WoW_fish_YOLO

WoW_fish_YOLO is an automated fishing bot for World of Warcraft utilizing a pre-trained YOLO model. This project is developed in Python with Ultralytics and has been trained on a dataset of 2000 screenshots from different regions of the game in the classic version.

Features

• 🎯 YOLO Object Detection: Real-time bobber detection with 99% accuracy
• 🔊 Audio Detection: Detects fish bite splash sounds
• 📊 Statistics Tracking: Session stats including success rate, fish/hour, detection rate
• 👀 Live Visualization: Real-time view of detections and bot status
• ⚙️ Configuration System: Save/load settings via JSON config file
• 🎨 Colored Logging: Clean, informative console output
• 🤖 Anti-Detection: Randomized delays and human-like mouse movement
• 🎣 Lure Support: Automatic lure application with timer

Requirements

• Python 3.12 or higher
• A good CPU (or a GPU with CUDA if you want to update the code to use CUDA)
• 1.8GB of disk space
• Working microphone for audio detection
• Linux system with wmctrl and xwininfo (for window management)

Installation

1. Clone the repository:

git clone <repository-url>
cd WoW_fish_YOLO

2. Install the required packages:

pip install -r requirements.txt

3. Install system dependencies (Linux):

sudo apt install wmctrl x11-utils

Configuration

First Run Setup

On first run, the bot will prompt you to configure:

• Fishing spell keybind
• Lure usage (optional)
• Lure keybind and duration

This configuration is saved to config.json for future use.

Manual Configuration

Edit config.json to customize settings:

{
    "fishing_key": "1",
    "lure_key": "2",
    "lure_duration_minutes": 10,
    "confidence_threshold": 0.5,
    "decibel_threshold": -45.0,
    "cast_duration_sec": 30,
    "viz_window_monitor": 2,
    "viz_window_width": 1200,
    "viz_window_height": 800,
    "viz_show_fps": true,
    "viz_show_audio_graph": true
}

Visualization Window Settings

• viz_window_monitor: Which monitor to display on (1=primary, 2=secondary, etc.)
• viz_window_width: Window width in pixels (default: 1200)
• viz_window_height: Window height in pixels (default: 800)
• viz_window_x: Override X position (null = auto-center on monitor)
• viz_window_y: Override Y position (null = auto-center on monitor)
• viz_show_fps: Show FPS counter (true/false)
• viz_show_audio_graph: Show real-time audio level graph (true/false)

Game Setup

1. Ensure World of Warcraft is open with window title "World of Warcraft"
2. Enable auto-loot in game settings
3. For best results:
   • Switch to first-person view
   • Hide UI (ALT + Z or ALT + W)

The setup should look like this:

Usage

Basic Usage

python bot.py

With Live Visualization

python bot.py --live-view

The live view window includes:

• Bobber Detection: Green bounding box with confidence score
• Crosshair: Red crosshair showing where the bot will click
• Status Display: Current bot state with color coding
• Audio Graph: Real-time audio level visualization with threshold line
• FPS Counter: Monitor performance
• Session Stats: Fish count, cast number, elapsed time

By default, the visualization window opens on your second monitor centered and in windowed mode (1200x800). You can customize this in config.json.

Debug Mode

python bot.py --debug

Custom Configuration

python bot.py --config my_config.json

All Options

python bot.py --help

Available options:

• --config, -c: Path to config file (default: config.json)
• --live-view: Enable live visualization window
• --debug: Enable debug mode with verbose logging
• --model: Path to YOLO model weights (overrides config)
• --no-config-prompt: Skip configuration prompts

Project Structure

WoW_fish_YOLO/
├── bot.py                  # Main bot application (refactored)
├── config.py               # Configuration management
├── logger.py               # Colored logging system
├── audio_detector.py       # Audio detection module
├── visualization.py        # Live view visualization
├── statistics.py           # Session statistics tracking
├── bot_states.py           # State management enums
├── monitor.py              # Window management (Linux)
├── config.json             # User configuration (auto-generated)
├── bot_old.py              # Original bot code (backup)
└── requirements.txt        # Python dependencies

How It Works

1. Detection: YOLO model detects fishing bobber in real-time screenshots
2. Audio Monitoring: Listens for splash sound when fish bites (>-45dB threshold)
3. Action: Automatically moves cursor and right-clicks to catch fish
4. Loop: Waits random delay, then recasts

Statistics

At the end of each session, you'll see statistics like:

==================================================
FISHING SESSION STATISTICS
==================================================
Session Duration:    1:23:45
Total Casts:         92
Fish Caught:         87
Failed Casts:        5
Success Rate:        94.6%

Bobber Detected:     91
Detection Failed:    1
Detection Rate:      98.9%

Avg Cast Time:       54.2s
Fish per Hour:       62.5
==================================================

Troubleshooting

Bot can't find game window

• Ensure WoW window title is exactly "World of Warcraft"
• Check that wmctrl is installed: wmctrl -l

Audio detection not working

• Check microphone is working: arecord -l
• Adjust decibel_threshold in config.json (try -40 or -50)

Bobber not detected

• Use --live-view to see what the bot sees
• Ensure first-person view and UI is hidden
• Try adjusting confidence_threshold in config

Model file not found

• Default path: ../runs/detect/train3/weights/best.pt
• Check archive/best.pt or specify with --model flag

Improvements & Future Work

• Training: Additional training with screenshots from different regions could enhance accuracy
• GPU Support: Add CUDA support for faster inference
• Cross-platform: Windows/Mac support for window management
• Advanced Anti-Detection: More sophisticated randomization patterns
• GUI: Optional graphical interface for configuration

Safety & Disclaimer

This bot is for educational purposes. Use at your own risk. Automated gameplay may violate game terms of service.