Many applications require off-chain generation of random numbers for efficiency, security, etc.

This class allows you to generate a stream of deterministic, high-quality,
cryptographically secure random numbers.

By seeding it with a Chainlink VRF result that is requested only once for the project,
it can be used to demonstrate that the random numbers are not cherry-picked.

C++98 and above compiler.

Just copy and paste and include include/chainrand.h into your project.

chainrand::CRNG crng("base10(<RNG_VRF_RESULT>)" "<RNG_SEED_KEY>");
// prints 10 determinstic random numbers between [0, 1)
for (int i = 0; i < 10; ++i) {
    std::cout << crng() << "\n";
}

Compile with optimizations -march=native flag to enable usage of AES instructions.

Current and future versions of this library will generate the same stream of random numbers from the same seed.

template <class Str> CRNG(Str seed);

Creates an instance of the crng initialized with the seed.

Parameters:

• seed If empty, defaults to the empty string "".

Example:

chainrand::CRNG crng("base10(<RNG_VRF_RESULT>)" + "<RNG_SEED_KEY>");

template <class T> T nextUint();
uint8_t nextUint8();
uint16_t nextUint16();
uint32_t nextUint32();
uint64_t nextUint64();

Returns: Returns a single uniform random number within [0, (1<<(sizeof(T)*8))-1].

double nextDouble();
double operator() ();
double random();

Returns a single uniform random number within [0,1). The numbers are in multiples of 2**-53.

Parameters: none

Returns: Returns a single uniform random number within [0,1).

float nextFloat();

Returns a single uniform random number within [0,1). The numbers are in multiples of 2**-24.

Parameters: none

Returns: Returns a single uniform random number within [0,1).

int64_t randrange(int64_t start, int64_t stop, int64_t step);
int64_t randrange(int64_t start, int64_t stop);
int64_t randrange(int64_t stop);

Returns a random integer uniformly distributed in [start, stop).
The integers are spaced with intervals of |step|.

Parameters:

• start The start of the range. (optional, default=0)
• stop The end of the range.
• step The interval step. (optional, default=1)

Returns:

A random integer uniformly distributed in [start, stop).

Examples:

int64_t r;
r = crng.randrange(3); // returns a random number in {0,1,2}
r = crng.randrange(-3); // returns a random number in {0,-1,-2}
r = crng.randrange(0, 6, 2); // returns a random number in {0,2,4}
r = crng.randrange(5, 0, 1); // returns a random number in {5,4,3,2,1}
r = crng.randrange(5, -5, -2); // returns a random number in {5,3,1,-1,-3}

int64_t randint(int64_t start, int64_t stop);
int64_t randint(int64_t stop);

Returns a random integer uniformly distributed in [start, stop].
The integers are spaced with intervals of |step|.

Parameters:

• start The start of the range. (optional, default=0)
• stop The end of the range.

Returns:

A random integer uniformly distributed in [start, stop].

Examples:

int64_t r;
r = crng.randint(3); // returns a random number in {0,1,2,3}
r = crng.randint(-3); // returns a random number in {0,-1,-2,-3}
r = crng.randint(-3, 1); // returns a random number in {-3,-2,-1,0,1}
r = crng.randint(3, -1); // returns a random number in {3,2,1,0,-1}

template <class ChoicePointer, class PopulationIterator, class WeightsIterator>
bool choose(ChoicePointer choicePointer,
            PopulationIterator populationBegin,
            PopulationIterator populationEnd,
            WeightsIterator weightsBegin,
            WeightsIterator weightsEnd);

Chooses a random element from the population.

ChoicePointer, PopulationIterator, WeightsIterator
can be plain old pointer types, or pointer-like classes.

If weights is not provided, every element of population will be equally weighted.

If weights are provided,
the first min(populationEnd - populationBegin, weightsEnd - weightsBegin)
elements of the population will be considered.

If the sum of the weights is less than or equal to zero,
every element of population will be equally weighted.

Parameters:

• choicePointer[out] A pointer to the choosen element.
• populationBegin An iterator to the start of the population.
• populationEnd An iterator to the end of the population.
• weightsBegin An iterator to the start of the population. (optional)
• weightsEnd An iterator to the end of the population. (optional)

Returns:

Whether an element has been choosen.

Examples:

std::vector<int> population;
population.push_back(1);
population.push_back(2);
population.push_back(3);
int choice;
if (crng.choose(choice, population.begin(), population.end())) {
    std::cout << choice << "\n";
}

std::vector<double> weights;
weights.push_back(10);
weights.push_back(1);
weights.push_back(0.1);
if (crng.choose(choice, population.begin(), population.end(), 
                weights.begin(), weights.end())) {
    std::cout << choice << "\n";
}

template <class ChoicePointer, class Population, class Weights>
bool choose(ChoicePointer choicePointer,
            const Population &population,
            const Weights &weights);

Chooses a random element from the population.

