aispec 0.1.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- checksums.yaml +7 -0
- data/LICENSE.txt +21 -0
- data/README.md +288 -0
- data/bin/aispec +6 -0
- data/lib/aispec/cli.rb +211 -0
- data/lib/aispec/configuration.rb +32 -0
- data/lib/aispec/core/assertion.rb +70 -0
- data/lib/aispec/core/assertions/deterministic.rb +147 -0
- data/lib/aispec/core/assertions/judge.rb +82 -0
- data/lib/aispec/core/contract.rb +105 -0
- data/lib/aispec/core/providers/anthropic.rb +77 -0
- data/lib/aispec/core/providers/base.rb +48 -0
- data/lib/aispec/core/providers/mock.rb +58 -0
- data/lib/aispec/core/providers/ollama.rb +64 -0
- data/lib/aispec/core/providers/open_router.rb +76 -0
- data/lib/aispec/core/providers/openai.rb +90 -0
- data/lib/aispec/core/providers/vllm.rb +86 -0
- data/lib/aispec/core/reporters/base.rb +34 -0
- data/lib/aispec/core/reporters/console.rb +44 -0
- data/lib/aispec/core/reporters/html.rb +148 -0
- data/lib/aispec/core/reporters/json.rb +47 -0
- data/lib/aispec/core/reporters/junit.rb +52 -0
- data/lib/aispec/core/reporters/markdown.rb +47 -0
- data/lib/aispec/core/runner.rb +83 -0
- data/lib/aispec/core/statistics/calculator.rb +120 -0
- data/lib/aispec/plugin.rb +24 -0
- data/lib/aispec/version.rb +5 -0
- data/lib/aispec.rb +34 -0
- metadata +128 -0
|
@@ -0,0 +1,120 @@
|
|
|
1
|
+
# frozen_string_literal: true
|
|
2
|
+
|
|
3
|
+
module AISpec
|
|
4
|
+
module Core
|
|
5
|
+
module Statistics
|
|
6
|
+
class Calculator
|
|
7
|
+
attr_reader :results
|
|
8
|
+
|
|
9
|
+
def initialize(results)
|
|
10
|
+
@results = results || []
|
|
11
|
+
end
|
|
12
|
+
|
|
13
|
+
def total_runs
|
|
14
|
+
@results.size
|
|
15
|
+
end
|
|
16
|
+
|
|
17
|
+
def total_assertions
|
|
18
|
+
@results.sum { |r| r[:assertion_results].size }
|
|
19
|
+
end
|
|
20
|
+
|
|
21
|
+
def passed_assertions
|
|
22
|
+
@results.sum { |r| r[:assertion_results].count(&:passed?) }
|
|
23
|
+
end
|
|
24
|
+
|
|
25
|
+
def failed_assertions
|
|
26
|
+
total_assertions - passed_assertions
|
|
27
|
+
end
|
|
28
|
+
|
|
29
|
+
def success_rate
|
|
30
|
+
return 0.0 if total_assertions.zero?
|
|
31
|
+
(passed_assertions.to_f / total_assertions * 100.0).round(1)
|
|
32
|
+
end
|
|
33
|
+
|
|
34
|
+
def confidence_interval
|
|
35
|
+
return 0.0 if total_assertions.zero?
|
|
36
|
+
|
|
37
|
+
# Wilson score interval z = 1.96 for 95% confidence
|
|
38
|
+
p = passed_assertions.to_f / total_assertions
|
|
39
|
+
n = total_assertions.to_f
|
|
40
|
+
z = 1.96
|
|
41
|
+
|
|
42
|
+
margin = z * Math.sqrt((p * (1 - p) + (z**2 / (4 * n))) / n) / (1 + (z**2 / n))
|
|
43
|
+
(margin * 100.0).round(1)
|
|
44
|
+
end
|
|
45
|
+
|
|
46
|
+
def mean_latency
|
|
47
|
+
return 0.0 if @results.empty?
|
|
48
|
+
(@results.sum { |r| r[:latency_ms].to_f } / @results.size).round(1)
|
|
49
|
+
end
|
|
50
|
+
|
|
51
|
+
def mean_cost
|
|
52
|
+
return 0.0 if @results.empty?
