).to_f
|
45
46
|
rescue Redis::CommandError => e
|
46
47
|
raise InsufficientDataError.new(e) # only CommandError possible
|
47
48
|
end
|
@@ -55,7 +56,7 @@ class RunningStat
|
|
55
56
|
|
56
57
|
# Resets the stat to reflect an empty dataset
|
57
58
|
def flush
|
58
|
-
redis.del(
|
59
|
+
redis.del(bucket_key)
|
59
60
|
end
|
60
61
|
|
61
62
|
private
|
@@ -64,15 +65,7 @@ class RunningStat
|
|
64
65
|
@redis || Redis.current
|
65
66
|
end
|
66
67
|
|
67
|
-
def
|
68
|
-
"#{BASE_KEY}:#{@data_bucket}
|
69
|
-
end
|
70
|
-
|
71
|
-
def mean_key
|
72
|
-
"#{BASE_KEY}:#{@data_bucket}:mean"
|
73
|
-
end
|
74
|
-
|
75
|
-
def sum_sq_diff_key
|
76
|
-
"#{BASE_KEY}:#{@data_bucket}:sum_sq_diff"
|
68
|
+
def bucket_key
|
69
|
+
"#{BASE_KEY}:#{@data_bucket}"
|
77
70
|
end
|
78
71
|
end
|
@@ -1,11 +1,10 @@
|
|
1
1
|
require 'digest/sha1'
|
2
|
+
require 'running_stat/redis_backend'
|
2
3
|
|
3
4
|
class RunningStat
|
4
5
|
module Lua
|
5
6
|
# Keys:
|
6
|
-
#
|
7
|
-
# mean - the mean of the dataset so far
|
8
|
-
# m2 - the running sum of the squares of the differences of the dataset
|
7
|
+
# bucket - the hash corresponding to the dataset metric
|
9
8
|
# Arguments:
|
10
9
|
# datum - the number to be added to the dataset
|
11
10
|
# Effects:
|
@@ -13,17 +12,15 @@ class RunningStat
|
|
13
12
|
# Returns:
|
14
13
|
# nothing
|
15
14
|
PUSH_DATUM = <<-EOLUA
|
16
|
-
local
|
17
|
-
local mean_key = KEYS[2]
|
18
|
-
local m2_key = KEYS[3]
|
15
|
+
local bucket_key = KEYS[1]
|
19
16
|
local datum = ARGV[1]
|
20
17
|
|
21
|
-
local mean = tonumber(redis.call("
|
18
|
+
local mean = tonumber(redis.call("HGET", bucket_key, "#{RedisBackend::MEAN_FIELD}")) or 0.0
|
22
19
|
local delta = datum - mean
|
23
20
|
|
24
|
-
local count = redis.call("
|
25
|
-
mean = redis.call("
|
26
|
-
redis.call("
|
21
|
+
local count = redis.call("HINCRBY", bucket_key, "#{RedisBackend::COUNT_FIELD}", 1)
|
22
|
+
mean = redis.call("HINCRBYFLOAT", bucket_key, "#{RedisBackend::MEAN_FIELD}", tostring(delta / count))
|
23
|
+
redis.call("HINCRBYFLOAT", bucket_key, "#{RedisBackend::M2_FIELD}", tostring(delta * (datum - mean)))
|
27
24
|
EOLUA
|
28
25
|
|
29
26
|
PUSH_DATUM_SHA1 = Digest::SHA1.hexdigest(PUSH_DATUM).freeze
|
@@ -1,11 +1,11 @@
|
|
1
1
|
require 'digest/sha1'
|
2
|
+
require 'running_stat/redis_backend'
|
2
3
|
require 'running_stat/insufficient_data_error'
|
3
4
|
|
4
5
|
class RunningStat
|
5
6
|
module Lua
|
6
7
|
# Keys:
|
7
|
-
#
|
8
|
-
# m2 - the running sum of the squares of the differences of the dataset
|
8
|
+
# bucket - the hash corresponding to the dataset metric
|
9
9
|
# Arguments:
|
10
10
|
# nothing
|
11
11
|
# Effects:
|
@@ -13,14 +13,14 @@ class RunningStat
|
|
13
13
|
# Returns:
|
14
14
|
# the sample variance of the dataset, as a stringified float
|
15
15
|
VARIANCE = <<-EOLUA
|
16
|
-
local
|
17
|
-
local m2_key = KEYS[2]
|
16
|
+
local bucket_key = KEYS[1]
|
18
17
|
|
19
|
-
local
|
18
|
+
local values = redis.call("HMGET", bucket_key, "#{RedisBackend::COUNT_FIELD}", "#{RedisBackend::M2_FIELD}")
|
19
|
+
local count = tonumber(values[1]) or 0
|
20
20
|
if count < 2 then
|
21
21
|
return redis.error_reply("#{InsufficientDataError::ERROR_STRING}")
|
22
22
|
else
|
23
|
-
local m2 = tonumber(
|
23
|
+
local m2 = tonumber(values[2]) or 0.0
|
24
24
|
return tostring(m2 / (count - 1))
|
25
25
|
end
|
26
26
|
EOLUA
|
@@ -0,0 +1,10 @@
|
|
1
|
+
class RunningStat
|
2
|
+
module RedisBackend
|
3
|
+
# fields within the metric's redis hash
|
4
|
+
COUNT_FIELD = :c # the dataset cardinality
|
5
|
+
MEAN_FIELD = :m # the arithmetic mean of the dataset
|
6
|
+
M2_FIELD = :m2 # the running sum of the squares of the differences of the dataset
|
7
|
+
|
8
|
+
# TODO: implement global redis set/get
|
9
|
+
end
|
10
|
+
end
|
data/lib/running_stat/version.rb
CHANGED
metadata
CHANGED
@@ -1,14 +1,14 @@
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
2
2
|
name: running_stat
|
3
3
|
version: !ruby/object:Gem::Version
|
4
|
-
version: 0.0
|
4
|
+
version: 0.1.0
|
5
5
|
platform: ruby
|
6
6
|
authors:
|
7
7
|
- Anuj Das
|
8
8
|
autorequire:
|
9
9
|
bindir: exe
|
10
10
|
cert_chain: []
|
11
|
-
date: 2016-01-
|
11
|
+
date: 2016-01-12 00:00:00.000000000 Z
|
12
12
|
dependencies:
|
13
13
|
- !ruby/object:Gem::Dependency
|
14
14
|
name: redis
|
@@ -89,6 +89,7 @@ files:
|
|
89
89
|
- lib/running_stat/invalid_data_error.rb
|
90
90
|
- lib/running_stat/lua/push_datum.rb
|
91
91
|
- lib/running_stat/lua/variance.rb
|
92
|
+
- lib/running_stat/redis_backend.rb
|
92
93
|
- lib/running_stat/version.rb
|
93
94
|
- running_stat.gemspec
|
94
95
|
homepage: https://www.github.com/anujdas/running_stat
|