.. DO NOT EDIT.
.. THIS FILE WAS AUTOMATICALLY GENERATED BY SPHINX-GALLERY.
.. TO MAKE CHANGES, EDIT THE SOURCE PYTHON FILE:
.. "_build/auto_examples/tutorial/03-Preparing_inputs.py"
.. LINE NUMBERS ARE GIVEN BELOW.

.. only:: html

    .. note::
        :class: sphx-glr-download-link-note

        :ref:`Go to the end <sphx_glr_download__build_auto_examples_tutorial_03-Preparing_inputs.py>`
        to download the full example code.

.. rst-class:: sphx-glr-example-title

.. _sphx_glr__build_auto_examples_tutorial_03-Preparing_inputs.py:


Preparing inputs
================

.. GENERATED FROM PYTHON SOURCE LINES 7-8

By default, the functions to be benchmarked with :func:`pyquickbench.run_benchmark` are expected to take as arguments a single integer ``n`` as in the following code.

.. GENERATED FROM PYTHON SOURCE LINES 8-37

.. code-block:: Python



    def comprehension(n):
        return ['' for _ in range(n)]

    def star_operator(n):
        return ['']*n

    def for_loop_append(n):
        l = []
        for _ in range(n):
            l.append('')
    
    all_funs = [
        comprehension   ,
        star_operator   ,
        for_loop_append ,
    ]

    n_bench = 12
    all_sizes = [2**n for n in range(n_bench)]

    pyquickbench.run_benchmark(
        all_sizes   ,
        all_funs    ,
        show = True ,
    ) 





.. image-sg:: /_build/auto_examples/tutorial/images/sphx_glr_03-Preparing_inputs_001.png
   :alt: 03 Preparing inputs
   :srcset: /_build/auto_examples/tutorial/images/sphx_glr_03-Preparing_inputs_001.png
   :class: sphx-glr-single-img





.. GENERATED FROM PYTHON SOURCE LINES 70-72

Most often however, the functions to be benchmarked take data as input that is more complex than a simple integer and a setup phase is needed. In the following example, we want to compare different implementations of array summation algorithms.


.. GENERATED FROM PYTHON SOURCE LINES 72-95

.. code-block:: Python


    import numpy as np
    import math

    def builtin_sum(x):
        return sum(x)

    def np_sum(x):
        return np.sum(x)

    def m_fsum(x):
        return math.fsum(x)

    all_funs = [
        builtin_sum ,
        np_sum      ,
        m_fsum      ,
    ]


    n_bench = 12
    all_sizes = np.array([2**n for n in range(n_bench)])








.. GENERATED FROM PYTHON SOURCE LINES 101-102

Here, we want the benchmark to measure the time taken by the different implementations as a function of the size of the input. The ``setup`` argument expects a callable that provides arguments for our different implementations to be timed. For instance, ``setup`` can return a dictionnnary ``kwargs``, so that each function to be benchmarked will be called as ``fun(**kwargs)``.

.. GENERATED FROM PYTHON SOURCE LINES 102-113

.. code-block:: Python


    def setup(n):
        x = np.random.random(n)
        return {'x' : x}

    pyquickbench.run_benchmark(
        all_sizes       ,
        all_funs        ,
        setup = setup   ,
        show = True     ,
    ) 



.. image-sg:: /_build/auto_examples/tutorial/images/sphx_glr_03-Preparing_inputs_002.png
   :alt: 03 Preparing inputs
   :srcset: /_build/auto_examples/tutorial/images/sphx_glr_03-Preparing_inputs_002.png
   :class: sphx-glr-single-img





.. GENERATED FROM PYTHON SOURCE LINES 117-118

This calling convention is why positional-only arguments are disallowed in :func:`pyquickbench.run_benchmark`. For instance, even though the following defines a legal Python 3.8+ function

.. GENERATED FROM PYTHON SOURCE LINES 118-124

.. code-block:: Python


    def pos_only_fun(n,/):
        return n

    print(pos_only_fun(42))





.. rst-class:: sphx-glr-script-out

 .. code-block:: none

    42




.. GENERATED FROM PYTHON SOURCE LINES 125-126

It is not allowed in :func:`pyquickbench.run_benchmark` since the following raises an error:

.. GENERATED FROM PYTHON SOURCE LINES 126-133

.. code-block:: Python


    try:
        print(pos_only_fun(n=42))
    except TypeError as err:
        print(f'TypeError: {err}')
    
    




.. rst-class:: sphx-glr-script-out

 .. code-block:: none

    TypeError: pos_only_fun() got some positional-only arguments passed as keyword arguments: 'n'




.. GENERATED FROM PYTHON SOURCE LINES 134-135

This comes with hardly any loss of generality since it is possible to wrap these positional-only arguments functions.

.. GENERATED FROM PYTHON SOURCE LINES 135-142

.. code-block:: Python


    def wrap_fun(n):
        return pos_only_fun(n)

    print(wrap_fun(42))
    print(wrap_fun(n=42))

    



.. rst-class:: sphx-glr-script-out

 .. code-block:: none

    42
    42





.. _sphx_glr_download__build_auto_examples_tutorial_03-Preparing_inputs.py:

.. only:: html

  .. container:: sphx-glr-footer sphx-glr-footer-example

    .. container:: sphx-glr-download sphx-glr-download-jupyter

      :download:`Download Jupyter notebook: 03-Preparing_inputs.ipynb <03-Preparing_inputs.ipynb>`

    .. container:: sphx-glr-download sphx-glr-download-python

      :download:`Download Python source code: 03-Preparing_inputs.py <03-Preparing_inputs.py>`

    .. container:: sphx-glr-download sphx-glr-download-zip

      :download:`Download zipped: 03-Preparing_inputs.zip <03-Preparing_inputs.zip>`


.. only:: html

 .. rst-class:: sphx-glr-signature

    `Gallery generated by Sphinx-Gallery <https://sphinx-gallery.github.io>`_