ReAct Agents Exploration

This is a simple exploration of ReAct Agents offered by LangChain Framework.

API used

Please note that the following API requires the use of API key to work and are not free.

Langchain's ChatOpenAi to chat with OpenAI's GPT-3.5-Turbo model (https://api.python.langchain.com/en/latest/chat_models/langchain_openai.chat_models.base.ChatOpenAI.html).

Environment file to edit

Please create an .env file with the following parameters. PYTHONPATH is required to be filled to ensure successful folder imports in project.

OPENAI_API_KEY=<YOUR API KEY>
PYTHONPATH = <Absolute path to the directory where this project is cloned>

# Optional if you are not using LangSmith for tracking llm utilisation related metrics
LANGCHAIN_API_KEY = <YOUR API KEY>
LANGCHAIN_TRACING_V2 = true
LANGCHAIN_PROJECT = <NAME FOR YOUR PROJECT>

For more information on Langsmith, refer to https://www.langchain.com/langsmith

Installation and execution

Please use Anaconda distribution to install the necessary libraries with the following command

conda env create -f environment.yml

Upon installation and environment exectuion, run the following to see the inner workings of Agent involving AgentAction and AgentFinish

python main_agent_scratchpad.py

Programming languages/tools involved

Python
Langchain
- ChatOpenAI
- PromptTemplate
- CallBacks
- Agent Schemas: AgentFinish and AgentAction

Acknowledgement and Credits