Handlers for IPythonDirective’s @doctest pseudo-decorator.
The Sphinx extension that provides support for embedded IPython code provides a pseudo-decorator @doctest, which treats the input/output block as a doctest, raising a RuntimeError during doc generation if the actual output (after running the input) does not match the expected output.
An example usage is:
.. ipython::
In [1]: x = 1
@doctest
In [2]: x + 2
Out[3]: 3
One can also provide arguments to the decorator. The first argument should be the name of a custom handler. The specification of any other arguments is determined by the handler. For example,
.. ipython::
@doctest float
In [154]: 0.1 + 0.2
Out[154]: 0.3
allows the actual output 0.30000000000000004 to match the expected output due to a comparison with np.allclose.
This module contains handlers for the @doctest pseudo-decorator. Handlers should have the following function signature:
handler(sphinx_shell, args, input_lines, found, submitted)
where sphinx_shell is the embedded Sphinx shell, args contains the list of arguments that follow: '@doctest handler_name’, input_lines contains a list of the lines relevant to the current doctest, found is a string containing the output from the IPython shell, and submitted is a string containing the expected output from the IPython shell.
Handlers must be registered in the doctests dict at the end of this module.
Simplistic converter of strings from repr to float NumPy arrays.
If the repr representation has ellipsis in it, then this will fail.
| Parameters: | s : str
|
|---|
Examples
>>> s = "array([ 0.3, inf, nan])"
>>> a = str_to_array(s)
Doctest which allow the submitted output to vary slightly from the input.
Here is how it might appear in an rst file:
.. ipython::
@doctest float
In [1]: 0.1 + 0.2
Out[1]: 0.3