lux-ai-query-builder

A LLM powered advanced search query builder

Running it

Create a file called "google_project.txt" Put the name of your google project in it. e.g. echo "name of your project" > google_project.txt

Then:

export FLASK_APP=ai-proxy.py
export FLASK_DEBUG=True
flask run --host localhost

Using it

Make a request with a q parameter to /api/translate/{scope}. The scope actually doesn't matter at the moment.

curl http://127.0.0.1:5000/api/translate/item\?q=17th+century+italian+paintings

will return

{"AND": [{"producedDate": "1600-01-01T00:00:00", "_comp": ">="}, {"producedDate": "1700-01-01T00:00:00", "_comp": "<"}, {"producedAt": {"partOf": {"name": "Italy"}}}, {"classification": {"name": "painting"}}], "_scope": "item"}

if your AI is feeling like it at the time.