Re: introduction distributed ptolemy
There was a mail configuration problem with the ptolemy-hackers list, so I'm forwarding a couple of bounced messages. Rather than trying to merge them, I'm sending them in the order they were originally sent. -Christopher From: Ferdinando Villa [EMAIL PROTECTED] Reply-To: [EMAIL PROTECTED] To: Bertram Ludaescher [EMAIL PROTECTED] Cc: Edward A Lee [EMAIL PROTECTED], Tobin Fricke [EMAIL PROTECTED], [EMAIL PROTECTED], [EMAIL PROTECTED], yang [EMAIL PROTECTED], [EMAIL PROTECTED] In-Reply-To: [EMAIL PROTECTED] References: [EMAIL PROTECTED] [EMAIL PROTECTED] [EMAIL PROTECTED] Content-Type: text/plain Organization: University of Vermont Message-Id: [EMAIL PROTECTED] Mime-Version: 1.0 X-Mailer: Ximian Evolution 1.4.6 Date: Mon, 14 Jun 2004 11:25:19 -0400 Content-Transfer-Encoding: 7bit On Mon, 2004-06-14 at 11:01, Bertram Ludaescher wrote: Ideally the ports of an actor should (or at least could) have multiple types: - the data type (including say XML Schema type), - the semantic type (e.g. a concept expression describing more formally what else might be known about the data flowing through the port) [[aside for Ferdinando: our reductionist/separatist approach does not preclude forever an integrated modeling solution - it's just bottom up to get sth useful soon/in finite time ;-]] Hi Bertram! one point - if you read reductionistic (which I probably wrote somewhere) as reductive you're misinterpreting me - we ecologists mostly see reduct. vs. holistic as complementary (with hierarchical thinking as a possible integrating framework) so when we say reductionistic, we mean exactly what you also mean... one GOOD way to look at the problem, usually the most practical, easier to study, while often the least conducive to synthetic understanding... but NOT separatist Philosophy aside, here's a more SEEK-specific provocation: don't you think that the conceptual/holistic/top-down approach is the one that needs the semantic types, while the workflow/reductionist/ptolemy approach would be just fine with just the storage/machine types of whatever virtual machine AMS will be? Also: where do the transformations belong (scaling, units)? I'd argue they belong mechanically to the reductionistic side - just like all other actors, created to calculate a concept - and if the user don't need to see them, it's not because we hide them up in the conceptual description, but because they're actors in the workflow, and the conceptual description that users work with is not the workflow. Maybe we're mixing the sides up somewhat, and if so, is this ok... or is it going to postpone the beautiful moment of clarity when we all realize that we've all been thinking the same thing all along? Cheers, ferdinand - the event consumption/production type (useful for scheduling a la SDF) - the communication type (through the Ptolemy/Kepler client, directly via say FTP or HTTP) etc At some levels of modeling ones does explicitely hide such detail from the modeler/user but at other levels this might be a good way of overcoming some scalability issues (if you have terabyte data streams you want them to go directly where they need to) A related problem of web servies (as actors) is that they send results back to the caller (Kepler) and don't forward them to the subsequent actor making large data transfers virtually impossible.. A simple extension to the web service model (anyone knows whether that's already done???) would allow for data to include *references* so that a process would be able return to Kepler just a reference to the result data and that reference would be passed on to the consuming actor who then understands how to derefernce it. This simple extension seems to be an easy solution to what we called before the 3rd party transfer problem: --[Actor A] --- [ Actor B] -- ... To stream large data set D from A to B w/o going through Ptolemy/Kepler one can simply send instead a handle D and then B, upon receiving D, understands and dereferenes it by calling the appropriate protocol (FTP/gridFTP, HTTP, SRB,...) Note that there are already explicit Kepler actors (SRBread/SRBwrite, gridFTP) for large data transfer. More elegant would it be to just send handles in the form, e.g., dereference(http://...ref-to-D..) Note that the special tag 'derefence' is needed since not every URL should be dereferenced (a URL can be perfectly valid data all by itself) It would be good if we could (a) define our extensions in line with web services extensions that deal with dereferencing message parts (if such exists) and (b) can work on a joint Kepler/Ptolemy/Roadnet/SEEK/SDM etc approach (in fact, Kepler is such a joint forum for co-designing this together..) Bertram PS Tobin: I recently met Kent and heard good news about ORB access in Kepler
Re: Ptolemy applets problems
Efrat Jaeger [EMAIL PROTECTED] writes: - I created the following applet, http://users.sdsc.edu/~efrat/workflows/queryPGAP.htm, very similar to the demo applets. Apparently it loads on some computers but on others I get a java.lang.NoClassDefFoundError: ptolemy/actor/gui/MoMLApplet error. On the same computers the existing demo applets load perfectly. I wasn't able to figure out what causes the exception. Does anyone have any suggestions? - Applets are notoriously tricky to get right. I looked over the source code and did not see anything obvious. There is some documentation at http://ptolemy.eecs.berkeley.edu/ptolemyII/ptIIlatest/ptII/doc/coding/applets.htm and there is a custom applet chapter in Volume 2 of the Ptolemy II design doc: http://ptolemy.eecs.berkeley.edu/papers/03/ptIIDesignSoftware The applet code you are using is the complex applet code that will work under Netscape 4.x. If you do not require support of Netscape 4.x, then you could use the much simpler applet .. /applet code. In fact, for debugging, I would use the simple applet code so that you are testing the same html code on the different platforms. Another trick is to look at the website logs and see what files are being downloaded. With Java 1.4.2, I noticed that if a class was loaded from a jar file, then it seemed like the applet class loader would not find individual .class files on the file system. The solution was to make sure and have the archive= tag include jar files that included all the .class files. I'm not sure about the details of this new behaviour, I poked around on the Sun website but did not see anything. When I tried to run your applet, I had to install Java 1.4.1_02, the applet would not run with Java 1.5.0-beta under Mozilla. Applets can have dynamic versioning or static versioning. Dynamic versioning permits a JRE greater or equal to a a certain version to be used. Static versioning requires that a certain version of JRE be used. The details are covered in: http://java.sun.com/products/plugin/versions.html From the title tag, it looks like your applet is based on one of the Ptolemy Eye applets. It looks like your applet and the Ptolemy II 3.0 SDF Eye applet are using dynamic versioning under IE, but static versioning under Mozilla? This would be a bug. The IE section says: classid=clsid:8AD9C840-044E-11D1-B3E9-00805F499D93 which http://java.sun.com/products/plugin/versions.html indicates is for dynamic versioning. The Netscape section says: application/x-java-applet;jpi-version=1.4.1_02 http://java.sun.com/products/plugin/versions.html indicates is for static versioning. The applet code shipped in Ptolemy II 3.0 in theory uses static version to restrict the version of the JVM can be used to 1.4.1_02. This was done because of a bug in 1.4.1 and 1.4.1_01 that machines that use a particular video chip set to hang. Since Professor Lee and I were affected by this bug, we restricted the JVM choice to 1.4.1_02. I tried to make the applet use dynamic versioning for Netscape so we would accept 1.4.1_02 or greater, but could not get this to work, it appeared to be a problem with the patch level. If possible, it is better to use dynamic versioning. The Ptolemy II 4.0 applets use dynamic versioning correctly to require Java 1.4.2 or later. In principle, Java 1.5.0-beta should work with these applets. In general, applets have restrictions: 1) Don't use the readFile() expression function in a model that is within an applet. This is because of restrictions in the applet class loader. 