[−][src]Crate vec_utils
vec-utils
This is an experimental crate that adds some helpful functionality to Vec<T>
, like map
and zip_with
. These functions allow you to transform a vec and try and reuse the allocation if possible!
use vec_utils::VecExt; fn to_bits(v: Vec<f32>) -> Vec<u32> { v.map(|x| x.to_bits()) } fn sum_2(v: Vec<f32>, w: Vec<f64>) -> Vec<f64> { v.zip_with(w, |x, y| f64::from(x) + y) }
But zip_with
is limited to taking only a single additional vector. To get around this limitation, this crate also exports some macros that can take an arbitrary number of input vectors, and in most cases will compile down to the same assembly as Vec::map
and Vec::zip_with
(sometimes with some additional cleanup code, but even then the macro solution is just as fast as the built-in version).
You can use the zip_with
and try_zip_with
macros like so,
use vec_utils::{zip_with, try_zip_with}; fn to_bits(v: Vec<f32>) -> Vec<u32> { zip_with!(v, |x| x.to_bits()) } fn sum_2(v: Vec<f32>, w: Vec<f64>) -> Vec<f64> { zip_with!((v, w), |x, y| f64::from(x) + y) } fn sum_5(a: Vec<i32>, b: Vec<i32>, c: Vec<i32>, d: Vec<i32>, e: Vec<i32>) -> Vec<i32> { zip_with!((a, b, c, d, e), |a, b, c, d, e| a + b + c + d + e) } fn mul_with(a: Vec<i32>) -> Vec<i32> { zip_with!((a, vec![0, 1, 2, 3, 4, 5, 6, 7]), |a, x| a * x) } fn to_bits_no_nans(v: Vec<f32>) -> Result<Vec<u32>, &'static str> { try_zip_with!(v, |x| if x.is_nan() { Err("Found NaN!") } else { Ok(x.to_bits()) }) }
You can use as many input vectors as you want, just put them all inside the input tuple. Note that the second argument is not a closure, but syntax that looks like a closure, i.e. you can't make a closure before-hand and pass it as the second argument. Also, you can't use general patterns in the "closure"'s arguments, only identifiers are allowed. You can specify if you want a move closure by adding the move keyword in from of the "closure".
use vec_utils::zip_with; fn add(a: Vec<i32>, b: i32) -> Vec<i32> { zip_with!(a, move |a| a + b) }
It also adds some functionality to reuse the allocation of a Box<T>
, using the BoxExt
/UninitBox
api.
use vec_utils::BoxExt; fn replace(b: Box<i32>, f: f32) -> Box<f32> { Box::drop_box(b).init(f) } fn get_and_replace(b: Box<i32>, f: f32) -> (Box<f32>, i32) { let (b, x) = Box::take_box(b); (b.init(f), x) }
Macros
try | Unwraps a result or propagates its error. |
try_zip_with | A macro to give syntactic sugar for |
zip_with | A wrapper around |
Structs
Input | A type that contains useful meta-data about a the Vec<_> that it was created from |
NoneError | The error type that results from applying the try operator ( |
Output | An write only buffer that may overlap with some input buffer
this allows reuse of that input buffer to turn it into a
|
UninitBox | An uninitialized piece of memory |
Traits
BoxExt | Extension methods for |
Try | A stable version of |
Tuple | A specialized const-list for emulating varaidic generics |
TupleElem | This trait abstracts away elements of the input stream |
VecExt | Extension methods for |
Functions
try_zip_with_impl | Does the work of the |