This is a streamlit build Chatbot UI which uses backend services created within SnapLogic, the first platform for generative integration.

Table of Contents

Using the chatbot
    Chat with one data domain
    Chat with multiple data domains
Managing Domain Data
    Overview of available namespaces
    Upload data into a new namespace
    Add data to an existing namespace
    Delete a namespace

Using the chatbot

Chat with one data domain

Choose a data domain you want to ask your questions against. Enter your query in the chat box. For example, if you're using a HR service chatbot, you might ask, "When do I got paid?".

Chat with multiple data domains

You also can choose more than one data domain to chat with. Switching on "use of multiple data domains" will let you pick multiple data domains.

When you select more than one data domain, you can ask questions that will highlight differences between them. For instance, if you choose the Lyft Annual Report and the Uber Annual Report as your data domains, you might ask, 'How do the service offerings in both reports compare?' or 'What is the difference in revenue between the two companies in 2022?'

Managing Domain Data

As this is a Retrieval-Augmented Generation (RAG) chatbot the data is stored in Pinecone Vector Database. With this chatbot you can manage the data and the corresponding namespaces. However, you need to authenticate first befor you can manage the domain data.

Overview of available namespace

In the Overview section all available namespaces are listed with the corresponding vector count.

Upload data into a new namespace

You can upload data into a new namespace by providing a name (for the namespace) followed by the selection of PDF file.

Add data to an existing namespace

You also can add new data to an existing namespace by selecting the appropiate namespace follwed by the selection of PDF file.

Delete a namespace

Existing namespaces can be deleted after confirmation. This process removes all vectors irreversible.