elbarhichi/Mouwajihi

Our project aims to implement a Retrieval-Augmented Generation (RAG) system to assist high school and college students in Morocco with academic and career orientation # OrienTech Chatbot ## Overview Our project aims to implement a Retrieval-Augmented Generation (RAG) system to assist high school and college students in Morocco with academic and career orientation. The solution consists of two main parts: 1. *Interview Phase*: An interactive session where the Language Learning Model (LLM) gathers information about the student's preferences, interests, and background. 2. *Guiding Phase*: Utilizing RAG, the chatbot processes documents to provide personalized academic and career guidance. ## Team EL BARHICHI Mohammed MEZIANY Imane EL YOUBI Asmae ## Solution Description We leverage RAG with documents extracted via web scraping from official Moroccan school and university websites. Our approach includes several optimizations: - *Enhanced Chunking*: Each chunk contains comprehensive information about a single field of study, including access routes, instead of using traditional chunking methods. - *Graph-Vector Database Integration*: We combine graph databases with vector databases to detect various possible pathways a student can take to reach a specific career goal. ## Architectures - *Graph Database Architecture (for the demo use case)*: - *Overall Project Architecture*: ## Repository Structure - *Data Directory*: Contains the scraped data post-chunking and the graph structure showing different paths to a given career. - *Scripts Directory*: Includes scripts for automating web scraping, graph creation, and path detection. - *Models Directory*: Houses notebooks for comparing different LLM approaches. - *Demos*: Contains the demo videos of the project - *App1 and App2*: Streamlit applications for launching the two parts of the solution (interviewing and guiding). ## Getting Started To get started with our project, follow these steps: 1. *Clone the Repository*: ```b sh git clone github.com cd mouwajihi ``` 2. *Install Dependencies*: ```b sh pip install -r requirements …