Hi I read on the py2exe wiki that the 0.83 line creates huge distributions because of trouble excluding backends that are not wanted. ( http://starship.python.net/crew/theller/moin.cgi/MatPlotLib ) I do need to have a small distribution on windows. Which version of matplotlib was the last to play nicely with py2exe? Has anyone had any luck with 0.83.3 and excluding tk and gtk backends? (I'm using wx) I'm also getting a strange dependancy on the unicode version of wxPython, when I'm using the non-unicode version, and that's adding another 5MB onto my distro. Thanks, -Jim
Using Matplotlib 0.87.3 (wxAgg backend), NumPy 0.9.8, SciPy 0.4.9,
wxPython unicode.2.6.3.2, python 2.4.3 the following setup.py worked
well with py2exe 0.6.5 on Widows XP:
#--------------------------------------------------------------
from distutils.core import setup
import py2exe
import glob
data =3D glob.glob(r'C:Documents and Settings\Apa\.matplotlib\*')
data.append(r'C:\Documents and Settings\Apa\.matplotlib\matplotlibrc')
includes =3D ['matplotlib.numerix.random_array',
'pytz.zoneinfo.UTC',
'scipy.misc.info']
excludes =3D ['Tkinter']
opts =3D {
"py2exe": {
'skip_archive': 1,
'includes': includes,
'excludes': excludes,
'packages': ['numpy'],
}
}
setup(
windows =3D [
{
"script": 'Wynne.py',
"icon_resources": [(1, 'THz.ico')]
}
],
data_files =3D [('matplotlibdata', data)],
options =3D opts
)
#--------------------------------------------------------------
Regards,
G=E9za
--=20
G=E9za Groma
Institute of Biophysics
Biological Research Center of the Hungarian Academy of Sciences
Temesv=E1ri krt. 62.
6701 Szeged, Hungary
phone: +36 62 599 620 fax: +36 62 433 133 cellular: +36 20 5648 303
On 6/28/06, James Carroll <mr...@gm...> wrote:
> Hi I read on the py2exe wiki that the 0.83 line creates huge
> distributions because of trouble excluding backends that are not
> wanted.
> ( http://starship.python.net/crew/theller/moin.cgi/MatPlotLib )
>
> I do need to have a small distribution on windows. Which version of
> matplotlib was the last to play nicely with py2exe?
>
> Has anyone had any luck with 0.83.3 and excluding tk and gtk backends?
> (I'm using wx)
>
> I'm also getting a strange dependancy on the unicode version of
> wxPython, when I'm using the non-unicode version, and that's adding
> another 5MB onto my distro.
>
> Thanks,
> -Jim
>
> Using Tomcat but need to do more? Need to support web services, security?
> Get stuff done quickly with pre-integrated technology to make your job ea=
sier
> Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronim=
o
> http://sel.as-us.falkag.net/sel?cmd=3Dlnk&kid=3D120709&bid=3D263057&dat=
=3D121642
> _______________________________________________
> Matplotlib-users mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-users
>
On 6/28/06, G=E9za Groma <gr...@nu...> wrote: > Using Matplotlib 0.87.3 (wxAgg backend), NumPy 0.9.8, SciPy 0.4.9, > wxPython unicode.2.6.3.2, python 2.4.3 the following setup.py worked > well with py2exe 0.6.5 on Widows XP: > I followed your example, and had some luck. Thanks. My final distribution is 47MB in 840 files. Does anyone have an example th= at does even better? Last October, I used numarray + matplotlib in an almost identical project, and managed to bundle it into 11MB and four files. But I can't remember exactly the combination