2009
04.14
04.14
Finally I added a quit button to the splash screen allowing you to quit the Cyberintegrator at the splash screen. This was something I always wanted to be able to do but was to lazy to add.
Complex problem-solving and data analysis via exploratory workflow creation.
Finally I added a quit button to the splash screen allowing you to quit the Cyberintegrator at the splash screen. This was something I always wanted to be able to do but was to lazy to add.
I’m in the process of doing some refactoring in the code. There will not be any visible new things for the Cyberintegrator but this should hopefully make things easier to maintain in the future. This refactoring means that the current unstable build is really unstable and should probably not be used. Hopefully later this week I’ll post an update that it is fixed.
To compile cyberintegrator from source code you can use the following:
1) check out from svn the automatic build project
svn co https://svn.ncsa.uiuc.edu/svn/cyberintegrator/trunk/hudson.build
2) copy the following as a script which should enable building of cyberintegrator. You might have to edit build.sh to point to your eclipse location.
export SVNPRODUCT="edu.uiuc.ncsa.cyberintegrator.rcp"
export PRODUCT="Cyberintegrator.product"
export FEATURELIST="edu.uiuc.ncsa.cyberintegrator.samplecontext"
#export CONFIGS="win32,win32,x86"
export CONFIGS="linux,gtk,x86_64"
sh ./build.sh
3) make sure you have zest installed in eclipse (help->software updates search for zest)
4) install svn pde build from http://sourceforge.net/projects/svn-pde-build/
You can now set properties for the engine used inside the Cyberintegrator server. Create a file called server.properties in the Cyberintegrator server folder. In this properties file you can set the following properties (same as a ThreadedEngine):
queue the file that contains the queue.
workers the number of worker threads (parallel steps execution).
onejvm should same JVM be used for the executors.
During the NEES demo I noticed that with large amounts of data it takes a while to generate the dataview, esp when updating tags etc. I spend some time to improve the performance of it. It will now less frequent do a full refresh (which can be done manually using F5). This should show a dramatic update in speed for the DataView. These changes will be pushed into the ToolView as well.
My plan is to try an write more frequent updates about what has happened. Some of the updates might be related to a jira tasks others might be updates that I did while working on the Cyberintegrator.
We have left the alpha stage and entered the beta stage with Cyberintegrator. One of the major changes that is now in Cyberintegrator is a new way of selecting what context to work with. You will see this at login time, a new combo box has been added to let you pick the context you want to work with. You can also add a new context definition at this time. In the cyberintegrator you can now see what context you are working with at the bottom in the status line, and from the File menu you can now quickly switch from one context to the next. The preferences to set up a context have been removed.
Instead of a step failing it is now possible for just a few outputs to be marked as failed. Allowing steps to continue whos inputs are not failed. Following is a list of all issues resolved in beta 1.
Over the last two weeks I have been working on the Cyberintegrator to fix numerous bugs that were introduced during the conversion of the context. Please update to this version if you use alpha 6.
One new feature that is introduced is the ability to save a whole context.
Following is a list of all issues resolved in alpha 7.
You can now download Cyberintegrator 2.0 alpha 6. This version includes a major rewrite of the way information is stored in context. As always you should not have to worry about this, the Cyberintegrator will automatically convert your data to the new version. However with this major change I would like to advice you to back up your context. To do this start your current installation of Cyberintegrator and go to Window->Preferences. Select Tupelo Preferences and next Tupelo Context. Check to see what context you are using, if using the local context go to the preferences for the local context and copy the directory this points to. Other contexts can be backed up but not as easy. The next release of Cyberintegrator will have a simple method to allow you to export/import a complete context.
The external executor is now capable of handling tools that have inputs that can appear multiple times, for example when combining images into a single image it was necessary before to create a tool that takes 2 inputs, merge those 2 inputs and rerun the tool with the output generated by the previous step and another image. To merge N images this would requre N-1 steps. Now the tool can specify that certain inputs can appear more than once. Now this requires 1 step to merge N images.
There is a new preference page that allows you to change the engines available to the Cyberintegrator as well modify the settings for the Cyberintgrator engines. There is now also a remote execution engine that allows you run steps on a remote machine. This requires the Cyberintegrator Server to be installed, as well as the use of a shared context.
Following is a list of all issues resolved in alpha 6.
You can now download Cyberintegrator 2.0 alpha 5. One of the new features in this version of the Cyberintegrator is the inclusion of a new executor, Human executor. This executor can be used to have the user perform an action, for example ask the user to take a picture and upload the image, or ask the user to run a tool that can not be encapsulated as a tool in the Cyberintegrator.
As of this version the Cyberintegrator is build using eclipse 3.4. This allowed us to add status messages at the bottom of the Cyberintegrator. This will show you who you are logged in as, as well as the number of jobs that are queued for the local engine and are running.
Following is a list of all issues resolved in alpha 5.