Dear SciPy developers, in my research field of Astronomy we often encounter timeseries with uneven temporal sampling. Such timeseries cannot be analyzed with the standard Fast Fourier Transform. The most frequently used tool to analyze such timeseries is the Lomb-Scargle periodogram, developed by Lomb (1976) and further extended by Scargle (1982). Currently, to my knowledge, SciPy does not include a routine to calculate such periodograms. Therefore I would like to contribute the following code which efficiently calculates the periodogram. It is licensed under the terms of the BSD open source license but I would be willing to release it under any other language if this is necessary for possible inclusion with SciPy. The routine is written in Fortran 95 and wrapped with f2py, but I would be willing to rewrite it in C if required. Also included is a Python test script. Please provide me with any feedback in case you would consider including this. Kind regards, Pim Schellart
On Mon, Dec 20, 2010 at 3:27 PM, Pim Schellart <P.Schellart@astro.ru.nl>wrote:
Dear SciPy developers,
in my research field of Astronomy we often encounter timeseries with uneven temporal sampling. Such timeseries cannot be analyzed with the standard Fast Fourier Transform. The most frequently used tool to analyze such timeseries is the Lomb-Scargle periodogram, developed by Lomb (1976) and further extended by Scargle (1982). Currently, to my knowledge, SciPy does not include a routine to calculate such periodograms. Therefore I would like to contribute the following code which efficiently calculates the periodogram. It is licensed under the terms of the BSD open source license but I would be willing to release it under any other language if this is necessary for possible inclusion with SciPy. The routine is written in Fortran 95 and wrapped with f2py, but I would be willing to rewrite it in C if required. Also included is a Python test script. Please provide me with any feedback in case you would consider including this.
Pim, I think an implementation of Lomb-Scargle and related algorithms would be a great addition to SciPy. With the 0.9 release coming soon, and also being holiday season, I'm not sure how soon anyone will be able to look at your code. Could you add this as an enhancement request to the Trac system? Here's the link: http://projects.scipy.org/scipy/wiki Create an account, and then create a new ticket. You can attach the code to the ticket. Best regards, Warren P.S. There is a python/numpy implementation here: http://www.astropython.org/snippet/2010/9/Fast-Lomb-Scargle-algorithm
On Wed, Dec 22, 2010 at 1:20 PM, Warren Weckesser <warren.weckesser@enthought.com> wrote:
On Mon, Dec 20, 2010 at 3:27 PM, Pim Schellart <P.Schellart@astro.ru.nl> wrote:
Dear SciPy developers,
in my research field of Astronomy we often encounter timeseries with uneven temporal sampling. Such timeseries cannot be analyzed with the standard Fast Fourier Transform. The most frequently used tool to analyze such timeseries is the Lomb-Scargle periodogram, developed by Lomb (1976) and further extended by Scargle (1982). Currently, to my knowledge, SciPy does not include a routine to calculate such periodograms. Therefore I would like to contribute the following code which efficiently calculates the periodogram. It is licensed under the terms of the BSD open source license but I would be willing to release it under any other language if this is necessary for possible inclusion with SciPy. The routine is written in Fortran 95 and wrapped with f2py, but I would be willing to rewrite it in C if required. Also included is a Python test script. Please provide me with any feedback in case you would consider including this.
Pim,
I think an implementation of Lomb-Scargle and related algorithms would be a great addition to SciPy. With the 0.9 release coming soon, and also being holiday season, I'm not sure how soon anyone will be able to look at your code. Could you add this as an enhancement request to the Trac system? Here's the link:
http://projects.scipy.org/scipy/wiki
Create an account, and then create a new ticket. You can attach the code to the ticket.
Best regards,
Warren
P.S. There is a python/numpy implementation here:
http://www.astropython.org/snippet/2010/9/Fast-Lomb-Scargle-algorithm
I would also be interested, but I am, like scipy, still on fortran 77, g77 on older MingW Josef
_______________________________________________ SciPy-Dev mailing list SciPy-Dev@scipy.org http://mail.scipy.org/mailman/listinfo/scipy-dev
participants (3)
-
josef.pktd＠gmail.com -
Pim Schellart -
Warren Weckesser