xtensor-stack / xtensor

C++ tensors with broadcasting and lazy computing
BSD 3-Clause "New" or "Revised" License
3.32k stars 396 forks source link

Output of example in ReadMe file is not matching with the current master #2790

Open faze-geek opened 4 months ago

faze-geek commented 4 months ago

Kindly checkout the Index Access example from https://github.com/xtensor-stack/xtensor?tab=readme-ov-file#basic-usage .

#include <iostream>
#include "xtensor/xarray.hpp"
#include "xtensor/xio.hpp"

int main(){
    xt::xarray<double> arr1
    {{1.0, 2.0, 3.0},
    {2.0, 5.0, 7.0},
    {2.0, 5.0, 7.0}};

    std::cout << arr1(0, 0) << std::endl;
}

Output

anurag@Anurags-MacBook-Air numpy % g++ -std=c++17 -o numpy src/numpy.cpp -Iinclude
anurag@Anurags-MacBook-Air numpy % ./numpy                                        
1           // wrong

1.0 is expected which is correctly shown in the ReadMe output.

I have also run the example on binder jupyter notebook to avoid any discrepancy caused locally.

spectre-ns commented 2 months ago

I suspect this is a rounding issue in std::cout if you use the following line it displays correctly:

   std::cout << std::fixed << std::setprecision(1) << arr1(0, 0);

This is not an xtensor issue. 1 or 1.0 are equal and that is the correct output.

faze-geek commented 2 months ago

I was not questioning the output, but with xtensor handling multiple dtypes getting 1. on print would be ideal. Let's make sure that the example prints in the same way as well to avoid any confusions.

It's a small refactor, maybe I can send in my first PR ?

  1. Update std::coutto use setprecision.
  2. Use a floating point number which won't be truncated like 1.5 instead of 1.0.
  3. Use xt::view(arr, 0, 0) for the indexing instead of arr(0, 0). This prints 1. as expected.