部署Django应用时遇ModuleNotFoundError: No module named 'scipy._lib'求助

Hey there, let's tackle this ModuleNotFoundError: No module named 'scipy._lib' issue you're facing on PythonAnywhere. I've seen similar problems with older scipy versions and virtual environment mismatches, so here are some targeted steps to fix it:

• Double-check your virtual environment is activated
  It's easy to accidentally run your Django server outside the virtual environment where you installed scipy, which would use PythonAnywhere's system Python instead of your custom libraries.

  1. Activate your virtual environment first:

     # If you used virtualenvwrapper
     workon your_virtual_env_name
     # Or if you created it with plain virtualenv
     source /path/to/your/virtualenv/bin/activate
     

  2. After activating, run pip3 list again to confirm scipy 1.1.0 is listed under this environment, then try python3 manage.py runserver once more.

• Reinstall scipy to fix incomplete installation
  Sometimes scipy installs can be partial due to network blips or missing system dependencies. Let's do a clean reinstall:

  # Uninstall existing scipy completely
  pip3 uninstall -y scipy
  # Reinstall the exact version you need
  pip3 install scipy==1.1.0
  

  If this still fails, try upgrading to a slightly newer version that's more compatible with Ubuntu 14.04 (scipy 1.2.3 works reliably with Python 3.4-3.6):

  pip3 install scipy==1.2.3
  

• Verify Python version consistency
  Older scipy versions like 1.1.0 have strict Python version compatibility. Make sure the Python version in your PythonAnywhere virtual environment matches your local machine:

  1. Check your local Python version: python3 --version
  2. Check PythonAnywhere's virtual environment version after activation:

     python3 --version
     

  If they don't match, create a new virtual environment with the correct Python version (e.g., for Python 3.6 on PythonAnywhere):

  mkvirtualenv --python=/usr/bin/python3.6 your_new_env
  

  Then reinstall all your dependencies including scipy.

• Install required system libraries
  Scipy relies on low-level math libraries that might be missing on Ubuntu 14.04. Run these commands in the PythonAnywhere console to install them:

  sudo apt-get update
  sudo apt-get install libblas-dev liblapack-dev gfortran
  

  After installing these system dependencies, reinstall scipy as mentioned earlier.

• Validate scipy's installation path
  Ensure scipy is installed in your virtual environment's site-packages directory:

  1. Activate your env, then run:

     pip3 show scipy
     

     Look for the Location field—it should point to something like /home/your_username/.virtualenvs/your_env/lib/pythonX.X/site-packages
  2. Confirm this path is in Python's system path:

     python3 -c "import sys; print('\n'.join(sys.path))"
     

  If the site-packages path isn't listed, your virtual environment activation might be broken. Try recreating the virtual environment from scratch.

内容的提问来源于stack exchange，提问作者Kirill Korolev