2) If possible, use the simple applet .../applet code and avoid the really complex javascript code. If you really think you need to support Netscape 4.x, consider looking at the web server logs for the number of Netscape 4.x users and evaluate the cost of supporting applets for those users. Note that Ptolemy II 4.0 supports Netscape 4.x. Perhaps the next release of Ptolemy II applets will not support Netscape 4.x? 3) The MoMLViewerApplet class does not permit look inside. It would be cool if we could use the Single Window interface created by Thales within an applet. To run the single window interface in Ptolemy II 4.0, do $PTII/bin/vergil -single The idea would be to create an applet that would use this interface 4) The code generator will generate applets for a model. See $PTII/doc/codegen.htm for somewhat out of date documentation. There will be a chapter in the design doc about how to use the applet generator. 5) For vergil applets, it might be better to create Web Start versions. I noticed that the Java Event Listener Tutorial is now using Web Start, see http://java.sun.com/docs/books/tutorial/uiswing/events/example-1dot4/ I started hacking up the applet code generator to generate both applets and Web Start configurations, but I never really finished that work BTW - If you use the controls parameter in your applet, then users can run the applet. To do this,
Ptolemy II 4.0-beta is available for download
. * Hierarchical classes, subclasses, and inner classes. Version 4.0 introduces actor-oriented classes, subclasses, and inner classes with inheritance. This is a major improvement in the modularity mechanisms available in actor-oriented design. A side effect of the class mechanism is that MoML files are now much smaller since they do not duplicate information defined in a base class. * Lifecycle management Version 4.0 introduces the first of a family of lifecycle management actors, which are actors that control the execution of associated models. This example uses one model to control repeated runs of another. Other Features that are new in Ptolemy II 4.0 * Statically Checked Unit System The new statically checked units system associates units with ports and checks compatibility in connections. * Higher-Order Components. Version 4.0 adds higher-order components, which are components (like actors) that operate on other components rather than directly on data. * Interactive decorations in Vergil diagrams. Actors can have custom icons, diagrams can be decorated with decorative elements, and parameters of the icons and decorative elements can be tied to values compute by the model. * Reconfiguration analysis A new capability helps to ensure the safe use of reconfiguration. Currently it is primarily used by the code generator to ensure that synchronous dataflow rate parameters are constant over the lifetime of the model. This analysis also enables parameterized static scheduling of some models where rate parameters change only between iterations of a sub-model. * Causality analysis A new (and still evolving) capability unifies causality analysis that is required by several domains. It provides a systematic framework for determining whether events at a particular input affect a particular output. * Copernicus The copernicus code generator has matured significantly: + Better handling of data types, reconfiguration, and attributes other than simple parameters. + Code Generation for the Giotto domain. + A primitive code generator for generating atomic actors from models. + Static analysis of reachable code for generating self-contained jar files. Plans - This release includes only the documentation for VisualSense, the Ptolemy User's Manual has not yet been updated. We are currently working on updating the documentation. My hope is that the Ptolemy II 4.0 production release will ship in mid June. Ptolemy II 4.0-beta is primarily for testing by seasoned developers and people who have used Ptolemy II before. Bugs The Ptolemy II 4.0-beta limitations are listed in: http://ptolemy.eecs.berkeley.edu/ptolemyII/ptII4.0/ptII4.0-beta/doc/ptII4.0.release.htm The Vergil welcome window has a link to a similar list. If you find bugs, check the limitations page inside Vergil, then check http://ptolemy.eecs.berkeley.edu/ptolemyII/ptII4.0/limitations.htm and then send email to [EMAIL PROTECTED] In your email, please be sure to include: * What operating system you are operating under (Windows XP sp1? Solaris 8?) * What version of Java you are running. (Run java -version, or use our Java Version Applet at http://ptolemy.eecs.berkeley.edu/ptolemyII/ptII4.0/JavaVersionApplet.htm * Whether the problem occurs in * Vergil * The local version of the web pages * The PtolemyII 4.0-beta website version of the web pages * The Web Start version -Christopher Christopher Hylands Brooks [EMAIL PROTECTED] University of California Programmer/Analyst Chess/Ptolemy/GSRC US Mail: 558 Cory Hall #1770 ph: 510.643.9841 fax:510.642.2739 Berkeley, CA 94720-1770 home: (F-Tu) 707.665.0131 (W-F) 510.655.5480 (office: 400A Cory) Posted to the ptolemy-hackers mailing list. Please send administrative mail for this list to: [EMAIL PROTECTED]
Re: I am using ptolemy II to develop an optical network simulator.
In comp.soft-sys.ptolemy, [EMAIL PROTECTED] (Kaizen) writes: This is for an academic project so the complexity of the simulator will not reach that of something commercial like Optiwave or VPI. VPI and Ptolemy Classic have a connection, http://ptolemy.eecs.berkeley.edu/conferences/99/opticalsystems.pdf Says VPI products comprise: *Photonic Transmission Design Suite (PTDS) based on Ptolemy 0.7 . . . The components that I am simulating include the transmitter, the fiber and the receiver. In the transmitter which I am working on now, there is a PRBS that goes thru a NRZ and that signal is modulated with that of a DFB laser. I don't know much about optical systems, so the above are just TLAs to me :-) What kind of domain should I be simulating this in, Hard to say. Usually network simulators use a timed domain, such as DE. However, it can be tricky to create large efficient DE models - it is hard to get the level of abstraction right. Someone else might have a suggestion I am currently doing it in SDF, but so far the entire simulation just slows down a lot when I leave it running for a while. 1. Make sure you are not using the Display actor, it eats up memory. 2. Try monitoring your memory usage. Is the java process consuming lots of memory? 3. There are some notes at $PTII/doc/coding/performance.htm, which is at http://ptolemy.eecs.berkeley.edu/ptolemyII/ptII4.0/ptII4.0-beta/doc/coding/performance.htm ( I see that the file has a .html bug which I will fix shortly) 4. Using a commercial tool like JProbe can help. They have a free trial. -Christopher Posted to the ptolemy-hackers mailing list. Please send administrative mail for this list to: [EMAIL PROTECTED]
Re: Compiling Ptolemy with Eclipse
Hmm, this sounds like an Eclipse bug tickled by Ptolemy adjusting
setting up the ptII/.classpath file.
