Atlantico Raspberry Pi port

This folder contains a Raspberry Pi (Zero 2 W) port of the Atlantico device code. The goal is to replicate the high-level behavior of the ESP32 firmware in Python so you can run development and testing on a Raspberry Pi.

Structure

• atlantico_rpi/ - the Python module with the core code:
  • mqtt_client.py - MQTT client wrapper (paho-mqtt)
  • model_util.py - model utilities (save/load, stubs for training/transform)
  • events.py - a tiny EventQueue helper for signaling heavy work to the main thread
  • device.py - high-level setup() / loop() placeholders
  • config.py - MQTT topics and file paths
• tests/ - unit and integration tests (pytest)
• device.json - example device definition used to derive the MQTT client id

Quick setup

1. Create and activate a Python virtual environment (recommended):

python -m venv .venv
source .venv/bin/activate

2. Install dependencies:

pip install -r requirements.txt

If you don't have a requirements.txt yet, at minimum install:

pip install pytest paho-mqtt

Tests

Run unit tests with:

pytest -q tests

Development environment (recommended)

On modern systems the project runs best with Python 3.11. If your default python is newer (for example Python 3.14) you can create a venv using a system Python 3.11 or use pyenv to install one:

# using system python3.11
python3.11 -m venv .venv
source .venv/bin/activate
python -m pip install --upgrade pip setuptools wheel
pip install -r requirements.txt

The integration test (tests/test_mqtt_integration.py) will attempt to connect to the broker configured in atlantico_rpi/config.py. If the broker is not reachable the test will be skipped.

device.json

The MQTT client will attempt to load device.json from the project root to derive the client_id (looks for keys client_id, client_name, device_id, name). A sample device.json is included.

MQTT behavior

• Topics mirror the ESP32 implementation and are defined in config.py.
• Metrics are sent/received as JSON. Models are transferred as binary payloads.
• The MQTT client runs paho's network loop in a background thread and dispatches incoming messages to registered callbacks. Heavy work should be signaled via the EventQueue rather than done directly in the callback.

Next steps

1. Implement ModelUtil.transform_data_to_model and training/prediction using TensorFlow.
2. Implement the device setup() and loop() to orchestrate training and networking.
3. Wire MQTT topic handlers to the EventQueue and implement the main thread worker.

If you want, I can: add a requirements.txt pinning paho-mqtt>=1.6, implement the topic handlers now, or convert the MQTT client to use the v2 callback API explicitly.