|
|
53
|
+
@results.sum { |r| r[:cost].to_f } / @results.size
|
|
54
|
+
end
|
|
55
|
+
|
|
56
|
+
def total_cost
|
|
57
|
+
@results.sum { |r| r[:cost].to_f }
|
|
58
|
+
end
|
|
59
|
+
|
|
60
|
+
def judge_scores
|
|
61
|
+
@results.flat_map do |r|
|
|
62
|
+
r[:assertion_results].select { |ar| ar.name == :judge && ar.metadata[:score] }.map { |ar| ar.metadata[:score].to_f }
|
|
63
|
+
end
|
|
64
|
+
end
|
|
65
|
+
|
|
66
|
+
def average_judge_score
|
|
67
|
+
scores = judge_scores
|
|
68
|
+
return nil if scores.empty?
|
|
69
|
+
(scores.sum / scores.size).round(2)
|
|
70
|
+
end
|
|
71
|
+
|
|
72
|
+
def variance_rating
|
|
73
|
+
scores = judge_scores
|
|
74
|
+
return "N/A" if scores.size < 2
|
|
75
|
+
|
|
76
|
+
mean = scores.sum / scores.size
|
|
77
|
+
var = scores.sum { |s| (s - mean)**2 } / scores.size
|
|
78
|
+
|
|
79
|
+
if var < 0.01
|
|
80
|
+
"Low"
|
|
81
|
+
elsif var < 0.05
|
|
82
|
+
"Moderate"
|
|
83
|
+
else
|
|
84
|
+
"High"
|
|
85
|
+
end
|
|
86
|
+
end
|
|
87
|
+
|
|
88
|
+
def recommendation
|
|
89
|
+
rate = success_rate
|
|
90
|
+
var = variance_rating
|
|
91
|
+
|
|
92
|
+
if rate >= 95.0 && var != "High"
|
|
93
|
+
"Stable"
|
|
94
|
+
elsif rate >= 80.0 || var == "Moderate"
|
|
95
|
+
"Needs Attention (Flaky)"
|
|
96
|
+
else
|
|
97
|
+
"Critical Failure"
|
|
98
|
+
end
|
|
99
|
+
end
|
|
100
|
+
|
|
101
|
+
def summary
|
|
102
|
+
{
|
|
103
|
+
total_runs: total_runs,
|
|
104
|
+
total_assertions: total_assertions,
|
|
105
|
+
passed_assertions: passed_assertions,
|
|
106
|
+
failed_assertions: failed_assertions,
|
|
107
|
+
success_rate: success_rate,
|
|
108
|
+
confidence_interval: confidence_interval,
|
|
109
|
+
mean_latency: mean_latency,
|
|
110
|
+
mean_cost: mean_cost,
|
|
111
|
+
total_cost: total_cost,
|
|
112
|
+
average_judge_score: average_judge_score,
|
|
113
|
+
variance: variance_rating,
|
|
114
|
+
recommendation: recommendation
|
|
115
|
+
}
|
|
116
|
+
end
|
|
117
|
+
end
|
|
118
|
+
end
|
|
119
|
+
end
|
|
120
|
+
end
|
|
@@ -0,0 +1,24 @@
|
|
|
1
|
+
# frozen_string_literal: true
|
|
2
|
+
|
|
3
|
+
module AISpec
|
|
4
|
+
class PluginDSL
|
|
5
|
+
def assertion(name, &block)
|
|
6
|
+
AISpec::Core::Assertion.register(name, &block)
|
|
7
|
+
end
|
|
8
|
+
|
|
9
|
+
def provider(name, provider_class)
|
|
10
|
+
AISpec::Core::Providers::Base.register(name, provider_class)
|
|
11
|
+
end
|
|
12
|
+
|
|
13
|
+
def reporter(name, reporter_class)
|
|
14
|
+
AISpec::Core::Reporters::Base.register(name, reporter_class)
|
|
15
|
+
end
|
|
16
|
+
end
|
|
17
|
+
|
|
18
|
+
class << self
|
|
19
|
+
def plugin(&block)
|
|
20
|
+
dsl = PluginDSL.new
|
|
21
|
+
dsl.instance_eval(&block) if block_given?
