Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
    Alexey Domozhirov
    @aldomozhirov

    Could you please help me with one more issue? I'm trying to run the probe using JVM run configuration (as described in readme.txt) with server running as separate app. So, the probe is managing to connect to server, but when trying to execute any "ask", I'm getting this in probe logs:

    Apr-23|14:07:23|ERR| Uplink connection failed.
    Apr-23|14:07:23|ERR|   +-+ Error reading data. -> (NCSocket.scala:111)
    Apr-23|14:07:23|ERR|     | o.a.n.common.socket.NCSocket.read -> (NCSocket.scala:111)
    Apr-23|14:07:23|ERR|     | o.a.n.common.util.NCUtils$$anon$3.run -> (NCUtils.scala:985)
    Apr-23|14:07:23|ERR|     +-+ Error deserialization data -> (NCUtils.scala:868)
    Apr-23|14:07:23|ERR|       | o.a.n.common.util.NCUtils$.deserialize -> (NCUtils.scala:868)
    Apr-23|14:07:23|ERR|       | o.a.n.common.socket.NCSocket.read -> (NCSocket.scala:106)
    Apr-23|14:07:23|ERR|       | o.a.n.common.util.NCUtils$$anon$3.run -> (NCUtils.scala:985)
    Apr-23|14:07:23|ERR|       +-+ [java.io.InvalidClassException] org.apache.nlpcraft.common.nlp.NCNlpSentence; local class incompatible: stream classdesc serialVersionUID = -2459378081725392973, local class serialVersionUID = -4180989482790036043 -> (ObjectStreamClass.java:689)
    Apr-23|14:07:23|ERR|         | o.a.n.common.util.NCUtils$.deserialize -> (NCUtils.scala:863)
    Apr-23|14:07:23|ERR|         | o.a.n.common.socket.NCSocket.read -> (NCSocket.scala:106)
    Apr-23|14:07:23|ERR|         | o.a.n.common.util.NCUtils$$anon$3.run -> (NCUtils.scala:985)

    Am I running the probe correctly or missing something?

    Aaron Radzinski
    @aradzinski
    is this ‘master’, MacOS 12.x and JDK 11?
    skhdl
    @skhdl
    Seems like you have different versions probe and server.
    For example, server started from IDE (master branch), and probe 0.7.4 from maven or something like this.
    Can it be?
    Alexey Domozhirov
    @aldomozhirov
    I have now MacOS 12.x and JDK 11. This is not master, but distribution of version 0.7.4. I've started server from IDE in one window and probe in another (the same IDE). Project was generated using gen-project from the same distribution.
    So, both server and probe started from IDE with the same Oracle JDK 11
    Alexey Domozhirov
    @aldomozhirov
    Hmm... Seems like I found the solution. I just started server from my project itself as Run Configuration (as written in readme.txt), but not from distribution (in the same IDE window). Now it's working.
    Aaron Radzinski
    @aradzinski
    Alexey - I would recommend to switch to master (upcoming 0.7.5). This release will bring some chnages that are not backward compatible (esprecially for intent DSL) + plenty of bug fixes
    Aaron Radzinski
    @aradzinski
    @rahul3 can you please review the fix NLPCRAFT-327 and confirm it?
    For some reason I can't assign the ticket to you back...
    Rahul Padmanabhan
    @rahul3
    @aradzinski I was working today and didn't have the time to do it sooner. I see the logic is already merged with master. Is any input from me still required?
    Aaron Radzinski
    @aradzinski
    @rahul3 Yes, the bug was fixed in the code (it has to be fixed in the code to affect all possible runtimes such as probe, server and CLI).
    @rahul3 Thanks for pointint this out!
    @rahul3 Rahul - take a look at NLPCRAFT-219, NLPCRAFT-186 and NLPCRAFT-330. These are Python related tickets - may be can takcle some of those!
    Rahul Padmanabhan
    @rahul3
    @aradzinski I initially wanted to take on NLPCRAFT-330. That is my original goal but I'm still trying to build a model in Scala to actually use NLPCraft to see what you require. I didn't particulary understand what you mean by native Python client. If I have an example of the functionality required, it would help me understand greatly.
    Rahul Padmanabhan
    @rahul3
    @aradzinski On a different note: The examples need to be built into the jar files for the examples in the documentation of the website to work. I foresee that you will eventually get a lot of data scientists curious about NLPCraft but who don't know much java and I see them struggling with this part.
    skhdl
    @skhdl

    Hi @rahul3 !
    Now everything works in the master branch, so you can continue to review examples.

