在Windows上pip安装Psychopy的报错及运行问题求助

Hey there, let's work through these frustrating import errors you're facing in your 64-bit Anaconda Spyder 2.7 environment. These issues almost always boil down to incompatible dependency versions—since Python 2.7 is end-of-life, newer library releases don't support it anymore. Here's what's worked for me and other developers in your situation:

1. Fix pyglet-related import errors (cannot import name key/logging) #

The pyglet library you have installed is likely too new for Python 2.7 and your Psychopy version. Roll back to a compatible release:

1. First, activate your Anaconda environment in the command prompt:

   conda activate your_spyder27_env_name
   

2. Install the last pyglet version that works with Python 2.7:

   pip install pyglet==1.3.2
   

   If conda throws an error, use the pip command above—it's more reliable for this specific version.

2. Reinstall a Python 2.7-compatible Psychopy version #

The latest Psychopy versions dropped support for Python 2.7 entirely. Uninstall your current installation and install the last supported release:

pip uninstall psychopy -y
pip install psychopy==1.85.6

This version is specifically built to work with Python 2.7 and avoids many dependency conflicts.

3. Resolve the PUINT NameError #

This error stems from mismatched numpy or pygame versions. Here's how to fix it:

• First, install a numpy version compatible with Python 2.7:

  conda install numpy==1.16.6
  

• If the error persists, reinstall a Python 2.7-friendly pygame release:

  pip install pygame==1.9.6
  

PUINT is a type definition that exists in these older, compatible library versions.

4. Double-check your environment setup #

• Confirm Spyder is using the correct interpreter: Go to Tools > Preferences > Python interpreter in Spyder, and make sure the path points to the Python executable in your Anaconda 2.7 environment.
• Restart Spyder completely after making any dependency changes—old cached imports can stick around and cause weird issues even after fixes are applied.

内容的提问来源于stack exchange，提问作者Ravi Mevada