February 2008 - IMP-dev

slight tweak of first patch
by Daniel Russel 15 Feb '08

15 Feb '08

1 0

Two minor cleanups
by Daniel Russel 15 Feb '08

15 Feb '08

First I added protection to the container macros to hide away certain implementation details. Second, I hadn't used the update_state function in Montecarlo since I had looked for states. I changed the name to update_states and fixed the optimizers.

1 0

no more trajectory
by Daniel Russel 14 Feb '08

14 Feb '08

I thought it went away before, but apparently not. This patch kills it.

1 0

Real optimized indexes iterator in ModelData
by Daniel Russel 14 Feb '08

14 Feb '08

I replaced the weird pseudo-iterator with a real one in ModelData (and fixed the optimizers). BTW, impem still doesn't find boost on my mac.

2 1

No unused stats
by Daniel Russel 14 Feb '08

14 Feb '08

I removed the unused stats code in ModelData.h. We can do the same things through ScoreStates (see MaxChangeScoreState) It also removes the flag which is set when particles are activated or deactived since that is not used any more. If we have one, it should be in Model anyway. And moves the vector include to after the IMP includes so it doesn't mask any missing vector includes in them.

1 0

Embed and IMP
by Daniel Russel 14 Feb '08

14 Feb '08

Can we have it so that IMP looks for EMBED with a normal looking directory structure? By this I mean it expects embedroot/include, embedroot/lib embedroot/lib/python-ver/site-packages). Currently I have to point scons to the copy of EMBed in its source dir, but it actually uses the installed copy for most things (since imppy.sh picks up environment variables when the tests are run). I need an installed copy so I can use it without pain in the NPC stuff. Thanks. I'll even overlook that IMP will get the python version wrong on my mac :-)

1 0

Specifying the location of EMBED and MODELLER
by Ben Webb 14 Feb '08

14 Feb '08

As of r376, I removed the compatibility-with-old-lab-people feature of checking for Modeller in $MODINSTALLSVN. If you want to build IMP with support for Modeller, either specify this on the command line, i.e. scons modeller=/my/modeller/location or (recommended) make a file called config.py in the top-level IMP directory and set it like any other Python variable in there: modeller="/my/modeller/location" The same is true for EM/emlib/EMBED; set its location with the embed variable: embed="/my/embed/location" Ben -- ben(a)salilab.org http://salilab.org/~ben/ "It is a capital mistake to theorize before one has data." - Sir Arthur Conan Doyle

1 0

boost check
by Daniel Russel 14 Feb '08

14 Feb '08

it fails on my mac even though I have boost installed (from macports). And I can't even run scons --help to see if there is a way to point it at the right place. More reason to respect environment variables.

2 7

Deleting modeldata
by Daniel Russel 14 Feb '08

14 Feb '08

ModelData was never deleted. It is now.

1 0

sconscript
by Daniel Russel 14 Feb '08

14 Feb '08

some montecarlo stuff got left off of the sconscript files

1 0