Trivia.NET

A networked, multiplayer trivia game built in Python using raw TCP sockets, inspired by Kahoot. Multiple players connect to a central server and compete to answer trivia questions in real time. The server tracks scores, enforces per-question time limits, and broadcasts a live leaderboard between rounds.

This was built as a networking assignment for INFO1112 (Operating Systems and Network Platforms) at the University of Sydney.

---

Features

• Client-server architecture over TCP with a custom JSON message protocol
• Newline-delimited message framing for reliable delivery over a stream socket
• Concurrent player handling on the server using select() for I/O multiplexing
• Graceful handling of mid-game disconnections (disconnected players remain on the leaderboard)
• Three client modes: manual input, auto-solve, and AI-powered (via Ollama)
• Auto-solve mode achieves 100% accuracy across all question types
• Time-limited answer windows enforced using multiprocessing on the client

---

Question Types

The server supports four question types, configurable via a JSON config file:

• Mathematics - Evaluates arithmetic expressions using the Shunting Yard algorithm (infix to Reverse Polish Notation)
• Roman Numerals - Converts roman numerals to their decimal equivalent
• Usable IP Addresses of a Subnet - Calculates the number of usable host addresses in a given CIDR subnet
• Network and Broadcast Address of a Subnet - Determines the network and broadcast addresses of a CIDR subnet using bitwise operations

---

Project Structure

.
├── server.py       # TCP server: manages players, questions, scoring, and game flow
├── client.py       # TCP client: connects to server and answers questions
├── questions.py    # Question generation logic (randomised)

---

Requirements

• Python 3.10 or higher
• requests library (for AI mode only)

Install dependencies:

pip install requests

For AI mode, Ollama must be running locally with a model pulled (e.g. ollama pull mistral).

---

Configuration

Both the server and client are configured via JSON files.

Server config example (server_config.json)

{
  "port": 7777,
  "players": 2,
  "question_types": [
    "Mathematics",
    "Roman Numerals",
    "Usable IP Addresses of a Subnet",
    "Network and Broadcast Address of a Subnet"
  ],
  "question_formats": {
    "Mathematics": "What is {}?",
    "Roman Numerals": "What is the decimal value of the roman numeral {}?",
    "Usable IP Addresses of a Subnet": "How many usable addresses are there in the subnet {}?",
    "Network and Broadcast Address of a Subnet": "What are the network and broadcast addresses of the subnet {}?"
  },
  "question_seconds": 10,
  "question_interval_seconds": 5,
  "ready_info": "Game starts in {question_interval_seconds} seconds!",
  "question_word": "Question",
  "correct_answer": "Correct!",
  "incorrect_answer": "Wrong answer.",
  "points_noun_singular": "point",
  "points_noun_plural": "points",
  "final_standings_heading": "Final standings:",
  "one_winner": "{} wins!",
  "multiple_winners": "It's a tie between {}!"
}

Client config example (client_config.json)

{
  "username": "player1",
  "client_mode": "you"
}

For AI mode:

{
  "username": "robot",
  "client_mode": "ai",
  "ollama_config": {
    "ollama_host": "localhost",
    "ollama_port": 11434,
    "ollama_model": "mistral:latest"
  }
}

---

Running the Game

1. Start the server:

python server.py --config server_config.json

The server waits until the configured number of players have joined before starting.

2. Start a client (in a separate terminal for each player):

python client.py --config client_config.json

3. Connect to the server:

Once the client starts, type:

CONNECT localhost:7777

4. Play:

Once all players have connected, the server sends a ready message and the game begins. Each question appears with a time limit. Type your answer and press Enter.

---

Client Modes

Mode	Description
you	Answer questions manually via standard input
auto	Answers automatically using code (100% accuracy on all question types)
ai	Sends questions to a local LLM via the Ollama API and submits its response

---

Commands (during a session)

Command	Description
CONNECT <host>:<port>	Connect to a server
DISCONNECT	Disconnect from the current server
EXIT	Quit the client

---

How the Protocol Works

All messages between client and server are JSON objects encoded in UTF-8, terminated by a newline character (\n). The server sends: READY, QUESTION, RESULT, LEADERBOARD, and FINISHED. The client sends: HI, ANSWER, and BYE.

Example question message from the server:

{
  "message_type": "QUESTION",
  "question_type": "Mathematics",
  "trivia_question": "Question 1 (Mathematics):\nWhat is 3 + 7?",
  "short_question": "3 + 7",
  "time_limit": 10
}

Example answer from the client:

{
  "message_type": "ANSWER",
  "answer": "10"
}