Announcement: jClustering, dynamic clustering framework

[José María Mateos]

Hi everybody,

jClustering is a general framework for dynamic imaging clustering (4D)
development built as an ImageJ plugin. If anyone is interested in this
area, here you can find development and use instructions, all the code
and binary downloads: https://github.com/HGGM-LIM/jclustering

All comments are welcome.

Best,

José María Mateos.

--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html

[dscho]

Hi José María,

On Tue, 25 Sep 2012, José María Mateos wrote:

> jClustering is a general framework for dynamic imaging clustering (4D)
> development built as an ImageJ plugin. If anyone is interested in this
> area, here you can find development and use instructions, all the code
> and binary downloads: https://github.com/HGGM-LIM/jclustering
>
> All comments are welcome.

This is great! Thanks for sharing this project with the community.

To make it easier to build and develop for others, I turned it into a
Maven project. For Eclipse users, this means that you can simply call
File>Import, then select 'Existing Maven Project...' and specify the
jclustering directory.  No need to search & download the appropriate
ImageJ or commons-math version, that is already part of the project
specifications. Other IDEs such as Netbeans and IntelliJ have equally
simple options to import Maven projects.

You can easily integrate my changes by accepting my pull request on
GitHub:

        https://github.com/HGGM-LIM/jclustering/pull/1

Ciao,
Dscho

P.S.: I also removed the javadocs from the repository because you can
easily generate up-to-date ones by calling 'mvn site'; the result will be
in target/site/apidocs/. This avoids having stale documentation in the
source code repository.

--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html

[José María Mateos]

On 25/09/2012 19:17, Johannes Schindelin wrote:

> This is great! Thanks for sharing this project with the community.
>
> To make it easier to build and develop for others, I turned it into a
> Maven project. For Eclipse users, this means that you can simply call
> File>Import, then select 'Existing Maven Project...' and specify the
> jclustering directory.  No need to search & download the appropriate
> ImageJ or commons-math version, that is already part of the project
> specifications. Other IDEs such as Netbeans and IntelliJ have equally
> simple options to import Maven projects.
>
> You can easily integrate my changes by accepting my pull request on
> GitHub:
>
> https://github.com/HGGM-LIM/jclustering/pull/1

Hello Johannes,

Thanks a lot for your help and your pull request. Currently I have a
small problem turning my project into a .jar file: as I would like
people to code new metrics and techniques for this framework, I've set
up an automatic class detection system that will automatically load
every class from the metrics/ and techniques/ directories. That would
allow external developers, even with little Java experience, to try
their own algorithms. I have not yet figured out how to do that inside a
.jar file, but I want you to know that I appreciate your suggestions
immensely and will incorporate them as soon as I can get my mind around
these issues. I hope the ImageJ user community finds this project useful
and aim to make it better.

Best,

José María Mateos.

--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html

[dscho]

Hi José María,

On Wed, 26 Sep 2012, José María Mateos wrote:

> On 25/09/2012 19:17, Johannes Schindelin wrote:
>
> > You can easily integrate my changes by accepting my pull request on
> > GitHub:
> >
> >  https://github.com/HGGM-LIM/jclustering/pull/1
>
> Thanks a lot for your help and your pull request. Currently I have a
> small problem turning my project into a .jar file: as I would like
> people to code new metrics and techniques for this framework, I've set
> up an automatic class detection system that will automatically load
> every class from the metrics/ and techniques/ directories.

Sorry, I did not realize. Will fix this (so that metrics/ and techniques/
are parsed even inside .jar files). That way, contributing for scientists
should get easier than before; we could even make a small screencast
describing how to set up Eclipse, clone the source code, and add a new
metric.

Once I am done, you should get a notification by GitHub saying so ;-)

Ciao,
Johannes

--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html

[ctrueden]

Hi José and everyone,

> I would like people to code new metrics and techniques for this
> framework, I've set up an automatic class detection system that will
> automatically load every class from the metrics/ and techniques/
> directories.

This is a case where ImageJ2 would be very helpful. The plugin framework is
extensible, so you can create your own style of plugin such as a "Metric"
or "Technique". The system automatically discovers available plugins, and
allows you to request all plugins of a particular type. With ImageJ2 you no
longer need to write code that scans for and loads such things.

