[Rdkit-discuss] Rings

[Evgueni Kolossov]

Hi Greg,

Is there any way to get separately total number of Rings and number of
SMALLEST rings?

Regards,
Evgueni

Re: [Rdkit-discuss] Rings

[Greg Landrum]

Dear Evgueni,

On Wed, Aug 20, 2008 at 11:46 AM, Evgueni Kolossov  wrote:
>
> Is there any way to get separately total number of Rings and number of
> SMALLEST rings?

The number of smallest rings is directly accessible through the SSSR
functions, the total number of rings is not. For the total number of
rings you'd need to do a breadth-first search and accumulate the rings
you encounter.

-greg

Re: [Rdkit-discuss] Rings

[Evgueni Kolossov]

So, RingInfo::numRings return the number of smallest rings?

2008/8/20 Greg Landrum 

> Dear Evgueni,
>
> On Wed, Aug 20, 2008 at 11:46 AM, Evgueni Kolossov 
> wrote:
> >
> > Is there any way to get separately total number of Rings and number of
> > SMALLEST rings?
>
> The number of smallest rings is directly accessible through the SSSR
> functions, the total number of rings is not. For the total number of
> rings you'd need to do a breadth-first search and accumulate the rings
> you encounter.
>
> -greg
>



-- 
Dr. Evgueni Kolossov (PhD)
ekolos...@gmail.com
Tel. +44(0)1628 627168
Mob. +44(0)7812070446

Re: [Rdkit-discuss] Rings

[Greg Landrum]

On Wed, Aug 20, 2008 at 9:17 PM, Evgueni Kolossov  wrote:
> So, RingInfo::numRings return the number of smallest rings?

Basically yes.
If you used the standard ring-finding code (called by the sanitization
algorithms that the standard molecule parsers unless you tell them not
to), then a molecule's ringInfo structure contains a symmetrized set
of smallest rings. This is basically the SSSR set but it contains an
extra ring in cases like cubane - C12C3C4C1C5C4C3C25 -  5 SSSR rings,
6 symmetrized SSR rings;  or adamantane - C1C2CC3CC1CC(C2)C3 - 3 SSSR
rings, 4 symmetrized SSR rings.

-greg

[Rdkit-discuss] Build Failure in RDKit May Release

[Andrew Fant]

I tried to build the May 2008 release tarball today on a centos 4.6 
system with gcc-3 and boost 1.36.0, and ran into a C++ problem that is 
beyond my ken.


I can get everything else to build, but in Code/Geometry/Wrap/Point.cpp, 
I get a boost error that is too long to cut and paste here.  The gist of 
it appears to be that there is a bad template definition for "operator+" 
somewhere in that source file.


If anyone is interested, I will gladly pass along a copy of the error. 
Otherwise if this error jogs anyone else's memory, that would be even 
better still.


Thanks,
Andy

Re: [Rdkit-discuss] Build Failure in RDKit May Release

[Greg Landrum]

Dear Andy,

On Wed, Aug 20, 2008 at 11:40 PM, Andrew Fant  wrote:
> I tried to build the May 2008 release tarball today on a centos 4.6
> system with gcc-3 and boost 1.36.0, and ran into a C++ problem that is
> beyond my ken.

I haven't tried boost 1.36.0 yet, so you're breaking new ground.

> I can get everything else to build, but in Code/Geometry/Wrap/Point.cpp,
> I get a boost error that is too long to cut and paste here.  The gist of
> it appears to be that there is a bad template definition for "operator+"
> somewhere in that source file.

The boost error messages are rather verbose, but with enough practice
you learn which parts to ignore and where the interesting pieces are.
Of course you shouldn't have to get that practice by interpreting the
errors in *my* code.

> If anyone is interested, I will gladly pass along a copy of the error.
> Otherwise if this error jogs anyone else's memory, that would be even
> better still.

Please send me the error message, I may be able to diagnose things
quickly. If I can't find a quick solution, I'll try a boost 1.36.0
install this weekend and see if I can figure out what's going on.

Best Regards,
-greg