Five Tree Wood AI

AI-powered aircon temperature prediction for Home Assistant.

About

Five Tree Wood AI is a machine learning model for predicting indoor temperature using environmental sensor data. It can be used as a standalone application or as a Home Assistant add-on.

Home Assistant Add-on Installation

Method 1: Add Repository (Recommended)

1. In Home Assistant, go to Supervisor → Add-on Store
2. Click the ⋮ menu → Repositories
3. Add this repository URL: https://github.com/Swamp-Ig/five-tree-wood-ai
4. Find "Five Tree Wood AI" in the add-on store and click Install
5. Configure the add-on (see Configuration section)
6. Start the add-on

Method 2: Manual Installation

1. Clone this repository to your Home Assistant addons folder
2. Restart Home Assistant Supervisor
3. The add-on will appear in your local add-ons list

Standalone Installation

1. Configuration

Copy the secrets template and configure your InfluxDB connection:

cp secrets.ini.template secrets.ini
# Edit secrets.ini with your actual InfluxDB credentials

2. Install Dependencies

Install the package in development mode with all dependencies:

# Install package with production dependencies
pip install -e .

# Or install with development dependencies
pip install -e .[dev]

Alternatively, you can still use requirements.txt:

pip install -r requirements.txt

3. Setup Code Formatting (Optional)

# Install git hooks for automatic formatting
python scripts/install_hooks.py

# Or manually format code anytime
python scripts/format_code.py

Package Structure

src/
└── five_tree_wood/
    └── ai/
        └── aircon/
            ├── train.py        # Model training functionality
            ├── predict.py      # Model prediction functionality
            └── __init__.py     # Package-level exports
scripts/
├── format_code.py      # Code formatting script
├── check_tools.py      # Tool availability checker
├── install_hooks.py    # Git hooks installer
└── pre_commit_hook.py  # Pre-commit hook implementation

Usage

Package-level imports

from five_tree_wood.ai.aircon import train_model, predict_temperature

# Train a model
model, features = train_model()

# Make a prediction
prediction = predict_temperature(
    time="2025-09-08T14:30:00",
    inside_temperature=22.5,
    outside_temperature=15.0,
    roof_temperature=18.0
)

Usage

Training the Model

# Using the installed script
aircon-ai train

# Or directly with Python
python src/main.py train

Making Predictions

# Using the installed script
aircon-ai predict now 22.5 15.0 18.0

# Or directly with Python
python src/main.py predict now 22.5 15.0 18.0

Running the API Server

# Using the installed script
aircon-ai api

# Or directly with Python
python src/main.py api

Direct Package Usage

# Simple package-level imports
from five_tree_wood.ai.aircon import train_model, predict_temperature

# Train a model
model, features = train_model()

# Make a prediction
prediction = predict_temperature(
    time="2025-09-08T14:30:00",
    inside_temperature=22.5,
    outside_temperature=15.0,
    roof_temperature=18.0
)

API Endpoints

• POST /predict - Make temperature predictions
• GET /health - Health check
• GET /info - API information
• POST /train - Train the model

Configuration

The application uses secrets.ini for sensitive configuration:

[influxdb]
url = http://your-influxdb-server:8086
token = your-influxdb-token
org = your-organization
bucket = your-bucket-name

Important: Never commit secrets.ini to version control. Use secrets.ini.template as a reference.

Development

Installation for Development

# Install in editable mode with dev dependencies
pip install -e .[dev]

# Or install dev tools separately
pip install black isort pylint pytest

Running Tests

# Using pytest directly
python -m pytest tests/ -v

# Or with setuptools
python -m pytest

Code Formatting

This project uses automated PEP8 code formatting with the following tools:

• Black: Code formatter for consistent style
• isort: Import statement organizer
• pylint: Style guide enforcement

Manual Formatting

# Format all code
black src tests

# Check with pylint
pylint src tests

License

MIT License - see the LICENSE file for details.

Format all code using our script

python scripts/format_code.py

Or use tools directly

python -m black src/ python -m isort src/ python -m pylint src/

#### Automatic Formatting

- **Git hooks**: Install with `python scripts/install_hooks.py`
- **VS Code**: Automatic formatting on save (configured in `.vscode/settings.json`)

#### Configuration

Tool configuration is centralized in modern Python standards:

- **pyproject.toml**: Black, isort, and pytest configuration
  # (flake8 configuration removed; pylint is now used)

## Node-RED Integration

Import the flow from `node-red-aircon-flow.json` to integrate with Node-RED.