xtensor-python
xtensor-python copied to clipboard
Python bindings for xtensor
Hi, The new added `xt::quantile` seems not to play nicely with `pytensor`. For me, the following : ```cpp xt::pytensor arr {{ 5.3, 4.3, 5.3 }, { 4.2, 4.2, 5.2 },...
xternsor-python is header only library, so prefer arch-independed paths cmake. References: * https://github.com/xtensor-stack/xtl/commit/d877d94836aff4d0f727acf3eaab8f4880ecb625 * https://github.com/xtensor-stack/xtensor/commit/7738389861044c9618c7d59fb7602f7dddc1df7f
Haven't tried it, but I'm assuming changes will be needed for numpy > 2.0 API?
Hello, I suppose it is a feature request. I want to call a python function provided by a user from my C++ code. The function argument is numpy array the...
Similar to https://github.com/xtensor-stack/xtensor-blas/pull/243 and done for the same reason (easier to integrate with `meson`).
Is there a way to install xtensor-python without conda?
Is there any current support for 64 bit indexing or plans to support it in the future?
- remove cmake\FindNumpy.cmake - provide fix find_package(Python COMPONENTS Interpreter Development.Module NumPy REQUIRED) in CMakeLists.txt
https://github.com/xtensor-stack/xtensor-python/blob/a7288ce1cfa8d1e58f91fd32fb3dad165625f20d/include/xtensor-python/xtensor_type_caster_base.hpp#L303