Should I run C++ from Python?

Question

Python

there is a function that receives point A and a list of points, calculates the Euclidean distance from point A to all points in the array
of points A is many and the array of points is very large, so here I have a bottleneck

, everything is complicated by the fact that the point is a dictionary that represents a multidimensional space (1.0 - 2k) and at point A many spaces may not be defined

transferring a dictionary to a sheet is not desirable

Is it worth making this function in C ++?
after a little searching, I realized that it is not so simple
. If it is, then what is preferable?
in Boost.Python I never found what to download, there is a ton of binaries
PyCXX is preferable, but it doesn’t compile,
otherwise what is poorly described in the standard library
If I make this function in C++ will the memory consumption increase?

Answer 1

There are options:
- use PyPy . JIT. speeds up code execution.
- NumPy is a library for scientific computing.
- Cython - A superset of python for easy writing extensions.

Answer 2

If so, what is preferable?

More ways:
- SWIG (supports many languages, including Python);
- ctypes - creates a library in C with the necessary functionality, collect it and call functions from the created dll/so/pyd;
— Python C API .

If I make this function in C++ will the memory consumption increase?

More than likely it will decrease. Variables in Python require more memory than variables in C (eg due to metadata, etc.).

Answer 3

Write a library in C with the desired function. Memory consumption will increase by the size of the resulting library, but I understand that you mainly need speed. As an option, write in Cython, in this case it will be almost the same as with hands in C.