petal-app-manager 0.1.2

A FastAPI interface for loading, managing and running Droneleaf Petal applications

Petal App Manager

A modular application framework for building and deploying "Petals" - pluggable components that can interact with various systems through a unified proxy architecture. Built on FastAPI, Petal App Manager provides a structured way to develop applications that need to interact with external systems like MAVLink devices, Redis, local databases, and more.

Overview

Petal App Manager serves as a backbone for developing modular applications. It:

• Provides a proxy system to interact with different backends (MAVLink, Redis, DynamoDB)
• Offers a plugin architecture for developing and loading "Petals"
• Handles HTTP, WebSocket, and MQTT (planned) endpoints automatically
• Manages the lifecycle of connections and resources

Dependencies

• Python 3.10+ and python3-dev package (for building some dependencies)

  sudo add-apt-repository ppa:deadsnakes/ppa --yes
  sudo apt update; apt-get update;
  sudo apt-get install python3.11 -y
  sudo apt-get install python3.11-dev
  

[!NOTE] You can change python3.11 to whatever version you like >=3.10

• Redis server (for caching and message passing)

  # Install Redis on Ubuntu/Debian
  sudo apt-get install redis-server
  
  # Start Redis service
  sudo systemctl start redis-server
  sudo systemctl enable redis-server  # Auto-start on boot
  

• Controller-dashboard setup

  The controller dashboard can be installed using

  hear-cli local_machine run_program --p controller_dashboard_prepare
  

• Additional dependencies based on specific petals

Installation

Dependencies Setup

• For building pymavlink from source, ensure GCC is used (see above)

export CC=gcc
pdm install -G prod

Installation From PyPI (recommended for users)

pip install petal-app-manager

You may run the server using

# Install and run with uvicorn
uvicorn petal_app_manager.main:app --port 9000

[!TIP] If you would like to run the server with logging to a file enabled: To a default location:

PETAL_LOG_TO_FILE=true uvicorn petal_app_manager.main:app --port 9000

Specify custom log file path

PETAL_LOG_TO_FILE=true PETAL_LOG_FILE_PATH=./app.log uvicorn petal_app_manager.main:app --port 9000

Set custom log level

PETAL_LOG_LEVEL=ERROR PETAL_LOG_TO_FILE=true PETAL_LOG_FILE_PATH=./app.log uvicorn petal_app_manager.main:app --port 9000

Development Installation (recommended for developers)

For development of petal-app-manager concurrently with your petal, it's recommended to use an editable installation.

1. First clone the petal-app-manager

[!NOTE] Please see the petal development guide first

```bash
git clone https://github.com/DroneLeaf/petal-app-manager.git
git clone https://github.com/DroneLeaf/petal-flight-log.git
git clone --recurse-submodules https://github.com/DroneLeaf/mavlink.git
cd petal-app-manager
```

2. Define your dev dependancies (i.e., your petal) in pyproject.toml as

   dev = [
       # your existing dependancies
       "-e file:///${PROJECT_ROOT}/../petal-flight-log/#egg=petal-flight-log",
       "-e file:///${PROJECT_ROOT}/../mavlink/pymavlink/#egg=leaf-pymavlink",
       # ...
       "-e file:///path/to/your/my-petal/#egg=my-petal"
   ]
   

[!NOTE] If you would like to develop mavlink or add user-defined mavlink messages, you may do so under your local clone of mavlink https://github.com/DroneLeaf/mavlink.git pymavlink will be available at /path/to/mavlink/pymavlink under the mavlink directory. You can then add it to pyproject.toml

dev = [
    "-e file:///path/to/pymavlink/#egg=leaf-pymavlink",
]

[!TIP] You may use relative paths intead of absolute paths like so (assuming your directories are one level higher than petal-app-manager)

cd .. # if in the petal-app-manager directory
git clone --recurse-submodules https://github.com/DroneLeaf/mavlink.git

and then add them to your dependancies under pyproject.toml

dev = [
    # existing petals
    "-e file:///${PROJECT_ROOT}/../mavlink/pymavlink/#egg=leaf-pymavlink",
    "-e file:///${PROJECT_ROOT}/../my-petal/#egg=my-petal",
]

3. Finally, you may install your dependancies in editable mode

   pdm install -G dev
   

4. You may now run the petal-app-manager server

   source .venv/bin/activate # to activate the pdm virtual environment in which everythign is installed
   uvicorn petal_app_manager.main:app --reload --port 9000
   

5. Test your endpoints:

   • Access your petal at: http://localhost:9000/petals/my-petal/hello
   • Check the API documentation: http://localhost:9000/docs

[!TIP] For debugging, you can use VSCode's launch configuration:

1. Add this to .vscode/launch.json:

   {
       "version": "0.2.0",
       "configurations": [
           {
               "name": "Petal App Manager",
               "type": "python",
               "request": "launch",
               "module": "uvicorn",
               "args": [
                   "petal_app_manager.main:app",
                   "--reload",
                   "--port", "9000"
               ],
               "jinja": true,
               "justMyCode": false
           }
       ]
   }
   

2. Start debugging with F5 or the Run and Debug panel

Project Structure

petal_app_manager/
├── __init__.py
├── main.py            # FastAPI application setup
├── api/               # Core API endpoints
├── plugins/           # Plugin architecture
│   ├── base.py        # Base Petal class
│   ├── decorators.py  # HTTP/WebSocket decorators
│   └── loader.py      # Dynamic petal loading
├── proxies/           # Backend communication
│   ├── base.py        # BaseProxy abstract class
│   ├── external.py    # MAVLink/ROS communication
│   ├── localdb.py     # Local DynamoDB interaction
│   └── redis.py       # Redis interaction
└── utils/             # Utility functions

How It Works

Proxy System

The framework uses proxies to interact with different backends:

• MavLinkExternalProxy: Communicates with PX4/MAVLink devices
• RedisProxy: Interfaces with Redis for caching and pub/sub
• LocalDBProxy: Provides access to a local DynamoDB instance

Proxies are initialized at application startup and are accessible to all petals.

Accessing the API

Once running, access:

• API documentation: http://localhost:9000/docs
• ReDoc documentation: http://localhost:9000/redoc
• Petal endpoints: http://localhost:9000/petals/{petal-name}/{endpoint}

Troubleshooting

Common Issues

• Redis Connection Errors:

  • Ensure Redis server is running: sudo systemctl status redis-server
  • Check default connection settings in main.py

• MAVLink Connection Issues:

  • Verify the connection string