A Multi-Dialog Conversational Approach for Service Mashup in IoT Environments

Branches

• master branch is the implementation of CoMMA and was used for the user study.
• single-utterance branch is the baseline agent that was compared with CoMMA.
• video-figure branch contains the realtime visualization tool (which is named realtime-vis.ipynb) and the dialogflow agent file.

Install

pip3 install -r requirements.txt

Testing Environment

• Python 3.5 with Flask==1.1.2
• Valid SSL certificate is required with filename of server.crt and server.key.
• Import the agent in your dialogflow console, and configure the webhook settings.
• Simply sudo ./app.py from your local environment.
• Simply ./app.py --help to see the usage.

Descriptions

• app.py : Webhook responses management.
• mashup.py : Class definitions for mashup and each service.
• identifier2ont.py : Converting user utterance to corresponding ontology definition.
• ont2nl.py : Converting ontology definition to corresponding natural langauge.

Functionality Tests

• mashup-test-generate.ipynb : Test if the mashup is properly generated from the dump interaction file.
• mashup-test-merge.ipynb : (Deprecated) Test if two mashup with the same trigger is properly merged.
• mashup-test-speak.ipynb : Test if the description for a mashup is properly generated from the dump interaction file.