Dear Barasa,
We have some good introductory tutorials in our Getting Started Guide. In
addition, the Users Guide is a useful reference. Both of these are
available from within Kepler's Help menu, and can also be downloaded in
several language translations here:
Hi Vicki,
A copy of your workflow, or a reduced version that illustrates the problem,
would be helpful.
The attached example workflow has three composites that simply delay for 2
seconds, and firing occurs for all three. How does your workflow differ
from this?
Matt
On Fri, Dec 27, 2013 at
Hi Kenneth --
Sure enough, you are right, it seems that saving a class does not work in
Kepler due to a bug. Once you have created a class in a workflow, you
should be able to right click on it and select 'Save Archive (KAR)...',
which allows you to save in a kar file of your choice in your 'My
I also took a look, and tried a larger array -- the attached workflow
processes 30K integers through an array and back out again, and works fine.
So there isn't an inherent limit at 20K elements. So I'm not quite sure
what's happening with yours, but maybe this will help to debug.
Matt
On
Hi Claas,
So I looked at your workflow and executed it, after getting the EML paths
properly configured for my system. The workflow ran as expected, and
produced the path to the data fram on output Display2. This is as
expected. I then added to the workflow by adding another RExpression
actor,
Hi,
I don't see a confluence module for Kepler in our system. Could you
clarify where you found out about this module, or what it's functionality
covers? In any case, for modules that don't ship with Kepler 2.3, you
should be able to install their associate suite from the Module Manager,
which
Albert --
Did you try running the command line shell with the -nogui and -runwf
arguments? Something like:
kepler.sh -nogui -runkar /home/mbj/myworkflow.kar
The location of kepler.sh will vary by platform, but if you get it on your
path, then you should be good to go. More details are here:
Wade --
Matlab is a bit trickier than R because the linkage to Kepler uses JNI via
a shared library (dll on Windows) for matlab, whereas we use a shell
invocation for R. So, for it to work, both the class path has to be right
and the dll has to match the OS and architecture. I don't have Windows
If all you are trying to accomplish is to get several actors that you
developed to work together, then putting them in the same suite is one
solution. An alternate solution is to make your suite that you are running
depend on the other modules that contain actors that you need. For
example, if
Luis --
If you install the optional Reporting suite, it includes a Workflow Run
Manager which provides a listing of all of your workflow runs, and includes
the execution time for each run along with a lot of other useful
information, as well as the ability to archive runs, return to earlier
runs,
The mailing list archives can be searched here;
http://www.mail-archive.com/kepler-users@kepler-project.org/
http://www.mail-archive.com/kepler-dev@kepler-project.org/
Regarding your R errors -- are you sure R is on your path and executing?
Can you execute any of the R demos that ship with
Josep,
If you already created the actor for Kepler 1.0, then for the most part it
will be able to work in 2.0 with some repackaging and minor redesign. The
actor API is still basically the same, but the way we package actors in
KARs and modules has changed. A perl actor would be generally
binab...@umail.iu.edu wrote:
Thanks for the very detailed response Matt, really appreciate it. I will
dig in further and post again if I encounter questions. Thanks again!!
On Tue, Dec 6, 2011 at 4:32 PM, Matt Jones jo...@nceas.ucsb.edu wrote:
Hi Bina ---
The answer differs for metadata
Hi Bina ---
The answer differs for metadata for data and metadata about processes.
For data, Kepler can utilize metadata from a wide variety of repositories
via the 'EcoGrid' SOAP web service interfaces, but mostly it uses the
Metacat repository run as part of the KNB network for now. That is
this year -- great stuff is afoot in the provenance world.
More details:
http://ipaw2012.bren.ucsb.edu/**index.php/Call_for_**Participationhttp://ipaw2012.bren.ucsb.edu/index.php/Call_for_Participation
Matt Jones
NCEAS, UC Santa Barbara
___
Kepler
Anne --
I thought your papers and approach were great when they originally came
out. Did the actors that you created end up in a released module in Kepler?
And if not, would you be willing to create such a module for a future
release?
Matt
On Wed, Nov 30, 2011 at 2:17 PM, Ngu, Anne Hee Hiong
Hi David --
Kepler is not directly funded by NCEAS, except that NCEAS contributes
in-kind support to helping to maintain the project. We've traditionally
supported Kepler on a variety of extramural grants, some of which are to
NCEAS (e.g., SEEK, REAP, SanParks), and others of which are to other
I believe that the folks at SDSC created the BrowserUI actor for just this
purpose -- you write an HTML page containing the input questions you want,
configure the BrowserUI to display it at the right point in the workflow,
and then gather the output from the configured ports. Ilkay probably has
Marc --
I tried this and it worked for me -- see attached Kepler workflow. I did
discover that one of the displays would not show if the file being copied
already exists in the destination directory -- it seems in that case that it
fails silently, which is arguably a bug.
