guillaumelenoir/wavepal

Frequency and time-frequency analysis for irregularly sampled time series

wavepal's Issues

freq_analysis(*args, computes_amplitude=True) throwing an error

Hello
for following command I am getting this error (please see below)
"x.freq_analysis(freqstep=0.0001,D=600.,percentile=percentile,n_moments=12,mywindow=7,computes_amplitude=True)"

############ command line outputs#########

Re-estimated D factor (WOSA): 600.1687
Preliminary steps for the WOSA periodogram:
100%|██████████| 9/9 [00:02<00:00, 4.23it/s]
Re-estimated frequency range: from 0.00175540168582 to 0.127055640806
Main loop, over the frequencies:
0%| | 0/1254 [00:00<?, ?it/s]
Traceback (most recent call last):
File "spectrum_by_wavepal_method.py", line 59, in
x.freq_analysis(freqstep=0.0001,D=600.,percentile=percentile,n_moments=12,mywindow=7,computes_amplitude=True)
File "build/bdist.linux-x86_64/egg/wavepal/Wavepal.py", line 1158, in freq_analysis
TypeError: norm() got an unexpected keyword argument 'axis'

please suggest me what to do?

Thanks
Phanindra

Python 3 Compatibility

Hello,

Thanks for your efforts for the project. I wondered if there is a Python 3 compatible version or some quick guidelines to make it Python 3 compatible?

Thank you.

Initial Update

Hi 👊

This is my first visit to this fine repo, but it seems you have been working hard to keep all dependencies updated so far.

Once you have closed this issue, I'll create seperate pull requests for every update as soon as I find one.

That's it for now!

Happy merging! 🤖

CARMA(p>1,q>=0) throwing an error at secind round

Hi there,
I am having uneven spaced time series with temporal resolution 0.1 year to 4.5year (1800 data points). I am very sure that my data has quasi-periodic and/or non-stationary periodicities.

Trying to apply CARMA(p,q) { eg: p>1, 1<=q<=p-1} and then wanna apply wavepal for wavelets.
This is my present error....!!!!!
Any help, please...
Thank you.

Note1: I Understand that the roots of Auto Regressive polynomial has Imaginary roots....!!
and also Importantly, I see Boost.Python.ArgumentError: Python argument types in
did not match C++ signature:

is this because of the BOOST version..?

Note2: I have used the commands that were presented in your example codes Lenoir 2018 paper-1.

############### Error..............................###############
Calculating PSD Lorentzian parameters...
Calculating coefficients of AR polynomial...
Calculating sigma...
Calculating log-likelihoods...
Decorrelation length (in number of samples) - Estimation: 7

SECOND ROUND: generates 70000 samples

---

Traceback (most recent call last):
File "spectrum_by_wavepal_method.py", line 60, in
x.carma_params(p=2,q=0,path_to_figure_folder=mypath,nmcmc=nmcmc,signif_level_type="n",make_carma_fig=False,nbins=15,dpi=400)
File "build/bdist.linux-x86_64/egg/wavepal/Wavepal.py", line 650, in carma_params
File "/usr/lib64/python2.7/site-packages/carmcmc/carma_pack.py", line 84, in run_mcmc
self.p, self.q, ntemperatures, False, nthin, init)
Boost.Python.ArgumentError: Python argument types in
carmcmc._carmcmc.run_mcmc_carma(numpy.int64, numpy.int64, vecD, vecD, vecD, int, int, int, bool, int, vecD)
did not match C++ signature:
run_mcmc_carma(int, int, std::vector<double, std::allocator >, std::vector<double, std::allocator >, std::vector<double, std::allocator >, int, int, int)
run_mcmc_carma(int, int, std::vector<double, std::allocator >, std::vector<double, std::allocator >, std::vector<double, std::allocator >, int, int, int, bool)
run_mcmc_carma(int, int, std::vector<double, std::allocator >, std::vector<double, std::allocator >, std::vector<double, std::allocator >, int, int, int, bool, int)
run_mcmc_carma(int, int, std::vector<double, std::allocator >, std::vector<double, std::allocator >, std::vector<double, std::allocator >, int, int, int, bool, int, std::vector<double, std::allocator >)

Regarding the installation in Ubuntu and macOS M1

Hi,

I was trying to install this package in Ubuntu and MACOS M1, but I couldn't. In the end, I got an error message regarding the carmcmc package that was not found. Do I need to install it separately? What do I do? Please help me with the installation of this in any machine. Any help would be appreciated.

Wishes,

Ajay

Test on data error

On Ubuntu 18, installed boost 1.59 and appears to build ok, but fails during tests, the last part of the output is:

SECOND ROUND: generates 16000 samples

---

Running sampler...
Number of steps added: 1
Number of tracked steps added: 1
Setting starting values...
Drawn from priors
...Initializing 0.2699
0.9582
-0.0003
-5.7993

Burning in... (8000 iterations)

0% 10 20 30 40 50 60 70 80 90 100%
|----|----|----|----|----|----|----|----|----|----|
**************************************************Average RAM Acceptance Rate is 0.275125
*

Sampling...

0% 10 20 30 40 50 60 70 80 90 100%
|----|----|----|----|----|----|----|----|----|----|

---

Calculating sigma...
Calculating log-likelihoods...
Traceback (most recent call last):
File "test.py", line 25, in
x.carma_params(make_carma_fig=True,nbins=20,dpi=400,path_to_figure_folder=path_to_figure_folder)
File "build/bdist.linux-x86_64/egg/wavepal/Wavepal.py", line 653, in carma_params
TypeError: assess_fit() takes at most 4 arguments (12 given)

how to supress legend, ylable,xlabel of frequency analysis output

I am unable to supress the legend, ylabel, xlabel texts in frequency analysis

"
plot_periodogram=x.plot_periodogram(xaxis='period')
plot_periodogram.suptitle("")
"
no use

Ubuntu installation with a Conda environment

Hi,

I've tried to install Wavepal on my Ubuntu machine in a conda environment that runs Python 2.7, but I've encountered numerous errors in the terminal stating that, amongst other warning, "error: Failed building wheel for acor", "error: could not build wheels or acor, which is required to install pyproject.toml-based projects". it is also stating to avoid running setup.py directly and to use pypa/build, while also stating that carmama and wavepal are not available despite them being downloaded onto the machine. Please let me know what I need to change in order for me to properly install the software.