[ome-devel] Re: [Fwd: OME analysis]

[Josiah Johnston]

Doug has been slow on the replies to us to. ;) I wouldn't take it  
personally, I think he's been busy with school and hanging out with his  
girlfriend (she recently visited him). You should be informed that he  
is no longer working on the OME project. He left at the end of last  
summer to pursue a PhD program. So while he is still a great source of  
wisdom on the Analysis Engine, he no longer has an obligation to the  
project.
Questions like this usually have the best chance of being answered if  
they are sent to the developer list:  
ome-devel at lists.openmicroscopy.org.uk

We're actually discussing a paper on the Analysis Engine, but that  
won't be out for a little while.
The documentation on the distributed AE is in the source tree of the  
main package, which you can get from http://cvs.openmicroscopy.org.uk/
	OME/src/perl2/OME/Analysis/Engine/README.DISTRIBUTED
A url to that is  
http://cvs.openmicroscopy.org.uk/horde/chora/co.php? 
r=1.7&f=OME%2Fsrc%2Fperl2%2FOME%2FAnalysis%2FEngine%2FREADME.DISTRIBUTED
As far as I know, the distributed AE hasn't been tested for a while,  
but worked just dandy the last time it was exercised. We are planning  
to use it in the Baltimore lab relatively soon, so it not be considered  
dead code.

Some folks in Dundee have been working on a species of distribution  
specialized for deconvolution.

General documentation on analysis modules, chains, etc can be found on  
the main website.  
http://www.openmicroscopy.org/concepts/analysis-modules.html

The state of the Analysis System is that yes, it works, but doesn't  
have a pretty interface for executing chains. You can extend it by  
writing new modules. The easiest way to get a feel for that that is to  
work off of examples in the xml directory OME/src/xml/OME/Analysis/.  
That's yet another thing we could use "Hello World" style introductions  
for.
Right now, it's pretty easy to write modules to wrap legacy matlab code  
or code that can be executed on a command line. The CLI handler hasn't  
been touched for a while, and won't work until it's brought up to  
speed. The matlab handler is under active development, and should work  
once you set the the OME-matlab connections. The best documentation we  
have of writing module definitions in xml is linked to from this page:
	http://www.openmicroscopy.org/api/xml/
Chains can be built by writing XML or by using the chain builder in  
Shoola.
For chains requiring little user input, the ome commander will work  
fine for executing them. Documentation on that is here
	http://www.openmicroscopy.org/system-admin/execute.html

Hopefully this will get you started. Thanks for your interest in the  
project. If you end up making extensions or writing new modules, please  
let us know. We can always use more code contributors, and like to know  
what people end up finding useful.

-Josiah

On Apr 13, 2005, at 1:15 AM, Francis Tang wrote:

> Dear Josiah,
>
> Please excuse me for mailing you like this.  I noticed you recently  
> checked in some changes to the /OME/src/perl2/OME/Analysis module and  
> thought you might be able to help me here.  I'd sent the email below  
> to Doug Creager a few days back but have yet to receive a reply.   
> Would you be able to help me here?
>
> Thanks in advance.
>
> Yours sincerely,
>
> Francis.
>
> -------- Original Message --------
> Subject: OME analysis
> Date: Sat, 09 Apr 2005 02:50:50 +0800
> From: Francis Tang <francis at bii.a-star.edu.sg>
> To: dcreager at alum.mit.edu
> CC: francis.tang at bii.a-star.edu.sg
>
> Dear Doug,
>
> My name is Francis Tang, and I am looking at installing OME for one of
> our projects we have here.
>
> I am especially interested in the analysis aspects of the OME project.
> There is documentation on the website regarding analysis chains and the
> analysis engine, and there is also mention in the 2.4 release note that
> distributed processing of image analyses is now supported.  Would you  
> be
> able to point me to further documentation about the current
> implementation status of the analysis engine?  To what extent is
> distributed execution supported?
>
> I look forward to hearing from you again soon.
>
> Yours sincerely,
>
> Francis.
>
> -- 
> Francis TANG, Post-Doctoral Research Fellow
> Bioinformatics Institute, BMSI, A-STAR, Singapore.
> Tel: +65 64788282  Fax: +65 64789048  Email: francis at bii.a-star.edu.sg
> Add: Matrix L7, Biopolis   WWW: http://www.bii.a-star.edu.sg/~francis/
>