Feat: Implement RAG pipeline for advanced conversational AI

[sfdamm]

This pull request introduces a comprehensive Retrieval-Augmented Generation (RAG) pipeline to power the AI assistant. This is a major update that enables the assistant to understand and answer questions based on a knowledge base of documents.

Key changes and features:

• End-to-End RAG Pipeline:

  • Implemented to handle the indexing, embedding, and retrieval of documents using a Weaviate vector store. VectorStoreService
  • Introduced to orchestrate the pipeline, taking user queries, retrieving relevant context from the vector store, and generating informed responses. LLMService
  • Added to construct the final response from the retrieved context. ResponseGenerator

• Document and File Handling:

  • The system now supports adding and deleting documents, which serve as the knowledge base for the RAG pipeline.
  • This allows the assistant's knowledge to be dynamically updated and managed.

• Enhanced Chat Functionality:

  • Developed a new chat controller () to manage the conversation flow. chat_controller.py
  • Integrated chat history to enable multi-turn, contextual conversations, allowing the assistant to remember previous parts of the dialogue.

• Bug Fixes and Stability:

  • Resolved a critical bug where a variable was missing in the , which previously caused response generation to fail. title``PromptTemplate
  • Ensured all new services are properly initialized and integrated.