8

I have installed Numpy successfully. But on the site , there is lot of things that I have to do such as building Numpy, Scipy, downloading ATLAS, LAPACK etc. I am really confused and even I have checked some of the other queries also. Still not able to understand.

Can anyone please explain me what exactly should I do? I have cygwin installed on my pc.

  1. How to buid Numpy? whats the use of building numpy?
  2. How to install Scipy?
  3. Whats ATLAS, LAPACK etc etc ?

Thank you so much(in advance).

Saullo G. P. Castro
  • 56,802
  • 26
  • 179
  • 234
Hemant
  • 619
  • 2
  • 6
  • 17
  • @Hermant What is it you want to do? Also, I use numpy/scipy in windows without cygwin with no problem. – KobeJohn Nov 11 '13 at 00:51
  • @Hermant. I should clarify - what do you want to do that isn't answered by the 3 answers already on here? Please explain what is not working or accept one of their answers. – KobeJohn Nov 11 '13 at 01:20

5 Answers5

5

I struggled with the same problem for a long time, and I eventually found an easy solution here`

Community
  • 1
  • 1
Ferdinando Randisi
  • 4,068
  • 6
  • 32
  • 43
3

You really should try Enthought Python Distribution. Just a single msi installation file, click a few Next and then you are done. They include lots of scientific packages in it, including numpy, scipy, matplotlib, etc., and also ipython environment.

Chunliang Lyu
  • 1,750
  • 1
  • 20
  • 34
  • Hi, thanks for this suggestion :) I read the same on scipy webpage also but i am still confused how to use it? as in do i need to uninstall my current Python idle (2.7.3)? or is this a kind of library ? Is it sumwhat different from the current Python IDLE (or shud i say the standard Python IDLE)? – Hemant Jan 14 '13 at 07:30
  • it just packages various python scientific libraries like numpy and scipy into a single installer. it can install python 2.7 for you, so you may first uninstall your current python. in my view, it just make the installation process much easier, you can also compile and install packages by yourself, but they already did the dirty work. anyway, you can just download their free version and give it a try. – Chunliang Lyu Jan 14 '13 at 12:03
0

Would WinPython an option for you? Unless necessary, I'd see if that covers my needs

jpmuc
  • 1,092
  • 14
  • 30
  • i really dont have ne idea abt WinPython. Can u jst tell me smthing abt this WinPython? – Hemant Jan 14 '13 at 07:27
  • The best thing is to check the project's website: http://code.google.com/p/winpython/. Basically, it offers all the functionality you might need and saves you from the pain of compiling all those libraries under Windows: http://scipy.github.com/building/windows.html – jpmuc Jan 14 '13 at 09:21
  • Hi, thanks for the suggestion :) I am definitely going to look out for it! But can u please tell me how different is WinPython from the basic or standard python IDLE that we use? and if possible can you just tell me what does this building of scipy, numpy mean? i am a beginner to python and i have to understand these concepts so that i can apply these concepts to learn Machine Learning in a better way. – Hemant Jan 14 '13 at 09:56
  • WinPython includes all the scientific packages you may need (numpy, matplotlib and all the scipy modules) plus and IDE similar to that of MatLab, where you can explore all variables used so far. Also it offers the concept of notebooks (you can create a session where you do some calculations and export all those actions so that someone else can reproduce them too). If you are familiar with MatLab this document might help you understand the meaning of numpy better http://www.scipy.org/NumPy_for_Matlab_Users – jpmuc Jan 16 '13 at 08:08
0

I did the following to get Scripy installed on my Windows virtualenv:

  1. I downloaded the file here: http://www.lfd.uci.edu/~gohlke/pythonlibs/#numpy and then selected the file: numpy-1.11.1+mkl-cp27-cp27m-win32.whl

. The file is applicable to me as I'm running Python 2.7 and I have a win32 machine. You can check this information on your side by running python in your cmd prompt. Mine says:

Python 2.7.9 (default, Dec 10 2014, 12:24:55) [MSC v.1500 32 bit (Intel)] on win32

The reason why we download numpy first is because scipy relates on it. Under scipy on the same website: http://www.lfd.uci.edu/~gohlke/pythonlibs/#scipy, it says:

Requires numpy+mkl.
Install numpy+mkl before installing scipy.
  1. Download the applicable scipy file for you, mine was this one: scipy-0.18.0-cp27-cp27m-win32.whl

Now, you can dump them into the root of your virtualenv folder. Send the following commands in your virtualenv:

pip install numpy-1.11.1+mkl-cp27-cp27m-win32.whl
pip install scipy-0.18.0-cp27-cp27m-win32.whl

Scipy will then install successfully.

Simon
  • 19,658
  • 27
  • 149
  • 217
0

the easiest way is in the following steps: Fixing scipy for python [ 2.n < python < 3.n ]

Download the necessary files from: http://www.lfd.uci.edu/~gohlke/pythonlibs/

Download the version of numpy+mkl (needed to run scipy) and then download scipy for your python type (2.n python written as 2n) or (3.n python written as 3n), n is a variable. Note you must know whether you have a 32bit or 64bit processor.

Create a directory somewhere on your computer, example [C:\DIRECTORY] to install the files numpy+mkd.whl and scipy.whl

Once both file are downloaded, find the location of the file on your computer and move it to the directory you created.

Example: first file installation is needed for scipy is in---> C:\DIRECTORY\numpy\numpy-0.0.0+mkl-cp2n-cp2nm-win_amd32.whl

Example: second file installation is in---> C:\DIRECTORY\scipy\scipy-0.0.0-cp2n-cp2nm-win_amd32.whl

Go to your command prompt and proceed the following example for a python version 2.n:

py -2.n -m pip install C:\DIRECTORY\numpy\numpy-0.0.0+mkl-cp2n-cp2nm-win_amd32.whl

should install

py -2.n -m pip install C:\DIRECTORY\scipy\scipy-0.0.0-cp2n-cp2nm-win_amd32.whl

should install

Test both modules on your python IDLE as following: import numpy import scipy

the modules are working if no errors are returned.

IFDAAS

Community
  • 1
  • 1