ax-devil-rtsp 0.1.0

Python library for RTSP streaming from Axis devices

ax-devil-rtsp

A Python library for RTSP streaming from Axis cameras with video and AXIS Scene metadata support.

The words 'AXIS Scene Metadata' is hereby called 'application data' in this project.

See also: ax-devil-device-api and ax-devil-mqtt for other Axis device management tools.

---

📋 Contents

• Feature Overview
• Installation
• Quick Start
• Testing
• Development Setup
• License

---

🔍 Feature Overview

Feature	Description	Python API
🔄 Combined Streaming	Simultaneous video and application data streaming (default)	RtspDataRetriever
📹 Video Only	Stream video frames without application data	RtspVideoDataRetriever
📊 Application Data (AXIS Scene Metadata) Only	Stream scene application data without video	RtspApplicationDataRetriever
⚡ Real-time Processing	Frame-by-frame processing with custom callbacks	on_video_data
🎯 RTP Extension Data	Access to ONVIF RTP extension data and timing information (enabled by default)	rtp_ext=True
🛠️ Axis URL Builder	Utility for constructing Axis-compatible RTSP URLs	build_axis_rtsp_url

---

📦 Installation

pip install ax-devil-rtsp

System Dependencies

On Linux, this library requires system packages for PyGObject and GStreamer. If they're missing, you'll see clear error messages with installation instructions.

Quick setup (Ubuntu/Debian):

# Check what's missing
python tools/dep.py --check

# Get installation commands
python tools/dep.py --install

---

🚀 Quick Start

Python API

from ax_devil_rtsp import RtspDataRetriever, build_axis_rtsp_url
from multiprocessing import freeze_support
import time

# Define callback functions
def on_video_data(payload):
    frame = payload["data"]
    diagnostics = payload["diagnostics"]
    print(f"Video frame: {frame.shape}, {diagnostics}")

def on_application_data(payload):
    xml_data = payload["data"]
    diagnostics = payload["diagnostics"]
    print(f"Application data: {len(xml_data)} bytes, {diagnostics}")

def on_error(payload):
    print(f"Error: {payload['message']}")

def main():
    # Build the RTSP URL or supply one directly
    # rtsp_url = "rtsp://username:password@192.168.1.90/axis-media/media.amp?analytics=polygon"
    rtsp_url = build_axis_rtsp_url(
        ip="192.168.1.90",
        username="username",
        password="password",
        video_source=1,
        get_video_data=True,
        get_application_data=True,
        rtp_ext=True,  # Enable RTP extension data
        resolution="640x480",
    )

    retriever = RtspDataRetriever(
        rtsp_url=rtsp_url,
        on_video_data=on_video_data,
        on_application_data=on_application_data,
        on_error=on_error,
        latency=100,
    )

    # Use context manager for automatic cleanup
    with retriever:
        print("Streaming... Press Ctrl+C to stop")
        while True:
            time.sleep(0.1)

if __name__ == "__main__":
    freeze_support()  # Needed on Windows when multiprocessing start method is 'spawn'
    main()

Note

Because ax-devil-rtsp forces the multiprocessing start method to 'spawn', your script's entry point must be guarded with if __name__ == "__main__": (as shown above). On Windows also call multiprocessing.freeze_support() before starting the retriever.

Video-only and application-data-only

from ax_devil_rtsp import (
    RtspVideoDataRetriever,
    RtspApplicationDataRetriever,
    build_axis_rtsp_url,
)

# Video-only
video_url = build_axis_rtsp_url(
    ip="192.168.1.90",
    username="username",
    password="password",
    video_source=1,
    get_video_data=True,
    get_application_data=False,
    rtp_ext=True,
)
with RtspVideoDataRetriever(
    rtsp_url=video_url,
    on_video_data=lambda p: print(p["diagnostics"]),
):
    ...

# Application data only
app_url = build_axis_rtsp_url(
    ip="192.168.1.90",
    username="username",
    password="password",
    video_source=1,
    get_video_data=False,
    get_application_data=True,
    rtp_ext=True,
)
with RtspApplicationDataRetriever(
    rtsp_url=app_url,
    on_application_data=lambda p: print(len(p["data"]))
):
    ...

### CLI Usage

The CLI offers two subcommands:

- `device` builds the RTSP URL from device details like IP address and
  credentials.
- `url` is for when you already have a complete RTSP URL. Options that modify
  the URL (for example `--resolution` or `--rtp-ext`) are only valid with the
  `device` command.

**Basic Usage (streams both video and application data):**
```bash
ax-devil-rtsp device --ip 192.168.1.90 --username admin --password secret

Using a complete RTSP URL:

ax-devil-rtsp url "rtsp://admin:secret@192.168.1.90/axis-media/media.amp?analytics=polygon"

Common Options:

# Custom resolution
ax-devil-rtsp device --ip 192.168.1.90 --username admin --password secret \
  --resolution 1280x720

# Different camera source
ax-devil-rtsp device --ip 192.168.1.90 --username admin --password secret --source 2

Specialized Modes:

# Video only (no application data overlay)
ax-devil-rtsp device --ip 192.168.1.90 --username admin --password secret --only-video

# Application data only (no video window)
ax-devil-rtsp device --ip 192.168.1.90 --username admin --password secret --only-application-data

# Disable RTP extension data
ax-devil-rtsp device --ip 192.168.1.90 --username admin --password secret --no-rtp-ext

For demo processing and lifecycle control, see: --enable-video-processing, --brightness-adjustment, and --manual-lifecycle (run ax-devil-rtsp device --help).

Environment Variables (Optional)

export AX_DEVIL_TARGET_ADDR=192.168.1.90
export AX_DEVIL_TARGET_USER=admin
export AX_DEVIL_TARGET_PASS=secret
export AX_DEVIL_USAGE_CLI=safe  # Set to "unsafe" to skip SSL verification

Set these variables to avoid passing --ip, --username, or --password each time you invoke the device command. AX_DEVIL_USAGE_CLI is shared with related ax-devil-* tools and kept here for consistency.

---

🧪 Testing

# Unit tests
pytest tests/unit/ -v

# Integration tests (local test servers)
pytest tests/integration/ -v

# Integration tests (real camera)
USE_REAL_CAMERA=true AX_DEVIL_TARGET_ADDR=192.168.1.90 pytest tests/integration/ -v

---

Note on GI/GStreamer

On Linux, PyGObject and GStreamer are system packages. Install them with your distro package manager before using this library (see Development Setup below). If they are missing, the library will show an error with install instructions.

Python Environment

python3 -m venv .venv
source .venv/bin/activate
pip install -e .[dev]

Helper Scripts

• tools/dep.py: Unified dependency management tool
  • Check dependencies: python tools/dep.py --check
  • Get install commands (Ubuntu/Debian): python tools/dep.py --install

---

📄 License

MIT License - See LICENSE file for details.

---

⚠️ Disclaimer

This project is independent and not affiliated with Axis Communications AB. For official resources, visit Axis developer documentation.