Context manager that creates a named temporary file and returns its name.

All parameters are passed on to ``tempfile.mkstemp``, see its documentation for more info.

Converts a NumPy array or matrix into a contiguous memoryview object that is suitable to be forwarded to the Graph constructor.

This is used internally to allow us to use a NumPy array or matrix directly when constructing a Graph.

Rescales a list of numbers into a given range.

`out_range` gives the range of the output values; by default, the minimum of the original numbers in the list will be mapped to the first element in the output range and the maximum will be mapped to the second element. Elements between the minimum and maximum values in the input list will be interpolated linearly between the first and second values of the output range.

`in_range` may be used to override which numbers are mapped to the first and second values of the output range. This must also be a tuple, where the first element will be mapped to the first element of the output range and the second element to the second.

If `clamp` is ``True``, elements which are outside the given `out_range` after rescaling are clamped to the output range to ensure that no number will be outside `out_range` in the result.

If `scale` is not ``None``, it will be called for every element of `values` and the rescaling will take place on the results instead. This can be used, for instance, to transform the logarithm of the original values instead of the actual values. A typical use-case is to map a range of values to color identifiers on a logarithmic scale. Scaling also applies to the `in_range` parameter if present.

Examples:

>>> rescale(range(5), (0, 8))
[0.0, 2.0, 4.0, 6.0, 8.0]
>>> rescale(range(5), (2, 10))
[2.0, 4.0, 6.0, 8.0, 10.0]
>>> rescale(range(5), (0, 4), (1, 3))
[-2.0, 0.0, 2.0, 4.0, 6.0]
>>> rescale(range(5), (0, 4), (1, 3), clamp=True)
[0.0, 0.0, 2.0, 4.0, 4.0]
>>> rescale([0]*5, (1, 3))
[2.0, 2.0, 2.0, 2.0, 2.0]
>>> from math import log10
>>> rescale([1, 10, 100, 1000, 10000], (0, 8), scale=log10)
[0.0, 2.0, 4.0, 6.0, 8.0]
>>> rescale([1, 10, 100, 1000, 10000], (0, 4), (10, 1000), scale=log10)
[-2.0, 0.0, 2.0, 4.0, 6.0]

Tries to interpret a string as an orientation value.

The following basic values are understood: ``left-right``, ``bottom-top``, ``right-left``, ``top-bottom``. Possible aliases are:

• ``horizontal``, ``horiz``, ``h`` and ``lr`` for ``left-right``
• ``vertical``, ``vert``, ``v`` and ``tb`` for top-bottom.
• ``lr`` for ``left-right``.
• ``rl`` for ``right-left``.

``reversed_horizontal`` reverses the meaning of ``horizontal``, ``horiz`` and ``h`` to ``rl`` (instead of ``lr``); similarly, ``reversed_vertical`` reverses the meaning of ``vertical``, ``vert`` and ``v`` to ``bt`` (instead of ``tb``).

Returns one of ``lr``, ``rl``, ``tb`` or ``bt``, or throws ``ValueError`` if the string cannot be interpreted as an orientation.

Returns consecutive pairs of items from the given iterable.

When `circular` is ``True``, the pair consisting of the last and first elements is also returned.

Example:

>>> list(consecutive_pairs(range(5)))
[(0, 1), (1, 2), (2, 3), (3, 4)]
>>> list(consecutive_pairs(range(5), circular=True))
[(0, 1), (1, 2), (2, 3), (3, 4), (4, 0)]
>>> list(consecutive_pairs([]))
[]
>>> list(consecutive_pairs([], circular=True))
[]
>>> list(consecutive_pairs([0]))
[]
>>> list(consecutive_pairs([0], circular=True))
[(0, 0)]

Safer variant of ``max()`` that returns a default value if the iterable is empty.

Example:

>>> safemax([-5, 6, 4])
6
>>> safemax([])
0
>>> safemax((), 2)
2

Safer variant of ``min()`` that returns a default value if the iterable is empty.

Example:

>>> safemin([-5, 6, 4])
-5
>>> safemin([])
0
>>> safemin((), 2)
2

Approximates the machine epsilon value for doubles.