Python: apk add numpy is not working on python:2.7.11-alpine

[lrodorigo]

Step to reproduce the problem:

FROM python:2.7.11-alpine

RUN echo @testing http://dl-cdn.alpinelinux.org/alpine/edge/testing >> /etc/apk/repositories && \
    apk add --update gfortran build-base py-numpy@testing

leads to:

Python 2.7.11 (default, May 12 2016, 18:32:26)
[GCC 5.3.0] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import numpy
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
ImportError: No module named numpy

Manually installing python from apk on clean Alpine docker image is working fine.

[yosifkit]

The python in the alpine image is built from source directly and is not configured the same as the python from apk and would therefore know nothing about the py-numpy apk package. Also, it seems impossible to pip install numpy without applying their same patch: http://git.alpinelinux.org/cgit/aports/tree/testing/py-numpy/numpy-1.11.0-musl.patch. :cry:

[tianon]

This appears to be fixed:

$ docker run -it --rm python:2 sh
# pip install numpy
Collecting numpy
  Downloading numpy-1.11.1-cp27-cp27mu-manylinux1_x86_64.whl (15.3MB)
    100% |████████████████████████████████| 15.3MB 156kB/s 
Installing collected packages: numpy
Successfully installed numpy-1.11.1
# python
Python 2.7.12 (default, Aug  8 2016, 18:07:45) 
[GCC 4.9.2] on linux2
Type "help", "copyright", "credits" or "license" for more information.
>>> import numpy
>>> 

[tianon]

I'm a dummy; tested python:2 instead of python:2-alpine. :disappointed:

[tianon]

Yep, confirmed still an issue with upstream numpy:

$ docker run -it --rm python:2-alpine sh
/ # apk add --no-cache gcc musl-dev
...
/ # pip install numpy
...
  gcc: numpy/core/src/multiarray/number.c
  gcc: numpy/core/src/multiarray/numpyos.c
  numpy/core/src/multiarray/numpyos.c:18:21: fatal error: xlocale.h: No such file or directory
  compilation terminated.
  numpy/core/src/multiarray/numpyos.c:18:21: fatal error: xlocale.h: No such file or directory
  compilation terminated.
  error: Command "gcc -fno-strict-aliasing -g -O2 -DNDEBUG -g -fwrapv -O3 -Wall -Wstrict-prototypes -fPIC -DHAVE_NPY_CONFIG_H=1 -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE=1 -D_LARGEFILE64_SOURCE=1 -Ibuild/src.linux-x86_64-2.7/numpy/core/src/private -Inumpy/core/include -Ibuild/src.linux-x86_64-2.7/numpy/core/include/numpy -Inumpy/core/src/private -Inumpy/core/src -Inumpy/core -Inumpy/core/src/npymath -Inumpy/core/src/multiarray -Inumpy/core/src/umath -Inumpy/core/src/npysort -I/usr/local/include/python2.7 -Ibuild/src.linux-x86_64-2.7/numpy/core/src/private -Ibuild/src.linux-x86_64-2.7/numpy/core/src/private -Ibuild/src.linux-x86_64-2.7/numpy/core/src/private -c numpy/core/src/multiarray/numpyos.c -o build/temp.linux-x86_64-2.7/numpy/core/src/multiarray/numpyos.o" failed with exit status 1

  ----------------------------------------
  Failed building wheel for numpy
...
    gcc: numpy/core/src/multiarray/number.c
    gcc: numpy/core/src/multiarray/numpyos.c
    numpy/core/src/multiarray/numpyos.c:18:21: fatal error: xlocale.h: No such file or directory
    compilation terminated.
    numpy/core/src/multiarray/numpyos.c:18:21: fatal error: xlocale.h: No such file or directory
    compilation terminated.
    error: Command "gcc -fno-strict-aliasing -g -O2 -DNDEBUG -g -fwrapv -O3 -Wall -Wstrict-prototypes -fPIC -DHAVE_NPY_CONFIG_H=1 -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE=1 -D_LARGEFILE64_SOURCE=1 -Ibuild/src.linux-x86_64-2.7/numpy/core/src/private -Inumpy/core/include -Ibuild/src.linux-x86_64-2.7/numpy/core/include/numpy -Inumpy/core/src/private -Inumpy/core/src -Inumpy/core -Inumpy/core/src/npymath -Inumpy/core/src/multiarray -Inumpy/core/src/umath -Inumpy/core/src/npysort -I/usr/local/include/python2.7 -Ibuild/src.linux-x86_64-2.7/numpy/core/src/private -Ibuild/src.linux-x86_64-2.7/numpy/core/src/private -Ibuild/src.linux-x86_64-2.7/numpy/core/src/private -c numpy/core/src/multiarray/numpyos.c -o build/temp.linux-x86_64-2.7/numpy/core/src/multiarray/numpyos.o" failed with exit status 1

