pysteps 0.2

[dnerini]

Since our first release in August, there have been a number of improvements and bug fixes that would justify a new release, in my opinion.

[pulkkins]

I agree on that. We need to prepare a changelog for the release. @dnerini, could you prepare a list of your changes since release 0.1 (just the most important ones). I will then merge that list with my own.

[dnerini]

OK, after the last pull request cf3273b from @aperezhortal, we are very close to issue pysteps 0.2. We are basically only missing the new pyVET module for optical flow. After that, I think we should have our new release.

This is my basic changelog since 0.1:

• Bug fixes.
• Performance and syntax improvements.
• New utility methods to adjust the domain by geographical coordinates.
• New data transformation methods: normal quantile, square-root and logarithmic transformations.
• New importer for MeteoSwiss' metranet data format.
• New shift and scale post-processing method.

[aperezhortal]

I added the VET module in the commit 257e80f. I tried it with the the example data and gives results similar to the LK in the nowcasting and in the motion fields.

I also made a several changes in the motion interface to remove the nested ifs and use a dict instead (little faster and cleaner). I did a unitary testing for many possible combinations to ensure that this change does not introduced any bug.

See the min diff in 257e80f#diff-e39de838352d1f6e85823ac4aa1ea32e.

My change log for since version 0.1 is:

• Replace the config module with a JSON pystepsrc configuration file.
• Add the Variational Echo Tracking method to the motion module.

[dnerini]

@pulkkins can you provide your changelog since 0.1? Once I have it, I will issue version 0.2.

[pulkkins]

Here is my tentative changelog for version 0.2:

• More sensible naming of several main modules: optflow -> motion, advection -> extrapolation and postproc -> postprocessing.
• Add missing documentation to several modules and methods.
• New intensity-scale verification method.
• New visualization methods: cartopy backend, exceedance probabilities and plotting motion fields on a map.
• New interface module for different FFT methods and allow the user to choose the method.
• Fix incorrect separation into components when using the bps method for adding perturbations to the advection field.
• More efficient computation of the FFT in the cascade decomposition by using rfft2/irfft2 instead of fft2/ifft2.
• Two different methods for adjusting the temporal autocorrelation coefficients in order to guarantee stationarity of the resulting AR(2) models.
• Modify the STEPS method to have the same number of positional arguments as the other methods in the nowcasts module.
• Move the deterministic S-PROG mask computation outside the stochastic part of the STEPS method to avoid repeated computations.

[dnerini]

We are almost ready for pysteps 0.2.
@pulkkins is updating the documentation and we still need to fix some minor bugs (see e.g. #29 ). Anything else that needs to be addressed before the release?

[dnerini]

I've drafted the new pysteps 0.2 release. Please, @pulkkins , @loforest and @aperezhortal have a look to it (you'll find it with the release link) before we publish it.

[loforest]

I extended a couple of sentences in the draft. It looks good!

[pulkkins]

Looks good! I did some small modifications. I think we are ready to release pysteps 0.2. Though, I did some more extensive modifications in the nowcasts.steps module (particularly commit 1784672). It needs to be checked that I didn't introduce any new bugs.

…

On Thu, Jan 3, 2019 at 11:43 AM Loris Foresti ***@***.***> wrote: I extended a couple of sentences in the draft. It looks good! — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#17 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AQTdZBpXshQuBJBbzBiy312AyXel74H7ks5u_dCpgaJpZM4XZbFy> .

[dnerini]

OK, then I will check once more than everything works and then publish the release we have drafted.

[dnerini]

I tested the current version with the pysteps/examples scripts and everything looks fine. I will publish the 0.2 release tomorrow, unless somebody has something against it.