Interview Assistant Backend

This project is a FastAPI backend that uses a fine-tuned Hugging Face transformer model to generate interview-style questions and answers from user input.
It applies text preprocessing (stopword removal, keyword extraction) before sending the input to the model to improve generation quality.

---

Features

• REST API built with FastAPI
• Text preprocessing (stopword removal, keyword filtering)
• Integration with a Hugging Face text-generation model
• Extracts structured JSON (question & answer) from model output

---

Project Structure

project-backend/
├── src/
│   └── main.py         # FastAPI application
├── .venv/              # Virtual environment
├── requirements.txt    # Python dependencies
└── README.md           # Documentation

---

Installation

1. Clone the repository:

   git clone https://github.com/your-username/interview-assistant-backend.git
   cd interview-assistant-backend

2. Create and activate a virtual environment:

   python -m venv .venv
   .venv\Scripts\activate   # On Windows
   source .venv/bin/activate # On Linux/Mac

3. Install dependencies:

   pip install -e .

Running the Server

Start the FastAPI app with Uvicorn:

cd src
uvicorn main:app

The server will start at: http://127.0.0.1:8000

---

API Usage

• Endpoint: /generate (POST)

Request Body:

{
  "text": "Can you tell me about React static decorators?"
}

Preprocessing applied:

react static decorators

Example Response:

{
  "question": "What is React Static Decorators?",
  "answer": "Decorators that change the behavior of React components but render the same after the render() call, ensuring React interoperability."
}

If parsing fails:

{
  "error": "Could not parse the generated text."
}

---

Requirements

• Python 3.9+
• FastAPI
• Uvicorn
• Transformers (Hugging Face)
• NLTK (for preprocessing)

Install everything with:

pip install -e .

---