当已经存在一个轮子时，Pip轮子正在构建一个新轮子

Question

我正试着在0.17.1为pandas做一个轮子。我希望它使用numpy版本的1.9.2。我已经在$PWD/wheelhouse中为该版本的numpy构建了一个轮子，还有一些其他pandas依赖项：

ls wheelhouse/
numpy-1.9.2-cp34-cp34m-linux_x86_64.whl python_dateutil-2.4.2-py2.py3-none-any.whl  pytz-2015.7-py2.py3-none-any.whl  six-1.10.0-py2.py3-none-any.whl

但是，当我告诉pandas进行构建时，即使我用--find-links通知它轮房文件夹，它仍然构建了一个新的numpy轮子

$ pip --version
pip 6.0.8 from /home/me/.pyenv/versions/3.4.3/lib/python3.4/site-packages (python 3.4)
$ pip wheel pandas==0.17.1 --find-links=$PWD/wheelhouse
Collecting pandas==0.17.1
  Using cached pandas-0.17.1.tar.gz
[... snipped, building stuff ...]
Collecting python-dateutil>=2 (from pandas==0.17.1)
  File was already downloaded /home/me/rebuild_numpy_py3/wheelhouse/python_dateutil-2.4.2-py2.py3-none-any.whl
Collecting pytz>=2011k (from pandas==0.17.1)
  File was already downloaded /home/me/rebuild_numpy_py3/wheelhouse/pytz-2015.7-py2.py3-none-any.whl
Collecting numpy>=1.7.0 (from pandas==0.17.1)
  Using cached numpy-1.10.2.tar.gz
    Running from numpy source directory.
Collecting six>=1.5 (from python-dateutil>=2->pandas==0.17.1)
  File was already downloaded /home/me/rebuild_numpy_py3/wheelhouse/six-1.10.0-py2.py3-none-any.whl
Skipping python-dateutil, due to already being wheel.
Skipping pytz, due to already being wheel.
Skipping six, due to already being wheel.
Building wheels for collected packages: pandas, numpy
  Running setup.py bdist_wheel for pandas
  Destination directory: /home/me/rebuild_numpy_py3/wheelhouse
  Running setup.py bdist_wheel for numpy
  Destination directory: /home/me/rebuild_numpy_py3/wheelhouse
Successfully built pandas numpy

$ ls wheelhouse/
numpy-1.10.2-cp34-cp34m-linux_x86_64.whl  numpy-1.9.2-cp34-cp34m-linux_x86_64.whl  pandas-0.17.1-cp34-cp34m-linux_x86_64.whl  python_dateutil-2.4.2-py2.py3-none-any.whl  pytz-2015.7-py2.py3-none-any.whl  six-1.10.0-py2.py3-none-any.whl

绑定到pandas上的版本是>=1.7.0，所以这里的轮子应该可以工作。那么，它为什么要造一个新轮子呢？我如何才能强制它使用现有的？

Answer 1

轮子不是这样工作的。pandas需要一个>= 1.7.0版本的numpy。您试图强制它查找1.9.2，尽管已经有了较新版本的numpy。即使你已经在你的仓库目录中有它，它也会检查PyPI的最新版本，因为这就是pandas的setup.py文件中声明的。