    Native Python client is REST access library for users (It will be more clear after examples review)
    Also you can look at existing Native Java client - https://github.com/apache/incubator-nlpcraft-java-client

    Aaron Radzinski
    @aradzinski
    As a general note - NLPCraft specifically DOES NOT catter to data sciennce crowd. It was specifically designed for commercial software development - and NOT for academic interests.
    NLPCraft fills the void created by dozens of Python frameworks that are unsuitable for the actual commercial software development and, in my personal opinion, stigmed the development and adoption of NLP in enterprise world
    Rahul Padmanabhan
    @rahul3
    That makes a lot of sense. I wasn't aware of the audience you intended NLPCraft to be for.
    Rahul Padmanabhan
    @rahul3
    There is an issue with the example on the website: https://nlpcraft.apache.org/first-example.html either the param --mdls=demo.LightSwitch needs to be changed to --mdls=demo.LightSwitchModel or the class name needs to be changed in the example code of LightSwitch.scala
    Otherwise we get a java.lang.ClassNotFoundException
    Aaron Radzinski
    @aradzinski
    Good catch! Can you please file the ticket?
    (so that we can track it)
    Rahul Padmanabhan
    @rahul3
    Will do within the next 24 hours.
    Rahul Padmanabhan
    @rahul3
    Rahul Padmanabhan
    @rahul3
    Question: I'm trying to use the java client by adding it as a maven dependency and maven cant find the artifact for 0.7.5. Is it on maven central?
    Could not find artifact org.apache.nlpcraft:nlpcraft-java-client:pom:0.7.5 in central (https://repo.maven.apache.org/maven2)
    skhdl
    @skhdl

    Hmm. Could you describe here or in the new ticket what are you doing and what is wrong with it?

    1. java client master branch is configured to nlpcraft '0.8.0'
      yes, it cannot have access to maven central, because last nlpcraft version which was uploaded to maven central is '0.7.5'
      "Cannot resolve org.apache.nlpcraft:nlpcraft:0.8.0 ()"

    So

    • if you want to work with java client master branch, you have to install locally main nlpcraft project 0.8.0
    • for 0.7.5 - best way to pull java client tag 'v0.7.5'

    But you have the another error "Could not find artifact org.apache.nlpcraft:nlpcraft-java-client:pom:0.7.5 in central (https://repo.maven.apache.org/maven2)"
    0.7.5 is uploaded on master (https://mvnrepository.com/artifact/org.apache.nlpcraft/nlpcraft/0.7.5, I have tested it rigth now)

    So, I try to understand how you got your error.

    Thanks, Sergey

    Rahul Padmanabhan
    @rahul3

    Hello Sergey, yes NLPCraft 0.7.5 is in maven central but I was trying to use the nlpcraft-java-client. I wanted to see the functionality of the java client because I want to build the python client. I followed the instruction at this link: https://nlpcraft.apache.org/download.html#java-client

    I just created a new maven project and added this is my pom.xml. nlpcraft 0.7.5 has no issues but the issue is with the java client at version 0.7.5:

    pom.xml

    <?xml version="1.0" encoding="UTF-8"?>
    <project xmlns="http://maven.apache.org/POM/4.0.0"
             xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
             xsi:schemaLocation="http://maven.apache.org/POM/4.0.0 http://maven.apache.org/xsd/maven-4.0.0.xsd">
        <modelVersion>4.0.0</modelVersion>
    
        <groupId>org.example</groupId>
        <artifactId>nlpcraft-testing</artifactId>
        <version>1.0-SNAPSHOT</version>
    
        <properties>
            <maven.compiler.source>11</maven.compiler.source>
            <maven.compiler.target>11</maven.compiler.target>
        </properties>
    
        <dependencies>
            <dependency>
                <groupId>org.apache.nlpcraft</groupId>
                <artifactId>nlpcraft</artifactId>
                <version>0.7.5</version>
            </dependency>
    
            <dependency>
                <groupId>org.apache.nlpcraft</groupId>
                <artifactId>nlpcraft-java-client</artifactId>
                <version>0.7.5</version>
            </dependency>
        </dependencies>
    </project>
    skhdl
    @skhdl

    Hi Rahul!
    There is a minor mistake on the site again.
    artifactId should be 'java-client' but not nlpcraft-java-client

    Thank you, ticket created - https://issues.apache.org/jira/browse/NLPCRAFT-332

    Note, please that for java client project review - maybe it is more useful to pull this project and look at its code
    (for 0.7.5 - git tag 'v0.7.5', for 0.8.0 - master branch, see my last message above)

    Rahul Padmanabhan
    @rahul3

    Thanks for clarifying. Sorry for all the basic newbie questions. I’m still familiarizing myself with the project.

    And yes, I cloned the client last week and have been going through the code.

    skhdl
    @skhdl
    Thank you for your feedback and documentation remarks/fixes
    Rahul Padmanabhan
    @rahul3
    image.png
    Another thing, I was building the weather bot in Java to test things out and I think there may be another issue. The darksky api is no longer accepting new users so getting a new token isn't possible.
    Aaron Radzinski
    @aradzinski
    Argh… that’s right
    DarkSky was going to close down eventually (Apple bought them)
    do you want to tackle this example’s migration to another weather provider?
    Rahul Padmanabhan
    @rahul3
    ah :( I was working on recreating this and ran into a bunch of issues, when I fixed them all, I got this: https://nlpcraft.apache.org/examples/weather_bot.html
    I could do that
    Aaron Radzinski
    @aradzinski
    that would be really helpful!
    just create the ticket and work on the pull request
    Rahul Padmanabhan
    @rahul3
    thanks, I found the client relatively ok to recreate in Python but I'll work on this first
    sounds good
    I may bother you all with dumb questions now and again: fair warning
    Aaron Radzinski
    @aradzinski
    the beuaty of open source is that you can work on what you really interested in :-)
    pick a problem that you like and have a go at it….
    Rahul Padmanabhan
    @rahul3
    will do
    :)
    Aaron Radzinski
    @aradzinski
    100% no problem