Object Classes

The fundamental classes that functions are built up from are described below.

Most of these classes are defined in mathics.builtin.base or mathics.core.expression.

The below are heuristics:

  • To define a Mathics constant based on a Sympy constant, e.g. Infinity use SympyConstant

  • To define Mathics constants based on a mpmath constant, e.g. Glaisure, use MPMathConstant

  • To define a Mathics constant based on a numpy constant, use NumpyConstant

  • To define a Mathics functions based on a Sympy function, e.g. Sqrt, use SympyFunction

  • To define a Mathics operator use UnaryOperator, PrefixOperator, PostfixOperator, or BinaryOperator depending on the type of operator that is being defined

  • To define a Mathics function which returns a Boolean value e.g. MatchQ use Test

  • To define a Mathics function that doesn’t fall into a category above, e.g. Attributes use Builtin

  • To define a Mathics variable e.g. $TimeZone or Mathics Symbols, e.g. True use Predefined

  • To define a Mathics atomic expression, e.g. ImageAtom use AtomicBuiltin

Atom Class

Recall that an Expression to be evaluated is kind of S-expression called and ExpressionList, where each list item is either itself an ExpressionList or an object in a class derived from Atom.

The Atom class we encountered earlier when describing the nodes that get created intially from a parse. However there are a few other kinds of Atoms or fundamention objects that can appear in an Evaluation list. These are

  • CompiledCode

  • Image

Symbol Class

Just above the Atom class is the Symbol which is an atomic element of an Expression. See Atomic Elements of Expressions.

Symbols are like Lisp symbols: you can think of them as starting out as a string name. But in contrast to a string, once set the symbol is set, it is immutable and indivisable.

Builtin class

A number of Mathics variables and functions are loaded when Mathics starts up, thousands of functions even before any Mathics packages are loaded. As with other Mathics objects like Atom and Symbol, Mathics variables and functions are implemented through Python classes.

The reason that we use a class for a Mathics variable or a Mathics function is so that we can give those Mathics object properties and attributes.

At the lowest level of the class hierarchy is Builtin.

Lets look at a simple one:

class Head(Builtin):
        <dd>returns the head of the expression or atom $expr$.

    >> Head[a * b]
     = Times
    >> Head[6]
     = Integer
    >> Head[x]
     = Symbol

    def apply(self, expr, evaluation):

        return expr.get_head()

In the above, we have not defined an evaluation() method explicitly so we get Expressions’s built-in evaluation() method.

A feature of the Builtin class is the convention that its provides a convention by which “apply” methods of the class can be matched using the method’s name which must start with “apply” and a pattern listed in the method’s doc string. This is used in the example above.

Here, Head has one paramater which is called expr. Note that in the Python method there is also expr variable it its method signature which is listed right after the usual self method that you find on all method functions.

At the end is an evaluation parameter and this contains definitions and the context if the method needs to evaluate expressions.

Predefined Class

Just above Builtin in the Mathics object class hierarchy is Predefined.

Some Mathics values like True are derived from Predefined. For example:

class True_(Predefined):
      <dd>represents the Boolean true value.

    name = "True"

In the above, note that the class name has an underscore (_) appended it. We do this so as not to conflict with the Python value True. The class variable name is used to associate the Mathics name.

A number of Mathics variables like $ByteOrdering are also derived directly from the Predefined class. Since Python class names cannot start with a dollar sign ($), we drop off the leading $, in the class name, and that gives us: ByteOrdering.

As with the True example shown above, the Mathics name is set using class variable name defined in the ByteOrdering class. For example:

class ByteOrdering(Predefined):
     <dd>returns the native ordering of bytes in binary data on your computer system.
   name = "$ByteOrdering"

 def evaluate(self, evaluation) -> Integer:
     return Integer(1 if sys.byteorder == "big" else -1)

The evaluate() function above is called to get the value of variable $ByteOrdering.


PrefixOperator and PostFixOperator

BinaryOperator and UnaryOperator

SympyConstant, MPMathConstant, and NumpyConstant

SympyFunction and MPMathFunction