Fyodor User Guide - Introduction

This chapter is a quick introduction to the basics of Fyodor.

Generators

Random data is created by generators, in Fyodor these are objects implementing the Generator interface:

public interface Generator<T> {
    public T next();
}

As you can see it's a simple interface, classes implementing it are expected to act as never-ending iterators over an infinite collection, each call to next() produces a random object of the specified type. For example new StringGenerator(22) will give me back a Generator that will create a random 22 character long string with each call to next():

Generator<String> generator = new StringGenerator(22);

for (int i = 0; i < 10; i++){
    System.out.println(generator.next());
}

{VVR}%:K<R0gR^82r>nZ\(
NS}2N/z7gCY&rmxMcK|4-7
COM\*E?jwdi5?tt%mvNI_#
;\^BO]7{0r^2q!C@x91\k/
oN=Dg+yl2R\Td49+ifk927
2lJ-]ICoRi6^mJ^$-?(t-E
ORnt91y-A`{l6@fjBOikr|
PzjvT(:%+a09wlpA>8>;_i
9@o!|Q+Zf;L`c<{Lp.7?uM
~V1fpQSYJiDN-iD7(^Q&gT

RDG class

The RDG class (it stands for Random Data Generators) is Fyodor's public API and what you should normally use to interact with it. It consists solely of static methods that return Generators so in the example above I'd actually call RDG.string(22) rather than new StringGenerator(22). The benefits of using this are

• using RDG instead of creating new instances of Generators directly lets us manage changes to the underlying implementations without breaking the public API, as well as gracefully deprecating parts of the codebase as needed.
• you can (and should) extend it to provide your own implementations of Generator<T> on top of those provided by Fyodor.
• a cleaner and simpler way for libraries in other JVM languages to wrap Fyodor.

Reproducibility

Using a single framework to generate all your random data needs has the big benefit of being able to reproduce a single scenario by reporting and manipulating the seed used to initialise the source of randomness. Setting up a lot of test data at random is all well and good but if you find you're getting intermittent errors with obscure edge-cases you're going to be want to be able to reproduce them and if your data is being generated from all kinds of implementations of randomness that are independent from each other you're not going to be able to do that.

Ranges

Range<T> is a Fyodor class that holds a lower and upper endpoint that is commonly used to create Generators that provide a range of random values. It's currently quite simple with all instances acting as closed ranges (i.e. inclusive of lower and upper endpoints), the simplest way to create one being the closed(T, T) static method. For instance RDG.string(Range.closed(10, 20)) will return a generator that creates a random string between 10 and 20 characters in length with each call to next():

Genrator<String> generator = RDG.string(Range.closed(10,20));

for (int i = 0; i < 10; i++){
    System.out.println(generator.next());
}

?z#2!qudyYEt&N
boYXS'yymAtj
Ikm,BD$|5oy;`0cK[
E,,u16c:;z
k|I{X=0rD&wN4W3Wz1pt
#E%:%t4w1!J
5yVUM76k8-VB}
WkNKrhV:VfqR\<-4:
-0([A5}kI@VW/5$fUV
aqth37yvdm8[~&

Next Page: Number Generators