|
|
22
|
+
end
|
|
23
|
+
end
|
|
24
|
+
end
|
data/lib/aispec.rb
ADDED
|
@@ -0,0 +1,34 @@
|
|
|
1
|
+
# frozen_string_literal: true
|
|
2
|
+
|
|
3
|
+
require_relative "aispec/version"
|
|
4
|
+
require_relative "aispec/configuration"
|
|
5
|
+
require_relative "aispec/plugin"
|
|
6
|
+
|
|
7
|
+
require_relative "aispec/core/contract"
|
|
8
|
+
require_relative "aispec/core/assertion"
|
|
9
|
+
require_relative "aispec/core/assertions/deterministic"
|
|
10
|
+
require_relative "aispec/core/assertions/judge"
|
|
11
|
+
|
|
12
|
+
require_relative "aispec/core/providers/base"
|
|
13
|
+
require_relative "aispec/core/providers/mock"
|
|
14
|
+
require_relative "aispec/core/providers/openai"
|
|
15
|
+
require_relative "aispec/core/providers/anthropic"
|
|
16
|
+
require_relative "aispec/core/providers/ollama"
|
|
17
|
+
require_relative "aispec/core/providers/open_router"
|
|
18
|
+
require_relative "aispec/core/providers/vllm"
|
|
19
|
+
|
|
20
|
+
require_relative "aispec/core/statistics/calculator"
|
|
21
|
+
require_relative "aispec/core/runner"
|
|
22
|
+
|
|
23
|
+
require_relative "aispec/core/reporters/base"
|
|
24
|
+
require_relative "aispec/core/reporters/console"
|
|
25
|
+
require_relative "aispec/core/reporters/json"
|
|
26
|
+
require_relative "aispec/core/reporters/markdown"
|
|
27
|
+
require_relative "aispec/core/reporters/junit"
|
|
28
|
+
require_relative "aispec/core/reporters/html"
|
|
29
|
+
|
|
30
|
+
require_relative "aispec/cli"
|
|
31
|
+
|
|
32
|
+
module AISpec
|
|
33
|
+
class Error < StandardError; end
|
|
34
|
+
end
|
metadata
ADDED
|
@@ -0,0 +1,128 @@
|
|
|
1
|
+
--- !ruby/object:Gem::Specification
|
|
2
|
+
name: aispec
|
|
3
|
+
version: !ruby/object:Gem::Version
|
|
4
|
+
version: 0.1.0
|
|
5
|
+
platform: ruby
|
|
6
|
+
authors:
|
|
7
|
+
- Wilbur Suero
|
|
8
|
+
bindir: bin
|
|
9
|
+
cert_chain: []
|
|
10
|
+
date: 1980-01-02 00:00:00.000000000 Z
|
|
11
|
+
dependencies:
|
|
12
|
+
- !ruby/object:Gem::Dependency
|
|
13
|
+
name: rexml
|
|
14
|
+
requirement: !ruby/object:Gem::Requirement
|
|
15
|
+
requirements:
|
|
16
|
+
- - ">="
|
|
17
|
+
- !ruby/object:Gem::Version
|
|
18
|
+
version: '3.0'
|
|
19
|
+
type: :runtime
|
|
20
|
+
prerelease: false
|
|
21
|
+
version_requirements: !ruby/object:Gem::Requirement
|
|
22
|
+
requirements:
|
|
23
|
+
- - ">="
|
|
24
|
+
- !ruby/object:Gem::Version
|
|
25
|
+
version: '3.0'
|
|
26
|
+
- !ruby/object:Gem::Dependency
|
|
27
|
+
name: bundler
|
|
28
|
+
requirement: !ruby/object:Gem::Requirement
|
|
29
|
+
requirements:
|
|
30
|
+
- - ">="
|
|
31
|
+
- !ruby/object:Gem::Version
|
|
32
|
+
version: '2.0'
|
|
33
|
+
type: :development
|
|
34
|
+
prerelease: false
|
|
35
|
+
version_requirements: !ruby/object:Gem::Requirement
|
|
36
|
+
requirements:
|
|
37
|
+
- - ">="
|
|
38
|
+
- !ruby/object:Gem::Version
|
|
39
|
+
version: '2.0'
|
|
40
|
+
- !ruby/object:Gem::Dependency
|
|
41
|
+
name: rake
|
|
42
|
+
requirement: !ruby/object:Gem::Requirement
|
|
43
|
+
requirements:
|
|
44
|
+
- - "~>"
|
|
45
|
+
- !ruby/object:Gem::Version
|
|
46
|
+
version: '13.0'
|
|
47
|
+
type: :development
|
|
48
|
+
prerelease: false
|
|
49
|
+
version_requirements: !ruby/object:Gem::Requirement
|
|
50
|
+
requirements:
|
|
51
|
+
- - "~>"
|
|
52
|
+
- !ruby/object:Gem::Version
|
|
53
|
+
version: '13.0'
|
|
54
|
+
- !ruby/object:Gem::Dependency
|
|
55
|
+
name: rspec
|
|
56
|
+
requirement: !ruby/object:Gem::Requirement
|
|
57
|
+
requirements:
|
|
58
|
+
- - "~>"
|
|
59
|
+
- !ruby/object:Gem::Version
|
|
60
|
+
version: '3.12'
|
|
61
|
+
type: :development
|
|
62
|
+
prerelease: false
|
|
63
|
+
version_requirements: !ruby/object:Gem::Requirement
|
|
64
|
+
requirements:
|
|
65
|
+
- - "~>"
|
|
66
|
+
- !ruby/object:Gem::Version
|
|
67
|
+
version: '3.12'
|
|
68
|
+
description: AISpec provides RSpec-like behavioral contract definition, execution,
|
|
69
|
+
and verification across AI providers for testing, CI, and production.
