dolfin team mailing list archive

[Johan Hake <johan.hake@xxxxxxxxx>]

On Wednesday November 10 2010 11:58:19 Nathan Borggren wrote:
> After upgrading to the newest version the error repeats with value_shape

Can you please report what the newest version means. The value_shape is 
introduced in the development version. It is not in the latests stable 
version.

> but is indeed fixed with the replacement pointed out by Stepan.  After
> making sure dim() was used I get two different errors, depending on if I
> use the demo posted in the documentation, or if I use the demo from the
> installation (which used the dim() method already).
> 
> Thanks,
>   Nathan
> 
> ##the error in the demo from the web documentation
> 
> $ python cahn.py
> Got expression dimension = 2
> Starting Newton solve.
> Calling FFC just-in-time (JIT) compiler, this may take some time.
> Traceback (most recent call last):
>   File "cahn.py", line 90, in <module>
>     solver.solve(problem, u.vector())
>   File "cahn.py", line 22, in F
>     assemble(self.L, tensor=b)
>   File "/usr/lib/python2.6/dist-packages/dolfin/fem/assemble.py", line 100,
> in assemble
>     common_cell=common_cell)
>   File "/usr/lib/python2.6/dist-packages/dolfin/fem/form.py", line 34, in
> __init__
>     (self._compiled_form, module, self.form_data) = jit(form,
> form_compiler_parameters, common_cell)
>   File "/usr/lib/python2.6/dist-packages/dolfin/compilemodules/jit.py",
> line 47, in mpi_jit
>     return local_jit(*args, **kwargs)
>   File "/usr/lib/python2.6/dist-packages/dolfin/compilemodules/jit.py",
> line 114, in jit
>     return jit_compile(form, parameters=p, common_cell=common_cell)
>   File "/usr/lib/python2.6/dist-packages/ffc/jitcompiler.py", line 64, in
> jit
>     return jit_form(object, parameters, common_cell)
>   File "/usr/lib/python2.6/dist-packages/ffc/jitcompiler.py", line 122, in
> jit_form
>     compile_form(preprocessed_form, prefix=jit_object.signature(),
> parameters=parameters)
>   File "/usr/lib/python2.6/dist-packages/ffc/compiler.py", line 145, in
> compile_form
>     oir = optimize_ir(ir, parameters)
>   File "/usr/lib/python2.6/dist-packages/ffc/optimization.py", line 38, in
> optimize_ir
>     oir_integrals = [_optimize_integral_ir(ir) for ir in ir_integrals]
>   File "/usr/lib/python2.6/dist-packages/ffc/optimization.py", line 56, in
> _optimize_integral_ir
>     oir = r.optimize_integral_ir(ir)
>   File "/usr/lib/python2.6/dist-packages/ffc/tensor/tensoroptimization.py",
> line 42, in optimize_integral_ir
>     ir["AK"][k] = (A0, GK, _optimize_tensor_contraction(A0.A0, rank))
>   File "/usr/lib/python2.6/dist-packages/ffc/tensor/tensoroptimization.py",
> line 80, in _optimize_tensor_contraction
>     return optimize(A0)
>   File "/usr/lib/python2.6/dist-packages/ferari/binary.py", line 390, in
> optimize_action
>     p = process( Adict )
>   File "/usr/lib/python2.6/dist-packages/ferari/binary.py", line 188, in
> process
>     G = get_graph( vecs , rho )
>   File "/usr/lib/python2.6/dist-packages/ferari/binary.py", line 72, in
> get_graph
>     (w,l) = dist( vecs[u] , vecs[v] )
>   File "/usr/lib/python2.6/dist-packages/ferari/binary.py", line 58, in rho
>     c = colinear( u , v )
>   File "/usr/lib/python2.6/dist-packages/ferari/binary.py", line 44, in
> colinear
>     uhat = util.unit_vector( u )
>   File "/usr/lib/python2.6/dist-packages/ferari/util.py", line 25, in
> unit_vector
>     raise RuntimeError, "divide by zero"
> RuntimeError: divide by zero

Could you try by uninstalling ferari:

  sudo apt-get remove ferari

You will probably be informed that the fenics package will be uninstalled bu 
that is ok. It is just a dummy package for all fenics packages.

> ##and the error in the demo from the install

What version of dolfin and ufl are you using. I a python shell do:

  import dolfin, ufl
  print dolfin.__version__, ufl.__version__

> /FEniCS/dolfin-doc/demo/pde/cahn-hilliard/python$ python demo.py
> *** Warning: The meaning of operator '+' has changed (no longer creating a
> mixed function space).

