@InProceedings{Marx/KBox/ICSC/2017,
Title = {{KBox}: {T}ransparently {S}hifting {Q}uery {E}xecution on {K}nowledge {G}raphs to the {E}dge},
Author = {Edgard Marx and Ciro Baron and Tommaso Soru and S\"oren Auer},
Booktitle = {11th IEEE International Conference on Semantic Computing, Jan 30-Feb 1, 2017, San Diego, California, USA},
Year = {2017}
}
KBox is an abbreviation for Knowledge Box. it is a data management framework designed to facilitate data deployment whether on the cloud, personal computers or smart devices (edge). KBox is designed to cover different aspects of data management. It can be used to describe how the data is built as well as its dependencies. The concept of a data management might be familiar to you if you are aware of software management tools such as npm for JavaScript, gem for Ruby, and NutGet for .NET. KBox is the option for data dependency management. The rationale behind KBox is to allow users to manage data dependency models for data-driven applications. That is, KBox helps you to publish as well as locate and install data models. Moreover, to make it easier to manage Knowledge Graphs, KBox has embedded one of the most popular RDF frameworks, Jena. With KBox users can natively share, deploy and query RDF Knowledge Graphs at scale.
- Why use KBox?
- What is possible to do with it?
- How can I use KBox?
- Installing KBox
- How can I execute KBox in command Line?
- How can I use KBox with python?
- How can I use KBox in my project?
- How can I list available knowledge bases?
- How can I query a published knowledge bases?
- How can I query multi-bases?
- Starting a SPARQL endpoint
- Querying a SPARQL endpoint
- Listing the resource's folder
- Changing the resource's folder
- Using KBox with Docker
- How can I publish my own dataset?
- How can I create my own custom build(jar) of KBox?
- Check our Wiki for old APIs
Systems usually deal with resources and knowledge that are often duplicated among several instances. For instance, when using the Stanford NLP library the resources and knowledge inside the library are duplicated among different applications. The idea is to have a common repository where users can share resources without duplication. In order to do that, we bring the RDF concept to bridge the gap among resource publishing, storing and locating.
With KBox you can share resources and knowledge among several applications, but not just that. In order to allow an easier knowledge dissemination, we have implemented Kibe library. The Kibe library allows applications to virtually install and query RDF knowledge bases. It takes around ~50 minutes to start querying DBpedia on your computer to avoid server overheads and faults.
You can use KBox either as a command-line program or a library in your application. It is easy to plug and use it.
- Installing via runnable jar file:
- Download the library here.
- Installing via pip:
- Install via pip:
pip install kbox
- Type the following:
kbox <command> [option]
Where [command] is:
* convert <directory|file> [<destFile>] [kb|zip] - convert the content of a directory (default kb).
kb - into a kb file. ps: the directory might contain only RDF compatible file formats.
zip - into a zip file.
* convert <file> [<destFile>] gzip - encode a given file.
* sparql <query> (kb <KB> | server <URL>) [install] [-json] - Query a given knowledge base (e.g. sparql "Select ..." kb "KB1,KB2")
- ps: use -install in case you want to enable the auto-dereference.
* server [port <port> (default 8080)] [subDomain <subDomain> (default kbox)] kb <kb-URL> [install] - Start an SPARQL endpoint in the given subDomain containing the given bases.
* server [port <port> (default 8080)] [subDomain <subDomain> (default kbox)] rdf <directories|files> [install [install]] - Start an SPARQL endpoint in the given subDomain containing the given RDF files.
* server [port <port> (default 8080)] [subDomain <subDomain> (default kbox)] target <target> - Start an SPARQL endpoint in the given subDomain containing the target RDF files.
* list [/p] - List all available KNS services and knowledge bases.
* list kns - List all available KNS services.
* install <URL> - Install a given resource.
* install kns <kns-URL> - Install a given KNS service.
* install kb <kb-URL> [version <version>] - Install a given knowledge base using the available KNS services to resolve it.
* install kb <kb-URL> file <kbFile> - Install a given kb file in a given Kb-URL.
* install kb <kb-URL> kns <kns-URL> [version <version>] - Install a knowledge base from a a given KNS service with the specific version.
* install [install] kb <kb-URL> rdf <directories|files> [version <version>] - Install a knowledge base from a a given RDF files with the specific version.
