Mailman 3 - SfePy - python.org

Re: [SciPy-User] ANN: SfePy 2009.4
by Robert Cimrman 25 Nov '09

25 Nov '09

Hi Osman, below is a hack that might help you. In terms of sfepy's Viewer, it should read: # Workaround for 'scene = gui.scene.mayavi_scene' e = mlab.get_engine() for scene in e.scenes: if scene.scene.render_window is gui.scene.scene_editor.render_window: print 'Yeah' break r. PS: Thanks, Gael! -------- Original Message -------- Subject: Re: [Fwd: Re: [SciPy-User] ANN: SfePy 2009.4] Date: Wed, 25 Nov 2009 15:20:40 +0100 From: Gael Varoquaux To: Robert Cimrman On Wed, Nov 25, 2009 at 03:11:39PM +0100, Robert Cimrman wrote: > is there a way to get a mayavi_scene from a MlabSceneModel instance in > mayavi 3.1.0? I guess something like > engine = mlab.get_engine() > scene = engine.<get the scene that corresponds to the MlabSceneModel instance> Not really something like that. It's a really design flaw, that we realized a bit late. Here is a hack (my_model.scene is the MlabSceneModel instance): for scene in e.scenes: if scene.scene.render_window is my_model.scene.scene_editor.render_window: print 'Yeah' HTH, Gaël

1 0

ANN: SfePy 2009.4
by Robert Cimrman 25 Nov '09

25 Nov '09

I am pleased to announce release 2009.4 of SfePy. Description ----------- SfePy (simple finite elements in Python) is a software, distributed under the BSD license, for solving systems of coupled partial differential equations by the finite element method. The code is based on NumPy and SciPy packages. Mailing lists, issue tracking, git repository: http://sfepy.org Home page: http://sfepy.kme.zcu.cz New documentation site: http://docs.sfepy.org/doc Many thanks to Logan Sorenson for the new documentation contents, and Vladimir Lukes for setting up the server. Highlights of this release -------------------------- - unified handling of user-defined functions (for defining subdomains, heterogeneous material properties, boundary conditions etc.) - greatly improved postprocessing and visualization capabilities, namely: - support for file sequences (evolutionary simulations) - animations (using ffmpeg) - automatic scalar bars - sfepy_gui.py: Mayavi2-based GUI to launch simulations Major improvements ------------------ Apart from many bug-fixes, let us mention: - quasistatic time stepping - graphical logging: - dynamic adding of data groups (new axes) to Log and ProcessPlotter - linear algebra: - reversed Cuthill-McKee permutation algorithm, graph in-place permutation - setting of parameter variables by a user-defined function - new tests and terms For more information on this release, see http://sfepy.googlecode.com/svn/web/releases/2009.4_RELEASE_NOTES.txt (full release notes, rather long). Best regards, Robert Cimrman

2 4

Re: update version of docs
by Vladimir Lukes 25 Nov '09

25 Nov '09

Hi, the docs pages at [1] were update to version 2009.4. Best, Vladimir [1] http://docs.sfepy.org/doc/ Logan Sorenson napsal(a): > Hi Robert, > > On Tue, Nov 24, 2009 at 7:30 AM, Robert Cimrman <cimr...(a)ntc.zcu.cz <mailto:cimr...@ntc.zcu.cz>> wrote: > > Hi, > > The docs at [1] need to be regenerated with the 2009.4 version. Can > somebody > with a working sphinx do that? > > > Done! The html is available from [2] (gh-pages branch) and you can check the result at [3]. Also, I disentangled the sphinx documentation into a separate branch (sphinx_docs) from the master branch at [2]. Maybe this way is easier to maintain? (Still getting the hang of git :) ) > > > > thanks, > r. > > > Best, > Logan > > > > [1] docs.sfepy.org/do <http://docs.sfepy.org/doc> > > [2] git://github.com/logansorenson/sfepy_doc2.git <http://github.com/logansorenson/sfepy_doc2.git> > [3] http://logansorenson.github.com/sfepy_doc2/ > > -- > > You received this message because you are subscribed to the Google Groups "sfepy-devel" group. > To post to this group, send email to sfepy...(a)googlegroups.com. > To unsubscribe from this group, send email to sfepy-devel...(a)googlegroups.com. > For more options, visit this group at http://groups.google.com/group/sfepy-devel?hl=en.

