[SciPy-Dev] Error while trying to build scipy: NotFoundError: nolapack/blas resources found

Sun Feb 25 15:34:13 EST 2018

As xoviat says, if you want all libs in your sandboxed environment then either use (1) conda, (2) a docker image, or (3) autotools with `PREFIX=<your-env>`, from easiest to hardest

Sent from Yahoo Mail on Android 

  On Sun, Feb 25, 2018 at 12:25 PM, xoviat<xoviat at gmail.com> wrote:   #yiv0637741737 #yiv0637741737 -- _filtered #yiv0637741737 {panose-1:2 4 5 3 5 4 6 3 2 4;} _filtered #yiv0637741737 {font-family:Calibri;panose-1:2 15 5 2 2 2 4 3 2 4;}#yiv0637741737 #yiv0637741737 p.yiv0637741737MsoNormal, #yiv0637741737 li.yiv0637741737MsoNormal, #yiv0637741737 div.yiv0637741737MsoNormal {margin:0in;margin-bottom:.0001pt;font-size:11.0pt;}#yiv0637741737 a:link, #yiv0637741737 span.yiv0637741737MsoHyperlink {color:blue;text-decoration:underline;}#yiv0637741737 a:visited, #yiv0637741737 span.yiv0637741737MsoHyperlinkFollowed {color:#954F72;text-decoration:underline;}#yiv0637741737 .yiv0637741737MsoChpDefault {} _filtered #yiv0637741737 {margin:1.0in 1.0in 1.0in 1.0in;}#yiv0637741737 div.yiv0637741737WordSection1 {}#yiv0637741737 
You’ve discovered that Python packages can depend on native system libraries, and that sharing libraries globally could be a problem. Of course, as with anything in computer science, you aren’t the first person to discover this problem.

If you’re looking for a package manager that can manage system libraries and Python packages, conda is your solution. However, you should consider the cost/benefit of learning conda vs. just keeping your head down and working on whatever issue you would like to work on.

Also note that those BLAS and LAPACK packages are for development only and should not be used in production. Distributors will use a more performant BLAS and LAPACK package that’s written in Assembly. Again, it’s a question of whether it’s worth your time to build a performant BLAS and LAPACK library vs. just using one that’s “good enough”  to get you through development.

From: Pinaki Nath Chowdhury
Sent: Sunday, February 25, 2018 2:15 PM
To: bwana.marko at yahoo.com; SciPy Developers List; xoviat at gmail.com
Subject: Re: [SciPy-Dev] Error while trying to build scipy: NotFoundError: nolapack/blas resources found

Thank you very much.

As you all pointed out, I needed to get BLAS and LAPACK development libraries and fortan compiler.

hence the issue was resolved by simply using:
`sudo apt-get install libblas-dev liblapack-dev gfortran`

But I am a bit worried that this command installed those development libraries in my main system and not in the virtualenv.

So just out of curiosity I wanted to ask, if there is some way in which we can completely install all packages in a virtualenv and then start coding? When I no more feel like working, I can simple delete the virtualenv.

Thanks

Pinaki

On Mon, Feb 26, 2018 at 12:32 AM, Mark Mikofski <bwana.marko at yahoo.com> wrote:

Hi Pinaki,

Building SciPy from source requires BLAS and LAPACK development libraries, and aFORTRAN compiler. That seems to me to be what your traceback is complaint about.

Can you use a wheel instead? There are wheels for every platform at pypi: https://pypi.python.org/pypi/scipy 

Can you tell us what platform you are using? If it is Ubuntu Linux try libblas-dev, liblapack-dev, libopenblas-dev, and gfortran by using `sudo apt install <package>`. In rhel/fedora the names are slightly different.

If you're in Mac osx, building from source then you probably will need homebrew to install the dev libraries and gcc (which different from clang that shops with xcode)

Try reading the dev docs on building SciPy from source here: https://scipy.github.io/devdocs/building/index.html

Another thought is if you are missing conda and pip, it's possible you have different sources of BLAS and LAPACK. Start your Python interpreter, import numpy and execute `numpy.show_config()` to see what numpy was built with. You will have to match the configuration for scipy.

Hope this helps?

Sent from Yahoo Mail on Android

On Sun, Feb 25, 2018 at 10:34 AM, Pinaki Nath Chowdhury

<pinakinathc at gmail.com> wrote:

Hi

I am trying to build Scipy package in virtualenv and getting the following error message:

```

Traceback (most recent call last):

  File "setup.py", line 467, in <module>

    setup_package()

  File "setup.py", line 463, in setup_package

    setup(**metadata)

  File "/home/pinaki/gsoc/local/lib/python2.7/site-packages/numpy/distutils/core.py", line 135, in setup

    config = configuration()

  File "setup.py", line 381, in configuration

    config.add_subpackage('scipy')

  File "/home/pinaki/gsoc/local/lib/python2.7/site-packages/numpy/distutils/misc_util.py", line 1024, in add_subpackage

    caller_level = 2)

  File "/home/pinaki/gsoc/local/lib/python2.7/site-packages/numpy/distutils/misc_util.py", line 993, in get_subpackage

    caller_level = caller_level + 1)

  File "/home/pinaki/gsoc/local/lib/python2.7/site-packages/numpy/distutils/misc_util.py", line 930, in _get_configuration_from_setup_py

    config = setup_module.configuration(*args)

  File "scipy/setup.py", line 15, in configuration

    config.add_subpackage('linalg')

  File "/home/pinaki/gsoc/local/lib/python2.7/site-packages/numpy/distutils/misc_util.py", line 1024, in add_subpackage

    caller_level = 2)

  File "/home/pinaki/gsoc/local/lib/python2.7/site-packages/numpy/distutils/misc_util.py", line 993, in get_subpackage

    caller_level = caller_level + 1)

  File "/home/pinaki/gsoc/local/lib/python2.7/site-packages/numpy/distutils/misc_util.py", line 930, 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

```

The following is details of my system:

```

python --verison : Python 2.7.12

pip freeze :

alabaster==0.7.10

Babel==2.5.3

certifi==2018.1.18

chardet==3.0.4

Cython==0.27.3

docutils==0.14

idna==2.6

imagesize==1.0.0

Jinja2==2.10

MarkupSafe==1.0

numpy==1.14.1

packaging==16.8

Pygments==2.2.0

pyparsing==2.2.0

pytz==2018.3

requests==2.18.4

six==1.11.0

snowballstemmer==1.2.1

Sphinx==1.7.1

sphinxcontrib-websupport==1.0.1

typing==3.6.4

urllib3==1.22

```

Hope the above information is sufficient.

Thanks

Pinaki

_______________________________________________
SciPy-Dev mailing list
SciPy-Dev at python.org
https://mail.python.org/mailman/listinfo/scipy-dev

_______________________________________________
SciPy-Dev mailing list
SciPy-Dev at python.org
https://mail.python.org/mailman/listinfo/scipy-dev

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scipy-dev/attachments/20180225/ca5719e9/attachment-0001.html>