Linux build and bug fix

Author: JohanMabille

docs/source/array_tensor.rst

@@ -0,0 +1,27 @@
+.. Copyright (c) 2016, Johan Mabille and Sylvain Corlay
+
+   Distributed under the terms of the BSD 3-Clause License.
+
+   The full license is in the file LICENSE, distributed with this software.
+
+Arrays and tensors
+==================
+
+``xtensor-python`` provides two container types wrapping numpy arrays: ``pyarray`` and ``pytensor``. They are the counterparts
+to ``xarray`` and ``xtensor`` containers.
+
+pyarray
+-------
+
+Like ``xarray``, ``pyarray`` has a dynamic shape. This means that you can reshape the numpy array on the C++ side and see this
+change reflected on the python side. ``pyarray`` doesn't make a copy of the shape or the strides, but reads them each time it
+is needed. Therefore, if a reference on a ``pyarray`` is kept in the C++ code and the corresponding numpy array is then reshaped
+in the python code, this modification will reflect in the ``pyarray``.
+
+pytensor
+--------
+
+Like ``xtensor``, ``pytensor`` has a static stack-allocated shape. This means that the shape of the numpy array is copied into
+the shape of the ``pytensor`` upon creation. As a consequence, reshapes are not reflected across languages. However, this drawback
+is offset by a more effective computation of shape and broadcast.
+

docs/source/usage.rst docs/source/basic_usage.rstdocs/source/usage.rst renamed to docs/source/basic_usage.rst

@@ -4,14 +4,11 @@
 
    The full license is in the file LICENSE, distributed with this software.
 
-Usage
-=====
-
 Basic Usage
------------
+===========
 
 Example 1: Use an algorithm of the C++ library on a numpy array inplace
-^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+-----------------------------------------------------------------------
 
 **C++ code**
 
@@ -58,7 +55,7 @@ Example 1: Use an algorithm of the C++ library on a numpy array inplace
 
 
 Example 2: Create a universal function from a C++ scalar function
-^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+-----------------------------------------------------------------
 
 **C++ code**
 
@@ -105,52 +102,3 @@ Example 2: Create a universal function from a C++ scalar function
      [-1.499227,  0.136731,  1.646979,  1.643002,  0.128456],
      [-1.084323, -0.583843,  0.45342 ,  1.073811,  0.706945]]
 
-
-Getting started with xtensor-cookiecutter
------------------------------------------
-
-`xtensor-cookiecutter`_ helps extension authors create Python extension modules making use of xtensor.
-
-It takes care of the initial work of generating a project skeleton with
-
-- A complete ``setup.py`` compiling the extension module
-- A few examples included in the resulting project including
-
-    - A universal function defined from C++
-    - A function making use of an algorithm from the STL on a numpy array
-    - Unit tests
-    - The generation of the HTML documentation with sphinx
-
-Usage
-^^^^^
-
-Install cookiecutter_
-
-.. code::
-
-    pip install cookiecutter
-
-After installing cookiecutter, use the xtensor-cookiecutter_:
-
-.. code::
-
-    cookiecutter https://github.com/QuantStack/xtensor-cookiecutter.git
-
-As xtensor-cookiecutter runs, you will be asked for basic information about
-your custom extension project. You will be prompted for the following
-information:
-
-- ``author_name``: your name or the name of your organization,
-- ``author_email`` : your project's contact email,
-- ``github_project_name``: name of the GitHub repository for your project,
-- ``github_organization_name``: name of the GithHub organization for your project,
-- ``python_package_name``: name of the Python package created by your extension,
-- ``cpp_namespace``: name for the cpp namespace holding the implementation of your extension,
-- ``project_short_description``: a short description for your project.
-  
-This will produce a directory containing all the required content for a minimal extension
-project making use of xtensor with all the required boilerplate for package management,
-together with a few basic examples.
-
-.. _xtensor-cookicutter: https://github.com/QuantStack/xtensor-cookiecutter
-.. _cookiecutter: https://github.com/audreyr/cookiecutter

docs/source/cookie_cutter.rst

@@ -0,0 +1,54 @@
+.. Copyright (c) 2016, Johan Mabille and Sylvain Corlay
+
+   Distributed under the terms of the BSD 3-Clause License.
+
+   The full license is in the file LICENSE, distributed with this software.
+
+Getting started with xtensor-cookiecutter
+=========================================
+
+`xtensor-cookiecutter`_ helps extension authors create Python extension modules making use of xtensor.
+
+It takes care of the initial work of generating a project skeleton with
+
+- A complete ``setup.py`` compiling the extension module
+- A few examples included in the resulting project including
+
+    - A universal function defined from C++
+    - A function making use of an algorithm from the STL on a numpy array
+    - Unit tests
+    - The generation of the HTML documentation with sphinx
+
+Usage
+-----
+
+Install cookiecutter_
+
+.. code::
+
+    pip install cookiecutter
+
+After installing cookiecutter, use the xtensor-cookiecutter_:
+
+.. code::
+
+    cookiecutter https://github.com/QuantStack/xtensor-cookiecutter.git
+
+As xtensor-cookiecutter runs, you will be asked for basic information about
+your custom extension project. You will be prompted for the following
+information:
+
+- ``author_name``: your name or the name of your organization,
+- ``author_email`` : your project's contact email,
+- ``github_project_name``: name of the GitHub repository for your project,
+- ``github_organization_name``: name of the GithHub organization for your project,
+- ``python_package_name``: name of the Python package created by your extension,
+- ``cpp_namespace``: name for the cpp namespace holding the implementation of your extension,
+- ``project_short_description``: a short description for your project.
+  
+This will produce a directory containing all the required content for a minimal extension
+project making use of xtensor with all the required boilerplate for package management,
+together with a few basic examples.
+
+.. _xtensor-cookicutter: https://github.com/QuantStack/xtensor-cookiecutter
+.. _cookiecutter: https://github.com/audreyr/cookiecutter

docs/source/index.rst

@@ -53,7 +53,9 @@ This software is licensed under the BSD-3-Clause license. See the LICENSE file f
    :caption: USAGE
    :maxdepth: 2
 
