Add support for field definition with a predicate

[fmbenhassine]

As of v3.8, in order to define a field, one should use the FieldDefinition API. This API provides 5 methods to define field properties (named, ofType, inClass, isAnnotatedWith and hasModifiers).

As a library user, this limited set of properties makes it impossible to use a custom condition on a field, for example field.isSynthetic().

As a library developer, I will need to update the API each time a user asks to add a custom field property.

A better approach is to use a Predicate<Field> as a replacement for FieldDefinition. Thanks to java.util.function.Predicate#and and java.util.function.Predicate#or, multiple predicates can be combined to define a field just like FieldDefinition, adding the flexibility to define custom conditions without having to change the API each time. Here is an example:

Predicate<Field> predicate = named("name").and(ofType(String.class)).and(inClass(Person.class));

Random Beans would provide common predicates and let the user define additional predicates as needed. A first work on this feature can be reviewed in branch predicates.

@PascalSchumacher WDYT?

BTW, this idea came from a situation where I wanted to exclude all types defined in a given package. I though about adding a TypeDefinition API allowing one to write something like:

EnhancedRandom enhancedRandom = new EnhancedRandomBuilder()
	        .exclude(TypeDefinition.type().inPackage("io.github.benas.randombeans.beans").get())
	        .build();

But then I found myself exposing all properties of a Class. I quickly realized that this was not the best approach.. Hence the idea of using predicates: No need to expose properties + ability to provide custom conditions.

[PascalSchumacher]

Good idea. 👍

[fmbenhassine]

@PascalSchumacher Thank you for your quick feedback!

Rebased and merged as 4929f12.