num_parser: a math interpreter and evaluator

num_parser allows you to easily parse strings into math expressions and evaluate them.

Features

• Binary and unary operators
• Supports multiple value types:
  • Bool,
  • Int,
  • Float,
  • Complex,
  • Vector
• Built-in functions
• Built-in constants
• User-defined functions: f(x,y) = xsin(y)+ysin(x)
• User-defined var: a = pi/2 or b = a+2
• Define you own functions with macros.
• Understands ambiguous syntax, like: g(x) = pisinx
• Recursion: f(x) = branch(x<=2, 1, f(x-1)+f(x-2))
• Serde support
• No panicking

Much more will be implemented in future releases!

Use Guide

Evaluating simple static expressions:

use num_parser::*;

assert_eq!(eval("2+2").unwrap(), Value::from(4));
assert_eq!(eval("sin(pi)").unwrap(), Value::from(0));
assert_eq!(eval("re(10+3i)").unwrap(), Value::from(10));

Using contexts:

use num_parser::*;

let mut context = Context::default();
// Declaring a function
let res = eval_with_mutable_context(
    "f(x) = branch(x<=2, 1, f(x-1) + f(x-2))",
    &mut context
).unwrap();

// Result is None
assert_eq!(res, None);
// Calling the function. We could just use eval_with_static_context at this point
let res = eval_with_mutable_context("f(10)", &mut context).unwrap();

assert_eq!(res, Some(Value::from(55)));

Values

Values are contained inside the Value enum, which provides useful functions to access the contained data:

use num_parser::Value;

let value = Value::Float(1.0);

assert_eq!(value.as_bool().unwrap(), true);
assert_eq!(value.as_int().unwrap(), 1);
assert_eq!(value.as_float().unwrap(), 1.0);
assert_eq!(value.as_complex().unwrap(), num::complex::Complex::new(1.0, 0.0));
assert_eq!(value.as_vector(), vec![Value::Float(1.0)]);

// Assign type implicitly:
let implicit = Value::from(1.0);

assert_eq!(value, implicit);

Note that, even thought the initial value was a float, it has been cast into ints and bools. This was possible since the value had no decimal part and it was a one. If these conditions were not met, the cast would have failed.

Operators

Binary operators:

Unary operators:

Functions

| branch | 3 (condition, true, false) | Returns the second argument if the condition is true, the third if it is false. | | sin | 1 | Returns the sine of the angle. | | cos | 1 | Returns the cosine of the angle. | | tan | 1 | Returns the tangent of the angle. | | asin | 1 | Returns the arcsine of the angle. | | acos | 1 | Returns the arccosine of the angle. | | atan | 1 | Returns the arctangent of the angle. | | sinh | 1 | Returns the hyperbolic sine of the angle. | | cosh | 1 | Returns the hyperbolic cosine of the angle. | | tanh | 1 | Returns the hyperbolic tangent of the angle. | | asinh | 1 | Returns the hyperbolic arcsine of the angle. | | acosh | 1 | Returns the hyperbolic arccosine of the angle. | | atanh | 1 | Returns the hyperbolic arctangent of the angle. | | re | 1 | Returns the natural part of the number. | | im | 1 | Returns the imaginary part of the number. | | polar | 1 | Returns the polar form (r, theta) of the complex number. | | arg | 1 | Returns the principal arg of the number. | | norm | 1 | Returns the length of the vector (re, im). |

Context

Contexts allows you keep track of user-defined functions and variables, as well as settings. They can be created as follows:

use num_parser::*;

let mut my_context = Context::new(settings::Rounding::Round(8));

Serde

You can use the optional feature serde_support to let all the public structs derive Serialize and Deserialize.

[dependencies]
num = { version = "<version>", features = [ "serde_support" ] }

License and contribution

num_parser is licensed under a MIT License.

Feel free to open issues and pull requests for any problems or ideas you come up with.