I've seen stack traces inside eclipse 2.1.2. Rather than
reinstalling, what I usually what I ended up doing was removing my
eclipse/workspace directory and trying various things.
You could try removing ptII/.classpath and seeing if the setup
completes. That worked for me.
I was able to build after hacking adding the following to .classpath.in:
classpathentry excluding=cm/ kind=src path=ptolemy/copernicus/c/test/
and I had to add
ptolemy/copernicus/c/test/cm/|
to the list of excluded files.
At one point, the Java build path window came up and I had
to remove jmf.jar and sound.jar
I later removed the jxta jar files because they had this
bogus ${PTII} inside of them. I'll fix configure momentarily.
This bug occurred because our devel tree contains the jxta jar files,
but we don't ship these jar files.
To remove the jxta jar files, right click on the ptII project, select
Properties - Java Build Path - Libraries
You may later want to exclude ptolemy/moml/jxta from the build.
Below is a copy of the .classpath.in that worked for me:
--start--
?xml version=1.0 encoding=UTF-8?
classpath
classpathentry excluding=cm/ kind=src path=ptolemy/copernicus/c/test/
classpathentry
excluding=**/CVS/|**/codeDoc/|config/|doc/coding/templates/|ptolemy/apps/agilent/|ptolemy/apps/anytimecontrol/|ptolemy/apps/charon/|ptolemy/apps/crazyboard/|ptolemy/apps/etherealSting/|ptolemy/apps/fullscreen/|ptolemy/apps/giotto/|ptolemy/apps/ia/|ptolemy/apps/kurt/|ptolemy/apps/jmf/|ptolemy/apps/kvm/|ptolemy/apps/lego/|ptolemy/apps/lejos/|ptolemy/apps/midp/|ptolemy/apps/mobileCoding/|ptolemy/apps/pendulum/|ptolemy/apps/rijndael/|ptolemy/apps/robot/|ptolemy/apps/security/|ptolemy/apps/sobel/|ptolemy/apps/softwalls/|ptolemy/apps/soot/|ptolemy/apps/throttle/|ptolemy/apps/tini/|ptolemy/apps/transmission/|ptolemy/apps/vergil/|ptolemy/apps/waba/|ptolemy/apps/webstart/|ptolemy/apps/xsl/|ptolemy/copernicus/c/test/|ptolemy/copernicus/c/test/cm/|ptolemy/domains/ct/demo/Saber/|ptolemy/domains/gr/lib/experimental/|ptolemy/domains/pdf/|ptolemy/plot/servlet/|vendors/@ECLIPSE_EXCLUDING_SRC@
kind=src path=/
classpathentry kind=con path=org.eclipse.jdt.launching.JRE_CONTAINER/
@ECLIPSE_CLASSPATHENTRY_LIBS@
classpathentry kind=output path=/
/classpath
--end--
I'll add a note to the website about this
-Christopher
Thanks for your help, I also tried the method described but I can not
get the project compiled, here is what I tried :
tar xfz ptII3.0.2.src.tar.gz
mv ptII3.0.2 ptII
export PTII=/home/franck/bin/my_PTII/ptII
cd $PTII
./configure
cd ~/bin/my_PTII
eclipse
File-New Project
Project Name : ptII
Directory: /home/franck/bin/my_PTII
Finish
-- An error occured while creating the java project
Reason : invalid project description
/home/franck/bin/my_PTII and /home/franck/bin/my_PTII/workspace overlap
If I change the directory with : /home/franck/bin/my_PTII/ptII
-- An error occurred while creating the java project
See error log for more details.
(java.lang.NullPointerException in the .log of Eclipse !)
I tried other thinks but I can only compile and run (make fast install)
without Eclipse !
Franck
On Mon, 2004-05-03 at 17:14, Christopher Hylands Brooks wrote:
With Eclipse, usually it is necessary to run configure so that the
.classpath file is properly set up. An alternative is to set up the
project class path by hand. Locally, our Eclipse users run configure,
so the by hand section of the doc could probably use some testing
and updating.
One issue is that Ptolemy II includes many packages that
use other external packages that might not be present, so setting the
classpath and excluding packages by hand can be time consuming.
-Christopher
Good morning,
I'm a software designer in the Pasteur Institute (Paris) and I'm
involved with a workflow definition and conception for sequence analy
sis
(mainly). Ptolemy II seems to correspond to the core functionalities
I
need and seems to be easy to customize (thanks to have made docs and
UML
diagrams !)
Unfortunatly I'm trying to make an eclipse project following the
instructions described but without success.
