The natural language queries are processed and transformed into SPARQL queries, which is used to lookup DBpedia to find a solution.
Install requirements using:
pip3 install -r req.txt
Create a folder data/
and place datasets there.
Run the standalone implementation using python3 Sparql.py "u<query>"
.
Otherwise, run the flask app using python3 app.py
and answer interact from the frontend.
Sample dataset used can be found at https://github.com/ag-sc/QALD/blob/master/7/data/qald-7-train-largescale.json .
To open the front end, simply open index.html
when inside the folder qald-interface
from a browser.
A partial implementation which uses Blazegraph hosted over a local network is available under kmst.py
.