DraftNexus AI 🛡️⚔️

DraftNexus-AI is a comprehensive tool for Mobile Legends: Bang Bang (MLBB) designed to assist in drafting, match logging, and analytics. It leverages machine learning to provide real-time hero recommendations based on team composition and enemy picks.

✨ Features

1. 🔮 Draft Recommender (Real-time)

• Role-Based Input: Dedicated inputs for Allied roles (Exp, Jungle, Mid, Roam, Gold).
• Flex Enemy Input: 5 flexible slots for Enemy picks.
• 🚫 Ban Support: 10 slots to exclude banned heroes from recommendations.
• Smart Suggestions:
  • Best Pick per Role: Displays the top recommended hero for each role.
  • Alternative Recommendations: Automatically suggests alternatives if the best hero is already picked (marked as (Alt)).
  • Real Data Restriction: Toggle to suggest only heroes present in your match_logs_real.csv history.

2. 📝 Match Logger

• Log match details: Teams, Winner/Loser, Game Duration, and metadata.
• Auto-Save to CSV (data/match_logs_real.csv) for future model training.
• Match History view with icons and details.

3. 📊 Analytics & Training

• Seed Data: seed_heroes.js to fetch latest hero stats and icons from API.
• Model Training: train_model.py to retrain the RandomForest model using your custom match logs.
• Visualization: Generate meta maps (visualize_analytics.py), power curves, and difficulty charts.

🚀 Getting Started

Prerequisites

• Python 3.10+
• Node.js (optional, for fetching fresh API data)

Installation

1. Clone the repository:

   git clone https://github.com/RfadnjdExt/DraftNexus-AI.git
   cd DraftNexus-AI

2. Install dependencies:

   pip install -r requirements.txt

Usage

Run the Main App:

streamlit run scripts/data_entry_app.py

Retrain Model (after adding new logs):

# 1. Generate/Augment Training Data
python scripts/generate_training_data.py

# 2. Train Model
python scripts/train_model.py

🛠️ Project Structure

• scripts/: Application logic, training scripts, and utilities.
• data/: CSV datasets (Base stats, Match logs, Meta performance).
• analysis_plots/: Generated analytics plots.

---

Powered by Scikit-Learn, Streamlit, and Pandas.