I'm now trying to install Ptolemy by hand following the chapter
Setting up for Ptolemy II By Hand described in doc/coding/eclipse.h
tm
and I have these errors :
1 - The project was not built since its classpath is incomplete. Cann
ot
find the class file for ptolemy.math.Complex
Re: Compiling Ptolemy with Eclipse
With Eclipse, usually it is necessary to run configure so that the .classpath file is properly set up. An alternative is to set up the project class path by hand. Locally, our Eclipse users run configure, so the by hand section of the doc could probably use some testing and updating. One issue is that Ptolemy II includes many packages that use other external packages that might not be present, so setting the classpath and excluding packages by hand can be time consuming. -Christopher Good morning, I'm a software designer in the Pasteur Institute (Paris) and I'm involved with a workflow definition and conception for sequence analysis (mainly). Ptolemy II seems to correspond to the core functionalities I need and seems to be easy to customize (thanks to have made docs and UML diagrams !) Unfortunatly I'm trying to make an eclipse project following the instructions described but without success. I'm now trying to install Ptolemy by hand following the chapter Setting up for Ptolemy II By Hand described in doc/coding/eclipse.htm and I have these errors : 1 - The project was not built since its classpath is incomplete. Cannot find the class file for ptolemy.math.Complex. Fix the classpath then try rebuilding this project. 2 - For the file /ptolemy/matlab/test/TestEngine.java : This compilation unit indirectly references the missing type ptolemy.math.Complex (typically some required class file is referencing a type outside the classpath) (Weird because I can see ptolemy.math in Eclipse) Can you say me if there is a mistake in the procedure described or I made something wrong. Thanks a lot ! Regards Franck Valentin --- - Posted to the ptolemy-hackers mailing list. Please send administrative mail for this list to: [EMAIL PROTECTED] Posted to the ptolemy-hackers mailing list. Please send administrative mail for this list to: [EMAIL PROTECTED]
Re: ANNOUNCE: galsC 0.1.0 (alpha) release
Elaine Cheong and Jie Liu announced availability of galsC 0.1.0 (alpha) -Christopher --- Forwarded Message Subject: [Ptolemy] ANNOUNCE: galsC 0.1.0 (alpha) release Date: Sun, 25 Apr 2004 16:18:30 -0700 (PDT) We are announcing the alpha release of galsC, a language and compiler for use with the TinyGALS programming model, which uses TinyOS as the underlying component model. galsC is compatible with TinyOS 1.x and nesC 1.1.1. You can use the galsC compiler to compile all your existing TinyOS programs. TinyGALS is a globally asynchronous, locally synchronous model for programming event-driven embedded systems, especially sensor networks. At the local level, software components communicate with each other synchronously via method calls. Components are composed to form actors. At the global level, actors communicate with each other asynchronously via message passing, which separates the flow of control between actors. A complementary model called TinyGUYS is a guarded yet synchronous model designed to allow thread-safe sharing of global state between actors without explicitly passing messages. The TinyGALS programming model is structured such that code for all inter-actor communication, actor triggering mechanisms, and access to guarded global variables can be automatically generated from a high level specification. By raising concurrency concerns above the level of TinyOS components, the TinyGALS programming model allows programmers to focus on the main tasks that the application must execute. Programs developed using this task-oriented model are thread safe and easy to debug. For more information and download instructions, please see: http://galsc.sourceforge.net/ To receive future announcements about galsC, please sign up here: http://lists.sourceforge.net/lists/listinfo/galsc-announce Please visit us at IPSN'04 (Berkeley, California, April 26-27, 2004): http://ipsn04.cs.uiuc.edu/ --- End of Forwarded Message Posted to the ptolemy-hackers mailing list. Please send administrative mail for this list to: [EMAIL PROTECTED]
Re: Reading a file from Matlab
In comp.soft-sys.ptolemy, [EMAIL PROTECTED] (Renooj Jacob) writes: I have a script file in matlab which solves a series of equations. I would like to access this matlab file from ptolemy and plot within ptolemy using the matlab expression. Is it possible? If not is there any other way. I'm not sure what the format of the script file is. You could read from the file using ExpressionReader, FileReader or LineReader and then you could use the String actors to process the data, though you might find it easier to create a custom actor to do custom plotting. You could then call the Matlab Expression actor to plot the data. Also i have a function file in matlab. i want to execute it from ptolemy. How to do that? I don't know much about Matlab, but I did create a simple test that reads in a function from a .m file. Is a function file a .m fle? Under Windows, running cd $PTII ptolemy/matlab/test $PTII/bin/vergil mFileTest.xml will bring up a model that reads mFileTest.m that defines a simple file. The model annotation says: This test uses a function defined in an m-file called mFileTest.m. The packageDirectories parameter of the MatlabExpression actor is set to include $PTII/ptolemy/matlab/test in the directories that are searched by Matlab. The _debugging parameter is set to 2 so that debugging output will appear on standard out. I also created a simple wireless model that uses a .m file at $PTII/ptolemy/domains/wireless/demo/MatlabWirelessSoundDetection/ Note that at this time you cannot use .m files with pure jar file based installers like Web Start. We could hack something up that would get the .m file from a jar file and copy it to a directory and add that directory to the packageDirectories parameter of the model. Also, it appears that gcc-5.3 has problems running the Matlab interface. Use gcc-5.2 instead Regards Renooj -Christopher Posted to the ptolemy-hackers mailing list. Please send administrative mail for this list to: [EMAIL PROTECTED]
Re: FOR LOOP
[EMAIL PROTECTED] (Renooj Jacob) writes: 1) How does one implement FOR LOOP in Ptolemy. There was a little bit of discussion about this in November. Searching google with: switch group:comp.soft-sys.ptolemy yields this post in comp.soft-sys.ptolemy Professor Lee wrote: The Switch actor will do this in DE... It will also do this in PN, without a timing model... Edward At 12:44 PM 11/19/2003 -0800, Christopher Hylands Brooks wrote: In comp.soft-sys, [EMAIL PROTECTED] (Dave from Dublin) writes: I am quite new to Ptolemy and I am working a little in the DE domain. I cannot seem to find an actor that will perform the equivalent of an IF statement, i.e. an actor that reads from a single input channel and chooses which output port to write to, based on the value read (e.g. 5 write to port1, else write to port2). Is there an actor to do this or is there another way to implement this kind of functionality? Any help greatly appreciated, This sounds a bit like Dynamic Dataflow (DDF), which is not supported in Ptolemy II. In SDF, having an actor that fires to a different port depending on its value makes it difficult to figure out the balance equations. You could try using a Comparator connected to a BooleanSelect or BooleanSwitch. Comparator is found under actor library - logic BooleanSelect and BooleanSwitch are found under actor library - boolean flow control - boolean flow control. Someone else might have a suggestion. -Christopher [EMAIL PROTECTED] (Renooj Jacob) writes: 2) I am trying to simulate Electromagnetic simulation using the Finite Difference time domain technique whose code is written in C language. I would like to do the same in Ptolemy. Is it possible? Which Director should i choose. I'm not sure. Take a look at th Continuous Time domain? -Christopher Posted to the ptolemy-hackers mailing list. Please send administrative mail for this list to: [EMAIL PROTECTED]
