[Joelib-help] Re: JOELib and CMLRSS and WWMM and OpenMolecules

[Peter Murray-R. <pm...@ca...>]

Greetings all,
I have been away from my mail for 24 hours and this all seems to be 
exploding! Fantastic. This is written before reading everything...

At 14:03 19/03/2004 +0100, E.L. Willighagen wrote:
>-----BEGIN PGP SIGNED MESSAGE-----
>Hash: SHA1
>
>On Friday 19 March 2004 12:33, Joerg K. Wegner wrote:
> > 1. I've again released a new JOELib version with sequential SAX parser
> > for HUGE CML files for uncompressed and compressed (ZIP) files (i've
> > found a workaround)!!!
>
>Excellent!
>
> > I think after this 'stable' CML standard support i can switch from SDF
> > to CML2 as standard format, because it can be compressed and is much
> > more verbose, which is recommended for storing descriptors (WITH KERNEL
> > informations).
> > https://sourceforge.net/projects/joelib/
>
>/me is happy

This sounds great. We are happy to explore how CML can support these type 
of applications. The challenge is not to write CML (which is easy) but to 
create a structure which is usable by others. So imagine that you got a CML 
file from someone you hadn't talked to - could you use it?

A good example is properties. Let's say Joerg creates something that looks 
like:
<molecule xmlns:joerg="http://de.uni-tuebingen.informatik/wegner/dict">
   <atomsSnipped/>
   <propertyList>
     <property>
       <scalar type="xsd:float" dictRef="joerg:clogp" 
title="cLogP>1.23</scalar>
     </property>
   </propertyList>
</molecule>

what should RssViewer do? Options might be
  1 display the molecule and omit the propertyList completely
  2 notify the user that there were properties but that RSS viewer couldn't 
view them
  3 list the values along with the dictRef, title *and* the xmlns. This 
identifies the property but doesn't explain it
  4 visit joerg's website and retrieve the dictionary entry.

I would suggest we aim for at least 2. However we are going into completely 
fresh semantic web territory - very exciting.

> > 2. I've added the 'attributearray' CML writer which is conform to the
> > OpenBabel array, only because its better layout in RSS files. I would be
> > happy if you will add a CML generation section for JOELib (analogue to
> > OpenBabel) to your CMLRSS documentation. Two properties in
> > joelib.properties need to be set:
> > joelib.io.types.ChemicalMarkupLanguage.output=attributearray
> > ## a first step to 'reproducable' descriptor calculation algorithms
> > 
> joelib.io.types.ChemicalMarkupLanguage.output.storeChemistryKernelInfo=false
> >
> > 3.1. I've published a 'QSAR, LBDD and SBDD'-RSS feed. I plan to release
> > from time to time (monthly) some algorithmic and QSAR, LBDD and SBDD
> > topics with literature references.

This is wonderful. Joerg, please liaise with YY so that (a) we notify 
people and (b) can include your name in the RSS providers at our ACS 
presentation. [This applies to any other list members - if you have a 
CMLRSS feed and let YY/Henry know in the next 2-3 days we will promote it.)

Again, we are promoting CMLRSS at the NeSC meeting - as announced earlier. 
Egon - please can you liaise with YY - within the next 3-4 days as to what 
version we can put on. (We can always announce a later one at the meeting).

Hope to post more later

P.

Peter Murray-Rust
Unilever Centre for Molecular Informatics
Chemistry Department, Cambridge University
Lensfield Road, CAMBRIDGE, CB2 1EW, UK
Tel: +44-1223-763069