IMP-dev March 2009

imp-dev@salilab.org

6 participants
26 discussions

This week in IMP
by Daniel Russel 06 Mar '09

06 Mar '09

- Helper functions for creating centroid and cover particles have been added since it is kind of painful and obscure to do on your own. The examples now use these helper functions. - a FixedParticleRefiner has been added for setting up ad-hoc refinement. It just returns a fixed list of particles and so means you can skip creating a hierarchy if you just have one particle you want to refine. - the atom module as been created and all atom and molecule related functionality has been moved there and the versions in core marked as deprecated. Please adjust your code accordingly. - there is a pdb reader and writer in the atom module - you can clone a hierarchy with bonds and atom information and all (but this is not tested yet, so check that the results make sense) - there is lots of new geometry: Planes, Spheres, Cylinders, spherical caps and such. When designing the interface of new code to be added to IMP, please: - look around for code that does something similar and to make the interfaces as similar as possible, this is the easiest way to make your code OK. At this point, the template coverage is pretty good. - make sure you follow the coding standards outlined on the web page. Read through those standards before you design your interface if you are not extremely familiar with them. - think about what should be part of the interface and what is an implementation detail. Hide implementation details in internal or anonymous namespaces (and if you don't know what these are and how they work, ask someone). Things should only be in the public interface if you know they will be used outside of the IMP library. - make sure all the tests pass with a non-fast build and with "deprecated=False" passed to scons to turn off deprecated classes and functions. - build the doc (scons doc > /dev/null) and make sure that there are no doc warnings from your code and that it all shows up in the resulting docs Anyway, things are filling out nicely in IMP :-)

1 0

breaking change with move to IMP.atom
by Daniel Russel 05 Mar '09

05 Mar '09

Molecule-related classes such as ResidueDecorator and AtomDecorator are now in IMP.atom. The IMP.core versions of most things have been left for the time being (but marked as deprecated). Unfortunately, technical difficulties with SWIG prevent that being done easily for ResidueDecorator and AtomDecorator and anything which uses them. The fix is just to replace core.AtomDecorator with atom.AtomDecorator (etc.) (a good excuse to learn sed if you don't already know it :-)

1 0

how do you call VectorD<X> from python ?
by Keren Lasker 04 Mar '09

04 Mar '09

thanks, keren.

2 4

is there a fast way to copy attributes from one particle to another ?
by Keren Lasker 04 Mar '09

04 Mar '09

thanks, keren.

2 4

Other optimizers and salilab only modules
by Daniel Russel 03 Mar '09

03 Mar '09

When trying to figure out why some of my optimizations were failing, I wrapped a number of optimizers available in GSL so they can be used in IMP. These include another implementation of CG, a quasi-newton optimizer and simplex (I mostly wanted CG, but the other came for free). We can't ship the wrappers (unless we want to ship a GPL-ed special edition of IMP along with the LGPL one), but they might be useful to people to experiment with. If others are interested, I'll add a salilab_modules directory in IMP for such in-house code which would not be released (we can figure out how to make it inaccessible to the outside world when such is needed).

2 1

IMP_LOG and show
by Keren Lasker 01 Mar '09

01 Mar '09

hi, Sometimes it is useful to use X.show() in logging ( for example transformation.show()) is there a way to do that? thanks, keren.

2 2

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

IMP-dev March 2009