A web demo for visualizing Semafor parses
Uses Flask as the web server.
Python requirements are in requirements.txt
To compile Coffeescripts:
coffee --watch --compile --output semviz/static/js/compiled/ semviz/static/coffeescripts/
Coffeescript tests require jasmine-node. To run:
jasmine-node --coffee semviz/static/spec
Stylesheets use Sass. To compile:
sass --watch semviz/static/css/style.scss:semviz/static/css/style.css
Make sure semviz.settings.SEMAFOR_HOME points to a valid installation of
SEMAFOR (>= v3.0-alpha-03).
Make sure SEMAFOR is running in server mode:
cd $SEMAFOR_HOME
java -Xms4g -Xmx4g -jar target/Semafor-3.0-alpha-03.jar model-dir:<directory-of-trained-model> port:4444
and semviz.settings.SEMAFOR_HOST and semviz.settings.SEMAFOR_PORT point to the running instance.
To start the (development) server:
./semviz/web_app.py