Class: PropCheck::Generator

Inherits:
Object
  • Object
show all
Defined in:
lib/prop_check/generator.rb

Overview

A ‘Generator` is a special kind of ’proc’ that, given a size an random number generator state, will generate a (finite) LazyTree of output values:

The root of this tree is the value to be used during testing, and the children are ‘smaller’ values related to the root, to be used during the shrinking phase.

Constant Summary collapse

@@default_size = 
10
@@default_rng = 
if RUBY_VERSION.to_i >= 3
  Random
else
  Random::DEFAULT
end
@@max_consecutive_attempts = 
100
@@default_kwargs = 
{ size: @@default_size, rng: @@default_rng,
max_consecutive_attempts: @@max_consecutive_attempts }

Class Method Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(&block) ⇒ Generator

Being a special kind of Proc, a Generator wraps a block.



26
27
28
 
# File 'lib/prop_check/generator.rb', line 26

def initialize(&block)
  @block = block
end

Class Method Details

.wrap(val) ⇒ Object

Creates a ‘constant’ generator that always returns the same value, regardless of ‘size` or `rng`.

Keen readers may notice this as the Monadic ‘pure’/‘return’ implementation for Generators.

>> Generators.integer.bind { |a| Generators.integer.bind { |b| Generator.wrap([a , b]) } }.call(size: 100, rng: Random.new(42))
=> [2, 79]



78
79
80
 
# File 'lib/prop_check/generator.rb', line 78

def self.wrap(val)
  Generator.new { LazyTree.wrap(val) }
end

Instance Method Details

#bind(&generator_proc) ⇒ Object

Create a generator whose implementation depends on the output of another generator. this allows us to compose multiple generators.

Keen readers may notice this as the Monadic ‘bind’ (sometimes known as ‘>>=’) implementation for Generators.

>> Generators.integer.bind { |a| Generators.integer.bind { |b| Generator.wrap([a , b]) } }.call(size: 100, rng: Random.new(42))
=> [2, 79]



90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
 
# File 'lib/prop_check/generator.rb', line 90

def bind(&generator_proc)
  # Generator.new do |size, rng|
  #   outer_result = generate(size, rng)
  #   outer_result.map do |outer_val|
  #     inner_generator = generator_proc.call(outer_val)
  #     inner_generator.generate(size, rng)
  #   end.flatten
  # end
  Generator.new do |**kwargs|
    outer_result = generate(**kwargs)
    outer_result.bind do |outer_val|
      inner_generator = generator_proc.call(outer_val)
      inner_generator.generate(**kwargs)
    end
  end
end

#call(**kwargs) ⇒ Object

Generates a value, and only return this value (drop information for shrinking)

>> Generators.integer.call(size: 1000, rng: Random.new(42))
=> 126



57
58
59
 
# File 'lib/prop_check/generator.rb', line 57

def call(**kwargs)
  generate(**@@default_kwargs.merge(kwargs)).root
end

#generate(**kwargs) ⇒ Object

Given a ‘size` (integer) and a random number generator state `rng`, generate a LazyTree.

Raises:



33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
 
# File 'lib/prop_check/generator.rb', line 33

def generate(**kwargs)
  kwargs = @@default_kwargs.merge(kwargs)
  max_consecutive_attempts = kwargs[:max_consecutive_attempts]

  (0..max_consecutive_attempts).each do
    res = @block.call(**kwargs)
    next if res.root == :"_PropCheck.filter_me"

    return res
  end

  raise Errors::GeneratorExhaustedError, ''"
  Exhausted #{max_consecutive_attempts} consecutive generation attempts.

  Probably too few generator results were adhering to a `where` condition.
  "''
end

#map(&proc) ⇒ Object

Creates a new Generator that returns a value by running ‘proc` on the output of the current Generator.

>> Generators.choose(32..128).map(&:chr).call(size: 10, rng: Random.new(42))
=> "S"



112
113
114
115
116
117
 
# File 'lib/prop_check/generator.rb', line 112

def map(&proc)
  Generator.new do |**kwargs|
    result = generate(**kwargs)
    result.map(&proc)
  end
end

#resize(&proc) ⇒ Object

Resizes the generator to either grow faster or smaller than normal.

‘proc` takes the current size as input and is expected to return the new size. a size should always be a nonnegative integer.

>> Generators.integer.resize{}



156
157
158
159
160
161
 
# File 'lib/prop_check/generator.rb', line 156

def resize(&proc)
  Generator.new do |size:, **other_kwargs|
    new_size = proc.call(size)
    generate(**other_kwargs, size: new_size)
  end
end

#sample(num_of_samples = 10, **kwargs) ⇒ Object

Returns ‘num_of_samples` values from calling this Generator. This is mostly useful for debugging if a generator behaves as you intend it to.



64
65
66
67
68
 
# File 'lib/prop_check/generator.rb', line 64

def sample(num_of_samples = 10, **kwargs)
  num_of_samples.times.map do
    call(**@@default_kwargs.merge(kwargs))
  end
end

#where(&condition) ⇒ Object

Creates a new Generator that only produces a value when the block ‘condition` returns a truthy value.



138
139
140
141
142
143
144
145
146
147
 
# File 'lib/prop_check/generator.rb', line 138

def where(&condition)
  map do |result|
    # if condition.call(*result)
    if PropCheck::Helper.call_splatted(result, &condition)
      result
    else
      :"_PropCheck.filter_me"
    end
  end
end

#with_config ⇒ Object

Turns a generator returning ‘x` into a generator returning `[x, config]` where `config` is the current `PropCheck::Property::Configuration`. This can be used to inspect the configuration inside a `#map` or `#where` and act on it.

>> example_config = PropCheck::Property::Configuration.new(default_epoch: Date.new(2022, 11, 22))
>> generator = Generators.choose(0..100).with_config.map { |int, conf| Date.jd(conf[:default_epoch].jd + int) }
>> generator.call(size: 10, rng: Random.new(42), config: example_config)
=> Date.new(2023, 01, 12)



129
130
131
132
133
134
 
# File 'lib/prop_check/generator.rb', line 129

def with_config
  Generator.new do |**kwargs|
    result = generate(**kwargs)
    result.map { |val| [val, kwargs[:config]] }
  end
end