petal-app-manager 0.1.0

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 3.8+
• python3-dev package (for building some dependencies)
• Redis server (for caching and message passing)
• Additional dependencies based on specific petals

Installation

From PyPI

pip install petal-app-manager

Development Installation

For development, it's recommended to use an editable installation where you define your local dependancies in pyproject.toml as

local = [
    "-e file:///path/to/your/my-petal/#egg=my-petal",
    "-e file:///path/to/pymavlink/#egg=leaf-pymavlink",
]

# Clone the repository
git clone https://github.com/your-org/petal-app-manager.git
cd petal-app-manager

pdm install -G dev -G local

Dependencies Setup

• Ensure python3-dev is installed (see above)

sudo apt-get install python3-dev
# or for specific Python versions
sudo apt-get install python3.11-dev

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

export CC=gcc
pdm install -G dev  # or pip install -e .

• Redis server must be running

# 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

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.

Petal Architecture

Petals are pluggable components that:

• Inherit from the Petal base class
• Use decorators (@http_action, @websocket_action) to define endpoints
• Are discovered and loaded automatically through Python entry points
• Have access to all registered proxies

Running the Server

Quick Start

# Install and run with uvicorn
pip install petal-app-manager
uvicorn petal_app_manager.main:app --reload

Developing Petals

Creating a New Petal

1. Create a new Python package with this structure

# Create a new directory for your petal
mkdir my-petal
cd my-petal

# Initialize a PDM project
pdm init
# Answer the prompts:
# - Python interpreter: Select your Python version (3.8+ recommended)
# - Project name: my-petal
# - Package name: my-petal
# - Version: 0.1.0
# - Description: My custom petal for Petal App Manager
# - License: MIT (or your preferred license)
# - Author: Your Name
# - Email: [e-mail]
# - Do you want to build this project for redistribution as a wheel? yes

# Create the source directory structure
mkdir -p src/my-petal

# Create the initial files
touch src/my-petal/__init__.py
touch src/my-petal/plugin.py

# Add petal-app-manager as a dependency
pdm add petal-app-manager

Now your project structure should look like:

my-petal/
├── pyproject.toml
└── src/
    └── my_petal/
        ├── __init__.py
        └── plugin.py

[!NOTE] For integrated development mode with petal-app-manager, add a reference to your local clone in pyproject.toml:

[tool.pdm.dev-dependencies]
dev = [
    "-e file:///path/to/petal-app-manager/#egg=petal-app-manager",
]

This enables you to make changes to both your petal and the framework simultaneously.

2. Define your petal in plugin.py

from petal_app_manager.plugins.base import Petal
from petal_app_manager.plugins.decorators import http_action, websocket_action
from petal_app_manager.proxies.redis import RedisProxy

class MyPetal(Petal):
    name = "my-petal"
    version = "1.0.0"
    
    @http_action(method="GET", path="/hello")
    async def hello_world(self):
        # Access proxies through self._proxies
        redis_proxy: RedisProxy = self._proxies["redis"]
        await redis_proxy.set("hello", "world")
        return {"message": "Hello World!"}

3. Register your petal in your petal's pyproject.toml

[project.entry-points."petal.plugins"]
my_petal = "my_petal.plugin:MyPetal"

4. Install your petal in development mode

pdm install --dev

Testing Your Petal

Once you've created your petal, you'll want to test it with the petal-app-manager:

1. Clone the petal-app-manager repository if you haven't already:

   git clone https://github.com/DroneLeaf/petal-app-manager.git
   cd petal-app-manager
   

2. Install your petal in development mode:

   • Add your petal to the pyproject.toml file

   # In petal-app-manager's pyproject.toml
   [dependency-groups]
   local = [
       # Add your petal in development mode
           "-e file:///path/to/your/my-petal/#egg=my-petal",  # Adjust path to your petal directory
   ]
   

   • From your petal-app-manager directory run the following

   pdm install -G dev -G local
   

   • petal-app-manager should automatically discover your petal through entry points defined in your petal's pyproject.toml

3. Run the server:

   uvicorn petal_app_manager.main:app --reload --port 9000
   

4. 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

Available Decorators

• @http_action(method="GET|POST", path="/endpoint"): Define HTTP endpoints
• @websocket_action(path="/ws/endpoint"): Define WebSocket endpoints
• @mqtt_action(topic="topic/name"): Define MQTT handlers (planned)

Example: FlightLogPetal

The FlightLogPetal demonstrates a comprehensive petal implementation:

• Downloads and manages flight logs from a PX4 drone
• Provides HTTP endpoints for retrieving and storing flight records
• Uses WebSockets for real-time download progress updates
• Interacts with multiple proxies (MAVLink, Redis, LocalDB)

Key endpoints include:

• GET /flight-records: Retrieve all flight records
• POST /save-record: Save a new flight record
• GET /available-px4-ulogs: List available ULog files from PX4
• GET /download-ulog-pixhawk: Download a ULog file from Pixhawk
• GET /get-px4-time: Get PX4 system time

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