[Numpy-discussion] Beginner Needing Help with Installation
scot.denhalter at gmail.com
Thu Nov 19 17:41:00 EST 2009
Yes, I am using the Snow Leopard OSX. Should I be coding through the Xcode
interface and not Python's IDLE shell?
You distinguish between system wide installation and user installation. You
may have seen from my post to Eric, that I found and installed .dmg
downloads for numpy and matplotlib. I have tried to figure out where these
installed themselves, but I haven't found them in the user folder. I assume
they have been installed system wide. Is this going to be a problem?
On Thu, Nov 19, 2009 at 2:57 PM, Pierre GM <pgmdevlist at gmail.com> wrote:
> On Nov 19, 2009, at 4:34 PM, Scot Denhalter wrote:
> > I am a beginning programmer who is reading Natural Language Processing
> with Python. The book provides tutorials for working with the NLTK, which
> needs numpy to run certain calculations. I have downloaded and installed
> Python 2.6. I have downloaded numpy 1.3.0, but I cannot figure out how to
> install it. I have read the Install.txt that is included, but I don't
> understand exactly what I need to do. The instructions are clear enough for
> those familiar with the programming world, but I am obviously not the
> document's target reader.
> > I am working with a MacBook Pro and do not have either fortran compiler
> installed. Can someone give me some guidance?
> Ciao Scott.
> First of all, don't panic...
> Mac OS 10.6.2 ? You have installed Xcode ? You're basically good to go.
> * Install nose (I use pip for the install), system-wide is fine.
> * Grab a fortran compiler here: http://r.research.att.com/tools/
> * Unzip your numpy sources archive
> * In the directory that gets created, enter in Terminal (or iTerm):
> python setup.py install --user 2>&1| tee build.log
> The --user is to make sure that you install numpy in your Users directory
> and not mess with the system
> The 2>&1 | tee build.log sends the messages you get during the compilation
> to a file build.log (that you gonna need if anything goes wrong).
> That should be enough. If it doesn't work, then post your build.log so that
> we can try and check where it fails.
> Let me know how it goes
> NumPy-Discussion mailing list
> NumPy-Discussion at scipy.org
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the NumPy-Discussion