[RFE] Postgres-based Vector Store support

[TamiTakamiya]

Is your feature request related to a problem? Please describe.
Currently road-core/service uses FaissVectorStore with pre-prepared data for RAG content.
We want to add Postgres-based Vector Store support with the pgvector Postgres
extension, which will enable more flexible RAG data handling.

Describe the solution you'd like
Use one of Postgres-based vector stores supported by llama-index.
The PoC shown below is using Supabase Vector Store

Describe alternatives you've considered
LangChain also provides various vectore stores.
However, I think using a vector store supported llama-index would be easier since the existing code was written for
the FaissVectorStore for llama-index.

Additional context
PoC

• Data preparation (aap-rag-content, built based on openshift/lightspeed-rag-content): https://github.com/ansible/aap-rag-content/tree/TamiTakamiya/pgvector
• Chatbot service (ansible-chatbot-service, a downstream of road-core/service): https://github.com/ansible/ansible-chatbot-service/tree/TamiTakamiya/pgvector

Demo video

Screenshare.-.2024-12-02.4_35_36.PM.mp4