.. module:: mlpy.wavelet

:py:mod:`mlpy.wavelet` - Wavelet Transform
==========================================

Padding
-------

.. autofunction:: pad


Discrete Wavelet Transform
--------------------------

Discrete Wavelet Transform based on the GSL DWT [Gsldwt]_.

For the forward transform, the output is the discrete wavelet transform
:math:`f_i \rightarrow w_{j,k}` in a packed triangular storage layout, where :math:`j`
is the index of the level :math:`j = 0 \dots J-1` and :math:`k` is the index
of the coefficient within each level, :math:`k = 0 \dots (2^j)-1`.
The total number of levels is :math:`J = \log_2(n)`. The output data 
has the following form,

.. math::

   (s_{-1,0}, d_{0,0}, d_{1,0}, d_{1,1}, d_{2,0}, \dots, d_{j,k}, ..., d_{J-1,2^{J-1}-1})

where the first element is the smoothing coefficient :math:`s_{-1,0}`,
followed by the detail coefficients :math:`d_{j,k}` for each level :math:`j`. 
The backward transform inverts these coefficients to obtain the original data. 

.. note::

   from GSL online manual (http://www.gnu.org/software/gsl/manual/)

.. autofunction:: dwt(x, wf, k, centered=False)
.. autofunction:: idwt(X, wf, k, centered=False)


Undecimated Wavelet Transform
-----------------------------

Undecimated Wavelet Transform (also known as stationary wavelet transform,
redundant wavelet transform, translation invariant wavelet transform,
shift invariant wavelet transform or Maximal overlap wavelet transform)
based on the "wavelets" R package.

.. autofunction:: uwt(x, wf, k, levels=0)
.. autofunction:: iuwt(X, wf, k)

.. autofunction:: uwt_align_h2
.. autofunction:: uwt_align_d4

Continuous Wavelet Transform
----------------------------

Continuous Wavelet Transform based on [Torrence98]_.

.. autofunction:: cwt
.. autofunction:: icwt

.. autofunction:: autoscales
.. autofunction:: fourier_from_scales
.. autofunction:: scales_from_fourier


.. [Torrence98] C Torrence and G P Compo. Practical Guide to Wavelet Analysis
.. [Gsldwt] Gnu Scientific Library, http://www.gnu.org/software/gsl/

Example (requires matplotlib)

.. code-block:: python

   >>> import numpy as np
   >>> import matplotlib.pyplot as plt
   >>> import mlpy.wavelet as wave
   >>> x = np.random.sample(512)
   >>> scales = wave.autoscales(N=x.shape[0], dt=1, dj=0.25, wf='dog', p=2)
   >>> X = wave.cwt(x=x, dt=1, scales=scales, wf='dog', p=2)
   >>> fig = plt.figure(1)
   >>> ax1 = plt.subplot(2,1,1)
   >>> p1 = ax1.plot(x)
   >>> ax1.autoscale_view(tight=True)
   >>> ax2 = plt.subplot(2,1,2)
   >>> p2 = ax2.imshow(np.abs(X), interpolation='nearest')
   >>> plt.show()

.. image:: images/cwt.png