Wrong documentation about Cartesian product constraint

[Alexander Keijzer]

The indexing of cartesian products is off by one. In the Python documentation and Rust docs the index starts at 0.
However, the code asserts that the index ends at the length of the vector (and therefore starts at 1):

optimization-engine/src/constraints/cartesian_product.rs

Line 140 in d9c6dde

 assert!(x.len() == self.dimension(), "x has wrong size"); 

And then loops until index-1 (and therefore my final index should be N not N-1):

optimization-engine/src/constraints/cartesian_product.rs

Line 146 in d9c6dde

 c.project(&mut x[j..i]); 

This should probably be changed in the docs as changing the code will break all implementations. However, starting the index at 1 is quite unintuitive when used in Rust and should be documented well.