GenAI Sharepoint RAG Chatbot

Objective: The primary goal of this demo was to showcase a RAG Chatbot utilizing data from a Sharepoint folder.

Process Overview:

Connection to Sharepoint: The system establishes a connection to a specific Sharepoint folder.
Data Acquisition: It downloads all the files contained within the folder.
Data Processing: The files are then processed differently based on their type. For instance, Word documents, PowerPoint presentations, and CSV files each have a unique processing method. The appropriate GenAI chunker & splitter is utilized for each file type.
Video File Handling: In the case of video files (MP4), the system extracts the audio, breaks it down into manageable parts, and uses OpenAI for transcription.
Data Storage: All processed information is stored in a VectorDB.
Chatbot Creation: Utilizing this data, a chatbot is created based on GPT-4 that can answer questions derived from the information in these files.

How to Run

.env file

Create a .env file with the following information:

OPENAI_API_KEY=<your open ai key>
OPENAI_API_TEMPERATURE=0.6
OPENAI_API_MODEL_NAME=gpt-4
OFFICE365_CLIENT_ID=<client id>
OFFICE365_CLIENT_SECRET=<client secret>
SHAREPOINT_SITE_URL=https://example.sharepoint.com/sites/myside
SHAREPOINT_FOLDER=Shared Documents/Engineering/genai

Installing Dependencies

pip3 install -r requirements.txt

Running

Running bootstrap.py will open a browser tab with the streamlit interface.

python3 bootstrap.py

Areas for Improvement

While the demo was successful, there are opportunities for enhancement. One significant aspect is the preparation of documents before processing. Encouraging users to provide more detailed information about the documents or leveraging GPT-4 to summarize the text could greatly enhance the chatbot's ability to provide insightful responses.

How to connect to SharePoint Online and SharePoint 2013 2016 2019 on premises with app principal

https://github.com/vgrem/Office365-REST-Python-Client/wiki/How-to-connect-to-SharePoint-Online-and-and-SharePoint-2013-2016-2019-on-premises--with-app-principal

Using Streamlit with Self-Signed SSL Certificate (Optional)

https://docs.streamlit.io/library/advanced-features/https-support

Generating a Self-Signed SSL Certificate for Localhost on MacOS

This guide provides steps to generate a self-signed SSL certificate specifically for https://localhost:8501 on MacOS, using OpenSSL.

Prerequisites

Ensure that OpenSSL is installed on your system. If not, you can install it using Homebrew.

Instructions

Step 1: Open Terminal

Open the Terminal application, which can be found in the Applications under Utilities, or searched for using Spotlight.

Step 2: Check for OpenSSL

Check if OpenSSL is installed by typing the following command in Terminal:

openssl version

If it's not installed, install it using Homebrew:

brew install openssl

Step 3: Create a Private Key

Generate a private key with:

openssl genrsa -out localhost.key 2048

Step 4: Create a Certificate Signing Request (CSR)

Generate a CSR using the private key:

openssl req -new -key localhost.key -out localhost.csr

When prompted for the "Common Name" (CN), enter localhost.

Step 5: Create a Configuration File

Create a file named localhost.conf with the following content:

[req]
distinguished_name = req_distinguished_name
req_extensions = v3_req

[req_distinguished_name]

[v3_req]
basicConstraints = CA:FALSE
keyUsage = nonRepudiation, digitalSignature, keyEncipherment
subjectAltName = @alt_names

[alt_names]
DNS.1 = localhost

Step 6: Generate the Self-Signed Certificate

Generate the certificate using your CSR, private key, and the configuration file:

openssl x509 -req -days 365 -in localhost.csr -signkey localhost.key -extfile localhost.conf -extensions v3_req -out localhost.crt

Step 7: Verify the Certificate

Check your certificate:

openssl x509 -text -noout -in localhost.crt

Step 8: Using the Certificate

The localhost.key is your private key, and localhost.crt is your self-signed certificate. These can be used in your server configuration.

Note

This self-signed certificate is only suitable for local development and testing. For production environments, consider using a certificate from a trusted Certificate Authority.

BenderScript / ShareBotAI