1 0

update version of docs
by Robert Cimrman 25 Nov '09

25 Nov '09

Hi, The docs at [1] need to be regenerated with the 2009.4 version. Can somebody with a working sphinx do that? thanks, r. [1] docs.sfepy.org/doc

2 2

Build error in sfepy/linalg/extmods and possible fix to sfepy/linalg/extmods/setup.py
by Logan Sorenson 20 Nov '09

20 Nov '09

Hi, I just did a git pull from [1]. I was getting the following build error (see (1) below) using the "distutils" method. It seems to be due to missing paths to sfepy/fem/extmods/* in sfepy/linalg/extmods, so I followed the example of sfepy/terms/extmods to put in the missing paths. This is available from [2] (it's on the master branch, should I have made a new branch for this?). It's a pretty quick fix to get the build to work, so I wouldn't be surprised if I'm overlooking something. :) After putting in the fix, I'm getting a test failure in tests/test_permutations.py (see (2)). I figure I missed something else in the distutils configuration. The "make" build method works before and after the fix. I know sfepy/linalg is new, so probably everything isn't in place for distutils building yet, but I thought I could at least help get the build working again. :) Thanks! Logan [1] git://git.sympy.org/sfepy.git [2] git://github.com/logansorenson/electrostatic_structural_test.git (1) logan@quantumdot:~/phoenix/projects/sfepy$ python setup.py build_ext --inplace ['sfepy', '/home/logan/phoenix/projects/sfepy/sfepy', '/home/logan/phoenix/projects/sfepy', '/home/logan/phoenix/projects/sfepy', '/usr/lib/python2.5/site-packages/numpydoc-0.2-py2.5.egg', '/usr/lib/python2.5/site-packages/scikits.umfpack-5.1.0-py2.5-linux-i686.egg', '/usr/lib/python2.5', '/usr/lib/python2.5/plat-linux2', '/usr/lib/python2.5/lib-tk', '/usr/lib/python2.5/lib-dynload', '/usr/local/lib/python2.5/site-packages', '/usr/lib/python2.5/site-packages', '/usr/lib/python2.5/site-packages/Numeric', '/usr/lib/python2.5/site-packages/PIL', '/usr/lib/python2.5/site-packages/gst-0.10', '/usr/lib/pymodules/python2.5', '/usr/lib/pymodules/python2.5/gtk-2.0', '/usr/lib/python2.5/site-packages/wx-2.8-gtk2-unicode'] non-existing path in 'sfepy/linalg/extmods': 'common_python.c' non-existing path in 'sfepy/linalg/extmods': 'array.i' non-existing path in 'sfepy/linalg/extmods': 'common.i' <snip> building extension "sfepy.linalg.extmods._rcm" sources swig: sfepy/linalg/extmods/rcm.i swig -python -Isfepy/linalg/extmods -o sfepy/linalg/extmods/rcm_wrap.c -outdir sfepy/linalg/extmods sfepy/linalg/extmods/rcm.i sfepy/linalg/extmods/rcm.i:9: Error: Unable to find 'types.h' sfepy/linalg/extmods/rcm.i:11: Error: Unable to find 'common.i' sfepy/linalg/extmods/rcm.i:12: Error: Unable to find 'array.i' error: command 'swig' failed with exit status 1 (2) tests/test_permutations.py --- test_rcm: failed! logan@quantumdot:~/phoenix/projects/sfepy$ ./runTests.py --debug tests/test_permutations.py <<< directory: tests, test files: 1 <<< tests/test_permutations.py sfepy: left over: ['__builtins__', '_filename', '__file__', 'TestCommon', '__name__', 'assert_', '__doc__', 'op'] >>> test instance prepared (1 test(s)) >>> <type 'exceptions.ImportError'> Traceback (most recent call last): File "./runTests.py", line 226, in <module> main() File "./runTests.py", line 217, in main run_tests(stats, dirname, [filename]) File "./runTests.py", line 148, in run_tests n_fail, n_total, test_time = run_test( conf_name, options ) File "./runTests.py", line 111, in run_test ok, n_fail, n_total = test.run( options.debug ) File "/home/logan/phoenix/projects/sfepy/sfepy/base/testing.py", line 38, in run ret = test_method() File "tests/test_permutations.py", line 11, in test_rcm from sfepy.linalg import rcm, permute_in_place, save_sparse_txt File "/home/logan/phoenix/projects/sfepy/sfepy/linalg/__init__.py", line 1, in <module> from extmods.rcm import rcm, permute_in_place File "/home/logan/phoenix/projects/sfepy/sfepy/linalg/extmods/rcm.py", line 28, in <module> import _rcm ImportError: /home/logan/phoenix/projects/sfepy/sfepy/linalg/extmods/_rcm.so: undefined symbol: g_error

