Implement the Add/Div/Mul/Sub blocks

[AlexSanches1]

No description provided.

[SamuelSarle]

LGTM otherwise but I suggest to change the while self.input.recv { ..., self.output.send(&res) } to while self.input.recv { ... } self.output.send(&result.unwrap_or(<default-value>)) with <default-value> being 0.0 for Add and Sub blocks, and 1.0 for Mul and Div blocks.
Reasoning for the defaults comes from the identity elements of addition or multiplication, e.g. adding/subtracting 0 does nothing and multiplying/diving by 1 does nothing and thus they're used as the values when the input set is empty, adding no numbers together equals 0 and multiplying no numbers together equals 1.

Also noting that as a future improvement, it's quite easy to switch these implementations to support any numerical types (f32,f64,u32,u64, etc.) by defining the blocks with generics as such:

pub struct Add<T: Message + core::ops::Add<Output = T> + num_traits::Zero = f64> { /* ... */ }
pub struct Sub<T: Message + core::ops::Sub<Output = T> + num_traits::Zero = f64> { /* ... */ }
pub struct Mul<T: Message + core::ops::Mul<Output = T> + num_traits::One = f64> { /* ... */ }
pub struct Div<T: Message + core::ops::Div<Output = T> + PartialEq + num_traits::One + num_traits::Zero = f64> { /* ... */ }

In summary what those generics are saying is, taking for example the Add block:

1. Anything that implements Message (so that we can send it over a port)
2. And implements core::ops::Add (so that we can call x + y on the types)
3. And implements num_traits::Zero (so that we can call T::zero() for the identity element as the starting value for the summing)
4. = f64, by default use f64 if nothing else was provided.

Let me know if you want the patch for those changes for this PR, otherwise I'll open another PR after this is merged.