Re: memory problem
This bounced because of configuration problem on the mail server, so I'm resending it. -Christopher --- Forwarded Message From: Stephen Andrew Neuendorffer [EMAIL PROTECTED] To: Nilgun Fescioglu Unver [EMAIL PROTECTED], [EMAIL PROTECTED], [EMAIL PROTECTED] Subject: [Ptolemy] Re: memory problem Date: Wed, 31 Mar 2004 08:55:16 -0800 The culprit is apparently the Display actor, which internally uses a JTextArea, which is a known memory hog: A quick google search turns up: I can't believe this exists in Swing, but I've found that when I loaded a 2MB file into JTextArea it creates literally 394,000+ java.lang.ref.Finalizer objects that hang around event after the JTextArea is released. What's worse is my memory usage spikes to 65MB! That's 30 times the size of the text string! I don't know what java.lang.ref.Finalizer object is, or why it gets created. After 6000 iterations, that's roughly 1.3 MB of text At 01:26 PM 3/30/2004 -0800, Nilgun Fescioglu Unver wrote: Hi, Sorry if you received this message for the second time.. I have a java.lang.OutOfMemory error problem with my model developed in Ptolemy II version 3.0.2. I am working on a model which uses an ArrayToken (of 180 tokens) and my actors change the value of the tokens in each iteration. As tokens are immutable I am creating new tokens in each iteration. During the run time after several iterations the system stops with an error java.lang.OutOfMemory. I thought some problem with garbage collection is creating the problem so I tried calling System.gc at the end of fire() function in my actors but there was no change in the number of iterations before getting the error. I tried to generate the same problem with the actors of Ptolemy and created a simple model using the ArrayExtract actor (the xml file is attached) and ran into the same problem after several iterations (6143 iterations on a computer with 512MB Ram, Windows XP operating system, java version 1.4.2 ). Could you help me to find out the reason of this problem? And how can I continue the simulation without running out of memory? Thanks, Nilgun Fescioglu Unver Attachment deleted Posted to the ptolemy-hackers mailing list. Please send administrative mail for this list to: [EMAIL PROTECTED]
Re: Random numbers
Kostas Oikonomou [EMAIL PROTECTED] writes: I'm trying to see if I can use the DE domain of Ptolemy II to simulate a simple semi- Markov process. The main obstacle I've run into is that I can't see any way of generating random numbers from distributions other than the exponential and normal. Is there some relatively easy way of doing this? (i.e. no Java code, because I don't know enough Java). I remember from a while ago that Ptolemy Classic had such facilities. I'm not sure. I don't know that much about random numbers. In Ptolemy II 3.0, there are the following actors that generate random numbers under actor library-random Bernoulli - An actor that outputs a random sequence of booleans. DiscreteRandomSource - An actor that produces tokens with a given probability mass function. Gaussian - An actor that outputs a random sequence with a Gaussian distribution. Rician - An actor that outputs a random sequence with a Rician distribution. Uniform - An actor that outputs a random sequence with a uniform distribution. Under actor library - sources - timed sources, there is PoissonClock - A Poisson process clock source. The expression language has random() and gaussian() functions: The functions random() and gaussian() shown in Table 5 on page 96 return one or more random numbers. With the minimum number of arguments (zero or two, respectively), they return a single number. With one additional argument, they return an array of the specified length. With a second additional argument, they return a matrix with the specified number of rows and columns. There is a key subtlety when using these functions in Ptolemy II. In particular, they are evaluated only when the expression within which they appear is evaluated. The result of the expression may be used repeatedly without re-evaluating the expression. Thus, for example, if the ivalue/i parameter of the iConst/i actor is set to random()/font, then its output will be a random constant, i.e., it will not change on each firing. The output will change, however, on successive runs of the model. In contrast, if this is used in an Expression actor, then each firing triggers an evaluation of the expression, and consequently will result in a new random number. random() and gaussian() are declared in ptolemy/data/expr/UtilityFunctions.java. You could add other functions to that file. You could also use the interface to Matlab, see http://ptolemy.eecs.berkeley.edu/ptolemyII/ptII3.0/ptII3.0.2/ptolemy/matlab/matlab.htm Someone else on the list might have other suggestions. -Christopher Posted to the ptolemy-hackers mailing list. Please send administrative mail for this list to: [EMAIL PROTECTED]
Re: Using Ptolemy II embedded in my product
Michael Doorley writes:
Hi Christopher
Thanks for adding me to the list.
1. Can I use Ptolemy II by calling methods in the API that would
create a model, populate it, run a simulation and extract the results?
As a bonus for the advanced user I would like to save the model as an
XML file that could be opened in Ptolemy II GUI tools (e.g. Vergil),
but that is not a priority.
Yes. See ptolemy/actor/gui/MoMLSimpleApplication.java and
ptolemy/actor/gui/PtExecuteApplication.java, which both instantiate and
run
a model.
I have looked at MoMLSimpleApplication.java and my problem with it is that
it seems to require a file as input. I am looking for a solution where I
can create and populate model in memory directly from my program. The
closest thing I have found is the Dining Philosophers demo of the CSP
domain, which does not seem to have a file dependency, unless I am missing
it.
I am currently trying build an equivalent of that program for DE. Do you
know if any similar sample already exists?
Thanks for your help,
Michael
Sorry, I thought you wanted to _read_ a MoML file.
You could do something similar to how the
ptII/ptolemy/domains/ct/demo/Helicopter/Helicopter.java
works.
You could create a similar model with domain polymorphic or DE
actors and using the DE Director
Note that if you call
System.out.println(exportMoML());
from within a TypedCompositeActor, then you will get the MoML output.
See also:
ct/demo/Lorenz/Lorenz.java
ct/demo/SigmaDelta/SigmaDelta.java
ct/demo/SquareWave/SquareWave.java
ct/demo/StickyMasses/StickyMasses.java
ct/demo/Thermostat/Thermostat.java
sdf/demo/Butterfly/Butterfly.java
For example, to run the last demo, you would do
java -classpath ptolemy.actor.gui.CompositeActorApplication -class
ptolemy.domains.sdf.demo.Butterfly.Butterfly
The shallow code generator will read in certain models and generate
java files.
(See
http://ptolemy.eecs.berkeley.edu/ptolemyII/ptIIlatest/ptII/doc/codegen.htm)
Below I ran the shallow code generator on
ptII/ptolemy/domains/de/lib/test/auto/Queue.xml
and added throws lines so that it would compile
The file is in the ptolemy/copernicus/shallow/cg/Queue
directory, so it is in the ptolemy.copernicus.shallow.cg.Queue package.
