ROS wrapper for rasa nlu. Mainly used for SPR robocup@home challenge.
Currently not a ROS package.

pip install sklearn_crfsuite sklearn --user
pip install testresources rasa_nlu spacy --user
pip install -U scikit-learn scipy sklearn-crfsuite --user
python -m spacy download en_core_web_md
python -m spacy link en_core_web_md en

• https://rasahq.github.io/rasa-nlu-trainer/

• Follow the installation process
• git clone https://github.com/walkingmachine/wonderland in your ROS workspace
• in wonderland repo : git checkout robocup2018
• catkin_make your workspace
• start wonderland : python manage.py runserver
• rosrun wm_nlu spr_nlu.py or
• rosrun wm_nlu gpsr_nlu.py

Currently, wm_nlu will query rasa_nlu to get the sentences entities. It will then query wonderland and return the answer to the answer.

Once the first version will be working, the wonderland part will be seperated from wm_nlu. Another package will be in charged of converting the ouput from wm_nlu to query wonderland database.