Failed to install python modules using pip install base on python:2.7.15-alpine3.8

[dodoflyy]

Hello, I want to install some python modules base on python:2.7.15-alpine3.8 but always failed.
Here is my Dockerfile :

FROM python:2.7.15-alpine3.8
# install python modules
COPY requirements.txt /build/RUN apk add --update curl gcc g++ \
    && rm -rf /var/cache/apk/* \
&& pip install --upgrade pip \
&& pip install setuptools --upgrade \
&& pip install -r /build/requirements.txt

Here is my requirements.txt :

scipy
pathlib
python-docx
pandas
mysql-connector-python==8.0.12
xlrd

There is too much warning or error information, I will post some of them.

      File "/usr/local/lib/python2.7/site-packages/numpy/distutils/misc_util.py", line 943, 
in _get_configuration_from_setup_py        config = setup_module.configuration(*args)
      File "scipy/linalg/setup.py", line 19, in configuration
        raise NotFoundError('no lapack/blas resources found')
    numpy.distutils.system_info.NotFoundError: no lapack/blas resources found
    
    ----------------------------------------
Command "/usr/local/bin/python -u -c "import setuptools, tokenize;__file__='/tmp/pip-install
-_nrjIx/scipy/setup.py';f=getattr(tokenize, 'open', open)(__file__);code=f.read().replace('\r\n', '\n');f.close();exec(compile(code, __file__, 'exec'))" install --record /tmp/pip-record-nYcbQS/install-record.txt --single-version-externally-managed --compile" failed with error code 1 in /tmp/pip-install-_nrjIx/scipy/

   #include "libxml/xmlversion.h"
                                 ^  compilation terminated.
  Compile failed: command 'gcc' failed with exit status 1
  creating tmp
  cc -I/usr/include/libxml2 -c /tmp/xmlXPathInitKognPA.c -o tmp/xmlXPathInitKognPA.o
  /tmp/xmlXPathInitKognPA.c:1:26: fatal error: libxml/xpath.h: No such file or directory
   #include "libxml/xpath.h"
                            ^
  compilation terminated.
  *********************************************************************************
  Could not find function xmlCheckVersion in library libxml2. Is libxml2 installed?
  *********************************************************************************
  error: command 'gcc' failed with exit status 1
  
  ----------------------------------------
  Running setup.py clean for lxml
  Failed building wheel for lxml

  pandas/_libs/testing.c: In function 'inittesting':
    ^~~~~~~
  pandas/_libs/internals.c: In function '__pyx_gb_6pandas_5_libs_9internals_12generator':  pandas/_libs/internals.c:9336:36: warning: '__pyx_t_23' may be used uninitialized in this 
function [-Wmaybe-uninitialized]           __pyx_cur_scope->__pyx_t_9 = __pyx_t_23;
           ~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~  pandas/_libs/internals.c:9333:36: warning: '__pyx_t_11' may be used uninitialized in this 
function [-Wmaybe-uninitialized]           __pyx_cur_scope->__pyx_t_5 = __pyx_t_11;
           ~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~  pandas/_libs/internals.c:9330:36: warning: '__pyx_t_6' may be used uninitialized in this f
unction [-Wmaybe-uninitialized]           __pyx_cur_scope->__pyx_t_4 = __pyx_t_6;
           ~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~  pandas/_libs/internals.c:9335:36: warning: '__pyx_t_15' may be used uninitialized in this 
function [-Wmaybe-uninitialized]           __pyx_cur_scope->__pyx_t_8 = __pyx_t_15;
           ~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~  pandas/_libs/internals.c:9334:36: warning: '__pyx_t_12' may be used uninitialized in this function [-Wmaybe-uninitialized]           __pyx_cur_scope->__pyx_t_7 = __pyx_t_12;           ~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~

I hope somebody knows why and how to fix. Thanks.

[tao12345666333]

see #318 .

[tao12345666333]

Alpine Linux doesn't support Python PEP513. So when you use Alpine Linux based image to install some packages, it needs to compile. (you must install gcc and other packages for this.)

[dodoflyy]

Alpine Linux doesn't support Python PEP513. So when you use Alpine Linux based image to install some packages, it needs to compile. (you must install gcc and other packages for this.)

Thanks for you reply. I go see #318 and install gcc, still don't work. I want to know what's other packages, where can I know the list? I can pip install from python:2-stretch.

[wglambert]

I can't figure out why Alpine won't at least install those packages from source. I would highly recommend in your use case to just use Debian. The packages you're installing probably eclipse the size benefit of using Alpine and it seems you have to install most if not all of those packages from source on Alpine which currently exhibit numerous issues preventing their installation, whereas Debian can use manylinux wheels that install in literally 10 seconds. I had thought that Alpine can use none wheels but they were not supported wheel on this platform when I specified them.

Also if that's your complete Dockerfile you'll probably want to remove gcc at the end of the RUN line as it's 76MB which is 17MB bigger than the image, also g++ as well which is 64MB.

If you want to continue troubleshooting this perhaps https://serverfault.com/questions/771211/docker-alpine-and-matplotlib is relevant to the compiling issues. I can't figure out why the packages won't build from source, perhaps it's dependencies? Regardless there's nothing we could change in the image to remedy this. You could try asking the Docker Community Forums, the Docker Community Slack, or Stack Overflow.

See also docker-library/docs#904 and #341 (comment) which is odd as I couldn't get the 2.7 none wheels to install when specifying their URL, and pip by default chose the .tar.gz