[EMAIL PROTECTED] 66% pwd
/vol/maury/maury2/cxh/ptII/ptolemy/copernicus/shallow/cg/Queue
[EMAIL PROTECTED] 67% cat CGQueue.java
/* CGQueue - Decompiled by JODE
* Visit http://jode.sourceforge.net/
*/
package ptolemy.copernicus.shallow.cg.Queue;
import ptolemy.actor.IOPort;
import ptolemy.actor.TypedCompositeActor;
import ptolemy.actor.TypedIOPort;
import ptolemy.actor.TypedIORelation;
import ptolemy.actor.gui.LocationAttribute;
import ptolemy.actor.gui.SizeAttribute;
import ptolemy.actor.lib.Clock;
import ptolemy.actor.lib.Test;
import ptolemy.data.expr.Parameter;
import ptolemy.domains.de.kernel.DEDirector;
import ptolemy.domains.de.kernel.DEIOPort;
import ptolemy.domains.de.lib.Queue;
import ptolemy.kernel.ComponentPort;
import ptolemy.kernel.CompositeEntity;
import ptolemy.kernel.attributes.VersionAttribute;
import ptolemy.kernel.util.Settable;
import ptolemy.kernel.util.Workspace;
import ptolemy.kernel.util.IllegalActionException;
import ptolemy.kernel.util.NameDuplicationException;
import ptolemy.moml.Documentation;
import ptolemy.moml.Vertex;
public class CGQueue extends TypedCompositeActor
{
public VersionAttribute A_createdBy;
public SizeAttribute A_vergilSize;
public LocationAttribute A_vergilLocation;
public DEDirector ADE;
public Parameter ADE_startTime;
public Parameter ADE_stopTime;
public Parameter ADE_stopWhenQueueIsEmpty;
public Parameter ADE_synchronizeToRealTime;
public Parameter ADE_isCQAdaptive;
public Parameter ADE_minBinCount;
public Parameter ADE_binCountFactor;
public ComponentPort PQueue_input;
public ComponentPort PQueue_output;
public ComponentPort PQueue_trigger;
public ComponentPort PClock_output;
public ComponentPort PClock_trigger;
public Parameter AClock_stopTime;
public Parameter AClock_period;
public Parameter AClock_offsets;
public Parameter AClock_values;
public Parameter AClock_numberOfCycles;
public Documentation AClock__doc;
public ComponentPort PClock2_output;
public ComponentPort PClock2_trigger;
public Parameter AClock2_stopTime;
public Parameter AClock2_period;
public Parameter AClock2_offsets;
public Parameter AClock2_values;
public Parameter AClock2_numberOfCycles;
public Documentation AClock2__doc;
public ComponentPort PTest_input;
public ComponentPort PTest_output;
public Parameter ATest_correctValues;
public Parameter ATest_tolerance;
public Parameter ATest_trainingMode;
public Vertex Arelation5_vertex1;
public CGQueue(Workspace workspace) throws IllegalActionException,
NameDuplicationException {
super(workspace);
__CGInit();
}
Re: Firing by events
The single window tabbed interface by Jerome Blanc and Benoit Masson of Thales might be useful here. This interface is in the ptII developer tree in $PTII/thales. The single window interface has not yet been publically released. To start it up from, run vergil -single. The ptII/thales/README says: For the single window mode, the principle is to catch the frames created on each Tableau and move its container into the main TabbedPane. In this mode, you can either create classic editors, or navigable graph editor where beside the actors library there is a browsable model tree (warning : this doesn't work with any state machine actor, because there is only a navigable ActorGraphFrame, and there should be other navigable Frame for modal models, FSM and interface automation actors. This could be solved by merging the NavigableActorGraphFrame features into the BasicGraph Frame, but as always we try to develop without modifying the original Ptolemy sources.) There is a specific effigy for navigable models (single window mode). To create this specific effigy, the MoML file should contain a property at the top level named _navigable (property name=3D_navigable /). It is created automatically for each new navigable graph editor (new menu), and could be added on existing models. So, perhaps this code could be used as the basis of an applet that would allow traversal of the model? The other thing I was working on was extending the applet code generator so that it would generate a Web Start setup that would consist of only the classes necessary to run the model. This work is still in progress. -Christopher Professor Lee wrote: At 05:00 PM 2/18/2004 -0800, Efrat Jaeger wrote: A different question is about applets. When I create an applet out of a model that contains a composite entity, is it possible to look inside the composite entity from within the applet? (for presenting and running the workflow on a web page). This has been on the to do list for some time... But it's not entirely clear what it should do... replace the current image in the web page? Open a window? Go to a new page? I'm inclined to think it should go to a new page, so the back button takes you back... Edward Posted to the ptolemy-hackers mailing list. Please send administrative mail for this list to: [EMAIL PROTECTED]
Re: Ptolemy II : reuse legacy application, distributed actors
BERGERO Didier [EMAIL PROTECTED] writes: I have 2 (newbie?) questions : * While waiting for e-ptolemy, with Ptolemy II and in the same simulation, how to use local actors and remote (distributed) actors ? : use of applets... I'm not sure what e-ptolemy is in this context. There are some actors that use jxta in $PTII/ptolemy/actor/lib/jxta We have some corba actors, though they have not been released. We are starting to look at Grid computing, but we do not have any results to report yet. Use of applets is covered in http://ptolemy.eecs.berkeley.edu/papers/03/ptIIDesignSoftware * and how to reuse legacy application e.g. how to wrap C,FORTRAN,ADA!! code to be an Ptolemy II actor ? Ptolemy II 3.0.2 includes an interface that uses JNI, see http://ptolemy.eecs.berkeley.edu/ptolemyII/ptII3.0/ptII3.0.2/jni/doc/JNIActorHelp.pdf See also the matlab actor in $PTII/ptolemy/matlab -Christopher Posted to the ptolemy-hackers mailing list. Please send administrative mail for this list to: [EMAIL PROTECTED]
Re: Is there a way to speed up the redraw?
