espressomd-users
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [ESPResSo-users] Espresso input/output with Python


From: ICP
Subject: Re: [ESPResSo-users] Espresso input/output with Python
Date: Wed, 28 Dec 2016 20:45:43 +0100

Can you please attach the output of cmake?


Kai Szuttor
Institute for Computational Physics
Universit├Ąt Stuttgart
Allmandring 3
Room 1.084
70569 Stuttgart, Germany
Phone:+49(0)711/685-67707

On 28 Dec 2016, at 19:47, Michael Winokur <address@hidden> wrote:

Hello Kai,

Thanks for your note.

As I noted Espresso's front end has really changed.

I now see why my cmake failed....by "source" it means and cmake ~/espresso-python and not ~/espresso-python/src
So now cmake and make work fine.

None of the tests succeed until I execute "make check".  Afterwards they do.

I am still not up to speed.

When I execute:
./pypresso testsuite/python/h5md.py

it yields

test_f (__main__.H5mdTest)
Test if forces have been written properly. ... skipped 'H5MD not compiled in, can not check functionality.'
test_pos (__main__.H5mdTest)
Test if positions have been written properly. ... skipped 'H5MD not compiled in, can not check functionality.'
test_vel (__main__.H5mdTest)
Test if velocities have been written properly. ... skipped 'H5MD not compiled in, can not check functionality.'

----------------------------------------------------------------------
Ran 3 tests in 0.000s

OK (skipped=3)

There is nothing useful in the
espresso-python/build/src/python/espressomd/io
or
import espressomd.h5md
espresso-python/build/src/python/espressomd/io/writer
and so from espressomd.io.writer import h5md
fails.

I must still be missing something.

Thanks,

Michael


On Tue, Dec 27, 2016 at 4:37 PM, ICP <address@hidden> wrote:
Hi Michael,

I dont know which guideline you followed but what should work is:
- cd into the espresso folder
- mkdir build
- cd build
- cmake ..
- make

You should be able to run the tests and use the h5md features. However if cmake prints some errors or missing libs (h5 lib e.g. should be the mpi parallelized one) you might not be able to use all features of espresso.

Cheers Kai

Kai Szuttor
Institute for Computational Physics
Universit├Ąt Stuttgart
Allmandring 3
Room 1.084
70569 Stuttgart, Germany
Phone:+49(0)711/685-67707

On 23 Dec 2016, at 22:03, Michael Winokur <address@hidden> wrote:

Hello Kai,

Thanks for the quick response.  There are still a few issues:

1.  My e-mail to espressomd-users is getting a rejection notice (but clearly it is getting through)
2.  The build using the python branch is a little cryptic because cmake breaks if I simply follow the directions.  After three minor changes (it is not clear that all three were needed) the source builds (I am on the learning curve with cmake but no matter).
3.  I presume that I must still customize myconfig.hpp
4.  After the build (no error) I note the the directory ~/core/src/python/espressomd/io/writer was missing h5md.py and __init__.py so this line "from espressomd.io.writer import h5md  # pylint: disable=import-error" in the test program failed.

I tried copying these files "by hand" from the python source branch into my working directory.  The above line worked and now there are also *.pyc files in this directory.

However the following line breaks:  "h5_file = h5md.H5md(filename="sample.h5", write_pos=True, write_vel=True,
                    write_force=True, write_type=True, write_mass=False)"
with this message