As the warning tells you we have changed the meaning of the + operator when 
dealing with FunctionSpace. You should now use '*'. It looks like your demos 
are out of sync with your fenics installation.

> Traceback (most recent call last):
>   File "demo.py", line 49, in <module>
>     q, v  = TestFunctions(ME)
> ValueError: need more than 1 value to unpack

You get this error because ME is not a mixed FunctionSpace.

> ##after fixing this and similar subsequent errors with q, v  =
> TestFunctions(ME), TestFunctions(ME) we get another error:

This will just create test and trial functions that are in the wrong 
FunctionSpace.

Johan

> $ python demo.py
> *** Warning: The meaning of operator '+' has changed (no longer creating a
> mixed function space).
> Traceback (most recent call last):
>   File "demo.py", line 62, in <module>
>     mu = factor*(2.0*c*(1.0-c)*(1.0-c) - 2.0*c*c*(1.0-c))
> TypeError: can't multiply sequence by non-int of type 'float'



> Thanks,
>    Nathan
> 
> On Wed, Nov 10, 2010 at 2:29 PM, Stepan Roucka <stepan@xxxxxxxxx> wrote:
> > I think you use dim() method instead:
> > 
> > class InitialConditions(Expression):
> >   def __init__(self):
> >       random.seed(2 + MPI.process_number())
> >   
> >   def eval(self, values, x):
> >       values[0] = 0.63 + 0.02*(0.5 - random.random())
> >       values[1] = 0.0
> >    
> >    def dim(self):
> >       return 2
> > 
> > The documentation of demos seems to be outdated?
> > The demos in the ppa packages use dim(), but in the
> > sources
> > 
> > http://bazaar.launchpad.net/~dolfin-core/dolfin/main/annotate/head%3A/dem
> > o/pde/cahn-hilliard/python/demo.py<http://bazaar.launchpad.net/%7Edolfin-
> > core/dolfin/main/annotate/head%3A/demo/pde/cahn-hilliard/python/demo.py>
> > value_shape is still used, I don't understand why.
> > 
> > Stepan
> > 
> > On 10 November 2010 18:48, Garth N. Wells <gnw20@xxxxxxxxx> wrote:
> > > On 10/11/10 17:44, N.A. Borggren wrote:
> > >> Hello,
> > >> 
> > >>     I am having some issues overloading eval for the class Expression
> > >> 
> > >> when I try to use vector valued functions.  For example the error
> > >> occurs for me in the Cahn-Hilliard demo copied directly from
> > >> http://www.fenicsproject.org/doc/demos/pde/cahn-
> > >> hilliard/python/documentation.html
> > >> 
> > >> $ python cahn.py
> > >> 
> > >> Got expression dimension = 0
> > >> 
> > >> Traceback (most recent call last):
> > >>   File "cahn.py", line 56, in<module>
> > >>   
> > >>     u.interpolate(u_init)
> > >>   
> > >>   File "cahn.py", line 10, in eval
> > >>   
> > >>     values[1] = 0.0
> > >> 
> > >> IndexError: index out of bounds
> > >> 
> > >> the problematic definition is:
> > >> 
> > >> class InitialConditions(Expression):
> > >>     def __init__(self):
> > >>         random.seed(2 + MPI.process_number())
> > >>     
> > >>     def eval(self, values, x):
> > >>         values[0] = 0.63 + 0.02*(0.5 - random.random())
> > >>         values[1] = 0.0
> > >>     
> > >>     def value_shape(self):
> > >>         return (2,)
> > >> 
> > >> What is this value_shape accomplishing?  The expression dimension
> > >> seems to still be zero.  How do I resolve this?
> > > 
> > > Which version of DOLFIN are you using?
> > > 
> > > Garth
> > > 
> > >> Thank you very much,
> > >> 
> > >>     Nathan Borggren
> > > 
> > > _______________________________________________
> > > Mailing list:
> > > https://launchpad.net/~dolfin<https://launchpad.net/%7Edolfin> Post to
> > >     : dolfin@xxxxxxxxxxxxxxxxxxx
> > > Unsubscribe :
> > > https://launchpad.net/~dolfin<https://launchpad.net/%7Edolfin> More
> > > help   : https://help.launchpad.net/ListHelp