In comp.soft-sys.ptolemy, luedeke [EMAIL PROTECTED] writes: I have the feeling that the redraw of a ptolemy.plot.Plot object is a bit slow, because doubled buffering or something else is not used, so the plot has to be recalulated each time a redraw is needed. Is there a way to speed up the redraw? Alexander Are you using the most recent version of Ptplot, Ptplot 5.3? Ptplot 5.3 is the version that shipped with Ptolemy II 3.0 There were some changes made for double buffering betweent Ptplot 5.2 and 5.3. Another change is that the lines are now drawn thicker. -Christopher Posted to the ptolemy-hackers mailing list. Please send administrative mail for this list to: [EMAIL PROTECTED]
Re: Point width on XYPlotter
In the XYPlotter, click on the plot format button in the upper right (this is the second from the right button that has two red impulse lines) Then select Marks: dots Save your model When you run it again, it will have a dot for the single point. The way this works is that the MoML file gets some plotml inside a configure element that sets the marks to dots -Christopher Christopher Hylands Brooks [EMAIL PROTECTED] University of California Programmer/Analyst Chess/Ptolemy/GSRC US Mail: 558 Cory Hall #1770 ph: 510.643.9841 fax:510.642.2739 Berkeley, CA 94720-1770 home: (F-Tu) 707.665.0131 (W-F) 510.655.5480 (office: 400A Cory) Hi, I'm trying to plot a single point on XYPlotter. How can I control the width of the point, so that I'll be able to view it? It is also necessary for plotting a sequence of points, if the connected variable is set to false. Thanks, Efrat --- - Posted to the ptolemy-hackers mailing list. Please send administrative mail for this list to: [EMAIL PROTECTED] Posted to the ptolemy-hackers mailing list. Please send administrative mail for this list to: [EMAIL PROTECTED]
Re: diva source code of newest version
Tristan Crecelius [EMAIL PROTECTED] writes: where can I find the newest source code of the diva lib. Last version used from Ptolemy is dated 21.04.2003 (diva.jar 1.035.788 Bytes) I've created a snapshot of the Diva tree from April 22, 2003 and made it available on the Ptolemy II 3.0.2 source download page http://ptolemy.eecs.berkeley.edu/ptolemyII/ptII3.0/index.htm Both .tar.gz and zip files are available directly: http://ptolemy.eecs.berkeley.edu/ptolemyII/ptII3.0/diva-22Apr03.tar.gz http://ptolemy.eecs.berkeley.edu/ptolemyII/ptII3.0/diva-22Apr03.zip It is not necessary to download the Diva source files to build Ptolemy II, Ptolemy II includes diva.jar. Download the diva .tar.gz or .zip if you are interested in extending the Ptolemy II GUI itself. 1. Download either diva-22Apr03.tar.gz (1.8 Mb), or diva-22Apr03.zip (2.5Mb) 2. Change to the Ptolemy II directory and uncompress them. For example: cd $PTII tar -zxf /tmp/diva-22Apr03.tar.gz 3. Follow the instructions in $PTII/diva/README.txt Note that this diva snapshot will build a slightly smaller jar file than what was shipped: 974201 bytes instead of 1035788 bytes. One difference is that a few test directories are not included. There are minor differences in a few .class files as well. I'm told that these primarily are caused by local variables being renamed in the version of diva.jar that was shipped with Ptolemy II 3.0.2 Let me know if you have problems. The diva home page is on http://www.gigascale.org/diva Note that the diva home page will move to http://embedded.eecs.berkeley.edu/diva at somepoint in the not to distant future. -Christopher Christopher Hylands Brooks [EMAIL PROTECTED] University of California Programmer/Analyst Chess/Ptolemy/GSRC US Mail: 558 Cory Hall #1770 ph: 510.643.9841 fax:510.642.2739 Berkeley, CA 94720-1770 home: (F-Tu) 707.665.0131 (W-F) 510.655.5480 (office: 400A Cory) Posted to the ptolemy-hackers mailing list. Please send administrative mail for this list to: [EMAIL PROTECTED]
Re: problems rebuilding Ptolemy classic from scratch
I have some notes for compiling Ptolemy II 0.7.2 devel under Linux at http://ptolemy.eecs.berkeley.edu/ptolemyclassic/pt0.7.2/linux.htm I'll see about adding a more prominent link to that page from the 0.7.1 page. xv not compiling is fairly common, you can run without it. xv is also available elsewhere. If pigiRpc is not present, then you probably have a compilation problem. Try cd $PTOLEMY make make.out and then look in make.out for lines that contain three *** -Christopher Hi again. Thanks all for your answers to my last question. i'm trying ti re build from scratch Ptolemy classic 0.7.1. I am with linux debian 3.0 (woody).I us e bash, not cshell. I got the sources to 0.7.1,, the tar-gz archives with the global src and others. Then I untar the archives, in $PTOLEMY, and foollow the instructi ons in the web page of Ptolemy project, doing make and make install. Then an er ror happened, something related with Tycho. Perhaps I hasn't got Tcl/tk or some thing about that. Then, I do: make everything Now, I got an error related with xv. I suppose that I didn't finish my installation or something doesn't work right, because when I try to start p igi from $PTOLEMY/bin/pigi, it says that pigi.Rpc doesn't exists or is not prep ared to be executable. I think I'm doing something wrong. Perphaps it would be easier if I get som e .deb files from ptolemy, but I don't know how to got them. Can I receive some help? Thanks again and regards. - This mail sent through IMP: http://horde.org/imp/ --- - Posted to the ptolemy-hackers mailing list. Please send administrative mail for this list to: [EMAIL PROTECTED] Posted to the ptolemy-hackers mailing list. Please send administrative mail for this list to: [EMAIL PROTECTED]
Re:
Hmm, I thought I sent a reply to this earlier this week, see below From: Christopher Hylands Brooks [EMAIL PROTECTED] To: Darine Abdelahad [EMAIL PROTECTED] Subject: Re: Ptplot mouse code Cc: [EMAIL PROTECTED] Date: Wed, 03 Sep 2003 09:04:41 -0700 Take a look at the zoom code in PlotBox.java It converts from window coordinates to plot coordinates. See also http://ptolemy.eecs.berkeley.edu/java/ptplot5.2/ptolemy/plot/doc/ptplotfaq.htm#ptplot potential features -Christopher Hi, I have a question: I want the user to be able to click on a point on the graph in order to obtain more information about it. To do so, I need the mouse click to retur n the coordinates of the point. So I tried to add to the Plot object an instance of a class that implements the MouseListener interface. And then in the mouseClicked functi on, the getX()and getY() methods of the MouseEvent are supposed to return the coordinates of the point where the user clicked. But these coordinates are given with respect to the window where the plot is found, whereas I need the coordinates with respect to the Xaxis and Yaxis of the plot (that is the coordinates assigned to the point by my application). So I was wondering how I can obtain these coordinates upon a mouse click. Can anyone help me with that? Thanx Posted to the ptolemy-hackers mailing list. Please send administrative mail for this list to: [EMAIL PROTECTED]
Re: first day with ptolemyII
In comp.soft-sys.ptolemy, cfk [EMAIL PROTECTED] writes:
Well, let me answer my own question and pose the next one.
It appears the webstart is an interactive demo only.
The Web Start version if fully functional with the exception
of the Jython actor, which uses some strange caching mechanism.
See below for details.
So, I downloaded the java run-time environment and installed it. I now
have in /usr/bin executables such as java, jar, javadoc and other
somewhat familiar things (although its been a few years since I
installed java). I also have a /usr/src/ptII3.0.2 directory and
was able to './configure', 'make' and 'make install'. The first
two completed with no errors, but 'make install' has troubles at
the end with '/usr/bin/jar: invalid option -- i' and no rule to
make target 'coding/condingConfig.jar' needed by 'docConfig.jar'.
I've seen this problem with Windows Cygwin jar before, where it is not
entirely compatible with the Sun jar.
Try putting the JDK bin in your path before /usr/bin.
I'll see about making a test for this in configure.
If you really want to use /usr/bin/jar, try editing
$PTII/mk/ptcommon.mk and change
(cd $(PTJAR_TMPDIR); $(JAR) -cvf tmp.jar .; $(JAR) -i tmp.jar)
to
(cd $(PTJAR_TMPDIR); $(JAR) -cvf tmp.jar .)
in three places.
