We observed an issue with the latest version of tifffile (2019.7.26.2), which happens when an older version of numpy is used as a dependency (v1.15.4):
$ pip install numpy==1.15.4
Collecting numpy==1.15.4
Downloading https://files.pythonhosted.org/packages/38/39/f73e104d44f19a6203e786d5204532e214443ea2954917b27f3229e7639b/numpy-1.15.4-cp37-cp37m-manylinux1_x86_64.whl (13.8MB)
|████████████████████████████████| 13.9MB 9.1MB/s
Installing collected packages: numpy
Found existing installation: numpy 1.17.4
Uninstalling numpy-1.17.4:
Successfully uninstalled numpy-1.17.4
Successfully installed numpy-1.15.4
builder@627e0aa2d14d:~$
builder@627e0aa2d14d:~$ ipython
Python 3.7.4 (default, Aug 13 2019, 20:35:49)
Type 'copyright', 'credits' or 'license' for more information
IPython 7.8.0 -- An enhanced Interactive Python. Type '?' for help.
In [1]: import imagecodecs_lite as imagecodecs
---------------------------------------------------------------------------
ValueError Traceback (most recent call last)
<ipython-input-1-368ac70f90eb> in <module>
----> 1 import imagecodecs_lite as imagecodecs
/opt/conda/lib/python3.7/site-packages/imagecodecs_lite/__init__.py in <module>
2 # imagecodecs_lite/__init__.py
3
----> 4 from ._imagecodecs_lite import __doc__, __version__
5 from ._imagecodecs_lite import *
__init__.pxd in init imagecodecs._imagecodecs_lite()
ValueError: numpy.ufunc size changed, may indicate binary incompatibility. Expected 216 from C header, got 192 from PyObject
builder@627e0aa2d14d:~$ pip install 'numpy>1.15.4'
Collecting numpy>1.15.4
Using cached https://files.pythonhosted.org/packages/9b/af/4fc72f9d38e43b092e91e5b8cb9956d25b2e3ff8c75aed95df5569e4734e/numpy-1.17.4-cp37-cp37m-manylinux1_x86_64.whl
Installing collected packages: numpy
Found existing installation: numpy 1.15.4
Uninstalling numpy-1.15.4:
Successfully uninstalled numpy-1.15.4
Successfully installed numpy-1.17.4
builder@627e0aa2d14d:~$ ipython
Python 3.7.4 (default, Aug 13 2019, 20:35:49)
Type 'copyright', 'credits' or 'license' for more information
IPython 7.8.0 -- An enhanced Interactive Python. Type '?' for help.
In [1]: import imagecodecs_lite as imagecodecs
In [2]:
In the snippet above, you can see it works well with the latest version of numpy (1.17.4), and it was also tested with the version 1.16, and worked fine. Should there be a pin of numpy>=1.16 in setup.py? Thank you.
We observed an issue with the latest version of tifffile (2019.7.26.2), which happens when an older version of numpy is used as a dependency (v1.15.4):
In the snippet above, you can see it works well with the latest version of numpy (1.17.4), and it was also tested with the version 1.16, and worked fine. Should there be a pin of numpy>=1.16 in
setup.py? Thank you.cc @cgohlke