🌸 Florr Powerful Tools

🤖 AI-Powered Automation Toolkit for florr.io — Smart AFK protection, auto-pathfinding, combat AI, and custom model training

---

🎯 What is This?

Florr Powerful Tools is a comprehensive automation framework for the popular florr.io browser game. It combines computer vision, deep learning, and path-planning algorithms to provide intelligent game assistance — without cheating, just making the game more enjoyable.

Whether you need AFK protection, want to auto-navigate maps, train custom AI models, or collect training data — this toolkit has you covered.

---

✨ Key Features

Feature	Description	Status
🛡️ AFK Protection	Auto-detect & solve AFK verification challenges	✅ Stable
🗺️ Auto Pathfinding	Lazy Theta* algorithm for intelligent navigation	✅ Stable
⚔️ Auto Combat	Smart mob detection, targeting & fighting	✅ Stable
🧠 AI Training	Build custom models with PyTorch + YOLO26	✅ Stable
📊 Data Collection	Harvest gameplay data for model training	✅ Stable
📸 Visual Testing	Browser screenshot + functional test harness	✅ New

---

🚀 Quick Start

Prerequisites

• Python 3.9–3.11
• Windows 10/11, macOS, or Linux
• (Optional) GPU for training

Installation

# Clone the repo
git clone https://github.com/PANP2010/florr_powerful_tools.git
cd florr_powerful_tools

# Create virtual environment
python -m venv venv
source venv/bin/activate        # Linux/macOS
# venv\Scripts\activate.bat     # Windows

# Install dependencies
pip install -r florr_assistant/requirements.txt

# Install Playwright for browser testing (optional)
pip install playwright
playwright install chromium

Run the Assistant

cd florr_assistant
python main.py

Run Tests

# Full test suite (screenshot + module tests)
bash browser_test_env/run_all_tests.sh

# Module tests only
python browser_test_env/module_tester.py

# Screenshot tests only
python browser_test_env/screenshot_tester.py --florr

---

🏗️ Project Architecture

florr_powerful_tools/
├── florr_assistant/             # 🧠 Main integrated assistant
│   ├── core/                    # Core services (engine, logger, config, platform)
│   ├── ui/                      # PyQt5 UI (main window, overlay, data collection)
│   └── modules/                 # Feature modules
│       ├── afk/                 # AFK detection & response
│       ├── pathing/             # Navigator + map classifier
│       ├── combat/              # Fighter + target selector
│       ├── data_collector/       # Training data collection
│       └── stats/               # Statistics gathering
│
├── browser_test_env/            # 🆕 Browser test harness (Playwright)
│   ├── screenshot_tester.py    # Headless screenshot capture
│   ├── module_tester.py         # Import & functionality tests
│   └── run_all_tests.sh         # One-command test runner
│
├── florr-auto-afk-main/         # Standalone AFK detection
├── florr-auto-pathing/          # Standalone pathfinding
├── florr-auto-framework-pytorch/ # AI training framework
├── florr-auto-sszone/           # Auto mob zone farming
│
├── upload_package/              # 📦 Kaggle training package
├── train_package/               # 🏋️ Local training package
└── florr_knowledge_base/        # 📚 Game wiki & data

---

🛡️ AFK Protection

Never get kicked for being idle again. The AFK module:

• Uses YOLO models (afk-det.pt, afk-seg.pt) to detect AFK windows (>90% accuracy)
• Plans a path through verification using an improved Dijkstra algorithm
• Responds to LLM-powered chat challenges
• Integrates via WebSocket browser extension (Tampermonkey)

# Quick test
cd florr-auto-afk-main
pip install -r py311-requirements.txt
python segment.py

---

🗺️ Auto Pathfinding

Built on the Lazy Theta* path-planning algorithm — smarter than A* because it accounts for line-of-sight:

