Skip to content

Latest commit

 

History

History
304 lines (214 loc) · 9.37 KB

installation.md

File metadata and controls

304 lines (214 loc) · 9.37 KB

Installation

Table of contents

Installing with pip

GENFIRE is hosted on PyPi, and thus can be easily installed with pip:

pip3 install genfire

You may need root privileges (sudo), depending on your environment. If this pip installation is successful, you can now open the GUI by running the launch.py script inside of the package

python3 /path/to/genfire/gui/launch.py

Alternatively, on Mac/Linux you should have access to the shortcut genfire which does the same thing (If anybody knows a reliable way to create a similar, installable shortcut command on Windows, I would love to hear about it)

genfire

Troubleshooting pip

If an error occurs when using pip to install the dependencies for GENFIRE, you may try manually installing that dependency with pip and then trying to install GENFIRE. For example, if you receive an error about PyQt5 not being installed, try manually installing it with pip install PyQt5. If this still fails, you can try instead installing with setup.py, as described below.

Installing with setup.py

If installing with pip fails, you can also try installing the package manually

Get the Source Code

To install GENFIRE as a python package, first download the source code. Then follow the procedure for your operating system.

Mac OS X

Python is preinstalled on Mac OS X, but it is generally a bad idea to alter the system python in /usr/bin as some programs depend on it. I would highly recommend using the package manager homebrew to create a python environment for you. The following commands will install homebrew and python3

$ /usr/bin/ruby -e "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/master/install)"
$ brew install python3

Once that's done, navigate to the top source code directory. That's the same folder as requirements.txt and setup.py

$ cd /path/to/GENFIRE

Install (most of the) dependencies using pip3, you may need to prepend the command with sudo depending on your system

$ pip3 install --upgrade pip
$ pip3 install -r requirements.txt

You will also need sip and PyQt5

$ brew install sip
$ brew install pyqt

You should now have all the dependencies necessary to run GENFIRE. To install it, make sure you are in the directory with setup.py and enter the follow. Again, you may need sudo

$ python3 setup.py install

If everything worked, you can now launch the gui from the command line with

$ genfire

You can also use the code as any other python package. Here's an example of launching the gui from within python

$ python3
import GENFIRE
GENFIRE.gui.launch.main()

(Optional) Turbo-charge GENFIRE with FFTW

GENFIRE can make use of pyFFTW, which wraps the FFTW library. I have tested a number of FFT routines including those in NumPy, SciPy, and pyFFTW, and found this to be the fastest one by a factor of 2-3.

The following are details for installing FFTW from source, but recently pip has begun to support precomiled libraries for the package, so you should first try the easy way with

pip3 install pyfftw

if this fails, then proceed with building the FFTW libraries by hand and then reinvoking pip as described below

In order for GENFIRE to use pyFFTW you must install the FFTW libraries. Download the source code, decompress it, and navigate into the unpacked directory from your terminal. pyFFTW needs FFTW to be compiled for all precision types, so you have to compile it three times. Use the following commands

$ ./configure --enable-threads --enable-shared
$ make
$ sudo make install
$ ./configure --enable threads --enable-shared --enable-float
$ make
$ sudo make install
$ ./configure --enable-threads --enable-shared --enable-long-double
$ make
$ sudo make install

Now, install pyFFTW with pip and and test that it worked

$ pip3 install pyfftw
$ python -c "import pyfftw"

If you don't receive an error, then it was successful. If you get an error along the lines of "ImportError: libfftw3l.so cannot open shared object file" then you need to set your DYLD_LIBRARY_PATH environmental variable so that pyFFTW can find the libraries

$ export DYLD_LIBRARY_PATH=/usr/local/lib:$DYLD_LIBRARY_PATH<

If the FFTW .so files are somewhere other than /usr/local/lib, then you should replace that part appropriately. To make this change permanent add the above line to the end of your ~/.bash_profile

Linux (Ubuntu 14.04)

Navigate to the source code directory. That's the same folder as requirements.txt and setup.py

$ cd /path/to/GENFIRE

Install (most of the) dependencies using pip

$ pip3 install --upgrade pip
$ sudo pip3 install -r requirements.txt

You will also need sip and PyQt5 this command should install both.

$ sudo apt-get install python3-pyqt5

You should now have all the dependencies necessary to run GENFIRE. To install it, make sure you are in the directory with setup.py and enter

$ sudo python setup.py install

If everything worked, you can now launch the gui from the command line with

$ genfire

You can also use the code as any other python package. Here's an example of launching the gui from within python

$ python3
import GENFIRE<br>
GENFIRE.gui.launch.main()

(Optional) Turbo-charge GENFIRE with FFTW

GENFIRE can make use of pyFFTW, which wraps the FFTW library. I have tested a number of FFT routines including those in NumPy, SciPy, and pyFFTW, and found this to be the fastest one by a factor of 2-3.

The following are details for installing FFTW from source, but recently pip has begun to support precomiled libraries for the package, so you should first try the easy way with

pip3 install pyfftw

if this fails, then proceed with building the FFTW libraries by hand and then reinvoking pip as described below

In order for GENFIRE to use pyFFTW you must install the FFTW libraries. Download the source code, decompress it, and navigate into the unpacked directory from your terminal. pyFFTW needs FFTW to be compiled for all precision types, so you have to compile it three times. Use the following commands

$ ./configure --enable-threads --enable-shared
$ make
$ sudo make install
$ ./configure --enable-threads --enable-shared --enable-float
$ make
$ sudo make install
$ ./configure --enable threads --enable-shared --enable-long-double
$ make
$ sudo make install

If you don't receive an error, then it was successful. If you get an error along the lines of "ImportError: libfftw3l.so cannot open shared object file" then you need to set your LD_LIBRARY_PATH environmental variable so that pyFFTW can find the libraries

$ export LD_LIBRARY_PATH=/usr/local/lib:$LD_LIBRARY_PATH<

If the FFTW .so files are somewhere other than /usr/local/lib, then you should replace that part appropriately. To make this change permanent add the above line to the end of your ~/.bashrc

Windows 10

For python3 on Windows 10 I would recommend using Anaconda from Continuum Analytics. It's a distribution of python that contains most of the packages used by GENFIRE all wrapped into a simple-to-use installer. Note you must use python3. Once you have python setup, open up a cmd prompt and navigate to the source directory. That's the same folder as requirements.txt and setup.py

$ C:\path\to\Anaconda\python3 setup.py install

If everything worked you can now launch the gui.oo

$ C:\path\to\Anaconda\python GENFIRE\gui\launch.py

You can also use the code as any other python package. Here's an example of launching the gui from within python

$ C:\path\to\Anaconda\python3
import GENFIRE
GENFIRE.gui.launch.main()

(Optional) Turbo-charge GENFIRE with FFTW

GENFIRE can make use of pyFFTW, which wraps the FFTW library. I have tested a number of FFT routines including those in NumPy, SciPy, and pyFFTW, and found this to be the fastest one by a factor of 2-3.

The following are details for installing FFTW from source, but recently pip has begun to support precomiled libraries for the package, so you should first try the easy way with

$ C:\path\to\Anaconda\Scripts\pip3 install pyfftw

if this fails, then consult the FFTW documentation

Installation Troubleshooting

If you have trouble installing PyQt5 or sip, consult their documentation If you have some problem with the "pip3 install -r requirements.txt" step, you can view the requirements.txt file to see the packages that are necessary, and try to install them one-by-one.