Activity for Peter Lin
-
Peter Lin committed [r3194] on SVN
improve length$ function
-
forgot to add rad-neq to the function group
-
fixed bug with rad-equals
-
adding method to make it easier to execute functions from java programs
-
updated jar with radix func
-
adding radix functions
-
finally found the bug with the not equals conditional element when there are multiple, silly logic error. didn't catch this since manners benchmark only has 1 rule that uses notEquals and it only uses 1 slot.
-
I'm stupid, the engine was correct, the rule I wrote to test joins with 2 not equal slots was matching too eagerly since the facts were in both sides duh!
-
some where it broke for manners benchmark, need to debug it over the weekend
-
brainfar, reverting the change, it was fine. the rule was wrong
-
fixed notequal join when there are multiple slots that are not equal
-
Here is the error from the IDE log when I click "new project" SEVERE [global] java.lang.ClassNotFoundException: javax.activation.DataContentHandlerFactory at java.base/jdk.internal.loader.BuiltinClassLoader.loadClass(BuiltinClassLoader.java:606) at java.base/jdk.internal.loader.ClassLoaders$AppClassLoader.loadClass(ClassLoaders.java:168) at java.base/java.lang.ClassLoader.loadClass(ClassLoader.java:522) at org.netbeans.ProxyClassLoader.loadClass(ProxyClassLoader.java:222) Caused: java.lang.ClassNotFoundException:...
-
Given M1 Macbook air defaults to using zsh now, I was able to download the zip version and get it to run. But after it starts, creating a new project fails. here's the steps to get it to run open terminal run "/bin/bash" run neuroph studio with ./neurophstudio --jdkhome [path to jdk home] in my case, I am running Azul zulu openjdk release for M1. The path zulu uses is /Library/Java/JavaVirtualMachines/zulu-15.jdk/Contents/Home/
-
Exception: java.lang.NoClassDefFoundError. java/util/jar/Pack200 I'm using Azul's openJDK build for M1 mac. openjdk version "15.0.1" 2020-10-20 OpenJDK Runtime Environment Zulu15.28+1013-CA (build 15.0.1+9)
-
I downloaded the zip version of release 2.98 and followed the docs for creating a basic test project. http://neuroph.sourceforge.net/image_recognition.html When I try to run the training wizard, it gives me a error no matter what value I use for Training set percent SEVERE [com.jme3.gde.core.scene.SceneApplication]: Uncaught exception thrown in Thread[Thread-13,6,main] java.lang.ClassCastException: org.neuroph.training.Variable cannot be cast to java.util.Stack at org.neuroph.training.SetWorkflowParametarsTask.execute(SetWorkflowParametarsTask.java:29)...
-
porting some fixes by dave from github back to sourceforge. I prefer subversion over git and often I curse Linus Torvald for creating GIT
-
updated the string functions to handle bindings
-
updated the jar and constants for a new release
-
changed JSONData to use servletContext to load the data file if the user puts them in WEB-INF
-
checking in unit test for the new JSON data
-
fixed bug with reading json data
-
adding support for JSON data to make it easier to startup a rule service
-
fixed silly bug with str-no-empty return to boolean
-
fixed the classloader in the case where the jars are in the WAR and not remote
-
almost done refactoring the ruleservice listener so it works with the latest version of tomcat.
-
due to how JSON works, we can't handle object inheritance. instead we have separate list for object and clips data.
-
forgot to change the servlet listener to read from .json file
-
updating the rule service to use JSON instead of XML. the version of xstream gets a warning in tomcat. time to change it
-
adding windows powershell script to make it easier to run
-
fixed a bug with modify function when the value for a slot is a binding that is declared in the action of the rule
-
updating the jar
-
fixed bug with macro generator for fields that are arrays
-
fixed bug with calling Java method with a binding
-
now modify can handle object fields that are objects
-
doing some clean up and added string-not-empty function
-
adding eager set contains function that mimics fuzzy matching. this way if the user's answer doesn't match the entries in the set exactly, the function has a better chance of matching.
-
-
adding batch for static objects
-
adding a new function to load java objects
-
adding the unix/mac shell version of start script
-
adding reference to Node object, this way we can associate a node to external objects
-
adding more examples to test graph query
-
forgot to update clear function to re declare the graph objects
-
fixed bug with working memory so it returns the right right memory for query joins. it works now
-
add watch so I can figure out why graph queries aren't working like I expected, there's a bug some where
-
getting rid of clone for graph queries
-
doing some clean up and refactoring of the defquery stuff. boy why did I make it so messy and why didn't I clean this up sooner?
-
fixed bug with query leftinput adapter node not getting set properly
-
it wasn't passing the right query parameters
-
graph query isn't handling queryParameterNode correctly yet
-
graph query compiler isn't correct yet, it's not handling the literal constraints correctly
-
changed the logic for some of the query nodes, since they need to propagate data now
-
graph query is almost working
-
now load-graph returns array of facts
-
graph query loads and load-graph almost ready
-
compile errors fixed, now it compiles the graph query, still need to test it
-
graph query isn't working yet, but making some progress
-
graph query compiler uses a different result node, it gets the last fact
-
adding support for compiling graph queries
-
improved print function to handle list and nested array
-
fixed a bug with defquery compiler
-
before I move forward, doing some experiments with regular defquery
-
adding GraphQuery, which extends defquery object, since it's basically the same functionality
-
starting to work on the graphQueryCompiler
-
started working on adding graph features to jamocha. something generic that supports basic nodes and edges
-
improved token-match so it can handle either object or slot
-
changed the order of the parameters to keep it consistent with set-contains and map-contains
-
some clean up of the new text functions
-
adding text functions
-
adding set-contains function
-
fixed bug with BoundParam so it resolves objects correctly
-
adding function to test if a map contains a string key
-
minor edits to the explanation. still need to i...
-
I didn't check these final changes in and just ...
-
-
-
-
funny, finally started to write the lessons lea...
1
