[aur-general] Plan of moving python(2)-pandas to community
Hi all, just wanted to let you know about my plan of moving Pandas to community, so they don't interfere with plans of other TUs or AUR maintainers. Pandas is popular package (approximately 50 votes for bot python 2 and python 3 version in AUR[1,2]) and is one of the basic packages for doing science in Python, it is vital part of so called "SciPy Stack"[3] (and actually last missing piece). I planned to do it lot earlier, but there were some remaining issues that were fixed by 0.14.0 release - most important finally official support for current release of Python 3.4[4]. Pandas has very few dependencies, but a lot of recommended and optional dependencies, all of which provide important pieces of functionality, to make the experience complete. That's why I will move in and maintain all it's dependencies as well, unless someone has anything against it. This includes: - bottleneck library [5,6] - statsmodels library [7,8] - patsy library [9,10] - set of libraries to read and write Excel files [11,12,13,14,15,16] The order of moving those packages will be following: - bottleneck because it is recommended dependency providing speed to lots of operations - libraries to deal with Excel, because it improves usability of library and is direct dependency - Pandas itself - Patsy, because it is dependency of statsmodels - statsmodels, because pandas is its dependency, and it is optdependency of pandas - so it comes last - Pandas again, but with statsmodels as optdepends and tests enabled I hope to push one update every few days. I will let the related packages settle in community for day or two and then push next ones, staring on 4'th June, hopefully finish with those before 17'th - it takes some time, mostly due to limited time I can spend of this particular move. Cheers, Andrzej. [1]: https://aur.archlinux.org/packages/python2-pandas/ [2]: https://aur.archlinux.org/packages/python-pandas/ [3]: http://www.scipy.org/stackspec.html [4]: http://pandas.pydata.org/pandas-docs/stable/whatsnew.html [5]: https://aur.archlinux.org/packages/python-bottleneck/ [6]: https://aur.archlinux.org/packages/python2-bottleneck/ [7]: https://aur.archlinux.org/packages/python-statsmodels/ [8]: https://aur.archlinux.org/packages/python-statsmodels/ [9]: https://aur.archlinux.org/packages/python-patsy/ [10]: https://aur.archlinux.org/packages/python2-patsy/ [11]: https://aur.archlinux.org/packages/python-openpyxl/ [12]: https://aur.archlinux.org/packages/python2-openpyxl/ [13]: https://aur.archlinux.org/packages/python-xlwt/ [14]: https://aur.archlinux.org/packages/python2-xlwt/ [15]: https://aur.archlinux.org/packages/python-xlrd/ [16]: https://aur.archlinux.org/packages/python2-xlrd/
Andrzej, Since you're now maintaining xlrd and xlwt, could you also grab my python2-xlutils package? python-xlutils should be deleted in the process. I had left it as a placeholder. -Aaron DeVore On Sun, Jun 1, 2014 at 7:18 AM, Andrzej Giniewicz <gginiu@gmail.com> wrote:
Hi all,
just wanted to let you know about my plan of moving Pandas to community, so they don't interfere with plans of other TUs or AUR maintainers.
Pandas is popular package (approximately 50 votes for bot python 2 and python 3 version in AUR[1,2]) and is one of the basic packages for doing science in Python, it is vital part of so called "SciPy Stack"[3] (and actually last missing piece). I planned to do it lot earlier, but there were some remaining issues that were fixed by 0.14.0 release - most important finally official support for current release of Python 3.4[4].
Pandas has very few dependencies, but a lot of recommended and optional dependencies, all of which provide important pieces of functionality, to make the experience complete. That's why I will move in and maintain all it's dependencies as well, unless someone has anything against it. This includes:
- bottleneck library [5,6] - statsmodels library [7,8] - patsy library [9,10] - set of libraries to read and write Excel files [11,12,13,14,15,16]
The order of moving those packages will be following:
- bottleneck because it is recommended dependency providing speed to lots of operations - libraries to deal with Excel, because it improves usability of library and is direct dependency - Pandas itself - Patsy, because it is dependency of statsmodels - statsmodels, because pandas is its dependency, and it is optdependency of pandas - so it comes last - Pandas again, but with statsmodels as optdepends and tests enabled
I hope to push one update every few days. I will let the related packages settle in community for day or two and then push next ones, staring on 4'th June, hopefully finish with those before 17'th - it takes some time, mostly due to limited time I can spend of this particular move.
Cheers, Andrzej.
[1]: https://aur.archlinux.org/packages/python2-pandas/ [2]: https://aur.archlinux.org/packages/python-pandas/ [3]: http://www.scipy.org/stackspec.html [4]: http://pandas.pydata.org/pandas-docs/stable/whatsnew.html [5]: https://aur.archlinux.org/packages/python-bottleneck/ [6]: https://aur.archlinux.org/packages/python2-bottleneck/ [7]: https://aur.archlinux.org/packages/python-statsmodels/ [8]: https://aur.archlinux.org/packages/python-statsmodels/ [9]: https://aur.archlinux.org/packages/python-patsy/ [10]: https://aur.archlinux.org/packages/python2-patsy/ [11]: https://aur.archlinux.org/packages/python-openpyxl/ [12]: https://aur.archlinux.org/packages/python2-openpyxl/ [13]: https://aur.archlinux.org/packages/python-xlwt/ [14]: https://aur.archlinux.org/packages/python2-xlwt/ [15]: https://aur.archlinux.org/packages/python-xlrd/ [16]: https://aur.archlinux.org/packages/python2-xlrd/
On 09.06.2014 02:15, Aaron DeVore wrote:
Since you're now maintaining xlrd and xlwt, could you also grab my python2-xlutils package? python-xlutils should be deleted in the process. I had left it as a placeholder.
Hi Aaron, No problem, I can take it over, but I will leave it in AUR for now, because python-xlutils isn't dependency of any package in community nor has required votes to go in by itself. I think than when both xlutils and xlwt supports Python3, Pandas might depend on it instead of separate xlrd and xlwt, so I will come back to it then. Thanks for maintaining it till now! Cheers, Andrzej.
participants (2)
-
Aaron DeVore
-
Andrzej Giniewicz