Crate mexprp [] [src]


A math expression parsing and evaluating library


  • f64 precision
  • multiple/arbitrary precision (somewhat incomplete)
  • low dependencies
  • custom variable contexts
  • custom function contexts
  • builtin constants and functions (eg pi, sin, max)
  • implicit multiplication
  • utf8-ready
  • support for multiple answers
  • complex numbers (somewhat incomplete)


There are three different ways to parse and evaluate an equation.

1. With eval()

This function parses and evaluates a string all at once with the default context. There's also an eval_ctx() function which takes a reference to a Context as well that will be used instead of the default Context. The type parameter can be anything that implements the Num trait. Some Num types support more operations than others. More info about Nums can be found in the Num module.

mexprp::eval::<f64>("10 / (2 + 3)"); // Ok(Answer::Single(2.0))

2. With Expression

Expression::parse() parses a string into a tree representation (a Term). It can also be parsed with a context with parse_ctx(), and it will store that context within it for future evaluations. It can also be evaluated with a reference to any other context with eval_ctx. It's important to ensure that the custom context contains any definitions the Expression depends on.

let expr: Expression<f64> = Expression::parse("3 ^ 4 / 9").unwrap();
let res = expr.eval(); // Ok(Answer::Single(9.0))

3. With Term

A Term is an Expression, but without the extra overhead of a context or the original string representation stored with it. It is literally a tree representing the equation by it's operations.

let term: Term<f64> = Term::parse("10 ^ -3").unwrap();
let res = term.eval(); // Ok(Answer::Single(0.001))

Answer Types

Evaluating an expression will return an Answer enum. An answer represents either a single value, or multiple. The most notable example of an operation that results in multiple answers is sqrt() which returns a positive and negative answer. Another obvious example is the ± operator. When implementing functions, it's important to handle each answer type when evaluating the arguments. More info about that and helper methods for it can be found in the documentation for the Answer enum.

Multiple Precisions

MEXPRP supports evaluating expressions with different precisions with the Num trait. Currently supported number types are

However, the implementation for certain types is incomplete. Only the f64 type fully implements all of the operations. Complex is the next best, but even it is still missing some. The others only implement a (small) subset of the functionality of the Num trait, and return a MathError::Unimplemented when an unsupported operation is attempted. It is hopeful that more functions will be implemented in the future, but some are very difficult to implement for arbitrary precision numbers.

The Complex number also supports selecting the precision to use with a Context. Set the precision field of the cfg field of a Context to set the precision to be used by Complex numbers.

For more info on the types, see the documentation for the num module.

To use another number type, change the type annotation(s) for your MEXPRP types.

extern crate rug;
use rug::Rational;
mexprp::eval::<Rational>("10/15"); // 2/3
let expr: Expression<ComplexFloat> = Expression::parse("(3 + 4i) × (6 - 3i)").unwrap();
let res = expr.eval(); // 30 + 15i

To set the precision of types that let you choose it (currently just

In case you don't want a dependency on rug, compile MEXPRP without the "rug" feature.

Using Contexts

You can evaluate expressions with custom variable and function definition's by defining a context. When defining custom functions, it's important to remember to parse the expression with the custom context, or else the parser will recognize your functions as variables instead. Expressions will store the context you parse them with, but you have to evaluate Terms with a reference to a context using Term::eval_ctx. For more info see the Context struct.

A Context also holds configuration values that define how MEXPRP parses and evaluates equations. These configuration values include enabling/disabling implicit multiplication, the precision to use for types that support selecting precisions (just Complex for now), and the behaviour of the sqrt() function. More info can be found in the API docs (check the context module).


pub use errors::EvalError;
pub use errors::MathError;
pub use errors::ParseError;
pub use num::Num;



All the errors


Number representation(s) This module contains the Num trait and its implementations.



Struct that holds configuration values used when evaluating expressions


A context holds values for variables and functions to be used in expressions. It is useful for both parsing and evaluation expressions. During parsing, all names will be treated as variables unless present in the Context the expression is being parsed with as functions at the time. The default context (created with new()) contains basic functions and constants such as sin, pi, etc, as well as the default configuration.


The main Expression struct. Contains the string that was originally requested to be parsed, the context the Expression was parsed with, and the Term the raw form was parsed as. For just the parsed version of the expression, use the Term enum.



An answer of an evaluatation. Can be either a single answer or multiple. This struct contains some helper methods for performing operations on single or multiple answers. The op method takes another Num, and a function with two Num arguments, itself and the other (as references). It performs that function on all combinations and returns an answer with all of the results in one. The unop function is similar but it performs an operation on only itself, without another value (unary operation).


The main representation of parsed equations. It is an operand that can contain an operation between more of itself. This form is the only one that can be directly evaluated. Does not include it's own context.



Implemented by functions defined in a context



Parse and evaluate a string


Parse and evaluate a string with the given context

Type Definitions


The result of an evaluation