Whilst originally designed to be used to represent and determine connection weights in StateNet, Expressions is a C# library used for writing and evaluating expressions in a given context.
Expressions can be used to write simple expressions:
var Expression = new ConstantInteger<TContext>(1) //An expression representing the int value 1As well as more complex nested expressions:
var Expression = new And<TContext>( //An expression following the usual boolean logic of the 'and' operator
new GreaterThan<int, TContext>( //An expression for the comparison operator >, this will evaluate to true as 4 > 1
new ConstantInteger<TContext>(4), new ConstantInteger<TContext>(1)),
new EqualTo<int, TContext>(new ConstantInteger<TContext>(1), new ConstantInteger<TContext>(1))); //An expression for the equality operator.To evaluate these expressions into the correct context, the Expressions library utilises the Interpreter design pattern; Every expression type has an .Interpret(context) method. When called on a given context, any non-terminal expressions will recursively call .Interpret(context) (passing through the same context) on the expressions within them until a terminal expression is reached. Terminal expressions can then be evaluated to their respective values and value types in the given context. These in turn are passed on to the containing non-terminal expressions until the expression is completely evaluated.
Building complex expressions manually can be a bit messy syntactically and so there is also an ExpressionFactory class.
public readonly ExpressionFactory<TContext> _ex = new ExpressionFactory<TContext>();
var one = _ex.Int(1); //Creates a new ConstantInteger expression with value 1
var true = new _ex.Bool(true); //Creates a new ConstantBool expression with value true
var SevenIsGreaterThanFive = new _ex.GreaterThan<int>(_ex.Int(7), _ex.Int(5)); //This is much tidier than the basic implementationA more complicated example highlights the improvement in conciseness:
public readonly ExpressionFactory<TContext> _expressions = new ExpressionFactory<TContext>();
var fibListExpression = _expressions.List(new int[] { 1, 1, 2, 3, 5, 8 });
for(int i = 0; i < 20; i )
{
fibListExpression = _expressions.List(_expressions.ConditionalList(_expressions.LessThan(_expressions.Last(fibListExpression), _expressions.Int(100)),
_expressions.Append(fibListExpression,
_expressions.Add(_expressions.First(_expressions.TakeLast(fibListExpression, _expressions.Int(2))),
_expressions.Last(_expressions.TakeLast(fibListExpression, _expressions.Int(2))))),
fibListExpression).Interpret(_context)); //An expression that when intepreted will add the next number in the fibonacci sequence to the list as long as that number is less than 100 and return the list as an expression or will just return the list as an expression if the next number is greater than 100.
}
var fibLessThan100 = fibListExpression.Interpret(_context); //Interpreting the expression above will return a list of containing the numbers in the Fibonacci sequence less than 100.In the above example operators on our expressions such as _expressions.Add() are to the left of a pair of expressions (lhs, rhs) that is acted on by an operator i.e
var addEx = _expressions.Add(_expressions.Int(2), _expressions.Int(2);Using the roman alphabet we are more used to reading from left to right and so the above expression can still be a little tricky to read, especially when the arguments are also complicated expressions. To improve this we also have a more fluent API for our operators. With this the above expression becomes:
var addEx = _expressions.Int(2).Add(_expressions.Int(2));Constant expressions of any type can be created using the generic ConstantExpression:
var ConstantExpression = new ConstantExpression<TType, TContext>(TType a); //An expression 'a' of some generic type And also using the ExpressionFactory:
public readonly ExpressionFactory<TContext> _expressions = new ExpressionFactory<TContext>();
var ConstantExpression = _expressions.Expression<TType>(TType a);We can also create various type specific constant expressions:
var ConstantFloatEx = _expressions.Float(3.14f); // An expression representing the float value 3.14
var ConstantColorEx = _expressions.Color(System.Drawing.Color.Red); // An expression representing the color red
var ConstantGuidEx = _expressions.Guid(Guid.NewGuid()); // An expression representing a constant guidWe can also make constant lists that can be used to represent expressions of lists of any generic type:
var ConstantListExpression = _expressions.List<TType>(TType[] list); //An expression representing a list of some generic typeArithmetic operations can act on expressions of any type with the GenericArithmeticOperators, though care must be exercised to ensure the operators are implemented on the given type properly:
var AddExpression = new Add<TType, TContext>(IExpression<TType, TContext> a, IExpression<TType, TContext> b); //An expression representing addition on the expressions a & b: a bWith the fluent API and ExpressionFactory we also have:
