I am building a matchmaking app that recommends dates between users. When a user searches for a date:
- The user triggers an https cloud function from the client
- This cloud function sends a request to a python server for date recommendations
- The python server uses a machine learning model to generate and return a list of json date recommendations
- The cloud function creates documents in a firestore database from the response
In order to generate recommendations the python server requires data (users, previous recommendations, searches, date venues etc), which is currently only stored in the firestore database.
I need a database the python server can query at runtime. So far, I have the following requirements:
- less than 500ms latency to retrieve data
- stays up to date in real time with firestore (I plan to do this with cloud functions)
- ability to perform complex queries
What database types or products suit this application?