51Degrees Device Detection Python  4.4

Developer Documentation | Available Properties

Introduction

This project contains 51Degrees Device Detection engines that can be used with the Pipeline API.

The Pipeline is a generic web request intelligence and data processing solution with the ability to add a range of 51Degrees and/or custom plug ins (Engines)

This git repository uses submodules, to clone the git repository run:

or if already cloned run the following to obtain all sub modules:

Dependencies

For runtime dependencies, see our dependencies page. The tested versions page shows the Python versions that we currently test against. The software may run fine against other versions, but additional caution should be applied.

Data

The API can either use our cloud service to get its data or it can use a local (on-premise) copy of the data.

Cloud

You will require a resource key to use the Cloud API. You can create resource keys using our configurator, see our documentation on how to use this.

On-Premise

In order to perform device detection on-premise, you will need to use a 51Degrees data file. This repository includes a free, 'lite' file in the 'device-detection-data' sub-module that has a significantly reduced set of properties. To obtain a file with a more complete set of device properties see the 51Degrees website. If you want to use the lite file, you will need to install GitLFS:

Then, navigate to 'fiftyone.devicedetection.onpremise/device-detection-cxx/device-detection-data' and execute:

Folders

• fiftyone_devicedetection - references both cloud and on-premise packages, contains generic Device-Detection Pipeline Builder.
  • examples - examples for switching between cloud and on-premise.
• fiftyone_devicedetection_cloud - cloud implementation of the engines.
  • tests - tests for the cloud engine and examples.
• fiftyone_devicedetection_onpremise - uses a native library which is built during setup to provide optimal performance and low latency device detections.
  • tests - tests for the on-premise engine.
• fiftyone_devicedetection_shared - common components used by the cloud and on-premise packages.
• fiftyone_devicedetection_examples - cloud and on-premise examples.
  • tests - tests for the examples.

Installation

PyPi

To install all packages, run:

The fiftyone-devicedetection package references both cloud and on-premise packages. If you do not want the on-premise engine or cannot meet the requirements for installing on-premise then install the cloud package on its own:

or

See the cloud and onpremise package readmes for more detail.

Build from Source

Device detection on-premise uses a native binary. (i.e. compiled from C code to target a specific platform/architecture) This section explains how to build this binary.

Pre-requisites

• Install C build tools:
  • Windows:
    • You will need either Visual Studio 2019 or the C++ Build Tools installed.
      • Minimum platform toolset version is v142
      • Minimum Windows SDK version is 10.0.18362.0
  • Linux/MacOS:
    • sudo apt-get install g++ make libatomic1
• If you have not already done so, pull the git submodules that contain the native code:
  • git submodule update --init --recursive

Build steps

It is recommended to use Pipenv when building packages. Pipenv will ensure all the required packages for building and testing are available e.g. cython and flask.

We use make to manage the build process. There are several targets for the make process. If unsure, you probably want to use the install target:

other make targets:

• cloud - only setup packages required for cloud
• onpremise - only setup packages required for onpremise
• test - run package tests
• clean - remove temporary files created when building extensions

Make can be installed on Windows. Alternatively, there is also a powershell setup script called setup.ps1. To run it, you may need to update the execution policy to allow unsigned scripts to execute:

Finally, if the make or powershell scripts fail, or you want to perform the steps manually for some other reason, these are the commands to build the native on-premise library and install the packages.

Examples

To run an example, navigate into the example module directory.

Then run an example:

Cloud

On-Premise