Am I The Asshole

Authors: Melissa Tro (Student ID: 571267), Jeff Nkatiah Edjekoomhene (Student ID: 582338)

This is a Streamlit app that uses the LangChain library to analyze and provide judgments on stories similar to those on the subreddit r/amitheasshole. The app evaluates the behavior of individuals involved in a given story and provides unbiased judgments and advice.

Installation

1. Clone the repository.
2. Install the required dependencies using pip install -r requirements.txt.

Usage

1. Set up your Pinecone environment by following the instructions provided in the Pinecone console.
2. Fill in the required secrets in secrets.toml.
3. Run the Streamlit app using streamlit run main.py.
4. Enter your OpenAI API key in the sidebar.
5. Describe the situation in the text area.
6. Click the "Submit" button to generate a verdict.

Deployed App

The app is deployed and can be accessed at https://chatbot-usw-ss23.streamlit.app/.

Components

LangChain

The LangChain library provides the following modules:

• llms: OpenAI language models integration.
• prompts: Prompt templates for chat conversations.
• chains: Language model chains for chat conversations.
• memory: Conversation memory management.
• embeddings.openai: OpenAI embeddings integration.
• vectorstores: Pinecone vector store integration.

Pinecone

Pinecone is used for semantic similarity search in the application. It provides the following features:

• Initialization of Pinecone with the API key and environment.
• Creation of a vector store using Pinecone.
• Similarity search and retrieval of relevant information.