Prompt Story Generator

Prompt Story Generator is an interactive storytelling system designed to generate and evaluate creative stories using advanced prompting techniques, including zero-shot, few-shot, and reflexion prompting. This system allows users to select a story title and genre or provide custom prompts, leveraging a fine-tuned GPT model to create tailored stories. It evaluates generated stories using robust metrics like ROUGE, BLEU, METEOR, and coherence, providing insights into the effectiveness of different prompting strategies.

Features

• Story Generation: Supports zero-shot, few-shot, and reflexion prompting techniques for diverse story outputs.
• Evaluation Metrics: Analyzes generated stories using ROUGE, BLEU, METEOR, and coherence metrics.
• Visualization: Graphically compares prompting strategies and their effectiveness.
• Interactive Interface: Simple web interface to select titles, genres, or provide custom prompts.

---

Getting Started

Prerequisites

1. Install Anaconda
2. Clone this repository:

   git clone https://github.com/srinivas-badiga/Story_Generation_and_Prompt_Evaluation.git

---

Installing Conda

To use this project, you will need Anaconda or Miniconda installed on your system. Follow these steps:

1. Download the installer:

   • Anaconda (full version with additional tools)
   • Miniconda (lightweight version)

2. Install Conda:

   • For Windows:
     • Double-click the downloaded installer and follow the on-screen instructions.
   • For macOS/Linux:
     • Open a terminal and run the following command (replace the filename with the actual downloaded file):

       bash Miniconda3-latest-Linux-x86_64.sh

     • Follow the prompts to complete the installation.

3. Verify the installation:

   conda --version

---

Setup Instructions

1. Navigate to the project directory:

   cd Story_Generation_and_Prompt_Evaluation

2. Create a virtual environment using conda:

   conda create -n test python=3.10
   conda activate test

3. Install pip and the required dependencies:

   conda install pip
   pip install -r requirements.txt

4. Download the necessary NLP model:

   python -m spacy download en_core_web_md

---

Running the Application

1. Run the Flask application:

   python app.py

2. Open your web browser and go to:

   http://127.0.0.1:5000
   

---

Project Workflow

1. User Input: Users can select a story title and genre or provide custom prompts through the web interface.
2. Story Generation: The backend processes the input and generates stories using the selected prompting technique.
3. Evaluation: The system evaluates the generated stories using metrics such as ROUGE, BLEU, METEOR, coherence and confidence scores to measure quality and relevance.
4. Visualization: Results are presented through comparative graphs and scores for user review.

Additionally, the system includes two distinct modules:

• Default Prompt Module: Generates stories using predefined prompts.
• Custom Prompt Module: Allows users to craft and generate stories based on their custom inputs.

---

Repository Structure

• app.py: Main Flask application.
• templates/: HTML templates for the web interface.
• static/: CSS and JavaScript files for styling and interactivity.
• data.csv: Dataset containing story titles and reference stories.
• requirements.txt: List of dependencies for the project.
• System Diagram.png: System Architecture.

---

Technologies Used

• Python 3.10
• Flask
• OpenAI GPT API
• spaCy
• Matplotlib

---

License

This project is open-source and available under the MIT License.

---

Acknowledgments

• Dataset: ROCStories-MOD
• LLM: Fine-tuned GPT model for storytelling.