On 2 September 2014 15:33, Nina Jeliazkova <jeliazkova.nina@gmail.com> wrote:



On 2 September 2014 15:11, gilleain torrance <gilleain.torrance@gmail.com> wrote:
Hi,

Ah, I forgot to cc the mail to your address - I never get the hang of
that on mailing lists.

Okay, so it's quite a mashup of different technologies then. So two options are:

1) Translate the CDK source with J2S, GWT, or similar.
2) Run the CDK as a back-end web service.

For the first one, I spent a bit of time yesterday looking at
Java->JavaScript translation (or 'transpiling'!) and one thing I
couldn't work out is how dependencies work. The CDK depends on a
number of other java libs - such as jgrapht - which would presumably
have to be also converted to JS.

As for the second, there has been other work in this area.
Ambit/OpenTox for example (http://ambit.sourceforge.net/index.html,
http://www.opentox.org) which has web services for various things,
including 2D layout of molecules. I don't know too much about this
project, and obviously it does a lot more than you require. A very
simple servlet that just takes a SMILES and returns some 2D layout as
(say) a mol file should be easy enough if that's all you need.


AMBIT has services for all functionality required (and more), and the recent version (2.5.8 released 31.08 , ambit2.war download at sourceforge.net ) has integration of ChemDoodle editor (and implementation of server side  layout and aromaticity as required by ChemDoodle sketcher). 




(try the rightmost button with the pen icon)


Oops, sorry, its' not ChemDoodle, but GGA Ketcher :) 

We've used ChemDoodle in a related project (http://www.xmetdb.org/xmetdb

Nina
 

These are REST services (CORS could be enabled and you could run the server and the client JS on different sites). 

I'd be glad to answer specific questions. 

Regards,
Nina
 
Any which way - it's nice to see an exploration of 2D/3D views like
this. Perhaps a crowdfunded approach to this would be successful.
Certainly molecular graphics usually look nice enough to get some
attention!

gilleain

On 9/2/14, Herman Bergwerf <hermanbergwerf@gmail.com> wrote:
> About CDK in MolView (http://molview.org)
>
> Sorry for the late response, I was not subscribed to the CDK devel mailing
> list so I haven't recievee your replies (until I found them on sourceforge)
>
> MolView is built on top of GLmol, ChemDoodle Web and JSmol.
> GLmol is used as primary render engine (fast), ChemDoodle is used for
> crystals (CIF is not yet implemented in GLmol) and JSmol is used as
> Canvas2D fallback and for advanced actions like force field minization,
> dipole calculation and measurements.
>
> I've looked at the CDK source but there are a lot of dependencies (eg.
> other packages in the DiagramGenerator)
>
> Can CDK also generate 3D coordinates? (without force field minimization but
> using templates etc.) Because that would be a great alternative to the
> Chemical Identifier Resolver (http://cactus.nci.nih.gov/chemical/structure)
>
> Maybe I should use J2S after all...
> I'm also thinking about a larger server-side part in MolView (but I'll need
> some crowdfunding) Then I could put CDK on a Tomcat server or something
> like that.
>
> Btw, MolView will be made open-source in the future!
>

------------------------------------------------------------------------------
Slashdot TV.
Video for Nerds.  Stuff that matters.
http://tv.slashdot.org/
_______________________________________________
Cdk-devel mailing list
Cdk-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/cdk-devel