Installing Ximilar Client
We have our own SDK client written in Python3.9+ (recommended). Before installing it, be sure to install Python
on your system first. You will need to have also a pip tools. You can optionally install the client directly from the repository.
Basic Installation Command
If you have all the prerequisities (python, opencv), the installation can be performed in two ways:
pip install ximilar-client
or manual installation from repository (with latest changes):
git clone https://gitlab.com/ximilar-public/ximilar-client.git
pip install -e ximilar-client
On some systems you will need to write pip3 install ximilar-client
instead of just pip
.
Installing prerequisities
Installing pip and python
MacOS
For mac systems, pip should be part of brew package. Just be sure to install homebrew (https://brew.sh/) and run following command:
brew install python3
Linux
Here are commands for the most popular ubuntu version (python should be already installed)
sudo apt update
sudo apt install python3-pip
Windows
Installing python and opencv is nicely described on this youtube video. Basically you need to download and install python from python.org
Installing Opencv-Python
Pip
The most easiest way to install python opencv is via pip again:
pip install opencv-python
You can optionally install opencv-python-contrib
, opencv-python-headless
or opencv-python-contrib-headless
. Headless versions are suitable for server usage with no graphical interface.
Pip - pick package
Sometimes your pip system will not install correct opencv version. On my Mac M2 (arm), I must find a correct package on pypi and then install it.
- Find correct version on https://pypi.org/project/opencv-contrib-python/#history
- Go to the version (for example 4.7.0.72) and click on
Download files
in the left menu. Then pick your distribution, in my case it wasopencv_contrib_python-4.7.0.72-cp37-abi3-macosx_11_0_arm64.whl
which is built for arm(m1/m2/...) processors. - Run pip or pip3 on downloaded whl package:
pip install your-downloaded-opencv-whl-package.whl
Install from opencv.org
Alternatively, you can try to install opencv directly from their page:
https://opencv.org/releases/
Building from source
The hardest way to install, however on some platforms necessary. This is the last option, for example here is tutorial for Jetson Nano:
https://qengineering.eu/install-opencv-on-jetson-nano.html
Here is way for installing opencv on M1 Macbooks:
https://moeenv.blog/?p=209&lang=en
Testing installation
After installing opencv (go to python command line) by writting python
or python3
in terminal/shell and test following code:
import cv2
print(cv2.__version__)