PengoSword123/emotionSteps

This is a FastAPI service that detects emotion in text and answers with a Retrieval-Augmented Generation (RAG) pipeline.

Setup (Windows PowerShell)

1. 1.Create and activate a virtual environment:

python -m venv .venv
. .venv\Scripts\Activate.ps1

1. 1.Install dependencies:

pip install -r requirements.txt

1. 1.(Optional) Set your Hugging Face token to use hosted models:

$env:HUGGINGFACEHUB_API_TOKEN = "hf_...your_token..."

1. 1.Run the API:

python app.py

1. 1.Test endpoints:
2. 2.Health: GET http://localhost:8000/health
3. 3.Process: POST http://localhost:8000/process with JSON body:

{ "message_text": "I am frustrated with my laptop." }

The service will return the detected emotion, an answer, and proof links or snippets from the knowledge base.