Apppublic

Gungz/quiz-generator-baidu-ernie

sourceHugging Facemitupdated 9mo agoView on Hugging Face

0likes

App README

🎓 Baidu ERNIE AI Quiz Generator

An intelligent, AI-powered tool that automatically generates multiple-choice quizzes from text content. Built with Gradio and powered by Baidu's ERNIE-5.0-Thinking-Preview model (accessed via OpenAI-compatible API).

✨ Features

—📄 Multi-Format Input: Extract text from PDF, TXT files, or live URLs.
—🧠 AI-Powered Generation: Uses Baidu's advanced ERNIE 5.0 model to generate context-aware questions.
—🎯 Adjustable Difficulty: Select between Easy, Medium, and Hard levels.
—📊 Customizable Length: Generate between 1 to 20 questions per session.
—💾 Export Options: Download generated quizzes as CSV (for Excel/LMS) or TXT (for printing).
—🎨 Interactive UI: Built with Gradio for a smooth, real-time user experience with progress tracking.

🛠️ Prerequisites

Before running the application, ensure you have Python 3.8+ installed.

Required API Key

This project uses the Baidu AI Studio API. You must have a Baidu account and create an API key:

1.Go to Baidu AI Studio.
2.Navigate to the API section to get your Access Token.

📦 Installation

1.Clone the repository (or download app.py):

bash

   git clone <your-repo-url>
   cd quiz-generator-baidu-ernie

1.Install dependencies:

bash

   pip install gradio openai pypdf pandas requests beautifulsoup4

1.Set Environment Variable: You must set your Baidu Access Token as an environment variable.

Linux/Mac:

bash

   export BAIDU_ACCESS_TOKEN="your_baidu_access_token_here"

Windows (CMD):

cmd

   set BAIDU_ACCESS_TOKEN="your_baidu_access_token_here"

Windows (PowerShell):

powershell

   $env:BAIDU_ACCESS_TOKEN="your_baidu_access_token_here"

🚀 Usage

Run the application using the following command:

bash

python app.py

Once the server starts, open the local URL provided in the terminal (usually http://127.0.0.1:7860).

How to Use the Interface

1.Select Source Type: Choose between "File Upload" (PDF/TXT) or "URL".
2.Upload/Paste:
3.If using File Upload: Drag and drop your document.
4.If using URL: Paste the website link.
5.Configure Quiz:
6.Number of Questions: Use the slider (1-20).
7.Difficulty: Select Easy, Medium, or Hard.
8.Generate: Click "✨ Generate Quiz".
9.Review & Download:
10.View the quiz in the Markdown preview pane.
11.Click "📥 CSV" or "📥 Text" to download the results.

& Configuration

You can modify the following constants at the top of app.py:

Constant	Description	Default
`BASE_URL`	The API endpoint for Baidu AI Studio	`https://aistudio.baidu.com/llm/lmapi/v3`
`MODEL_NAME`	The specific model version to use	`ernie-5.0-thinking-preview`

🧠 Technical Details

—LLM Integration: The code uses the OpenAI Python client library configured to point to Baidu's API endpoint, making it compatible with standard OpenAI SDK patterns while using Baidu's infrastructure.
—Text Extraction:
—PDFs: Uses PyPDF2.
—Web Scraping: Uses Requests and BeautifulSoup to strip navigation/scripts for clean text.
—Output Structure: The LLM is prompted to return strict JSON formatted as:

json

  [
    {
      "question": "...",
      "options": ["A) ...", "B) ...", "C) ...", "D) ..."],
      "answer": "A",
      "explanation": "..."
    }
  ]

📁 Project Structure

.
├── app.py              # Main application logic and Gradio UI
├── requirements.txt    # Python dependencies
└── README.md           # This file

⚠️ Troubleshooting

—Error reading PDF: Ensure the PDF is text-based (not a scanned image) and not encrypted.
—Authentication Error: Check if BAIDU_ACCESS_TOKEN is set correctly and has not expired.
—Empty Text from URL: Some websites block scrapers. The tool uses a standard User-Agent, but complex JavaScript-heavy sites may not render fully.

📝 License

This project is open-source and available for educational purposes.