Details about the algorithm can be found in I. Segalovich «A fast morphological algorithm with unknown word guessing induced by a dictionary for a web search engine», MLMTA-2003, Las Vegas, Nevada, USA.
The wrapper's code in under MIT license, but please remember that Yandex.MyStem is not open source and licensed under conditions of the Yandex License.
The wrapper should at least work on Ubuntu Linux 12.04+, Windows 7+ (+ people say it also works on OS X).
<dependency>
<groupId>ru.stachek66.nlp</groupId>
<artifactId>mystem-scala</artifactId>
<version>0.1.6</version>
</dependency>
Only mystem 3.{0,1} are supported currently. Please create issues for compatibility troubles and other requests.
Probably the most important thing to remember when working with mystem-scala is that you should have just one MyStem instance per mystem/mystem.exe file in your application.
import java.io.File
import ru.stachek66.nlp.mystem.holding.{Factory, MyStem, Request}
object MystemSingletonScala {
val mystemAnalyzer: MyStem =
new Factory("-igd --eng-gr --format json --weight")
.newMyStem(
"3.0",
Option(new File("/home/coolguy/coolproject/3dparty/mystem"))).get()
}
object AppExampleScala extends App {
MystemSingletonScala
.mystemAnalyzer
.analyze(Request("Есть большие пассажиры мандариновой травы"))
.info
.foreach(info => println(info.initial + " -> " + info.lex))
}
import ru.stachek66.nlp.mystem.holding.Factory;
import ru.stachek66.nlp.mystem.holding.MyStem;
import ru.stachek66.nlp.mystem.holding.MyStemApplicationException;
import ru.stachek66.nlp.mystem.holding.Request;
import ru.stachek66.nlp.mystem.model.Info;
import scala.Option;
import scala.collection.JavaConversions;
import java.io.File;
public class MyStemJavaExample {
private final static MyStem mystemAnalyzer =
new Factory("-igd --eng-gr --format json --weight")
.newMyStem("3.0", Option.<File>empty()).get();
public static void main(final String[] args) throws MyStemApplicationException {
final Iterable<Info> result =
JavaConversions.asJavaIterable(
mystemAnalyzer
.analyze(Request.apply("И вырвал грешный мой язык"))
.info()
.toIterable());
for (final Info info : result) {
System.out.println(info.initial() + " -> " + info.lex() + " | " + info.rawResponse());
}
}
}
The references to this repository are highly appreciated, if you use our work.
@misc{alekseev2018mystemscala,
author = {Anton Alekseev},
title = {mystem-scala},
year = {2018},
publisher = {GitHub},
journal = {GitHub repository},
howpublished = {\url{https://github.com/alexeyev/mystem-scala/}},
commit = {the latest commit of the codebase you have used}
}
If you do cite it, please do not forget to cite the original algorithm's author's paper as well.
Anton Alekseev [email protected]
- Vladislav Dolbilov, @darl
- Mikhail Malchevsky
- @anton-shirikov
- Filipp Malkovsky
- @dizzy7