* install kn <kn-URL> [format [version <version>]] - Install a given knowledge base using the available KNS services to resolve it.
* remove kns <kns-URL> - Remove a given KNS service.
* info <URL> format <format> version <version>]] - Gives the information about a specific KB.
* locate <URL> - returns the local address of the given resource.
* locate kb <kb-URL> version <version>] - returns the local address of the given KB.
* locate kn <kn-URL> format version <version>]] - returns the local address of the given KB.
* search <kn-URL-pattern> [format <format> [version <version>]] [/p] - Search for all kb-URL containing a given pattern.
* r-dir - Show the path to the resource folder.
* r-dir <resourceDir> - Change the path of the resource folder.
* version - display KBox version.
** Note: If you want to get the results of above commands as a JSON output, you can append -o json parameter at the
end of each command.
Let's look at the list command as an example,
kbox list
KBox KNS Resource table list
##############################
name,format,version
##############################
http://purl.org/pcp-on-web/dbpedia,kibe,c9a618a875c5d46add88de4f00b538962f9359ad
http://purl.org/pcp-on-web/ontology,kibe,c9a618a875c5d46add88de4f00b538962f9359ad
When you append -o json parameter at the end, the result will be look like this,
kbox list -o json
{
"status_code": 200,
"message": "visited all KNs.",
"results": [
{
"name": "http://purl.org/pcp-on-web/dbpedia",
"format": "kibe",
"version": "c9a618a875c5d46add88de4f00b538962f9359ad"
},
{
"name": "http://purl.org/pcp-on-web/ontology",
"format": "kibe",
"version": "c9a618a875c5d46add88de4f00b538962f9359ad"
}
]
}
This -o jsonsub command is only applicable with list, install, remove, info, locate, search, r-dir & version commands.
- Install KBox pip package
pip install KBox
-
Import the kbox package (
from kbox import kbox). -
Execute any KBox command with execute() function as follows.
kbox.execute('<kbox_command>')
**Note: execute() method will return a string output which contains the result of the executed command.
You can see more details about KBox python package from here
- KBox is distributed over Maven.
- You can add KBox to your project by importing the desired library:
- Add the following dependency on your project:
<dependency>
<groupId>org.aksw.kbox</groupId>
<artifactId>kbox.kibe</artifactId>
<version>v0.0.1-alpha3-RC16</version>
</dependency>
- Add the internal AKSW repository to your pom file:
<repositories>
<repository>
<id>maven.aksw.internal</id>
<name>University Leipzig, AKSW Maven2 Repository</name>
<url>http://maven.aksw.org/archiva/repository/internal</url>
</repository>
...
</repositories>
- Rock it.. ;-)
You can list the available knowledge graphs with list command:
kbox list
Knowledge base table list
http://dbpedia.org/3.9/en/full
http://dbpedia.org/3.9/en
http://dbpedia.org/2015-10/en
http://dbpedia.org/2015_10/en/full
http://dbpedia.org/2015-10/en/full
http://dbpedia.org/ontology
http://www.w3.org/1999/02/22-rdf-syntax-ns
https://www.w3.org/2000/01/rdf-schema
...
Weeeelll... it's quite easy.
Remember the commands listed on 'How can I execute KBox in command Line'.
It's just about executing the command line below. Remember to add -install, so the knowledge base is automatically dereferenced.
kbox sparql "Select (count(distinct ?s) as ?n) where {?s ?p ?o}" kb "https://www.w3.org/2000/01/rdf-schema" install
------
| n |
======
| 32 |
------
Or using the java API...
KBox.query("Select (count(distinct ?s) as ?n) where {?s ?p ?o}", true, new URL("https://www.w3.org/2000/01/rdf-schema"));
You might want to setup the model before starting to execute multiple queries on it:
Model model = KBox.createModel(new URL("https://www.w3.org/2000/01/rdf-schema"));
KBox.query("Select (count(distinct ?s) as ?n) where {?s ?p ?o}", model);
It's very easy, as you just need to add the knowledge base you want to query separated by commas as the command below:
In the given example, we query two knowledge bases, https://www.w3.org/2000/01/rdf-schema and http://xmlns.com/foaf/0.1.
kbox sparql "Select (count(distinct ?s) as ?n) where {?s ?p ?o}" kb "https://www.w3.org/2000/01/rdf-schema,http://xmlns.com/foaf/0.1" install
-------
| n |
=======
| 123 |
-------
Or using the java API...