public readonly ExpressionFactory<TContext> _expressions = new ExpressionFactory<TContext>();
var SubtractExpression = _ex.Expression<TType>(a).Subtract(_ex.Expression<TType>(b)); //An expression representing subtraction on the expressions a & b: a - b
var MultiplyExpression = _ex.Expression<TType>(a).Multiply(_ex.Expression<TType>(b));; //An expression representing multiplication on the expressions a & b: a * bAbove we can see that the type can be inferred by the operator but if we want to be more explicit again there are also type specific variations of these operators:
var AddFloatExpression = _expressions.Float(2.72f).AddFloat(_expressions.Float(1.41f)); //An expression representing addition of two floats: 2.72 1.41
var SubtractDecimalExpression = _expressions.Decimal(2.6m).SubtractDecimal(_expressions.Decimal(1.9m)); //An expression respresenting subtraction of the right float from the left: 2.6 - 1.3
var MultiplyDoubleExpression =_expressions.Double(1.2).MultiplyDouble(_expressions.Double(3.4)); //An expression representing the multiplication of two doubles: 1.2 * 3.4There is also the special case of string concatenation that can be considered as the addition operator acting on string expressions:
var ConcatStringExpression = _expressions.String(foo).ConcatString(_expressions.String(bar)); //An expression representing the concatenation (addition) of two string expressions: 'foo' 'bar'Similarly to the arithmetic operators, expressions with boolean relational operators can be made on any given type, though - again - care must be taken to ensure the operators are properly implemented on the type:
var GreaterThanExpression = new GreaterThan<TContext>(IExpression<TType, TContext> a, IExpression<TType, TContext> b); // An expression representing the comparison 'a > b'Using the the fluent API and ExpressionFactory:
public readonly ExpressionFactory<TContext> _expressions = new ExpressionFactory<TContext>();
var LessThanExpression = _expressions.Expression<TType>(a).LessThan(_expressions.Expression<TType>(b)); //An expression representing the comparison a < b
var GreaterThanOrEqualToExpression = _expressions.Expression<TType>(a).GreaterThan(_expressions.Expression<TType>(b)); // An expression representing the comparison a >= b
var LessThanOrEqualToExpression = _expressions.Expression<TType>(a).LessThanOrEqualTo(_expressions.Expression<TType>(b)); // An expression representing the comparison a <= bSimilarly, expressions with boolean equality operators can be made on any given type:
var EqualToExpression = _expressions.Expression<TType>(a).EqualTo(_expressions.Expression<TType>(b)); //An expression represent the comparison 'a == b'
var NotEqualToExpression = _expressions.Expression<TType>(a).NotEqualTo(_expressions.Expression<TType>(b)); //An expression represent the comparison 'a != b'For boolean expressions we have the usual boolean logical operators. Using the fluent API and ExpressionFactory:
public readonly ExpressionFactory<TContext> _expressions = new ExpressionFactory<TContext>();
var OrExpression = _expressions.Bool(true).Or(_expressions.Bool(false)); //An expression representing the boolean expression 'true OR false'
var NotExpression = _expressions.Bool(true).Not(); //An expression representing the boolean expression 'NOT true'
var AndExpression = _expressions.Bool(true).And(_expressions.Bool(false)); //An expression representing the boolean expression 'true AND false'
var XOrExpression = _expressions.Bool(true).XOr(_expressions.Bool(false)); //An expression representing the boolean expression 'true XOR false'There are also the All and Any operations that are equivalent to the boolean logic operations NAND and NOR, respectively:
var AllExpression = _expressions.Bool(true).All(_expressions.Bool(true), _expressions.Bool(false)); //An expression respresenting the boolean expression 'true AND true AND false'
var AnyExpression = _expressions.Bool(true).Any(_expressions.Bool(true), _expressions.Bool(false)); //An expression respresenting the boolean expression 'true OR true OR false'For list expressions we also have some of the usual list operations. Using the fluent API and ExpressionFactory:
public readonly ExpressionFactory<TContext> _expressions = new ExpressionFactory<TContext>();
var list1 = new TType[] { a, b };
var list2 = new TType[] { c, d };
var ConcatListExpression = _expressions.List(list1).ConcatList(list2); //A list expression representing the concatenation of two list expressions 'list1 list2'
var FirstExpression = _expressions.List(list1).First(); //An expression representing the first item in the list
var LastExpression = _expressions.List(list1).Last(); //An expression representing the last item in the list
var TakeFirstExpression = _expressions.List(list1).TakeFirst(_expressions.Int(n)); //A list expression of the first n items in 'list1'
var TakeLastExpression = _expressions.List(list1).TakeLast(_expressions.Int(m)); //A list expression of the last m items in 'list1'
var CountExpression = _expressions.List(list1).Count(); //An integer expression representing the number of items in the list