CrystalPlan

Experiment planning tool for elastic neutron diffraction single-crystal experiments.

Installing Locally for Development

The prerequisites for installing CrystalPlan are that you need the following libraries available on your machine:

- vtk
- wxpython == 3.0
- xtensor
- xtensor-python
- pybind11
- Mayavi
- gcc > 4.8 (c++14 support)

The easiest way to get all of these dependencies in one place is to use a conda env. The code for this project can be installed using pip from the git repo:

pip install -e .  

This will install the required python dependencies and compile the C++ extensions.

Running

CrystalPlan can be run by executing the following in the top level folder of the git repo.

python crystalplan.py

Running using Docker

The easiest way to run CrystalPlan is to grab the docker container and run it using xhost:

docker pull samueljackson92/crystal-plan

On OSX you can then run:

docker run --rm -it -e DISPLAY=docker.for.mac.localhost:0 --privileged -v /tmp/.X11-unix:/tmp/.X11-unix crystal-plan

The command for linux may differ a little in the arguments supplied to enable X11 forwarding from the container.

Packaging

For packaging CrystalPlan for distribution to users and IDAaaS the PyInstaller library is used. This tool finds all of the dependencies for the project and packs them into a single folder that can be zipped and sent to the user. PyInstaller supports all three operating systems and a different .spec file is included in the repo for each one.

CentOS / RHEL7

The easiest way is to grab a copy of the crystal-plan-package docker image. Running this will generate a fully packaged version of CrystalPlan that can be shipped to IDAaaS, isis-compute, or directly to end users.

The docker container can be run as follows:

docker run -v $PWD:/CrystalPlan crystal-plan-dev

Where $PWD is the top level of your local crystal plan git repo. This will be mounted to the container and then built for CentOS. The output of PyInstaller will be directed into ./dist. The packaged application can be run by extracting the folder from the tar ball and executing the crystalplan binary.

OSX

You'll need install all the relevant dependencies to a windows machine. The easiest way I've found to do this is with conda where you can conda install pretty much everything you need. You'll also need a compiler capable of compiling at least C++14.

Then you're ready to python setup.py build_ext in the git repo to build the C++ extensions. Once you've done this you should be able to run CrystalPlan on OSX.

pythonw crystalplan.py

To package it for OSX you can run PyInstaller directly:

pyinstaller crystalplan-osx.spec -y

This will output the package application into the dist folder. To create a .dmg for crystal plan you can use the dmgbuild tool (available on PyPI). Then you can run:

dmgbuild -s ../settings.py "CrystalPlan" crystal-plan-osx.dmg

Which will output a dmg for CrystalPlan that can be shared.

Windows

You'll need install all the relevant dependencies to a windows machine. The easiest way I've found to do this is with conda where you can conda install pretty much everything you need. You'll also need to install MS visual studio

= 2015 and run the following commands in the command prompt prior to building the C++ extensions for crystal plan.

"%VS140COMNTOOLS%\..\..\VC\vcvarsall.bat" x64
set DISTUTILS_USE_SDK=1
set MSSdk=1

Then you're ready to python setup.py build_ext in the git repo to build the C++ extensions. Once you've done this you should be able to run CrystalPlan on Windows:

python crystalplan.py

To package it for windows you can run PyInstaller directly:

pyinstaller crystalplan-windows.spec -y

The output can be found in the dist folder. The contents is self contained and can be zipped and sent instrument scientists and other users.