GreenMine

This is the open-source repository for GreenMine, introduced in the paper titled: Visual Text Mining with Progressive Taxonomy Construction for Environmental Studies

Reproduction Guide

Since reproducing the system from scratch costs API tokens for querying GPT models, you can start directly from Step 3 with the pre-generated data.

Disclaimer: For privacy reasons, we can not release the original data as shown in the paper. Therefore, the data available here is synthetically generated by gpt-4o-mini, and are only provided to demonstrate the visual components of the system, rather than for practical usage.

1. Generating Synthetic Data (Optional)

2. Preprocessing (Optional)

3. Launching Back-end

1. Request for an OpenAI's api key, and then create a file called openai_api_key (no file extension needed) under the directory server/. Copy and paste your api key into the file.
2. Under the directory server/, execute the following command:

pip install -r requirements.txt
flask run

3. The backend should be launched at localhost:5000. You can check by accessing localhost:5000/test/ in your browser, you should see Hello GreenMine showing on the web page.

4. Launching Front-end

1. Install Node.js
2. Under the repository root directory, execute the following command:

npm i
npm run dev

3. The front-end should be launched at localhost:5173. You can access it in your browser.