2 2

place for sphinx docs
by Robert Cimrman 18 Nov '09

18 Nov '09

Hi, the sphinx docs, thanks to Logan, are covering now pretty much everything the old documentation wikis did. They are at [1] now, but we should put them to a more permanent(-looking) location. Any suggestions how to proceed? Just a few ideas: We have the 'sfepy.org' domain, so it would be nice to have the docs somewhere under it. Or we could create the 'sfepy' github user and put the docs to the 'doc' repo, resulting in http://sfepy.github.com/doc/. The latter can be done immediately, for the former my web knowledge is too limited. r. [1] http://logansorenson.github.com/sfepy_doc2/

1 1

Electrostatic-elastic coupling
by Logan Sorenson 05 Nov '09

05 Nov '09

Hi, I was wondering how to implement the electrostatic-elastic coupling in SfePy. Basically, I think this coupling is described by the Maxwell stress tensor [1] which gives the stress due to the electrostatic field (setting the magnetic field to 0 in the quasistatic approximation) on the surface of a conductor. I think the coupling can be handled in a similar manner with the piezoelectric coupling. For example, in input/piezo.py, the equations are: equations = { '1' : """- %f * dw_mass_vector.i1.Y( inclusion.density, v, u ) + dw_lin_elastic_iso.i1.Y( inclusion.lam, inclusion.mu, v, u ) - dw_piezo_coupling.i1.Y2( inclusion.coupling, v, phi ) = 0""" % omega_squared, '2' : """dw_diffusion.i1.Y( inclusion.dielectric, psi, phi ) + dw_piezo_coupling.i1.Y2( inclusion.coupling, u, psi ) = 0""", } I think there should be a similar set of equations for the electrostatic-elastic coupling case, but the integration should be done over the interface of the elastic material and the vacuum regions. I think that can be handled by adding a 2D integral in the problem definition. What I'm missing is how to define the coupling term in SfePy. I think the form of the term is basically the same as dw_surface_ltr in sfepy_manual.pdf. Also, it looks like the gradient of the electric potential can be obtained from cachesBasic. Is it possible to use the gradient of the electric potential information to calculate the Maxwell stress tensor using appropriate vector operations (divergence and dot product), along with the dielectric tensor? Would this be doable through the dw_surface_ltr term? Or is it better to implement a new coupling term (based off dw_surface_ltr) to handle this (and how would one go about that :) )? Hope everything is clear! Any advice is greatly appreciated! Thanks, Logan [1] http://en.wikipedia.org/wiki/Maxwell_stress_tensor

2 7

Updated Sphinx Documentation
by Logan Sorenson 05 Nov '09

05 Nov '09

Hi, I finally got some time to update the sphinx documentation. The final result is at [1] and the git repo can be cloned from [2]. Basically, so far I've just pulled stuff from the google pages and converted into sphinx format and did some polishing. I think it's a good start, but definitely needs more work! For example, the tutorial right now is more "theoretical", and may not be "practical" enough for a new user to understand...probably this can be separated into a simpler tutorial and then an advanced tutorial or reference section, especially the problem description format. Also, I haven't had time to look into genDocs.py yet... Thanks! Logan [1] http://logansorenson.github.com/sfepy_doc2 [2] git://github.com/logansorenson/sfepy_doc2.git

