Hassle-free questions generation for NUS educators
Team Bloopers recognises the busy schedules of university educators. With Intelli-exam, we attempt to employ our proprietary RAG Pipeline, integrated with h2oGPTe's API to run their LLM model for generation of exam questions.
Question types include MCQs or open-ended, with educators having the flexibility to execute two query types:
-
Sample Questions: Useful in generating consistent exam question format throughout every semesters, without repeating questions
-
Additional Context: Useful in generating questions that rely on context, especially for courses that are continuously advancing in landscape
Our vectorised database stores syllabus content from NUS courses. With our RAG pipeline, we minimise the chances of the LLM hallucinating with inaccurate information, while at the same time ensuring questions generated stay within the syllabus of the courses.
- Intelli-Exam/
├── client/
│ ├── src
│ │ └── ...
│ ├── package.json
│ └── package-lock.json
├── server/
│ ├── folder_manager
│ │ └── temp_folder_manager.py
│ ├── models
│ │ └── qna_generation_model.py
│ ├── pdf_reader
│ │ └── pdf_reader.py
│ ├── rag
│ │ ├── database_setup.py
│ │ └── rag_retrieval.py
│ ├── requirements.txt
│ └── server.py
├── .gitignore
└── README.md-
React: Frontend library for building interactive UI components
-
Flask: Provides a RESTful API backend, allowing communication between the frontend and backend.
-
MongoDB Atlas: A cloud-based NoSQL database, for vector search capabilities.
-
H2OGPTE: H2OGPTE, developed by H2O.ai, provides the large language model used in this project via API calls.
-
Ensure you have Python and Node.js installed on your system:
- Python: You can download Python from the official website. Make sure to add Python to your system's PATH during installation.
- Node.js: You can download Node.js from the official website. Follow the installation instructions for your operating system.
-
Clone the repository:
git clone https://github.com/AY2324S2-DSA4213-TeamBloopers/Intelli-Exam.git cd Intelli-Exam -
Install React dependencies:
cd client npm install -
Install Python dependencies in a virtual environment:
For MacOS/Linux:
cd ../server python3 -m venv venv source venv/bin/activate pip install -r requirements.txt
For Windows
cd ../server python3 -m venv venv venv/Scripts/activate pip install -r requirements.txt -
In the same terminal, run the server:
python3 server.py
-
Without closing the current terminal, create a new terminal and start the react instance:
cd client npm start
The website is then run at: http://localhost:3000/
Intelli-Exam should already have all the required content within the syllabus uploaded in its database. If you can find the NUS module from the "NUS Module" drop down selector, you can generate questions from the module.
Under "Type Of Upload", select the type of documents that are going to be used. It can either be sample questions or context. For sample questions, the questions generated will attempt to follow the style of questions in the file. For context, the questions generated will attempt to form questions using the complimentary information provided whilst still focusing on the syllabus.
Upload the files by presing "Browse files". Take note that the Intelli-Exam currently accepts pdf files only.
You can select the number of MCQ and open-ended questions you want to generate, between 0 to 20.
Finally, press submit and wait for the excel file to download on your browser!
-
Due to copyright issues, we currently only support content and question generation for DSA4213. Addition course codes in the dropdown are meant to be placeholders, selecting them for generation will execute for DSA4213.
-
Due to our batching strategy when generating question and answers pair, do expect to wait for a few minutes generating beyond tens of questions.