EDGE_AI

A reproducible workspace for EDGE_AI analysis, Binance data extraction, and the migration of the EMA 20/50 strategy to Jesse.

This research workspace includes:

• Pre-downloaded historical data.
• Backtest artifacts.
• Helper scripts to reproduce the strategy.
• Migration and diagnostic documentation.

---

What stays in Git

Files and directories that are part of the public repository:

• README.md
• README_MIGRATION.md
• scripts/
• data/binance/BTCUSDT_1h.csv
• outputs/backtest/
• .gitignore

Note: The content within outputs/backtest/ is intentionally versioned to document the specific results obtained during research.

What stays out of Git

Items that should not be published:

• .venv/
• .env and .env.* (except for examples)
• nofx/
• jesse/
• __pycache__/, *.pyc, *.pyo, *.pyd
• .DS_Store, Thumbs.db
• Temporary files and logs

These paths are already managed by the .gitignore file.

---

Structure

EDGE_IA/
├── README.md
├── README_MIGRATION.md
├── .gitignore
├── data/
│   └── binance/
│       └── BTCUSDT_1h.csv
├── outputs/
│   └── backtest/
│       ├── README.md
│       ├── equity_curve.csv
│       ├── summary.json
│       └── trades.csv
└── scripts/
    ├── bootstrap_jesse.sh
    ├── prepare_backtest.sh
    └── run_ema_crossover_backtest.py

---

How to Reproduce

1. Local Workspace Backtest

To validate the dataset and run the auxiliary backtest, execute:

cd /Users/User/Desktop/NFX_IA
bash scripts/prepare_backtest.sh

2. Separate Jesse Environment

Jesse has been moved to an exclusive folder outside of this repository:

• /Users/User/Desktop/jesse

If you want to use the Jesse dashboard, run the bootstrap from its specific directory:

cd /Users/User/Desktop/jesse
bash scripts/bootstrap_jesse.sh

---

Publication Notes

Before pushing this repository to GitHub:

1. Ensure that EDGE_IA/ and jesse/ internal system files are not added to the commit.
2. Keep data/binance/BTCUSDT_1h.csv only if you intend to publish the reproducible dataset.
3. Keep outputs/backtest/ only if you want to share the results and the "paper trail" of the experiment.
4. For a lighter repository, remove data/ and outputs/ and update this documentation accordingly.

---

Project Results

The baseline backtest followed a straightforward strategy:

• Asset: BTCUSDT
• Timeframe: 1h
• Entry: EMA 20 crossing above EMA 50.
• Exit: EMA 20 crossing below EMA 50.
• Shorting: Disabled.
• Stop Loss / Take Profit: None.
• Fees: Configurable.

Full migration details can be found in README_MIGRATION.md.