ct6502 / wavelets Goto Github PK

I suggest adding the topics wavelet, wavelets in the About section.

Dear Chris,

I think I may have stumbled over a minor bug in the python
function waveletFunctions.py: If you run the code on a time series of uneven
length (e.g. len(x)=101), it returns "ValueError: operands could not be
broadcast together with shapes (101,) (100,)" as the Fourier transform f
and the daughter have different sizes. You may check this via

import numpy as np
from waveletFunctions import wavelet

x = np.random.randn(101)
dt = 1.0
w = wavelet ( x, dt )

The problem is in the construction of the wave vector k. Here, the last
entry of the kplus vector is always omitted (kplus[0:-1]). However, in
case of uneven length of the time series, the Fourier transform returns
no value for the Nyquist frequency and therefore the last entry of wave
vector kplus should be used as well. I found that the following lines
replacing lines 115-119 solve this problem:

#....construct wavenumber array used in transform [Eqn(5)]
kplus = np.arange(1, n/2+1)
kplus = (kplus * 2 * np.pi / (n * dt))
kminus = (-(kplus[0:(n-1)/2])[::-1])
k = np.concatenate(([0.], kplus, kminus))

Here, I simply replaced kplus[0:-1] by kplus[0:(n-1)/2]. The np.fix
function is actually not needed as dividing by integer automatically
takes care of rounding down.

Best regards,
Mitch

Hello, thank you very much for your contribution to enable me to use Python for wavelet analysis. I have a question that I would like to ask, and I wonder if you are comfortable answering it.
In your example data, the time granularity is seasonal, i.e. 0.25 years, and it is continuous full year data. Now I have a time series with a daily granularity, but only observations are available from 1 February to 30 November each year. There are two ways I can think of to do wavelet analysis on such data.

1. fill all the days of the year for which there is no data with zeros, and then perform a wavelet analysis.
2. Define the number of days in the year and define the period for which data is actually available as a year, i.e. 303 days. Afterwards perform a wavelet analysis and apply the definition of a year to all cycles of the analysis.
   My personal preference is to use the second approach, but is this theoretically feasible?

Dear Developers,

Thank you for your code! I have been using it for chemistry applications, it was great!
I have been using your code by customizing it for my purpose in the past using the fortran release. Since now g77 compiler has been discontinued, I would like to use gfortran to compile it. I think there might be issues with the cfftpack.f and chisqr.f libraries.
Is there any updated version that can be compiled with gfortran or some suggestion to compile the code with new compilers?

Thank you very much in advance