MultiCoT

Overview

MultiCoT is a repository that demonstrates Chain-of-Table reasoning on multiple tables powered by LangGraph.

Try the hosted demo: https://cyqiq-cot-demo-c27oq5sqda-uc.a.run.app/

Installation

Follow these steps to set up MultiCoT:

Clone the Repository

git clone https://github.com/CYQIQ/MultiCoT.git

Install Dependencies Navigate to the repository directory and run:
```
pip install -r requirements.txt
```

Configure API Keys Create a .env file in the root directory. Add your OpenAI API key and LangChain API details as follows:

OPENAI_API_KEY="..."
LANGCHAIN_API_KEY="..."
LANGCHAIN_TRACING_V2="..."
LANGCHAIN_ENDPOINT="..."
LANGCHAIN_PROJECT="..."

Usage

For detailed instructions on how to run MultiCoT, refer to the CYQIQ_COT.ipynb Jupyter Notebook within the repository.

Data

The data directory includes an example dataset comprised of 6 CSV files necessary to run the notebook. To run MultiCoT with your own data, change the data loading logic and the table descriptions in CYQIQ_COT.ipynb

Help

Feel free to email us at hello@cyqiq.ai

Acknowledgments

Big thanks to @hwchase17 and team LangChain for their amazing guides.
Big thanks to @zaturalma2 for helping put the notebook together.