Regards,
Curtis


On Wed, Sep 26, 2012 at 10:15 AM, Johannes Schindelin <
[hidden email]> wrote:

> Hi José María,
>
> On Wed, 26 Sep 2012, José María Mateos wrote:
>
> > On 25/09/2012 19:17, Johannes Schindelin wrote:
> >
> > > You can easily integrate my changes by accepting my pull request on
> > > GitHub:
> > >
> > >  https://github.com/HGGM-LIM/jclustering/pull/1
> >
> > Thanks a lot for your help and your pull request. Currently I have a
> > small problem turning my project into a .jar file: as I would like
> > people to code new metrics and techniques for this framework, I've set
> > up an automatic class detection system that will automatically load
> > every class from the metrics/ and techniques/ directories.
>
> Sorry, I did not realize. Will fix this (so that metrics/ and techniques/
> are parsed even inside .jar files). That way, contributing for scientists
> should get easier than before; we could even make a small screencast
> describing how to set up Eclipse, clone the source code, and add a new
> metric.
>
> Once I am done, you should get a notification by GitHub saying so ;-)
>
> Ciao,
> Johannes
>
> --
> ImageJ mailing list: http://imagej.nih.gov/ij/list.html
>

--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html

[José María Mateos]

On 26/09/2012 17:37, Curtis Rueden wrote:
> This is a case where ImageJ2 would be very helpful. The plugin framework is
> extensible, so you can create your own style of plugin such as a "Metric"
> or "Technique". The system automatically discovers available plugins, and
> allows you to request all plugins of a particular type. With ImageJ2 you no
> longer need to write code that scans for and loads such things.

Ooooh, that would be a dream. In any case, that brings up a question:
once ImageJ2 is released, will the regular ImageJ be deprecated? If that
is the case, then I guess it makes little sense to try to fix my current
code, as it will have to be ported anyway.

Best,

José.

PS: I'm glad my plugin is getting good suggestions from you guys.

--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html

[ctrueden]

Hi José,

> In any case, that brings up a question: once ImageJ2 is released, will the
> regular ImageJ be deprecated? If that is the case, then I guess it makes
> little sense to try to fix my current code, as it will have to be ported
> anyway.

I believe that Wayne's plan is to keep working on the IJ1 codebase. With
ImageJ2, we are trying to maintain backward compatibility with ImageJ 1.x
as much as possible. We expect that some people will continue using IJ1 for
years to come. Since ImageJ2 includes a copy of ImageJ1, any new IJ1
features and bugfixes will be automatically available in IJ2 as well.

That said, ImageJ2's architecture offers more power and flexibility when
developing plugins, UI-independence and headless support, server-side
analysis, better interoperability with other software, and much more. Our
goal is to eventually migrate e.g. all the Fiji plugins over to IJ2, though
it is certainly not something that will happen overnight.

It is also possible to "mix and match" IJ1 and IJ2 code (though with a few
caveats) to ease migration of existing code.

In short, you will certainly not "have to" port your code to IJ2—but there
will be advantages to doing so. If you want to get a jump start on
exploring ImageJ2 now, we are happy to help, offer advice on how to get
started, etc. Or if you would rather wait until ImageJ2 comes out of beta,
that is understandable, and there will certainly be more documentation by
then.

Regards,
Curtis


On Wed, Sep 26, 2012 at 10:53 AM, José María Mateos <[hidden email]>wrote:

> On 26/09/2012 17:37, Curtis Rueden wrote:
>
>> This is a case where ImageJ2 would be very helpful. The plugin framework
>> is
>> extensible, so you can create your own style of plugin such as a "Metric"
>> or "Technique". The system automatically discovers available plugins, and
>> allows you to request all plugins of a particular type. With ImageJ2 you
>> no
>> longer need to write code that scans for and loads such things.
>>
>
> Ooooh, that would be a dream. In any case, that brings up a question: once
> ImageJ2 is released, will the regular ImageJ be deprecated? If that is the
> case, then I guess it makes little sense to try to fix my current code, as
> it will have to be ported anyway.
>
> Best,
>
> José.
>
> PS: I'm glad my plugin is getting good suggestions from you guys.
>
>
> --
> ImageJ mailing list: http://imagej.nih.gov/ij/list.**html<http://imagej.nih.gov/ij/list.html>
>

--
ImageJ mailing list: http://imagej.nih.gov/ij/list.html