Add ndtri, igam, chdtr #3
Conversation
|
Thanks a lot for this pull request @lericson! Are you planning to add more functions? What is the scope of the changes you would like to make in the repository? |
|
Oh, I’m using this to make a decently fast particle filter in Python for a robot. We’ll see how much more I need, probably not that much for a while.
Ideally we would find a better way to add the CFFI boilerplate; the process is a bit tedious at the moment, and I don’t see why it has to be!
… On 29 Nov 2016, at 11:43, Juan Luis Cano Rodríguez ***@***.***> wrote:
Thanks a lot for this pull request @lericson <https://github.com/lericson>! Are you planning to add more functions? What is the scope of the changes you would like to make in the repository?
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub <#3 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAC3NV3gEVF1QJosMB99amqJ66rxCBlZks5rDAHtgaJpZM4K7nW7>.
|
Exactly! I think it would be nice to automate at least part of this process, but now I don't have the time resources to do it. In any case, when you think this pull request is ready to review please ping me, we can check the code together and I will be happy to merge :) |
|
Other than that, adding automation functionality can go in a separate PR. |
|
Also it’d be nice if we could somehow generate ufuncs, so you don’t have to create array versions of every single-variate function.
… On 29 Nov 2016, at 11:47, Juan Luis Cano Rodríguez ***@***.***> wrote:
Ideally we would find a better way to add the CFFI boilerplate; the process is a bit tedious at the moment, and I don’t see why it has to be!
Exactly! I think it would be nice to automate at least part of this process, but now I don't have the time resources to do it. In any case, when you think this pull request is ready to review please ping me, we can check the code together and I will be happy to merge :)
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub <#3 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAC3NRJqomdHA4VHrAjpI2NZAJ7SXL37ks5rDALUgaJpZM4K7nW7>.
|
|
Seems to be a lot of work:
https://docs.scipy.org/doc/numpy/user/c-info.ufunc-tutorial.html <https://docs.scipy.org/doc/numpy/user/c-info.ufunc-tutorial.html>
… On 29 Nov 2016, at 11:54, Ludvig Ericson ***@***.***> wrote:
Also it’d be nice if we could somehow generate ufuncs, so you don’t have to create array versions of every single-variate function.
> On 29 Nov 2016, at 11:47, Juan Luis Cano Rodríguez ***@***.*** ***@***.***>> wrote:
>
> Ideally we would find a better way to add the CFFI boilerplate; the process is a bit tedious at the moment, and I don’t see why it has to be!
>
> Exactly! I think it would be nice to automate at least part of this process, but now I don't have the time resources to do it. In any case, when you think this pull request is ready to review please ping me, we can check the code together and I will be happy to merge :)
>
> —
> You are receiving this because you were mentioned.
> Reply to this email directly, view it on GitHub <#3 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAC3NRJqomdHA4VHrAjpI2NZAJ7SXL37ks5rDALUgaJpZM4K7nW7>.
>
|
|
Or perhaps this can be used instead http://numba.pydata.org/numba-doc/0.29.0/user/vectorize.html but requires a totally different approach. |
|
Wow, great idea.
… On 29 Nov 2016, at 12:09, Juan Luis Cano Rodríguez ***@***.***> wrote:
Or perhaps this can be used instead http://numba.pydata.org/numba-doc/0.29.0/user/vectorize.html <http://numba.pydata.org/numba-doc/0.29.0/user/vectorize.html> but requires a totally different approach.
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub <#3 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAC3NaQyHrS_dYdFL_ZPeFOTBt_p_Nv2ks5rDAf-gaJpZM4K7nW7>.
|
|
I just realized that the project already requires numba, so it's fine to use it to vectorize the functions. At the end of this PR there should be:
|
|
And by the way, I can focus on providing Travis CI integration so we can monitor the progress of this: #4 |
|
😟 |
|
It might be nice to put all the |
Fear not, we'll get there 😄 |
|
This project is not maintained anymore, please check out https://github.com/limix/ncephes. |
No description provided.