BioImage Wiki (internal)Contents:
See also: BioimageInstallation, BioimageOntology
Bioimage Documentation
BioImage Documentation There are currently two versions of the BioImage site running. The main site (www.bioimage.org) runs on milos (www.milos.zoo.ox.ac.uk) and the development site runs on andros (devel.bioimage.org) For an overview of the site architecture, see ![[WWW]](/wiki/modern/img/moin-www.png){kind=small}
http://www.bioimage.org/pub/simon/OntologiesInBioImage.html
Both sites use tomcat as the servlet container, and apache as the webserver. Compilation and upload of code to both sites is managed with an ant build script
Up to date documentation for the BioImage site can be created by running the documentation target in the ant build script on the current cvs version of the BioImage code. The code is available at http://paros.zoo.ox.ac.uk/viewcvs/viewcvs.cgi/ and the ant build file is in the root directory
The cvs contains all code required for running the site, including dependant jar files. The site requires a postgres database to be running - properties are set in the bioimage.properties file accessed through the devel.bioimage.org/admin - see one of the team for the password.
This is the documentation for the CurrentVersion as at Nov 2003
The packages currently divide up as follows
org.bioimage has basic functionality - admin, stylesheet checking, and error redirects
org.bioimage.owlbase has the owlbase code
org.bioimage.struts handles setup - loading properties and logon checking
org.bioimage.struts.form - holds the stylesheet form only
org.bioimage.struts.retrieval.* contain the retrieval and dipslay packages in standard struts layout (forms, actions and logic)
org.bioimage.struts.submission.* holds the submission code
org.bioimage.utils holds utilities that are not necessarily part of the bioimage core site
The core of the system is the bioimage ontology - which can be found at bioimage/static/ontology/bioimage.daml. Some of the modelling decisions involved in creating the BioImage Ontology are discussed in
There is some discussion of media storage in
StorageofMedia.doc: Strategy for media storage (draft)
Videofiles.htm: Video file size issues and reiser
BioImageDB.vsd: v0.1 visio
BioImageDB.zip: BioImage model v0.3 (zipped html)
bioimage_db.zip: v0.9
These models draw on ideas from the BBC SMEF model
SMEFDataModelv1.7.pdf: BBC business data model
But are more directly derived from the AAF data model http://aaf.sourceforge.net/docs/aafObjectModel.pdf
see also http://www.aafassociation.org
and also the ABC ontology at http://milos2.zoo.ox.ac.uk/twiki/pub/Bioimage/PapersByOthersWeLike/lagoze-final.pdf
== (Compiling) ===
Problems compiling jena.jar may be due to a bug in ant (out of memory errors) Ant needs to run in a separate JVM to avoid this problem.
1. Profiling with eclipse
Some notes on setting up profiling. Use the eclipse profiler from http://sourceforge.net/projects/eclipsecolorer/
There are instructions for setting this up at http://eclipsecolorer.sourceforge.net/index_profiler.html
However, these instructions do not address the problem of setting up the profiler when eclipse is running under windows xp and tomcat is running on a linux server. To do this you must:
1) add the following lines to catalina.sh on the linux box before JAVA_OPTS add MAINCLASS="org.apache.catalina.startup.Bootstrap"
and then
if [ "$1" = "profile" ]; then shift
if [ "$1" = "heap" ]; then JAVA_OPTS="$JAVA_OPTS -![[XrunProfilerDLL]](/wiki/modern/img/moin-inter.png){kind=small}
3,10,0 -DPROFILER_PROFILE_HEAP=1 -Xbootclasspath/a:/var/tomcat5/bin/jakarta-regexp.jar:/var/tomcat5/bin/profiler_trace.jar:/var/tomcat5/bin/commons-lang.jar -DPROFILER_PACKAGE_FILTER=A%MAINCLASS%:Msun.:Mcom.sun.:Mjava.:Mjavax.:Porg.bioimage. -DPROFILER_TIMING_METHOD=1 " shift fi
if [ "$1" = "cpu" ]; then JAVA_OPTS="$JAVA_OPTS -XrunProfilerDLL:1 -Xbootclasspath/a:/var/tomcat5/bin/jakarta-regexp.jar:/var/tomcat5/bin/profiler_trace.jar:/var/tomcat5/bin/commons-lang.jar -DPROFILER_PACKAGE_FILTER=A%MAINCLASS%:Msun.:Mcom.sun.:Mjava.:Mjavax.:Porg.bioimage. -DPROFILER_TIMING_METHOD=1 " shift fi fi
start tomcat with a line like ./bin/catalina.sh profile cpu start
2) You need to copy libProfilerDLL.so into the jre/lib/i386 subdirectory on the linux box. You also need to make sure that jakarta-regexp.jar, profiler_trace.jar and commons-lang.jar are in tomcat5/bin
3) After installing the profiler plugin In the Configurations section, highlight "Remote Profiler" and click the New button. Set the "Name:" text field to something appropriate. Hit the Browse... button and select the project that you are working on. Finally, under "Host:" put the address of the machine which will be running tomcat - eg devel.bioimage.org
4)also installed ProfilerDLL?.dll from root plugin folder into bin folder of your JRE installation on the machine running eclipse - though I'm not sure this was necessary.
-- ChrisCatton 2005-05-11 16:23:39