Hierarchical data and multimodal RAG using Semantic Chunking and Langchain

This repo is built as a demo for the Hierarchical data table which overlaps to different pages without header. It uses Langchain, Semantic Chunking, Azure Document intelligence and AI Search Please insert your file and try the lab

Prerequisites

Before running the notebook, make sure you have the following installed:

1. Python: Install the latest version of Python from the official website.
2. Jupyter Notebook: Install Jupyter Notebook using the following command in PowerShell:

   pip install jupyter
   

To replace the multi_page_table.pdf with your document, follow these steps:

1. Open the existing document.
2. Locate the multi_page_table section.
3. Replace the content of the multi_page_table.pdf path with the path of the existing document and change the name references
4. Save the changes.

Steps

1. Clone the repository:

   git clone <repository_url>
   

2. Navigate to the project directory:

   cd <project_directory>
   

3. Create a virtual environment (optional but recommended):

   python -m venv venv
   

4. Activate the virtual environment:

   .\venv\Scripts\activate
   

5. Install the required dependencies:

   pip install -r requirements.txt
   

6. Install Jupyter Notebook: pip install jupyter

7. Launch Jupyter Notebook:

   jupyter notebook
   

8. In your web browser, navigate to the notebook file (lab.ipynb) and open it.

9. Run the notebook cells one by one to execute the demo , replace the document location

That's it! You have successfully run the demo Python notebook in Windows PowerShell.