3 stable releases

1.0.2 Aug 12, 2022

#1871 in Parser implementations

MIT license

155KB
3.5K SLoC

num_parser: a math interpreter and evaluator

crate license docs

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:

Operator Description Precedence
^ Exponentiation 90
/ Division 70
* Multiplication 70
% Modulo 70
+ Sum 60
- Subtraction 60
< Less than 50
> Greater than 50
<= Less or equal to 50
>= Greater or equal to 50
== Equal to 40
!= Not equal to 40
&& Logical AND 30
|| Logical OR 20
, Aggregation. Creates vectors 10
= Assignment. Used for functions and vars declarations 0

Unary operators:

Operator Description Precedence
! Logical NOT 80
- Negation 60

Functions

Function Parameters Amount Description
min >=1 Returns the minimum value.
max >=1 Returns the maximum value.
floor 1 Returns the greatest lower integer.
ceil 1 Returns the lowest greater integer.
round 1 Returns the rounded integer.
ln 1 Returns the natural log of the number.
log 2 (base, arg) Returns the logarithm of the number with the specified base.
exp 1 Returns e^(arg).
rand 2 (min, max) Returns a random float between the two number specified.
abs 1 Returns the absolute value of a number.
sqrt 1 Returns the square root of a number.
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::*;

// Generate the default context
let mut default = Context::default();

// Generate a custom context
let mut custom = Context::new(
    settings::Rounding::NoRounding,
    settings::AngleUnit::Degree,
    settings::DepthLimit::NoLimit
);

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.

Dependencies

~2MB
~37K SLoC