2 7

This is a test
by Test 05 Nov '09

05 Nov '09

This is just a test message Delete it pls

1 0

time for a release?
by Robert Cimrman 23 Oct '09

23 Oct '09

Hi, the Git version now contains a lot of interesting changes and new features compared to the latest released version (2009.3), see the would-be RELEASE_NOTES below. There are still some problems to be addressed, though, listed here: Known problems: --------------- - undocumented new features: - 'functions' keyword and change to materials definition - new application option 'use_equations' to select equations - homogenization is completely undocumented - chunk_size must be greater than the number of elements (-> 1 chunk only) for the "materials in physical QPs" to work properly - isfepy is broken for multiple view() calls! - sfepy_gui.py not updated for new Viewer capabilities (animations etc.) - is the installation within femhub ok?? (Viewer changed a lot...) While I will try to fix the problems before releasing, I am not sure we shall wait till all of them are fixed, what do you think? I would also appreciate some help, apart from the usual testing. While there is probably nobody familiar enough with the internals (or am I wrong?), the topic accessible to some of you is the documentation, namely the 'functions' keyword and change to materials definition - if you feel like fixing this at [1], go ahead, please. I would also like to set-up a sphinx-based documentation site for sfepy, like [2] or [3], probably using github. Any volunteers out there? I have made some attempts to learn how to do it, but have currently no time to go on. And yes, new functions added to sfepy are documented and use the sphinx-compatible format, following the numpy's standards. [1] http://code.google.com/p/sfepy/wiki/ProblemDescription [2] http://docs.scipy.org/doc/ [3] http://docs.sympy.org/ Feel free to ask anything related to the features described below. Yours truly, r. Would-be RELEASE_NOTES: ----------------------- - major branches merged: - 'ulf' branch: updated Lagrangian (UL) formulation - 'functions' branch: - unified passing extra arguments to boundary condition, material, and region functions - physical quadrature point generation - unified/improved handling of material parameters in terms: - all material parameters defined in physical quadrature points - all terms updated, some terms were coalesced into one - 'porous' branch: homogenized porous media - input file keywords: - new 'functions' keyword - simplifications & unifications: - results of all time steps of an evolutionary problem can be saved to a single HDF5 file - enable passing variables data to ProblemDefinition.solve() - runTests.py: allow multiple test files as command line arguments - Viewer.call_mlab() split and refactored - short syntax for periodic boundary conditions - simplified input file syntax of materials - postprocessing and visualization: - using FileSource class abstracts the particular format for storing results: - VTK, HDF5 supported now - support for file sequences (evolutionary simulations) - time step selection for HDF5 (single) and VTK (sequence) files - animations (using ffmpeg) - minimalistic ViewerGUI - show scalar bars - watch results file (HDF5) and add time steps as they are saved - listing data ranges works offscreen, summary for file sequence - sfepy_gui.py: Mayavi2-based GUI to launch simulations - changes aimed at interactive work: - Domain, Region, Field creation refactoring - data probing - postprocessing mode: - read a previously probed data from the probe text file, re-plot them, and integrate them along the probe - graphical logging: - dynamic adding of data groups (new axes) to Log and ProcessPlotter - many bug fixes, namely: - fix import_file() for multiple imports - miscellaneous updates: - quasistatic time stepping - new zero-order elements: 3_4_P0, 3_8_Q0 - more elastic tensor construction functions: - elastic tensor from Young's modulus and Poisson's ratio - elastic tensors for use in mixed formulation - setting of parameter variables by a user-defined function - new terms: - dw_volume_wdot_scalar_eth (exponential decay dot product convolution term) - dw_biot_eth (exponential decay Biot convolution term) - dw_lin_elastic_eth (exponential decay elastic convolution term) - updated Lagrangian (UL) formulation terms: - dw_ul_bulk_penalty, dw_ul_he_neohook, dw_ul_he_mooney_rivlin

2 19