Changes between Initial Version and Version 1 of AdvancedInstall


Ignore:
Timestamp:
May 16, 2016, 3:04:57 PM (4 years ago)
Author:
flip
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • AdvancedInstall

    v1 v1  
     1= Advanced Install =
     2
     3These instructions for installing Vespa are for users who are
     4comfortable using the command line and finding and installing
     5Python packages.
     6
     7== Install Python ==
     8
     9Vespa will work with any 64-bit Python 2.7. Some possible sources for
     10Python are [https://www.python.org Python.org] (of course!),
     11[http://conda.pydata.org/miniconda.html Continuum Miniconda],
     12[https://www.continuum.io/downloads Continuum Anaconda],
     13or, if you're on Linux, the Python from your package manager.
     14
     15Vespa's dependencies are listed in the file
     16[source:/trunk/vespa/requirements.txt vespa/requirements.txt]. If you're
     17familiar with `pip`, please note that it's '''not''' possible to use this
     18requirements file with pip's `-r` option because not all of Vespa's
     19dependencies are pip-friendly.
     20
     21== Check Your Python ==
     22
     23Download [export:/trunk/get_python_info.py get_python_info.py] and
     24run it with this command --
     25{{{
     26python get_python_info.py 1
     27}}}
     28
     29It will tell you whether or not Vespa is compatible with your Python.
     30
     31== Upgrade pip ==
     32
     33Ensure you're running the latest version of pip. (Older versions of pip
     34have trouble with some of the steps in this guide.)
     35
     36{{{
     37python -m pip install --upgrade pip
     38}}}
     39
     40== Install Vespa's Easy Dependencies ==
     41
     42These packages are easy to install with pip on all the platforms
     43Vespa supports --
     44{{{
     45pip install packaging pydicom==0.9.9 pygamma hlsvdpro vespa-suite
     46}}}
     47
     48== Install Numpy, Scipy, and Matplotlib ==
     49
     50Under '''OS X and Linux''', you can install Numpy, Scipy, and matplotlib via `pip` (on OS X) and the
     51package manager (on Linux).
     52
     53On '''Windows''', we suggest installing Numpy and Scipy from the
     54[http://www.lfd.uci.edu/~gohlke/pythonlibs/ wheels provided by Christoph Gohlke].
     55Once Numpy and Scipy are installed, you can `pip install matplotlib`.
     56
     57== Install PyWavelets ==
     58
     59Under '''Windows''', you can install this with `pip install pywavelets==0.3.0`.
     60
     61Under '''Linux''', get it from your package manager.
     62
     63Under '''OS X''', use the
     64[https://scion.duhs.duke.edu/vespa/analysis/wiki/PyWaveletsOsx PyWavelets 0.3.0 wheel that we provide].
     65
     66== Install wxPython ==
     67
     68Under '''Linux''', get wxPython 3 if your package manager offers it, or
     69wxPython 2.8 otherwise. It might be under a name like 'python-wxgtk'.
     70
     71Under '''Windows and OS X''', install wxPython 3 from one of the packages
     72provided on [http://wxPython.org wxPython.org].
     73
     74== Install FFTW3 (Linux Only) ==
     75
     76'''Linux''' users, use your package manager to install the FFTW3 runtime
     77library. Look for it under a name like `libfftw3` or just `fftw`.
     78
     79
     80== Finish the Vespa Install ==
     81
     82Run this command --
     83{{{
     84python -m vespa.check_dependencies
     85}}}
     86
     87If that goes OK, run this command --
     88{{{
     89python -m vespa.create_shortcuts
     90}}}