From: Yaqiang W. <yaq...@gm...> - 2019-03-21 15:09:23
|
I found Stewori has add a star to JythonLab repository. So I am thinking again the necessary for providing an integrated Jython package. Readme and license problem could be sovled by adding more contents or files on them. Unfortunately, the pip with Jython was not installled succesfully in my computer that means I can not test the pip ability for packages with java components. NumJy is a pure Java solution which is more easy to be used in various computer systems. I tested the big array multiply speed of NumJy and NumPy in my computer. At first calculation NumPy is faster than NumJy significantly, but after, I guess the JIT optimization effect, NumJy even runs little faster than NumPy. Of course I believe NumPy should be much faster than NumJy in most other functions. I also mentioned ND4J project using JNI and native BLAS for high speed, which is valuable to be tried as an optional backend of NumJy in the future. On Thu, Mar 21, 2019 at 4:52 PM Yaqiang Wang <yaq...@gm...> wrote: > Hi Jeff, > > I am glad for your positive feedback. For JythonLab I want to mimic > Anaconda for easy usage, but you are right it's not a good way in present > status. So I will remove it from github. All other packages could be added > in site-packaged seperately with some dependencies relationship. > > Regards > Yaqiang > > On Thu, Mar 21, 2019 at 3:46 PM Jeff Allen <ja...@fa...> wrote: > >> Hi Yaqiang, >> >> this looks like really useful work. Thank-you for sharing it. I had >> thought of attempting a numjy myself, but I have my hands full with the >> Jython core. >> >> A good way to distribute your work is through one or more packages a >> user can install in their own Jython site-packages. I see you have >> separate repositories for the parts, and that looks right. These days, >> for pure Python and CPython extensions, people go to PyPI. I don't know >> if that can be made to work for packages with compiled Java components, >> but it would be interesting to try. >> >> The pre-installed approach in jythonlab is perhaps not the best way to >> provide these as a bundle, as it stands. (Anaconda may be a good thing >> to mimic here, but that's a very complex product.) The immediate problem >> is partly GitHub's fault. It presents README.txt as your front page and >> LICENSE.txt as your licence. But these are the Jython ones, and so they >> don't apply to the repo overall. Maybe also look at customisation >> through a modified site.py that announces what one has just launched. >> >> Jeff Allen >> >> On 21/03/2019 05:03, Yaqiang Wang wrote: >> > JythonLab (https://github.com/Yaqiang/jythonlab) is a Jython >> > distribution package with scientific computation and visialization >> > packages: >> > >> > NumJy (https://github.com/Yaqiang/numjy) for multi-dimensional array >> > operation and computation. >> > PlotJy (https://github.com/Yaqiang/plotjy) for 2D/3D plotting. >> > DataframeJy (https://github.com/Yaqiang/dataframejy) provide DataFrame >> > object for Jython. >> > DatasetJy(https://github.com/Yaqiang/datasetjy) Read multiple >> > scientific data formats: Netcdf, Grib, HDF and many others. >> > EditorJy (https://github.com/Yaqiang/editorjy) for Jython code editing >> > and running. >> > >> > In JythonLab, above packages were added in Lib->site-pakages folder. >> > jythonlab.bat file was added in bin folder for start EditorJy desktop. >> > The packages are still in starting status. >> > >> > Hope the wok will be useful for some jython users. >> > >> > Regards >> > Yaqiang >> > >> >> >> _______________________________________________ >> Jython-dev mailing list >> Jyt...@li... >> https://lists.sourceforge.net/lists/listinfo/jython-dev >> > > > -- > ************************************************* > Dr. Yaqiang Wang > Chinese Academy of Meteorological Sciences (CAMS) > 46, Zhong-Guan-Cun South Avenue > Beijing, 100081 > China > > yaq...@gm... > > www.meteothink.org > ************************************************** > -- ************************************************* Dr. Yaqiang Wang Chinese Academy of Meteorological Sciences (CAMS) 46, Zhong-Guan-Cun South Avenue Beijing, 100081 China yaq...@gm... www.meteothink.org ************************************************** |