that I used. If I try to use a zip file (comment out the skip_archive) to reduce size and filecount, I get the error: File "matplotlib\backends\backend_wxagg.pyc", line 21, in ? File "matplotlib\figure.pyc", line 5, in ? File "matplotlib\artist.pyc", line 4, in ? File "matplotlib\transforms.pyc", line 215, in ? File "matplotlib\_transforms.pyc", line 1, in ? File "matplotlib\numerix\__init__.pyc", line 52, in ? File "numarray\__init__.pyc", line 42, in ? File "numarray\numarrayall.pyc", line 2, in ? File "numarray\generic.pyc", line 1120, in ? File "numarray\numarraycore.pyc", line 16, in ? File "numarray\_sort.pyc", line 12, in ? File "numarray\_sort.pyc", line 10, in __load File "numpy\__init__.pyc", line 35, in ? File "numpy\_import_tools.pyc", line 173, in __call__ File "numpy\_import_tools.pyc", line 68, in _init_info_modules File "<string>", line 1, in ? File "numpy\random\__init__.pyc", line 3, in ? File "numpy\random\mtrand.pyc", line 12, in ? File "numpy\random\mtrand.pyc", line 10, in __load File "numpy.pxi", line 32, in mtrand AttributeError: 'module' object has no attribute 'dtype' and even in the working example where I do skip_archive, I get the warnings= : The following modules appear to be missing ['FFT', 'LinearAlgebra', 'MA', 'MLab', 'Matrix', 'Numeric', 'Pyrex', 'Pyrex= .Comp iler', 'RandomArray', 'backends.draw_if_interactive', 'backends.new_figure_= manag er', 'backends.pylab_setup', 'backends.show', 'cephes', 'fcompiler.FCompile= r', ' fcompiler.show_fcompilers', 'genapi', 'lib.add_newdoc', 'numerix.ArrayType'= , 'nu merix.Complex', 'numerix.Complex32', 'numerix.Complex64', 'numerix.Float', = 'nume rix.Float32', 'numerix.Float64', 'numerix.Int', 'numerix.Int16', 'numerix.I= nt32' , 'numerix.Int8', 'numerix.NewAxis', 'numerix.UInt16', 'numerix.UInt32', 'n= umeri x.UInt8', 'numerix.absolute', 'numerix.add', 'numerix.allclose', 'numerix.a= lltru e', 'numerix.arange', 'numerix.arccos', 'numerix.arccosh', 'numerix.arcsin'= , 'nu merix.arcsinh', 'numerix.arctan', 'numerix.arctan2', 'numerix.arctanh', 'nu= merix .argmax', 'numerix.argmin', 'numerix.argsort', 'numerix.around', 'numerix.a= rray' , 'numerix.arrayrange', 'numerix.asarray', 'numerix.asum', 'numerix.bitwise= _and' , 'numerix.bitwise_or', 'numerix.bitwise_xor', 'numerix.ceil', 'numerix.cho= ose', 'numerix.clip', 'numerix.compress', 'numerix.concatenate', 'numerix.conjug= ate', 'numerix.convolve', 'numerix.cos', 'numerix.cosh', 'numerix.cross_correlat= e', ' numerix.cumproduct', 'numerix.cumsum', 'numerix.diagonal', 'numerix.divide'= , 'nu merix.dot', 'numerix.equal', 'numerix.exp', 'numerix.fabs', 'numerix.fft.ff= t', ' numerix.fft.inverse_fft', 'numerix.floor', 'numerix.fmod', 'numerix.fromfun= ction ', 'numerix.fromstring', 'numerix.greater', 'numerix.greater_equal', 'numer= ix.hy pot', 'numerix.identity', 'numerix.indices', 'numerix.innerproduct', 'numer= ix.is contiguous', 'numerix.less', 'numerix.less_equal', 'numerix.log', 'numerix.= log10 ', 'numerix.logical_and', 'numerix.logical_not', 'numerix.logical_or', 'num= erix. logical_xor', 'numerix.matrixmultiply', 'numerix.maximum', 'numerix.minimum= ', 'n umerix.mlab.amax', 'numerix.mlab.amin', 'numerix.mlab.cov', 'numerix.mlab.d= iff', 'numerix.mlab.hanning', 'numerix.mlab.rand', 'numerix.mlab.std', 'numerix.