AI Smart Home Simulator

Minimal TCP client/server simulator built with asyncio.

At this stage, the project supports basic TCP communication between client and server. The server listens on port 9999, and settings can be changed in config.toml.

---

Requirements

• Python 3.11+
• Linux or Windows
• Docker (optional, for containerized run)

---

Local Setup (venv)

Linux / macOS

python3 -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
python src/main_server.py
python src/main_client.py

Windows

python -m venv .venv
.venv\Scripts\Activate.ps1
pip install -r requirements.txt
python src/main_server.py
python src/main_client.py

Deactivate venv:

deactivate

---

Frontends (CLI or Web)

The client supports two interchangeable frontends. The default is the CLI.

Start the server first, then pick a frontend for the client:

# CLI (default)
python src/main_client.py

# Web frontend (Django)
python src/main_client.py --frontend web

Once the web frontend starts, open http://127.0.0.1:8000 in a browser to add devices, view their live state, and update parameters. Override the bind address with the WEB_HOST / WEB_PORT environment variables if needed.

In the CLI, use Show scheduler tasks to inspect automations currently stored in the server scheduler.

Both frontends reuse the exact same client models and controllers; only the view layer differs (src/smart_home/client/views/cli.py vs. src/smart_home/client/views/web/).

---

Regenerate Protobuf (after changing message.proto)

python -m grpc_tools.protoc -I src --python_out=src src/smart_home/proto/v1/message.proto

If needed, install the generator first:

pip install grpcio-tools

---

Testing

If needed, install pytest first:

pip install pytest

Run all tests:

python -m pytest tests

AI Provider: Gemini

The server can use Gemini through the Gemini API. Create a Gemini API key in Google AI Studio, then put it in a local .env file:

cp .env.example .env

Edit .env:

AI_PROVIDER=gemini
GEMINI_API_KEY=your-key
GEMINI_MODEL=gemini-3.5-flash

Then start the server:

python src/main_server.py

The default AI config is in config.toml:

[ai]
provider = "gemini"
gemini_model = "gemini-3.5-flash"

Run only protobuf pytest tests:

python -m pytest tests/test_protobuf_pytest.py

Run protobuf serialization script test:

python tests/test_protobuf.py

---

Docker: Start Server + Client

1. Start only the server in background:

docker compose up -d server

2. Start client session:

docker compose run --rm client

3. In the client prompt (Client >), type your message and press Enter.

4. (Optional) In another terminal, watch server logs:

docker compose logs -f server

5. Stop everything:

docker compose down

Setup script

The root setup scripts can run the Docker automatically.

Linux, macOS, WSL, or Git Bash:

bash setup.sh

Windows PowerShell:

.\setup.ps1

If Docker is not running, the scripts wait and retry until Docker becomes available.