AI Travel Assistent for Vietnam

This project implements a hybrid AI travel Assistant that provides personalized travel recommendations for Vietnam. It leverages a combination of semantic search using Pinecone and a knowledge graph using Neo4j to deliver comprehensive and contextually relevant travel itineraries. The AI components are powered by NVIDIA's API for generating embeddings and chat responses.

Project Overview

The AI travel Assistant is designed to understand a user's travel query, retrieve relevant information from a vector database and a graph database, and then generate a detailed travel plan. The project is structured as a series of Python scripts that handle data processing, database population, and the main chat application.

Key Features

• Hybrid Search: Combines semantic search for finding relevant entities and graph-based search for discovering connections and relationships between them.
• RAG Pipeline: Implements a Retrieval-Augmented Generation (RAG) pipeline to provide grounded and informative responses.
• Stateful Agent: Uses LangGraph to create a stateful and modular agent, making the logic easy to follow, maintain, and extend.
• NVIDIA Integration: Utilizes NVIDIA's API for high-quality embeddings and chat models.
• Data Visualization: Includes a script to visualize the knowledge graph stored in Neo4j.

Getting Started

To get the AI travel Assistant up and running, you will need to set up the required services, configure your API keys, and run the data loading scripts.

Prerequisites

• Python 3.10+
• Docker
• An active internet connection

1. Clone the Repository

First, clone this repository to your local machine:

git clone https://github.com/utkarsh-284/AI-Travel-Assistant.git

2. Install Dependencies

Install the required Python packages using pip:

pip install -r requirements.txt

3. Set Up API Keys and Configuration

You will need API keys from NVIDIA and Pinecone, as well as credentials for your Neo4j database.

1. NVIDIA API Key:

   • Obtain your NVIDIA API key from build.nvidia.com.

2. Pinecone API Key:

   • Log in to your Pinecone account to get your API key.

3. Neo4j Database:

   • Run a Neo4j container using the following Docker command. This will start a Neo4j instance with the necessary ports exposed and authentication set to neo4j/password.

   docker run --restart always --publish=7474:7474 --publish=7687:7687 --env NEO4J_AUTH=neo4j/password neo4j:2025.10.1

4. Configure the Project:

   • Create a config.py file in the root of the project.
   • Add your API keys and database credentials to the config.py file, using config_example.py as a template.

4. Data Loading

Before you can start chatting with the assitant, you need to load the travel data into Pinecone and Neo4j.

1. Upload to Pinecone:

   • Run the pinecone_upload.py script to generate embeddings for the dataset and upload them to your Pinecone index.

   python pinecone_upload_nvidia.py

2. Load into Neo4j:

   • After starting the Neo4j Docker container, run the load_to_neo4j.py script to populate the graph database.

   python load_to_neo4j.py

5. Running the Chat Assistant

Once the data is loaded, you can start the AI travel Assistant:

python hybrid_chat_langgraph.py

The script will prompt you to enter your travel questions. Type exit to quit the application.

(Optional) Visualizing the Graph

To visualize the knowledge graph, you can run the visualize_graph.py script. This will generate an HTML file (neo4j_viz.html) that you can open in your browser.

python visualize_graph.py

Project Structure

• config.py: Configuration file for API keys and database credentials.
• hybrid_chat_langgraph.py: The main chat application that runs the RAG pipeline.
• improvements.md: A document detailing the architectural and performance enhancements.
• load_to_neo4j.py: Script to load the travel dataset into the Neo4j graph database.
• pinecone_upload_nvidia.py: Script to upload the dataset with embeddings to Pinecone.
• requirements.txt: A list of all the Python dependencies.
• vietnam_travel_dataset.json: The raw travel data in JSON format.
• visualize_graph.py: Script to generate a visualization of the Neo4j graph.

Contact

Utkarsh Bhardwaj

Contact: ubhardwaj284@gmail.com

Publish Date: 1st November, 2025