= mlab. svd', 'numerix.multiply', 'numerix.negative', 'numerix.nonzero', 'numerix.n= ot_eq ual', 'numerix.nx', 'numerix.ones', 'numerix.outerproduct', 'numerix.pi', '= numer ix.power', 'numerix.product', 'numerix.put', 'numerix.putmask', 'numerix.ra= nk', 'numerix.ravel', 'numerix.repeat', 'numerix.reshape', 'numerix.resize', 'nu= merix .searchsorted', 'numerix.shape', 'numerix.sin', 'numerix.sinh', 'numerix.si= ze', 'numerix.sometrue', 'numerix.sort', 'numerix.sqrt', 'numerix.subtract', 'nu= merix .swapaxes', 'numerix.take', 'numerix.tan', 'numerix.tanh', 'numerix.trace',= 'num erix.transpose', 'numerix.typecode', 'numerix.typecodes', 'numerix.where', = 'nume rix.which', 'numerix.zeros', 'numpy.Complex', 'numpy.Complex32', 'numpy.Com= plex6 4', 'numpy.Float', 'numpy.Float32', 'numpy.Float64', 'numpy.Inf', 'numpy.In= finit y', 'numpy.Int', 'numpy.Int16', 'numpy.Int32', 'numpy.Int8', 'numpy.UInt16'= , 'nu mpy.UInt32', 'numpy.UInt8', 'numpy.absolute', 'numpy.add', 'numpy.amax', 'n= umpy. amin', 'numpy.arange', 'numpy.array', 'numpy.asarray', 'numpy.asarray_chkfi= nite' , 'numpy.cast', 'numpy.concatenate', 'numpy.conjugate', 'numpy.diag', 'nump= y.dot ', 'numpy.exp', 'numpy.extract', 'numpy.eye', 'numpy.finfo', 'numpy.float_'= , 'nu mpy.floor', 'numpy.greater_equal', 'numpy.hstack', 'numpy.identity', 'numpy= .imag ', 'numpy.inf', 'numpy.infty', 'numpy.insert', 'numpy.iscomplexobj', 'numpy= .isfi nite', 'numpy.log', 'numpy.log2', 'numpy.logical_not', 'numpy.matrix', 'num= py.mg rid', 'numpy.newaxis', 'numpy.ogrid', 'numpy.ones', 'numpy.outerproduct', '= numpy .pi', 'numpy.poly1d', 'numpy.product', 'numpy.r_', 'numpy.ravel', 'numpy.re= al', 'numpy.reshape', 'numpy.sign', 'numpy.single', 'numpy.sqrt', 'numpy.subtrac= t', ' numpy.sum', 'numpy.take', 'numpy.transpose', 'numpy.typecodes', 'numpy.type= name' , 'numpy.where', 'numpy.zeros', 'numpy_distutils', 'numpy_distutils.command= .buil d_flib', 'numpy_distutils.command.cpuinfo', 'numpy_distutils.cpuinfo', 'num= py_di stutils.fcompiler', 'pkg_resources', 'pre', 'setuptools', 'setuptools.comma= nd', 'setuptools.command.egg_info', 'matplotlib.numerix.Float', 'matplotlib.nume= rix.F loat32', 'matplotlib.numerix.absolute', 'matplotlib.numerix.alltrue', 'matp= lotli b.numerix.asarray', 'matplotlib.numerix.ceil', 'matplotlib.numerix.equal', = 'matp lotlib.numerix.fromstring', 'matplotlib.numerix.put', 'matplotlib.numerix.r= avel' , 'matplotlib.numerix.sqrt', 'matplotlib.numerix.take', 'matplotlib.numerix= .tran spose', 'matplotlib.numerix.where', 'numarray.Complex', 'numarray.Complex32= ', 'n etc... Help me please! Thanks, -Jim
On 2006年6月28日, James Carroll apparently wrote:=20 > My final distribution is 47MB in 840 files. Does anyone=20 > have an example that does even better?=20 I think PyInstaller is supposed to help with this problem: http://pyinstaller.hpcf.upr.edu/ But I have not used it. Cheers, Alan Isaac
