This repository contains a dockerized version of the great SeCo Language Analysis Services Play application. The pre-built image is also available in Docker Hub.
$ docker run --rm -p 19990:9000 -t -i milankinen/seco-las
# "dockerhost" hostname alias points to your docker host machine's address
$ curl 'http://dockerhost:19990/las/baseform?text=Terve+maailmaan!&locale=fi'
# =>
# "terve maailma !"By default, the application uses 2G memory which is the absolute minimum requirement -
smaller amounts cause OutOfMemoryError exceptions due to big sizes of the used models.
Memory can be configured by using LAS_MEMORY environment variable. The value is integer
describing Java process memory in megabytes, e.g
$ docker run --rm -e LAS_MEMORY=4096 -p 19990:9000 -t -i milankinen/seco-las See the official documentation.
The dockerfile and build files are licensed under MIT
However, note that the dependencies inside the container have different licenses:
| Dependency | License |
|---|---|
| seco-lexicalanalysis and seco-lexicalanalysis-play | MIT |
| seco-hfst | Apache-2.0 |
| HFST models | GPLv3 |
| mate-tools | GPLv3 |
| marmot | GPLv3 |