terminate called after throwing an instance of 'std::domain_error'
  what():  Class 'ScriptInterface::Writer::H5mdScript' not found.
[romano5:16742] *** Process received signal ***
[romano5:16742] Signal: Aborted (6)
[romano5:16742] Signal code:  (-6)
[romano5:16742] [ 0] /lib/x86_64-linux-gnu/libpthread.so.0(+0x11100)[0x7f014eb57100]
[romano5:16742] [ 1] /lib/x86_64-linux-gnu/libc.so.6(gsignal+0xcf)[0x7f014deb5fdf]
[romano5:16742] [ 2] /lib/x86_64-linux-gnu/libc.so.6(abort+0x16a)[0x7f014deb740a]
[romano5:16742] [ 3] /usr/lib/x86_64-linux-gnu/libstdc++.so.6(_ZN9__gnu_cxx27__verbose_terminate_handlerEv+0x15d)[0x7f014724a04d]
[romano5:16742] [ 4] /usr/lib/x86_64-linux-gnu/libstdc++.so.6(+0x8f016)[0x7f0147248016]
[romano5:16742] [ 5] /usr/lib/x86_64-linux-gnu/libstdc++.so.6(+0x8f061)[0x7f0147248061]
[romano5:16742] [ 6] /usr/lib/x86_64-linux-gnu/libstdc++.so.6(+0x8f279)[0x7f0147248279]
[romano5:16742] [ 7] /home/winokur/Espresso/espresso-utd/core/src/script_interface/libEspressoScriptInterface.so(_ZN15ScriptInterface19ScriptInterfaceBase11make_sharedERKNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEENS0_14CreationPolicyE+0xa0b)[0x7f0149dd070b]
[romano5:16742] [ 8] /home/winokur/Espresso/espresso-utd/core/src/script_interface/libEspressoScriptInterface.so(_ZN15ScriptInterface23ParallelScriptInterfaceC1ERKNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEE+0x1eb)[0x7f0149df540b]
[romano5:16742] [ 9] /home/winokur/Espresso/espresso-utd/core/src/script_interface/libEspressoScriptInterface.so(_ZN15ScriptInterface19ScriptInterfaceBase11make_sharedERKNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEENS0_14CreationPolicyE+0x10d)[0x7f0149dcfe0d]
[romano5:16742] [10] /home/winokur/Espresso/espresso-utd/core/src/python/espressomd/script_interface.so(+0x101c0)[0x7f0145e041c0]
[romano5:16742] [11] /usr/bin/python(+0xed414)[0x55aad6ef6414]
[romano5:16742] [12] /usr/bin/python(PyObject_Call+0x43)[0x55aad6ef08f3]
[romano5:16742] [13] /usr/bin/python(PyEval_EvalFrameEx+0x60ff)[0x55aad6f09c6f]
[romano5:16742] [14] /usr/bin/python(PyEval_EvalCodeEx+0x255)[0x55aad6f02345]
[romano5:16742] [15] /usr/bin/python(+0x1161a8)[0x55aad6f1f1a8]
[romano5:16742] [16] /usr/bin/python(PyObject_Call+0x43)[0x55aad6ef08f3]
[romano5:16742] [17] /usr/bin/python(+0x12c54e)[0x55aad6f3554e]
[romano5:16742] [18] /usr/bin/python(PyObject_Call+0x43)[0x55aad6ef08f3]
[romano5:16742] [19] /usr/bin/python(+0x12c287)[0x55aad6f35287]
[romano5:16742] [20] /usr/bin/python(+0xed414)[0x55aad6ef6414]
[romano5:16742] [21] /usr/bin/python(PyObject_Call+0x43)[0x55aad6ef08f3]
[romano5:16742] [22] /usr/bin/python(PyEval_EvalFrameEx+0x60ff)[0x55aad6f09c6f]
[romano5:16742] [23] /usr/bin/python(PyEval_EvalCodeEx+0x255)[0x55aad6f02345]
[romano5:16742] [24] /usr/bin/python(PyEval_EvalCode+0x19)[0x55aad6f020e9]
[romano5:16742] [25] /usr/bin/python(+0x1293af)[0x55aad6f323af]
[romano5:16742] [26] /usr/bin/python(PyRun_InteractiveOneFlags+0x199)[0x55aad6e8c745]
[romano5:16742] [27] /usr/bin/python(PyRun_InteractiveLoopFlags+0xc6)[0x55aad6e8c499]
[romano5:16742] [28] /usr/bin/python(+0x65bfc)[0x55aad6e6ebfc]
[romano5:16742] [29] /usr/bin/python(Py_Main+0x6c1)[0x55aad6eddfb1]
[romano5:16742] *** End of error message ***

So clearly something is still amiss.

How should I proceed?

Thanks again,

Michael


On Fri, Dec 23, 2016 at 2:51 AM, Kai Szuttor <address@hidden> wrote:
Hi Michael,

please use the python branch if you want the latest features of the
python interface. There is an in-core implementation of the hdf5 format
which is documented and also has a sample script.
Please keep in mind that you have to install the hdf5-plugin for VMD
available on github in order to be able to read the written h5 files of
espresso.

As far as I know the imd stuff is not ported and won't be ported to
python, instead we have an openGL and mayavi visualization (also
available in the python branch).

Cheers,

Kai
On Thu, Dec 22, 2016 at 04:46:25PM -0600, Michael Winokur wrote:
>    Hello,
>
>    I am starting to make sure I can use some basic functions and it seems
>    that the best way at present to transfer data to VMD is by the hdf5
>    read/write functions.A  I've been trying to get hdmd5_Write.py to work but
>    this sample file generates an error message:A
>
>    Traceback (most recent call last):
>    A  File "<stdin>", line 35, in <module>
>    A  File "espressomd/h5md.pyx", line 247, in
>    espressomd.h5md.write_to_h5.pos
>    (/home/winokur/Espresso/espresso-master/build/src/python/espressomd/h5md.cpp:7376)
>    AttributeError: 'System' object has no attribute 'n_part'
>
>    My version of Espresso is from the "master" git download.A
>
>    Also, will the imd communication protocol be ported to Python?
>
>    Thanks,
>
>    Michael



reply via email to

[Prev in Thread] Current Thread [Next in Thread]