In the future, could
This is a known bug, reported here:
http://bugzilla.ecoinformatics.org/show_bug.cgi?id=5098
Kepler should not need connectivity, and it should be using a locally-cached
version of the MOML DTD. For some reason TBD the code that loads the DTD as
a local resource is not being called or finding the
bug 5098.
On Mon, Apr 11, 2011 at 10:51 AM, Matt Jones jo...@nceas.ucsb.edu wrote:
This is a known bug, reported here:
http://bugzilla.ecoinformatics.org/show_bug.cgi?id=5098
Kepler should not need connectivity, and it should be using a
locally-cached version of the MOML DTD. For some reason
Yes -- you can set the max heap size by giving -Xmx4096m as an argument to
the java command that you use to start kepler. Set whatever maximum heap
size you need, but also note that this may limit some memory-intensive
actors from working properly. I presume that you are using a shell script
to
at 12:43 PM, Matt Jones mbjones.89 at gmail.com wrote:
Hi Fatemeh,
Kepler 1.0.0 is obsolete, and has been deprecated for a long time. You
should upgrade to kepler 2 (currently 2.1, but soon to be 2.2).
Nevertheless, I got an old version of kepler 1.0.0 up and running on
windows XP, and here's
Dear Madhavi,
That is a fascinating, and very pointed, question. As an open source
project, and as a software system that explicitly has built an
infrastructure to allow 3rd parties to build their own modules, Kepler is
nicely situated for 3rd parties to build high quality extensions that work
Madhavi,
Some Kepler users have definitely used Kepler at scales like this --
Norbert's work on plasma fusion simulations that generate 800GB of data in a
30 hour simulation run under Kepler control comes to mind. However, the
performance at large scale like this will entirely depend on the
Hi John,
I think you missed option 4:
- The list is filled with busy people and it slipped through the cracks.
Sorry.
But, to the point of your questions:
1) When we first shipped Kepler 1.0.0, there were two windows installers:
one with R included, and one without to keep the download sizes
Yes, use a PortParameter instead, and it can be set on an incoming channel,
which overrides the parameter setting.
Matt
On Thu, Dec 2, 2010 at 10:04 AM, lgomes at inf.puc-rio.br wrote:
Hi,
suppose I have an actor which has a parameter and I want that another
actor sets its value, is it
David --
When you are executing from a KAR file, you need to use the '-runkar'
option, rather than the '-runwf' option. Try:
$ /usr/local/Kepler-2.0/kepler.sh -runkar
/home/dlebauer/pecan/kepler/pecan.kar
I'm not sure about the other errors about directories not existing --
someone else will
Thanks, Thomas, for the feedback. I'll ask David to take a look at the issue
you report and see if there's an improvement to the build that is needed.
Regards,
Matt
On Fri, Aug 13, 2010 at 5:40 AM, Tomasz ?ok tzok at man.poznan.pl wrote:
Hi again,
I have just fixed this issue. Previously I
Chris,
I replicated these bugs, and so I have filed defect reports that we will
work on for a future release. I will have to ask the development team to
assess the difficulty in fixing these. The new bugs are:
For the comma in the filename issue:
check functionality of different
Kepler actors, etc.?
Second, I'll try to use separate class loader for my module. I've already
found the docs about it.
Thank you for answers,
Tomek
Matt Jones jones at nceas.ucsb.edu:
I think it would be a good idea to upgrade these jar files to more
I think it would be a good idea to upgrade these jar files to more recent
releases, assuming that testing shows no adverse effects on the modules that
depend on them. As these are in very low-level modules, there may be many
different modules that depend on them, so we would need to test several
Yes, I see I can replicate this on my machine, and it happens even if they
are the same type, or the input type is a subclass of the output type. I
filed a bug report (http://bugzilla.ecoinformatics.org/show_bug.cgi?id=5136).
Thanks for pointing it out.
Matt
On Mon, Aug 9, 2010 at 3:30 PM,
Dear Ming,
We have not yet released the reporting suite -- but we are in the fina
testing stages to do so. We expect to release it very soon, and when we do
you will see the reporting suite in the module manager and will be able to
activate it. We'll plan on sending an announcement when the
if you're interested.