KBox.query("Select (count(distinct ?s) as ?n) where {?s ?p ?o}",
true, new URL("https://www.w3.org/2000/01/rdf-schema"),
new URL("http://xmlns.com/foaf/0.1"));
You might want to setup the model before starting to execute multiple queries on it:
Model model = KBox.createModel(new URL("https://www.w3.org/2000/01/rdf-schema"),
new URL("http://xmlns.com/foaf/0.1"));
KBox.query("Select (count(distinct ?s) as ?n) where {?s ?p ?o}", model);
Yes, you can!!
kbox server kb "https://www.w3.org/2000/01/rdf-schema,http://xmlns.com/foaf/0.1" install
Loading Model...
Publishing service on http://localhost:8080/kbox/query
Service up and running ;-) ...
If you successfully instantiate your server, now you can access the web client at http://localhost:8080 as shown bellow:
by console
kbox server "http://localhost:8080/kbox/query" sparql "Select (count(distinct ?s) as ?n) where {?s ?p ?o}"
-------
| n |
=======
| 123 |
-------
or JAVA API
URL url = new URL("Select (count(distinct ?s) as ?n) where {?s ?p ?o}");
ServerAddress serverURL = new ServerAddress("http://localhost:8080/kbox/query");
try(ResultSet rs = KBox.query(sparql, serverURL);) { // do not forget to close the result set.
...
}
kbox r-dir
Your current resource directory is: kbox/dir/path
You might get problems with permissions using KBox in shared environments. Therefore, you can change the resource directory of KBox to your desired one.
kbox r-dir new/path
You can also use KBox Docker container by following the steps below:
- Install Docker in your machine (Consult the guide for more details).
sudo apt-get update
sudo apt-get -y install docker-engine
- Pull KBox from AKSW hub repository.
docker pull aksw/kbox
- Run it...
docker run aksw/kbox sparql "Select (count(distinct ?s) as ?n) where {?s ?p ?o}" kb "https://www.w3.org/2000/01/rdf-schema" install
------
| n |
======
| 32 |
------
- Start a server using docker:
docker run -p 8080:8080 aksw/kbox server kb "https://www.w3.org/2000/01/rdf-schema" install
Loading Model...
Publishing service on http://localhost:8080/kbox/query
Service up and running ;-) ...
- You can also query your endpoint from a KBox docker container, but do not forget to use the '--network host' flag:
docker run --network host aksw/kbox -sparql "Select ?s {?s ?p ?o} limit 5" -server "http://localhost:8080/kbox/sparql"
------
| n |
======
| 32 |
------
- You can also cache the installed knowledge bases, so you don't need to download them multiple times:
mkdir my-cache-dir/.kbox
docker run --name kbox --mount type=bind,source=my-cache-dir/.kbox,target=/root/.kbox aksw/kbox -server -kb "https://www.w3.org/2000/01/rdf-schema" -install
Although you can create your own KNS service and publish your datasets, currently KBox does not allow you to directly publish content to make them available to other users. In order to do that, you must contact us.
Contact: [email protected]
Information needed:
- Dataset Label;
- Dataset Version;
- The Publisher: Your or your organization's email/URL;
- The Creator: Who has created the Knowledge base, e.g. DBpedia -> http://dbpedia.org;
- The License: the dataset license;
- The URL where the Knowledge graph file can be dereference (please create the file using KBox
createIndexcommand); - The Dataset URI name: the URI name that will be used by users to dereference your dataset;
- The Dataset description: Give us a few words to help others to know what your dataset is about;
- Tell us one reason why KBox is awesome. :-)
If you want to do some modifications or changes to the KBox and want to use the KBox with those modifications, you have to create your own custom build (jar). To do that follow these steps,
- Open a terminal on the KBox home directory
- Issue the following command in the terminal to build the project.
mvn clean install - Then navigate to
kbox.kibe/targetdirectory - In there you will be able to see a jar file named similar to kbox.kibe-x.x.x-........-jar-with-dependencies.jar
- This is the jar that contains all the modifications you did.
![dependabot[bot] avatar](https://avatars.githubusercontent.com/in/29110?v=4 "dependabot[bot]")
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent