Page Menu
Home
HEPForge
Search
Configure Global Search
Log In
Files
F19244716
00README.txt
No One
Temporary
Actions
View File
Edit File
Delete File
View Transforms
Subscribe
Award Token
Flag For Later
Size
2 KB
Referenced Files
None
Subscribers
None
00README.txt
View Options
The NPStat python interface uses SWIG to wrap the C++ API, matplotlib
for plotting, numpy for some interfaces, and wxPython for GUIs.
To rebuild the python API, make sure you have a recent swig version
(3.0.12 is known to work). Install python development files and numpy. Look
at the Makefile in this directory and edit a few variables at the beginning
so that they correspond to your system settings. Then type "make distclean"
and "make". Note that, due to multiple instantiations of various templates,
the .cc file generated by swig is very large. If you use g++, you will need
a machine with at least 8GB of memory in order to compile it.
Installing wxPython: see http://wxpython.org. In Fedora Linux (which
I am currently using as my development machine) the installation can
be done simply by saying "yum install wxPython".
With NPStat, you will want to use matplotlib with the "wxagg" backend.
This backend is not necessarily going to come stock with your standard
matplotlib installation. This is how you can install matplotlib from
source on a Linux box to get this backend:
1) Install various "development" packages such as wxPython-devel,
tk-devel, etc. for the matplotlib backends you are planning to use.
2) Get the latest stable matplotlib release from
http://matplotlib.org/downloads.html
If you want to make WXAgg to be default matplotlib backend (as I normally
do), copy setup.cfg.template in the matplotlib distribution directory into
"setup.cfg" and edit it (put "backend = WXAgg" line there). Then follow
the installation instructions w.r.t prerequisites. Once all prerequisites
are installed, what I usually do is just run
python setup.py install --user
3) Optionally, install ipython (see http://ipython.org/)
4) Take a look at the beginning of the "npstat_browser_test.py" script.
You might want to change variables "plotting_styles_dir" and "ipython_mode".
5) Run "python npstat_browser_test.py". Alternatively, if you have set
the "ipython_mode" variable to "True" in the previous step, start ipython
shell and run "import npstat_browser_test" at the prompt.
The NPStat python interface is under development, so don't be surprised
if something does not work as described on your machine.
File Metadata
Details
Attached
Mime Type
text/plain
Expires
Tue, Sep 30, 4:44 AM (21 h, 3 m)
Storage Engine
blob
Storage Format
Raw Data
Storage Handle
6560513
Default Alt Text
00README.txt (2 KB)
Attached To
Mode
rNPSTATSVN npstatsvn
Attached
Detach File
Event Timeline
Log In to Comment