> > My final distribution is 47MB in 840 files. Does anyone > > have an example that does even better? > > I think PyInstaller is supposed to help with this problem: > http://pyinstaller.hpcf.upr.edu/ > But I have not used it. > > Cheers, > Alan Isaac Hi Alan, PyInstaller looks intriguing, but the traffic on their mailing list makes me think that it doesn't have enough users (& therefore developers) yet. I want to stay on the well-traveled-paths where I can find them. I think for now I'll use NSIS w/ py2exe to create a single executable that unpacks to a temporary directory and launches itself when it's double clicked. I've used this sucessfully in the past. ( something like this: http://wiki.wxpython.org/index.cgi/SmallApp ) I think what I'm really hoping for is to find a way to include just the core functionality of matplotlib, without pulling in numpy and scipy and all the bells and whistles. Of course, the real solution to my problem is to just start using all the bells and whistles (I'm happy to see the mathtext module that will work on systems that don't have TeX!) -Jim On 6/28/06, Alan G Isaac <ai...@am...> wrote: > On 2006年6月28日, James Carroll apparently wrote: > > > > > Using Tomcat but need to do more? Need to support web services, security? > Get stuff done quickly with pre-integrated technology to make your job easier > Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo > http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642 > _______________________________________________ > Matplotlib-users mailing list > Mat...@li... > https://lists.sourceforge.net/lists/listinfo/matplotlib-users >
On 2006年6月29日, James Carroll wrote: > PyInstaller looks intriguing, but the traffic on their > mailing list makes me think that it doesn't have enough > users (& therefore developers) yet. I want to stay on the > well-traveled-paths where I can find them. Understood. But do not overlook that PyInstaller is really the current life of Gordon McMillan's Python Installer, so it really is not a new kid on the block. Cheers, Alan Isaac
> Understood. But do not overlook that PyInstaller is really > the current life of Gordon McMillan's Python Installer, > so it really is not a new kid on the block. I gave PyInstaller a try, and it seemed to be smart and easy at first, and created a single exe that was just 8MB. The problem is that when I run the exe, It gives me the following errors before it crashes: - Fatal Error: MSVCR71.DLL! - Fatal Error: colud not be extracted. - Fatal Error: Error loading Python DLL - a message box pointing to python24.dll in a temp dir. - Fatal Error: - "Please tell microsoft about this problem." [Don't send] moving msvcr71.dll and python24.dll to the same directory didn't change anything. I'll put some more time into trying it, but my first impression isn't good. Meanwhile, I've removed numarray entirely from the project, and am using matplotlib.rcParams['numerix'] = 'numpy' and py2exe + nsis is working pretty well. I have a working single executable that is 12MB, but it takes about 18 seconds to decompress & run. If I turn off the NSIS lzw compression, then it it is a 46MB executable, and runs in about 12 seconds. I might be trying to create my own C++ wxWidgets top level application and embed wxPython & matplot lib in that. (has anyone ever had luck doing this?) Thanks, -Jim
On 2006年6月30日, James Carroll apparently wrote:=20 > I gave PyInstaller a try, and it seemed to be smart and easy at first, an= d=20 > created a single exe that was just 8MB. The problem is that=20 > when I run the exe, It gives me the following errors before it=20 > crashes:=20 > - Fatal Error: MSVCR71.DLL!=20 > - Fatal Error: colud not be extracted.=20 http://pyinstaller.hpcf.upr.edu/cgi-bin/trac.cgi/ticket/36 hth, Alan Isaac