The -i option is used to generate an index for the jar file which
make applications and applets start faster.
I added /usr/src/ptII3.0.2/bin to my path in .bashrc. I also added
an export of PTII=/usr/src/ptII3.0.2. At this point, if I invoke
vergil, I can see the error couldnt find 'java/lang/Object'. I
have been here before and I think it is the lack of a setting for
JAVAHOME and CLASSPATH. Unfortunately, I dont know where
java/lang/Object is, so setting the CLASSPATH and JAVAHOME are
mysterious right now.
Seems like your Java installation is broken and not finding
any the runtime jar file that contains java.lang.Object.
I don't have a good solution for this, I don't have a Linux machine in
front of me.
Under Solaris, Object.class can be found in:
[EMAIL PROTECTED] 9% jar -tvf /opt/local/j2sdk1.4.1_02/jre/lib/rt.jar | grep
java/lang/Object.class
1428 Thu Feb 20 14:58:48 PST 2003 java/lang/Object.class
[EMAIL PROTECTED] 10%
However, adding rt.jar to the classpath should not be necessary.
The way that I would proceed is to poke around on the
Sun tutorial and make sure that your JVM works
http://java.sun.com/docs/books/tutorial/getStarted/cupojava/index.htmld
Also, you should be able to start vergil by hand by including
lib/diva.jar
Something like the following should work:
java -classpath ${PTII}:${PTII}/lib/diva.jar ptolemy.vergil.VergilApplication
I am hoping that someone who uses java a bit more then I have over
the last couple of years would consider helping me navigate this
minefield a bit.
Yep, the classpath can be a bear.
Charles
At this point, I can navigate to /usr/src/ptII3.0.2/bin
cfk [EMAIL PROTECTED] wrote in message
news:[EMAIL PROTECTED]
Gentlemen:
I downloaded the webstart and installed the ptolemyII 3.0.2
application ion a RedHat8 system and am currently enjoying
reading through the examples. But, I am puzzled by a *real*
newbie question. How do I get back to running the application
when I exit it. At this point, I can see that it
is running from a new .javaws directory and I can see http and
lastAccessed in that directory. But what do I invoke to startup
ptolemy when I exit and reboot the comptuer.
The Java Web Start Application Manager is a standalone application
that lists the Web Start applications that are available.
http://java.sun.com/j2se/1.4.2/docs/guide/jws/Readme.html#install
Under Windows, there is a standalone Java Web Start Application under
Start - All Programs - Java Web Start
Under Solaris (for me_, there is standalone Java Web Start Application zip
file installer at
/opt/local/j2sdk1.4.1_02/jre/javaws-1_2_0_02-solaris-sparc-i.zip
For now, I'll leave it on for a few days until I figure it out.
Guess everyone has to be a newbie at something on a regular
basis in order to get anywhere.
-Christopher
Christopher Hylands Brooks [EMAIL PROTECTED] University of California
Programmer/Analyst Chess/Ptolemy/GSRC US Mail: 558 Cory Hall #1770
ph: 510.643.9841 fax:510.642.2739 Berkeley, CA 94720-1770
home: (F-Tu) 707.665.0131 (W-F) 510.655.5480 (office: 400A Cory)
Posted to the ptolemy-hackers mailing list. Please send administrative
mail for this list to: [EMAIL PROTECTED]
Re: help me on image filtering
On 7/23, in comp.soft-sys.ptolemy, Paul.Shumacher wrote:
Hi,
I am new to Ptolemy so please forgive me if this question has been
raised already.
I am interested in doing various operations on 2-D images/frames. For
example, using a
1-D FIR filter to smooth or find edges. I have tried the ImageSequence
actor followed
by MatrixToSequence or ImagePartition, but I get the following error:
Conversion is not supported from ptolemy.data.IntToken '3' to the type
[int].
I am using Ptolemy II 3.0. Any suggestions?
Well, you could use the new JAIEdgeDetection actor to do this,
there is a test at
http://ptolemy.eecs.berkeley.edu/ptolemyII/ptII3.0/ptII3.0-beta/ptolemy/actor/lib/jai/demo/test/JAIEdgeDetection.xml
If you install Ptolemy II 3.0 under Windows with the built in
JRE, the Java Advanced Imaging API will be included in the JRE.
If you are building from source, you would need to install
the Java Advanced Imaging API and rerun configure
cd $PTII
rm config.*; ./configure
cd ptolemy/actor/lib/jai
make
If you start up Ptolemy II 3.0 and do File - Open URL
and cut and paste the above URL in, the model will open for you.
You will need to change the JAIImageReader fileOrURL path
to an image on your local disk.
When you run the image, edge detection will occur.
The actors you refer to are part of the Vector Quantization
actors in sdf/lib/vq. These actors are somewhat special purpose, see
the sdf/demo/HTVQ/HTVQ.xml demo for an example use.
You should be able to create a model that has
ImageSequence - ImagePartition - ImageDisplay
1. ImageSequence reads in a series of images and outputs
IntMatrices.
If you use it, you may need to change the imageURLTemplate
from
/ptolemy/domains/sdf/lib/vq/data/seq/missa/missa***.qcf
to
ptolemy/domains/sdf/lib/vq/data/seq/missa/missa***.qcf
(I fixed this in the release and development trees)
2. ImagePartition reads in IntegerMatrices, not integers, so this could
be part of the problem in your example
Change the partitionColumns to 88 and
the partitionRows to 72
3. Add a SDFDirector and set the iterations to 100
When you run it, the ImageDisplay should pop up.
You may need to mouse on the ImageDisplay actor to get it to display
each frame.
I think the problem you were having with
ImageSequence - MatrixToSequence - ImagePartition
is that ImagePartition expects an IntMatrix, not a sequence of
Integers.
You could do
ImageSequence -
MatrixToSequence - Quantizer - Round
-SequenceToMatrix - ImageDisplay
1. ImageSequence needs the imageURLTemplate adjusted as per above.
2. In MatrixToSequence and SequenceToMatrix, set the rows to 144
and the columns to 176
4. Change the Quantizer to {-1.0, 40.0} to Quantize the values
-Christopher
Christopher Hylands Brooks [EMAIL PROTECTED] University of California
Programmer/Analyst Chess/Ptolemy/GSRC US Mail: 558 Cory Hall #1770
ph: 510.643.9841 fax:510.642.2739 Berkeley, CA 94720-1770
home: (F-Tu) 707.665.0131 (W-F) 510.655.5480 (office: 400A Cory)
Posted to the ptolemy-hackers mailing list. Please send administrative
mail for this list to: [EMAIL PROTECTED]