|
|
70
|
+
email:
|
|
71
|
+
- wilbur.himself@gmail.com
|
|
72
|
+
executables:
|
|
73
|
+
- aispec
|
|
74
|
+
extensions: []
|
|
75
|
+
extra_rdoc_files: []
|
|
76
|
+
files:
|
|
77
|
+
- LICENSE.txt
|
|
78
|
+
- README.md
|
|
79
|
+
- bin/aispec
|
|
80
|
+
- lib/aispec.rb
|
|
81
|
+
- lib/aispec/cli.rb
|
|
82
|
+
- lib/aispec/configuration.rb
|
|
83
|
+
- lib/aispec/core/assertion.rb
|
|
84
|
+
- lib/aispec/core/assertions/deterministic.rb
|
|
85
|
+
- lib/aispec/core/assertions/judge.rb
|
|
86
|
+
- lib/aispec/core/contract.rb
|
|
87
|
+
- lib/aispec/core/providers/anthropic.rb
|
|
88
|
+
- lib/aispec/core/providers/base.rb
|
|
89
|
+
- lib/aispec/core/providers/mock.rb
|
|
90
|
+
- lib/aispec/core/providers/ollama.rb
|
|
91
|
+
- lib/aispec/core/providers/open_router.rb
|
|
92
|
+
- lib/aispec/core/providers/openai.rb
|
|
93
|
+
- lib/aispec/core/providers/vllm.rb
|
|
94
|
+
- lib/aispec/core/reporters/base.rb
|
|
95
|
+
- lib/aispec/core/reporters/console.rb
|
|
96
|
+
- lib/aispec/core/reporters/html.rb
|
|
97
|
+
- lib/aispec/core/reporters/json.rb
|
|
98
|
+
- lib/aispec/core/reporters/junit.rb
|
|
99
|
+
- lib/aispec/core/reporters/markdown.rb
|
|
100
|
+
- lib/aispec/core/runner.rb
|
|
101
|
+
- lib/aispec/core/statistics/calculator.rb
|
|
102
|
+
- lib/aispec/plugin.rb
|
|
103
|
+
- lib/aispec/version.rb
|
|
104
|
+
homepage: https://github.com/wilburhimself/aispec
|
|
105
|
+
licenses:
|
|
106
|
+
- MIT
|
|
107
|
+
metadata:
|
|
108
|
+
homepage_uri: https://github.com/wilburhimself/aispec
|
|
109
|
+
source_code_uri: https://github.com/wilburhimself/aispec
|
|
110
|
+
changelog_uri: https://github.com/wilburhimself/aispec/blob/main/CHANGELOG.md
|
|
111
|
+
rdoc_options: []
|
|
112
|
+
require_paths:
|
|
113
|
+
- lib
|
|
114
|
+
required_ruby_version: !ruby/object:Gem::Requirement
|
|
115
|
+
requirements:
|
|
116
|
+
- - ">="
|
|
117
|
+
- !ruby/object:Gem::Version
|
|
118
|
+
version: 3.0.0
|
|
119
|
+
required_rubygems_version: !ruby/object:Gem::Requirement
|
|
120
|
+
requirements:
|
|
121
|
+
- - ">="
|
|
122
|
+
- !ruby/object:Gem::Version
|
|
123
|
+
version: '0'
|
|
124
|
+
requirements: []
|
|
125
|
+
rubygems_version: 4.0.7
|
|
126
|
+
specification_version: 4
|
|
127
|
+
summary: The engineering framework for behavioral contracts for LLM applications.
|
|
128
|
+
test_files: []
|