How to run in localhost server compiling from the newest source code

[ShomyLiu]

First I git clone the source code from https://github.com/dbpedia-spotlight/dbpedia-spotlight But I got a 0.70.jar after maven the project. However it can run successful with the command:

java -jar **********

So I find this repo, and clone to my computer and then use maven to package a jar: spotlight-1.0-jar-with-dependencies.jar

But when I run the jar, it got an error:

Error: Could not find or load main class org.dbpedia.spotlight.web.rest.Server

Did I forget something to deal with? Thanks

BTW, According to my test, the version 0.7 has a quite faster speed than the 1.0. I annotate a same text file, one cost 110 seconds and the other is 260 seconds.

[darioloetscher]

Hey,

You just picked the wrong jar. The jar you want to run is rest/target/rest-1.0-jar-with-dependencies.jar .

Then you can follow the normal instruction here https://github.com/dbpedia-spotlight/dbpedia-spotlight/wiki/Run-from-a-JAR .

As u can see, they just renaming the rest-1.0-jar-with-dependencies.jar to dbpedia-spotlight-nightly-build.jar in there CI http://jenkins.dbpedia-spotlight.org/job/spotlight/lastBuild/console .

[ShomyLiu]

@darioloetscher Hi, Thanks. It does work with the jar in rest/target/rest-1.0-jar-with-dependencies.jar. But why is the annotate speed in version 1.0 is quite slower than the version 0.7, costing almost 2X time.

[sandroacoelho]

Hi @darioloetscher, thank you for the answer.

Hi @ShomyLiu: our version 1.0 is version 0.7 without our first approach with Lucene. Btw, I am going to check it. Thank you for reporting it

[ShomyLiu]

@sandroacoelho OK, Looking forward to your checking result.

[Skunnyk]

@ShomyLiu can you post more informations about your performances problems between 0.7 and 1.0 ? Maybe you are hit by the "100-Continue" thing, see https://github.com/dbpedia-spotlight/dbpedia-spotlight/issues/450#issuecomment-329505073

[ShomyLiu]

@Skunnyk Thanks!
It seems that the problem is caused by 100-Continue. And as the #450 says, use a nginx server behind spotlight rest server, which means I can start multiple spotlight instances ,and use a upstream for loading balance like below:

 upstream localhost {
        server localhost:2222;  #a spotlight instance server
        server 1ocalhost:2223; #another spotlight instance server
        ....;
    }

    server {
        listen 80;
        location / {
            proxy_pass http://localhost;
        }
    }

Sorry for that I am new to nignx server, So would you mind checking the code ? Thanks

[Gautamshahi]

Hi, @ShomyLiu Do you have the jar file, can you please provide the link? On the given website its not working and nobody is helping.

[ShomyLiu]

@Gautamshahi Hi, I have uploaded the spotlight 1.0 jar in : https://drive.google.com/file/d/1vFwVCzoY9R9rBPpfmgD9LlXpW6KdEc7-/view?usp=sharing

As for the language jar (i.e, en.tar.gz) you can download from https://sourceforge.net/projects/dbpedia-spotlight/files/2016-10/en/model/

then you can run your jar according to https://github.com/dbpedia-spotlight/dbpedia-spotlight/wiki/Run-from-a-JAR

tar en.tar.gz
java -jar rest-1.0-jar-with-dependencies.jar path/to/model/folder/en_2+2 http://localhost:2222/rest

In fact, you can maven the spotlight 1.0 jar from the source code https://github.com/dbpedia-spotlight/dbpedia-spotlight-model: Be sure that installing java and maven

mvn clean
mvn install

this will cost some time in the first time for downloading some jars. After finishing this part, the jar will be in rest/target/rest-1.0-jar-with-dependencies.jar

[Gautamshahi]

Thank you for your effort.

I tried both ways but getting the error of load main class, one of the screenshot.

[ShomyLiu]

@Gautamshahi Hi, from the error information, it seems that your target folder DISI_3rd and 4th has blank space. You can rename your folder that doesn't have spaces or escape wiht \

[Gautamshahi]

Thank you, any suggestion?

[ShomyLiu]

@Gautamshahi From your error information, it's out of memory while loading the english corpus model. You can set more memory for jvm via -xms =xmx, which depends on the memory in your computer, such as:

java -Xms8G -Xmx8G -jar ........

[Gautamshahi]

Thank you, Yes It needs more RAM. I laptop has 6GB only.