TruthTrail lets you perform endless new true/false-quizzes from a markdown file you provide. From the markdown file, the program will generate an always new statement and ask you to answer whether it is true or false. The program will then tell you if you were correct or not.
If you need, you could ask for a detailed explanation of the statement.
The application selects sections from your provided markdown file randomly, weighted by the length of the sections. It also stores the times it has already selected a section, so that it will, over multiple sessions, select all sections equally often.
After selected a section, the application will use OpenAI's GPT 4 Turbo to generate a new statement from the section's text.
If you add a new section to the markdown file, the application will automatically include it in the selection process and set the times it has already selected it to 0. So in the next session, the new section will be selected with a higher probability than the other sections until the times it and other sections have been selected equal out.
The application counts your correct answers and provides you a grade that refreshes with every new statement. The grade is based on the German school grading system, and a pass with a 4.0 begins at 75% correct answers, due to being correct by just guessing has a 50% chance.
You can override the points that were given to you by the program, if you think the application was wrong.
You also can reset the points and start over.
The markdown file can be formatted as any usual markdown file, but with one exception according to the sections: Do prevent to use one subsection without a headline at the beginning of a section if this section also contains a subsection with a headline. The application will otherwise oversee the first subsection and not include it in the selection process.
✅ Do:
# Section 1
## Subsection 1.1
Text for Subsection 1.1
## Subsection 1.2
Text for Subsection 1.2❌ Don't:
# Section 1
Text for Section 1 <-- ❌ prevent this, because the application will oversee it
## Subsection 1.1
Text for Subsection 1.1
## Subsection 1.2
Text for Subsection 1.2The markdown file in the app is referenced to as source.md at the same level as the server file truthtrail.js. You can change the file path in the server file if you want to use another file. The file source.md, given with the source code, is a sample file with some sections and subsections about the field of Community Psychology.
In truthtrail.js, the prompts to instruct the AI to generate a new statement are defined. Currently, the prompts instruct the AI to behave like a professor of Community Psychology. You can change the prompt to instruct the AI to behave like a professor of another field.
TruthTrail has a server and a client part. The server part is a Node.js script and the client part is a simple HTML file with JavaScript.
To install the server part, you need to have Node.js installed. Then you can install the server part by setting your working directory to the directory where the truthtrail.js file is located and then running the following command in the terminal:
npm installTruthTrail uses the OpenAI API to generate new statements. You need to use a .env file to provide the API key. The .env file should be located in the same directory as the truthtrail.js file. The content of the .env file should look like this:
OPENAI_API_KEY="<your-api-key>"You can get an API key from OpenAI by signing up for their service.
The client part is a Progressive Web App (PWA). You can install it on your device by opening the client in your web browser and then clicking on the install button in the address bar.
To start the server part, you need to set your working directory to the directory where the truthtrail.js file is located. Then you can start the server by running the following command in the terminal:
node truthtrail.jsThe server will then start and listen on port 8080.
To start the client part, you need to start the http-server in the same directory. You can do this by running the following command in the terminal:
http-serverThe server will then start and listen on port 8081. You can then open the client by entering http://localhost:8081 in your web browser.
This project is licensed under the MIT License - see the LICENSE file for details.
Copyright 2024 Prof. Dr. Mike Hammes, Germany, psymeets.ai