I put on the server (centos 6) in a virtual environment (virtualenv) PyQt5: pip install pyqt5 . Trying to write something using it. On the very first line from PyQt5.Qt import Qt an ImportError: cannot import name 'Qt' error ImportError: cannot import name 'Qt' . I just call python generate.py (third python of course). What could be the problem?

  • Can it help? Pip install --upgrade virtualenv - ⷶ ⷩ ⷮ ⷪ ⷩ
  • @Anton no, even did not help. And where does it all this? What is the difference in the version of the package virtualenv? He stands in the system. Put recently ... - PECHAPTER
  • one
    Try from PyQt5.QtCore import Qt - prusanov
  • @prusanov hmm, the error has changed "ImportError: /lib64/libc.so.6: version` GLIBC_2.14 'not found ". - PECHAIR
  • @DarkByte, just kind of found somewhere that the virtualenv update helped. True, there was some other package, not pyqt5. It is a pity that it did not work - ⷶ ⷩ ⷮ ⷪ ⷩ

1 answer 1

In order not to bother with CentOS with the compilation of the libraries necessary for Qt to work, you can conda install pyqt .

An example that puts a miniconda assembly into a centos: 6 docker container creates a separate environment (pyqt5) and puts the pyqt package into it (the default version):

 $ docker run -it --rm centos:6 # mkdir miniconda && cd miniconda # curl -O https://repo.continuum.io/miniconda/Miniconda3-latest-Linux-x86_64.sh # bash Miniconda3-latest-Linux-x86_64.sh -b # export PATH="/root/miniconda3/bin:$PATH" # python -V # -> Python 3.6 # conda create -n pyqt5 # conda install -n pyqt5 pyqt # source activate pyqt5 # python -c 'from PyQt5.Qt import Qt'
  • I have never worked with a conda. How does he even differ from virtualenv? And why do you think that he will be better? And I’m generally afraid of the docker as a fire ... I’m probably not going to risk contacting him. Too hard. - PECHAIR
  • @DarkByte: 1- docker I used to demonstrate behavior on a pure CentOS system (you indicated in the question that you have a server on CentOS). If you do not need a docker, you can not use it. 2- conda allows you to isolate packages written in different languages ​​(in the case of Qt, this is C ++), and not only on pure Python as in the case of virtualenv (in this case, you have to install Qt yourself). Installing packages is a task with many dimensions, so almost all cases are extreme, everything is on the border. In different cases, a variety of options can be. I showed one of the workers - jfs