.. currentmodule:: gwpy.plot

##################
Log scales in GWpy
##################

GWpy overrides the default :mod:`axis tick formatters <matplotlib.ticker>`
for logarithmic scales to one with the following logic:

- ``0.1 <= x <= 1000`` and ``x`` has less than 2 decimal places, just
  represent with ``'%s'``
- otherwise use the default formatting ``'%s^{%.2f}' % (base, logx)``

This results in plots that have slightly nicer default ticks:

.. plot::
   :include-source:
   :context: reset

   >>> import gwpy  # <-- import anything from gwpy to active custom formatter
   >>> import numpy
   >>> from matplotlib import pyplot
   >>> fig = pyplot.figure()
   >>> ax = fig.gca()
   >>> ax.plot(numpy.arange(.1, 5, step=.1))
   >>> ax.set_xscale('log')
   >>> ax.set_yscale('log')
   >>> fig.show()

Here you can see the X-axis formats with the updated ticks, showing '1'
and '10' instead of '10 :sup:`0`' and '10 :sup:`1`', while the Y-axis --
because it has ticks below 0.01 -- uses the standard scientific notation.

The new log formatter can be disabled by manually setting the major and
minor formatters for log-scaled axes to
:class:`~matplotlib.ticker.LogFormatterMathtext`:

.. plot::
   :include-source:
   :context: close-figs

   >>> from matplotlib.ticker import LogFormatterMathtext
   >>> ax.yaxis.set_major_formatter(LogFormatterMathtext())