Population, Weight can be std::vector,
or classes with the operator[] and size() methods.

If weights is not provided, every element of population will be equally weighted.

If weights are provided,
the first min(population.size(), weights.size())
elements of the population will be considered.

If the sum of the weights is less than or equal to zero,
every element of population will be equally weighted.

Parameters:

• choicePointer[out] A pointer to the choosen element.
• population A vector/array-like container of elements.
• weights A vector/array-like container of weights. (optional)

Returns:

Whether an element has been choosen.

Examples:

std::vector<int> population;
population.push_back(1);
population.push_back(2);
population.push_back(3);
int choice;
if (crng.choose(choice, population)) {
    std::cout << choice << "\n";
}

std::vector<double> weights;
weights.push_back(10);
weights.push_back(1);
weights.push_back(0.1);
if (crng.choose(choice, population, weights)) {
    std::cout << choice << "\n";
}

template <class CollectedIterator, class PopulationIterator, class WeightsIterator>
size_t sample(CollectedIterator collectedBegin,
              PopulationIterator populationBegin,
              PopulationIterator populationEnd,
              size_t k,
              WeightsIterator weightsBegin,
              WeightsIterator weightsEnd);

Chooses k random elements from the population without replacement.

CollectedIterator, PopulationIterator, WeightsIterator
can be plain old pointer types, or pointer-like classes.

If k is more than the length of the population, only k elements will be returned.

If weights is not provided, every element of population will be equally weighted.

If weights are provided,
the first min(populationEnd - populationBegin, weightsEnd - weightsBegin)
elements of the population will be considered.

If the sum of the weights is less than or equal to zero,
every element of population will be equally weighted.

Parameters:

• collectedBegin[out] An iterator to the collected results.
• populationBegin An iterator to the start of the population.
• populationEnd An iterator to the end of the population.
• k The number of elements to choose. (optional, default=1)
• weightsBegin An iterator to the start of the weights. (optional)
• weightsEnd An iterator to the end of the weights. (optional)

Returns:

The number of elements choosen.

Examples:

const int k = 2;
std::vector<int> population, collected(k);
population.push_back(1);
population.push_back(2);
population.push_back(3);

crng.sample(collected.begin(), population.begin(), population.end(), k);

for (int i = 0; i < k; ++i)
    std::cout << collected[i] << " ";
std::cout << "\n";

std::vector<double> weights;
weights.push_back(10);
weights.push_back(1);
weights.push_back(0.1);
crng.sample(collected.begin(), population.begin(), population.end(), k, 
            weights.begin(), weights.end());

for (int i = 0; i < k; ++i)
    std::cout << collected[i] << " ";
std::cout << "\n";

template <class Collected, class Population, class Weights>
size_t sample(Collected &collected,
        const Population &population,
        size_t k,
        const Weights &weights);

Chooses k random elements from the population without replacement.

The Population and Weights can be std::vector,
or classes with the operator[] and size() methods.

If k is more than the length of the population, only k elements will be returned.

If weights is not provided, every element of population will be equally weighted.

If weights are provided,
the first min(populationEnd - populationBegin, weightsEnd - weightsBegin)
elements of the population will be considered.

If the sum of the weights is less than or equal to zero,
every element of population will be equally weighted.

Parameters:

• collected[out] A vector/array-like container of elements.
• population A vector/array-like container of elements.
• k The number of elements to choose. (optional, default=1)
• weights A vector/array-like container of weights. (optional)

Returns:

The number of elements choosen.

Examples:

const int k = 2;
std::vector<int> population, collected(k);
population.push_back(1);
population.push_back(2);
population.push_back(3);

crng.sample(collected.begin(), population.begin(), population.end(), k);

for (int i = 0; i < k; ++i)
    std::cout << collected[i] << " ";
std::cout << "\n";

std::vector<double> weights;
weights.push_back(10);
weights.push_back(1);
weights.push_back(0.1);
crng.sample(collected.begin(), population.begin(), population.end(), k, 
            weights.begin(), weights.end());

for (int i = 0; i < k; ++i)
    std::cout << collected[i] << " ";
std::cout << "\n";

template <class RandomAccessIterator>
void shuffle(RandomAccessIterator begin, RandomAccessIterator end);

Shuffles the elements in-place.

RandomAccessIterator can be plain old pointer type, or pointer-like class.

Parameters:

• begin[in/out] An iterator to the start of the sequence.
• end[in/out] An iterator to the end of the sequence.

template <class Vector> 
void shuffle(Vector &v);

Shuffles the elements in-place.

Vector can be std::vector,
or class with the operator[] and size() methods.

Parameters:

• v[in/out] A vector/array-like container of elements.

double gauss(double mu, double sigma);

Normal distribution, also called the Gaussian distribution.

Parameters:

• mu The mean. (optional, default=0.0)
• sigma The standard deviation. (optional, default=1.0)

Returns:

A random number from the Gaussian distribution.

MIT