-   usage
+   basic_usage
+   array_tensor
+   cookie_cutter
 
 .. toctree::
    :caption: API REFERENCE

include/xtensor-python/pyarray.hpp

@@ -272,13 +272,13 @@ namespace xt
     template <class T>
     inline void pyarray<T>::init_from_python()
     {
-        m_shape = inner_shape_type(reinterpret_cast<size_type*>(PyArray_SHAPE(python_array())),
-                                   static_cast<size_type>(PyArray_NDIM(python_array())));
-        m_strides = inner_strides_type(reinterpret_cast<size_type*>(PyArray_STRIDES(python_array())),
-                                       static_cast<size_type>(PyArray_NDIM(python_array())));
+        m_shape = inner_shape_type(reinterpret_cast<size_type*>(PyArray_SHAPE(this->python_array())),
+                                   static_cast<size_type>(PyArray_NDIM(this->python_array())));
+        m_strides = inner_strides_type(reinterpret_cast<size_type*>(PyArray_STRIDES(this->python_array())),
+                                       static_cast<size_type>(PyArray_NDIM(this->python_array())));
         m_backstrides = backstrides_type(*this);
-        m_data = container_type(reinterpret_cast<pointer>(PyArray_DATA(python_array())),
-                                static_cast<size_type>(PyArray_SIZE(python_array())));
+        m_data = container_type(reinterpret_cast<pointer>(PyArray_DATA(this->python_array())),
+                                static_cast<size_type>(PyArray_SIZE(this->python_array())));
     }
 
     template <class T>

include/xtensor-python/pycontainer.hpp

@@ -234,7 +234,7 @@ namespace xt
     template <class D>
     inline void pycontainer<D>::fill_default_strides(const shape_type& shape, strides_type& strides)
     {
-        strides_type::value_type data_size = 1;
+        typename strides_type::value_type data_size = 1;
         for(size_type i = strides.size(); i != 0; --i)
         {
             strides[i - 1] = data_size;
@@ -258,8 +258,8 @@ namespace xt
     template <class D>
     inline bool pycontainer<D>::check_(pybind11::handle h)
     {
-        int type_num = detail::numpy_traits<T>::type_num;
-        return PyArray_Check(h.ptr()) && PyArray_EquivTypenums(PyArray_TYPE(h.ptr()), type_num);
+        int type_num = detail::numpy_traits<value_type>::type_num;
+        return PyArray_Check(h.ptr()) && PyArray_EquivTypenums(PyArray_TYPE(python_array()), type_num);
     }
 
     template <class D>
@@ -269,7 +269,8 @@ namespace xt
             return nullptr;
 
         int type_num = detail::numpy_traits<value_type>::type_num;
-        auto res = PyArray_FromAny(ptr, PyArray_DescrFromType(type_num), 0, 0, NPY_ARRAY_ENSUREARRAY, nullptr);
+        auto res = PyArray_FromAny(ptr, PyArray_DescrFromType(type_num), 0, 0, 
+                                   NPY_ARRAY_ENSUREARRAY | NPY_ARRAY_FORCECAST, nullptr);
         return res;
     }
 

include/xtensor-python/pytensor.hpp

@@ -223,22 +223,22 @@ namespace xt
         m_shape = shape;
         m_strides = strides;
         compute_backstrides();
-        m_data = container_type(reinterpret_cast<pointer>(PyArray_DATA(python_array())),
-                                static_cast<size_type>(PyArray_SIZE(python_array())));
+        m_data = container_type(reinterpret_cast<pointer>(PyArray_DATA(this->python_array())),
+                                static_cast<size_type>(PyArray_SIZE(this->python_array())));
     }
 
     template <class T, std::size_t N>
     inline void pytensor<T, N>::init_from_python()
     {
-        if(PyArray_NDIM(python_array()) != N)
+        if(PyArray_NDIM(this->python_array()) != N)
             throw std::runtime_error("NumPy: ndarray has incorrect number of dimensions");
 
-        std::copy(PyArray_DIMS(python_array()), PyArray_DIMS(python_array()) + N, m_shape.begin());
-        std::transform(PyArray_STRIDES(python_array()), PyArray_STRIDES(python_array()) + N, m_strides.begin(),
+        std::copy(PyArray_DIMS(this->python_array()), PyArray_DIMS(this->python_array()) + N, m_shape.begin());
+        std::transform(PyArray_STRIDES(this->python_array()), PyArray_STRIDES(this->python_array()) + N, m_strides.begin(),
                 [](auto v) { return v / sizeof(T); });
         compute_backstrides();
-        m_data = container_type(reinterpret_cast<pointer>(PyArray_DATA(python_array())),
-                                static_cast<size_type>(PyArray_SIZE(python_array())));
+        m_data = container_type(reinterpret_cast<pointer>(PyArray_DATA(this->python_array())),
+                                static_cast<size_type>(PyArray_SIZE(this->python_array())));
     }
 
     template <class T, std::size_t N>