    ----------------------------------------
Command "/usr/local/bin/python2 -u -c "import setuptools, tokenize;__file__='/tmp/pip-build-ctFB0g/numpy/setup.py';exec(compile(getattr(tokenize, 'open', open)(__file__).read().replace('\r\n', '\n'), __file__, 'exec'))" install --record /tmp/pip-ROVRNB-record/install-record.txt --single-version-externally-managed --compile" failed with error code 1 in /tmp/pip-build-ctFB0g/numpy/

[tianon]

A really, really hacky workaround is to do something like ln -s locale.h /usr/include/xlocale.h, which allows me to successfully install numpy:

/ # ln -s locale.h /usr/include/xlocale.h
/ # pip install numpy
Collecting numpy
  Using cached numpy-1.11.1.zip
Building wheels for collected packages: numpy
  Running setup.py bdist_wheel for numpy ... done
  Stored in directory: /root/.cache/pip/wheels/10/1d/d8/608bdeaee5aad9e278259fa68d21a29c995a9117a250f7be26
Successfully built numpy
Installing collected packages: numpy
Successfully installed numpy-1.11.1

[bmcilw1]

This is still not working.
docker run python:3-alpine /bin/sh -c "apk update && pip install --upgrade pip && pip install numpy"

EDIT: I got it working using posted comments and will shortly post a link to a dummy TensorFlow application used to test.

[AlJohri]

@tianon Perhaps the base image can include this line?

RUN ln -s /usr/include/locale.h /usr/include/xlocale.h

Or at least include something in the README saying that this is needed for installing numpy?

[AlJohri]

Although I guess this is technically backward incompatible because all Dockerfiles that have manually put this line will run into an error saying that the locale file already exists.

[wglambert]

$ docker run -it --rm python:2-alpine sh
Unable to find image 'python:2-alpine' locally
2-alpine: Pulling from library/python
911c6d0c7995: Already exists 
17a85618b748: Pull complete 
3fc0dee7ce0d: Pull complete 
ed0218841895: Pull complete 
Digest: sha256:fbf776db54e0cf14a03ce7737c011c0f14c4626b7cda53580535c3140ce9c2a9
Status: Downloaded newer image for python:2-alpine

/ # apk add --no-cache gcc musl-dev
fetch http://dl-cdn.alpinelinux.org/alpine/v3.7/main/x86_64/APKINDEX.tar.gz
fetch http://dl-cdn.alpinelinux.org/alpine/v3.7/community/x86_64/APKINDEX.tar.gz
(1/13) Installing binutils-libs (2.30-r1)
(2/13) Installing binutils (2.30-r1)
(3/13) Installing gmp (6.1.2-r1)
(4/13) Installing isl (0.18-r0)
(5/13) Installing libgomp (6.4.0-r5)
(6/13) Installing libatomic (6.4.0-r5)
(7/13) Installing pkgconf (1.3.10-r0)
(8/13) Installing libgcc (6.4.0-r5)
(9/13) Installing mpfr3 (3.1.5-r1)
(10/13) Installing mpc1 (1.0.3-r1)
(11/13) Installing libstdc++ (6.4.0-r5)
(12/13) Installing gcc (6.4.0-r5)
(13/13) Installing musl-dev (1.1.18-r3)
Executing busybox-1.27.2-r11.trigger
OK: 110 MiB in 43 packages
/ # pip install numpy
Collecting numpy
  Using cached https://files.pythonhosted.org/packages/d5/6e/f00492653d0fdf6497a181a1c1d46bbea5a2383e7faf4c8ca6d6f3d2581d/numpy-1.14.5.zip
Building wheels for collected packages: numpy
  Running setup.py bdist_wheel for numpy ... done
  Stored in directory: /root/.cache/pip/wheels/8e/6f/1c/e089e6eef4d32bdd8a9ab9dfed2da700ebd383c3eb28a37b7d
Successfully built numpy
Installing collected packages: numpy
Successfully installed numpy-1.14.5

Issue is resolved, closing