website - http://www.dbpedia-spotlight.org
demo service - http://demo.dbpedia-spotlight.org
Since v1.0, DBpedia Spotlight was split into two versions, under the same API, as follow:
DBpedia-Spotlight-Model: described in Improving Efficiency and Accuracy in Multilingual Entity Extraction
DBpedia-Spotlight-Lucene: described in DBpedia Spotlight: Shedding Light on the Web of Documents
This important movement was the way that we found to deliver faster fixes and new releases, providing solutions for each annotation approach.
Our first achievement is related with licensing. DBpedia Spotlight Model is now full compliance with Apache 2.0. It means that you can use it without any commercial restrictions.
We are so excited because there's even more great news to come.
If you require any further information, feel free to contact us via dbpedia@infai.org. We are already very excited to spend time with you on further community meetings and to publish new DBpedia releases.
Keep annotating,
All the best
DBpedia Spotlight looks for ~3.5M things of unknown or ~320 known types in text and tries to link them to their global unique identifiers in DBpedia.
The requirements to compile the DBpedia Spotlight project are as follows:
The process to compile the DBpedia Spotlight code to produce the corresponding jar files are as follows:
export JAVA_HOME=/path/of/the/jdk1.8/
- only required if the default JDK is differentmvn package
- this process will take some timeEach module (core, index, rest) will produce the corresponding jar file in the "target" directory, e.g., core/target or index/target. The jar file containing the DBpedia Spotlight annotation service will be in the rest module directory (rest/target/rest-1.1-jar-with-dependencies.jar
).
To install the DBpedia Spotlight in the local maven repository, replace the mvn package
command with the mvn install
command.
To test the produced jar file, download a language model from the DBpedia Databus, uncompress it and execute the following command:
java -Dfile.encoding=UTF-8 -Xmx10G -jar rest-1.1-jar-with-dependencies.jar /path/to/the/uncompress/language/model http://0.0.0.0:2222/rest
When the DBpedia Spotlight annotation service is ready, it could be queried through a curl command as follows:
curl http://0.0.0.0:222/rest/annotate --data-urlencode "text=Barack Obama was the president of the United States" - H "Accept: text/turtle"
Go to our Demonstration page, copy+paste some text and play with the parameters to see how it works.
https://api.dbpedia-spotlight.org/{LANGUAGE}/annotate
You can use our demonstration Web Service directly from your application.
curl https://api.dbpedia-spotlight.org/en/annotate \
--data-urlencode "text=President Obama called Wednesday on Congress to extend a tax break
for students included in last year's economic stimulus package, arguing
that the policy provides more generous assistance." \
--data "confidence=0.35"
or for JSON:
curl https://api.dbpedia-spotlight.org/en/annotate \
--data-urlencode "text=President Obama called Wednesday on Congress to extend a tax break
for students included in last year's economic stimulus package, arguing
that the policy provides more generous assistance." \
--data "confidence=0.35" \
-H "Accept: application/json"
If you need service reliability and lower response times, you can run DBpedia Spotlight in your own In-House Server based on Docker.
Models and raw data for most languages are available here.
If you use DBpedia Spotlight in your research, please cite the following paper:
@inproceedings{isem2013daiber,
title = {Improving Efficiency and Accuracy in Multilingual Entity Extraction},
author = {Joachim Daiber and Max Jakob and Chris Hokamp and Pablo N. Mendes},
year = {2013},
booktitle = {Proceedings of the 9th International Conference on Semantic Systems (I-Semantics)}
}
All the original code produced for DBpedia Spotlight Model is licensed under Apache License, 2.0.
More documentation is available from the DBpedia Spotlight wiki.
Check the FAQ here