Taira is a conversational AI chatbot capable of reading and interacting with various types of documents such as PDF, Word, Excel files, and webpages. Powered by a Retrieval-Augmented Generation (RAG) model, it provides answers based on document contents using in-context learning techniques, including few-shot learning and chain-of-thought reasoning.
-
Supports multiple document types:
- Word (DOCX)
- Excel (XLSX)
- Webpages (via URLs)
-
Chat history: Taira remembers the conversation history and provides context-aware responses.
-
Vector Database: Creates a vectorized database from document contents for efficient document retrieval and question answering.
-
In-context learning: Uses few-shot learning and chain-of-thought reasoning to answer questions based on the document contents.
-
Clone the repository:
git clone https://github.com/your-repo/taira.git cd taira -
Install the dependencies:
pip install -r requirements.txt
-
Set up environment variables:
- Create a .env file and add the OPENAI_API_KEY key with your OpenAI API key:
OPENAI_API_KEY=your_openai_api_key
- Run the application:
python src\main.py
- Loading Documents: Taira can process the following document types:
- PDF: Loaded using the PyPDFLoader.
- Word (DOCX): Loaded using Docx2txtLoader.
- Excel: Loaded with UnstructuredExcelLoader.
- Webpages: URL input is processed with WebBaseLoader.
- Interaction:
- After loading a document or webpage, you can interact with Taira by asking questions.
- Taira retrieves relevant information from the loaded documents using a vector database.
- It answers questions based on the document content or chat history.
After loading a document, you can ask questions like:
- User: "What are the key points in the report?"
- Taira: "The report outlines the major findings, focusing on market trends, consumer behavior, and product preferences."
For any questions or clarifications, please contact Raza Mehar at [raza.mehar@gmail.com].