Cleanup temporary shape allocation

Author: SylvainCorlay

README.md

@@ -28,15 +28,14 @@ conda install -c conda-forge xtensor-python
 
 xtensor-python offers two container types wrapping numpy arrays inplace to provide an xtensor semantics
 
- - ``pytensor``
- - ``pyarray``.
+ - `pytensor`
+ - `pyarray`.
 
 Both containers enable the numpy-style APIs of xtensor (see [the numpy to xtensor cheat sheet](http://xtensor.readthedocs.io/en/latest/numpy.html)).
 
- - On the one hand, ``pyarray`` has a dynamic number of dimensions. Just like numpy arrays, it can be reshaped with a shape of a different length (and the new shape is reflected on the python side).
+ - On the one hand, `pyarray` has a dynamic number of dimensions. Just like numpy arrays, it can be reshaped with a shape of a different length (and the new shape is reflected on the python side).
 
- - On the other hand ``pytensor`` has a compile time number of dimensions, specified with a template parameter. Shapes of ``pytensor`` instances are stack allocated, making ``pytensor`` a significantly
-faster expression than ``pyarray``.
+ - On the other hand `pytensor` has a compile time number of dimensions, specified with a template parameter. Shapes of `pytensor` instances are stack allocated, making `pytensor` a significantly faster expression than `pyarray`.
 
 ### Example 1: Use an algorithm of the C++ standard library on a numpy array inplace.
 

include/xtensor-python/pyarray.hpp

@@ -397,7 +397,7 @@ namespace xt
 
         this->m_ptr = tmp.release().ptr();
         init_from_python();
-        std::copy(rhs.data().begin(), rhs.data().end(), this->data().begin());
+        std::copy(rhs.data().cbegin(), rhs.data().cend(), this->data().begin());
     }
 
     /**
@@ -425,6 +425,7 @@ namespace xt
     inline pyarray<T>::pyarray(const xexpression<E>& e)
         : base_type()
     {
+        // TODO: prevent intermediary shape allocation
         shape_type shape = forward_sequence<shape_type>(e.derived_cast().shape());
         strides_type strides = make_sequence<strides_type>(shape.size(), size_type(0));
         compute_strides(shape, layout_type::row_major, strides);

include/xtensor-python/pytensor.hpp

@@ -306,7 +306,7 @@ namespace xt
         : base_type()
     {
         init_tensor(rhs.shape(), rhs.strides());
-        std::copy(rhs.data().begin(), rhs.data().end(), this->data().begin());
+        std::copy(rhs.data().cbegin(), rhs.data().cend(), this->data().begin());
     }
 
     /**

include/xtensor-python/pyvectorize.hpp

@@ -9,6 +9,8 @@
 #ifndef PY_VECTORIZE_HPP
 #define PY_VECTORIZE_HPP
 
+#include <type_traits>
+
 #include "pyarray.hpp"
 #include "xtensor/xvectorize.hpp"
 
@@ -20,13 +22,13 @@ namespace xt
     {
         xvectorizer<Func, R> m_vectorizer;
 
-        template <class F>
+        template <class F, class = std::enable_if_t<!std::is_same<std::decay_t<F>, pyvectorizer>::value>>
         pyvectorizer(F&& func)
             : m_vectorizer(std::forward<F>(func))
         {
         }
 
-        inline pyarray<R> operator()(const pyarray<Args>&... args)
+        inline pyarray<R> operator()(const pyarray<Args>&... args) const
         {
             pyarray<R> res = m_vectorizer(args...);
             return res;