• ✅ 8-map support: Ocean, Desert, Jungle, Garden, Hel, Anthell, Sewers, and more
• ✅ Real-time player position detection via HSV color filtering
• ✅ Stuck detection & recovery with random directional nudges
• ✅ Configurable target maps and navigation intervals

# Quick test
cd florr-auto-pathing
python main.py

---

⚔️ Auto Combat

Fight smarter, not harder:

• YOLO mob detection across 77+ mob types
• Priority-based targeting with rarity weighting (Mythic 3.7×, Ultra 1.8×, Legendary 0.6×)
• Auto-equip switching and petal management
• Sandstorm detection & avoidance

# Quick test
cd florr-assistant
python -c "from modules.combat import Fighter; print('Combat module loaded')"

---

🧠 AI Training Framework

Train your own custom models:

Model	Architecture	Input	Output
Base	3-layer MLP	73-dim state	5-dim action
Attention	Multi-head attention	73-dim state	5-dim action

Training stack: PyTorch + Ultralytics YOLO26 + NumPy

# Kaggle / Google Colab (free GPU)
!python scripts/train_cloud.py --mode all --num-samples 5000 --epochs 100 --batch 16

# Local (needs GPU)
python scripts/train_cloud.py --mode all --num-samples 10000 --epochs 150 --batch 32

YOLO26 vs YOLOv8:

Metric	YOLOv8s	YOLO26s	Improvement
mAP50-95	44.8%	48.6%	+3.8%
Params	11.2M	9.5M	-15%
CPU speed	baseline	+43%	faster

---

🧪 Testing

The project includes a browser test harness powered by Playwright:

# Run everything
bash browser_test_env/run_all_tests.sh

# Just module import tests
python browser_test_env/module_tester.py

# Screenshot tests
python browser_test_env/screenshot_tester.py --florr
python browser_test_env/screenshot_tester.py --url https://github.com

Test coverage:

Category	Status
Core modules (logger, config, engine, events)	✅ Pass
Feature modules (AFK, pathing, combat, data)	✅ Pass
OpenCV (cv2)	✅ Pass
PyTorch	⚠️ Optional
YOLO (ultralytics)	⚠️ Optional
GUI (PyQt5)	⚠️ Needs display

---

⚙️ Configuration

Main config: config/default.yaml

modules:
  afk:
    enabled: true
    idle_threshold: 10
    detection_interval: 3

  pathing:
    enabled: true
    target_map: "ocean"
    avoid_danger: true

  combat:
    enabled: false
    safe_distance: 200
    attack_distance: 150

ui:
  theme: "dark"   # light / dark
  language: "zh-cn"  # en-us / zh-cn

---

🔧 Development

Add a new module

from florr_assistant.modules.base import BaseModule

class MyModule(BaseModule):
    name = "my_module"
    version = "1.0.0"
    description = "My awesome feature"
    priority = 50  # Lower = runs later
    dependencies = []  # e.g. ["map_classifier"]

    def _on_start(self):
        self._log("My module started")

    def _on_tick(self):
        pass  # Called every loop iteration

    def _on_stop(self):
        self._log("My module stopped")

---

📜 Tech Stack

Category	Technologies
Language	Python 3.9+
Computer Vision	OpenCV, PIL, NumPy
Deep Learning	PyTorch, Ultralytics (YOLO26)
GUI	PyQt5, Tkinter
Automation	PyAutoGUI, PyWin32, mss
Web	FastAPI, WebSocket
Testing	Playwright, pytest

---

🤝 Contributing

Contributions are welcome! Please:

1. Fork the repo
2. Create a feature branch (git checkout -b feature/AmazingFeature)
3. Commit your changes (git commit -m 'Add AmazingFeature')
4. Push to the branch (git push origin feature/AmazingFeature)
5. Open a Pull Request

---

📄 License

GPL v3 — see LICENSE for details.

---

🙏 Acknowledgements

• Shiny-Ladybug — original florr-auto-afk project
• Ultralytics — YOLO framework
• The florr.io community 💐

---

Made with ❤️ for the florr.io community