This docker container allows you to build, install and run the rstWeb Rhetorical Structure Theory annotation tool (Zeldes 2016) in a docker container.
Specifically, this will run the version 2.0.3 of rstWeb with an additional
REST API for importing, exporting and converting rs3 files (to PNG images).
Unlike rstweb-docker,
this container runs the "local" version of rstWeb.
Note: The rs3 files you upload, edit and store in rstWeb will
only exist as long as the container is running. In order to persist the
data, you would need to bind-mount
an empty SQLite database into /opt/rstweb/rstweb.db.
To build, install and run rstweb-service, simply type:
docker run -p 9000:8080 -ti nlpbox/rstweb-service
Inside the container, rstWeb (and the added rstweb-service REST API)
listens on port 8080, but we'll map it to port 9000 of our host system
(localhost / 127.0.0.1).
Now, you can create a project called rst-workbench in the running
rstWeb instance. We'll use httpie for the examples.
http POST "http://127.0.0.1:9000/api/projects/rst-workbench"
To list the currently stored projects, call:
http GET "http://127.0.0.1:9000/api/projects"
To import the rs3 file /path/to/source.rs3 into the project rst-workbench
and store it there under the name target.rs3, type:
http --form POST "http://127.0.0.1:9000/api/documents/rst-workbench/target.rs3" \
rs3_file@/path/to/source.rs3
Afterwards, you can open the document target.rs3 in your browser
under the URL [http://127.0.0.1:9000/open] and edit it in the Structure
editor. Don't forget to save your changes using the save button!
You can either download the edited file by clicking on the </> xml
button in the Structure editor or retrieve it programmatically (this is
what rst-workbench does under the hood):
http GET "http://127.0.0.1:8080/api/documents/rst-workbench/target.rs3"
Zeldes, Amir (2016). rstWeb - A Browser-based Annotation Interface for Rhetorical Structure Theory and Discourse Relations. In: Proceedings of NAACL-HLT 2016 System Demonstrations. San Diego, CA, 1-5.