Cheers,
Matt
-- Forwarded message --
Date: Wed, May 12, 2010 at 1:33 AM
Subject: WANDS'10 call for participation -- please help distribute
To: Matt Jones jones at nceas.ucsb.edu
*** WANDS: The 1st International Workshop on Workflow Approaches to New
Data-centric
Hi Tom,
Kepler 2.0.0 is being prepared for release, and is going through testing.
For that purpose, we are developing a series of release candidates. These
will not be supported after 2.0.0 comes out, but any issues you find with
them would be valuable to report to kepler-dev to improve the
Harini,
The expression actor simply evaluates the expression, and places the result
on the output port. So, if you have two ports (a1, a2) and the incoming
value to a1 is 3 and incoming value of a2 is 5, then the expression 'a1'
will output 3 on the output port, and the expression 'a1 + a2' will
Edward and Christopher covered most of the differences between Kepler and
Ptolemy, especially with respect to the 1.0 release. The current
development version of Kepler that is leading to our 2.0 release includes a
number of additional major differences, some of which include:
1) Module-based
I've learned that SDSC seems to be migrating away from the SRB and instead
setting up a new iRODS service. Certainly that may affect the accounts that
people use, and it may also have implications for whether the SRB actors
continue to function properly -- I'm not sure if the iRODS server uses
In addition to what Chad said, a group in Japan worked on internationalizing
Kepler for scientists in Japan (contact Akiko Ogawa on this list). They
might be interested in contributing to improvements that allow for general
internationalization. They also found it was useful to translate the
Hi Barbara,
You need the DDF director for your Boolean switch actors to work, so stick
with that.
Looking at your workflow, I think you need to remove the DDF director from
the inner Check Range subworkflow. This allows the outer workflow to
control the execution with its DDF director. When I
Hi Ufuk,
There is an example in the R chapter of the User Guide (available from
the Help menu in Kepler 1.0.0, or from the kepler-docs subversion
repository). See page 348, section 11.4.1.1.5 (Example Five: Using
Ports Configured as Multiports ). Its not clear from your description
exactly
I also believe that Ilkay Altintas and Daniel Crawl have implemented a
'smart-rerun' feature for at lease some specific scenarios (e.g,
specific restrictions on actors maintaining state) that allows
workflows to be rerun from provenance, or to pick up an aborted
workflow partway through a run.
Dan Crawl seems to have checked in some fixes to the Kepler XSLT actor
that allow it to accept XMLToken as input and no longer writes the
temp files to disk. Does this meet your needs, Chris? For now you
can only access these changes from the development build, but it will
be released with the
Hi Chris,
We've done a few wfs that used GIS operations, mainly for processing
tasters before data analysis. We created a few utility actors for GIS.
One wraps The GDAL library and enables reprojection, clipping, and
image format conversions. There is another actor to handle various
Years ago we (Ilkay) worked on a web start version of kepler that we
tried for a while. It was fairly difficult to set up and maintain,
especially because of a variety of jar signing issues, and we didn't
see many benefits of that approach over a plain installer. So we
never formally pursued it,
Hi Michael,
The MatlabExpression actor calls the local shared library for Matlab,
which it needs to be able to find. If you read the actor
documentation in the Actor Reference Manual (or from the Help menu in
Kepler), it describes how to set the appropriate environment variables
before running
Jose,
That is the Expression actor. You can find it by searching for
Expression, and you can enter any expression that fits the expression
language. There is a detailed description in section 8.1 of the
Kepler user manual starting on page 238. There is also a more
detailed treatment of the
Dear Jos?,
There are many other good workflow systems in existence, but I'd have
to agree with you that Kepler sounds like a good fit. Taverna is a
good system, but is heavily centered on executing web services. We've
done a lot of work in Kepler to create various versions of a command
line
As an alternative, if compiling the source isn't something you want to
undertake, maybe Dan could package up a KAR file for you to import into your
installed version of Kepler. Dan, do you think this would be feasible?
Matt
On Wed, Oct 15, 2008 at 8:59 AM, Daniel Crawl crawl at sdsc.edu wrote:
project, and others.
Matt Jones,
on behalf of the Kepler project members
--
~
Matt Jones
jones at nceas.ucsb.edu
National Center for Ecological Analysis and Synthesis (NCEAS)
UC Santa Barbara http://www.nceas.ucsb.edu/ecoinformatics
Lots of us would like to have this working, and it is right on the top
of the list for some of the SEEK developers. But it is complicated and
we are rearchitecting the system to support archive file formats that
should make it far easier for us to add in new actors or workflows that
are both
52 matches
Mail list logo