Build and Install¶

Install binary packages¶

We provide Conda binary packages for some cameras. This is, by far, the easiest way to get started with LImA! For instance:

Install first lastest miniconda3 (https://docs.conda.io/en/latest/miniconda.html)

Install mamba package in your “base” environment to speed up your future installations, the default conda installer is very slow, so we prefer to use mamba:

conda install mamba -c conda-forge

Install now the Lima camera package (e.g basler) at the same time you create the new environment for your Lima installation:

mamba create -n basler -c conda-forge -c esrf-bcu lima-camera-basler

would install a fully loaded Lima and all its dependencies with the Basler camera plugin and SDK. The camera comes as a python module but is also C++ development package that includes header files and CMake package config files.

If you need to run the Python Tango device server you should install the Tango camera package:

mamba create -n basler -c conda-forge -c esrf-bcu lima-camera-basler-tango

Build from source¶

First, you need to Get the Source. Two methods are provided to build LImA from source:

• using our install script that aims to hide the complexity of CMake;

• using CMake directly for developers who are already acquainted with the tool and need the extra flexibility.

Using scripts¶

The install scripts will run CMake to compile and/or install.

It accepts input arguments (see below) but it also uses a configuration file scripts/config.txt. Feel free to update this file for setting a permanent configuration for your own installation.

For Linux:

[sudo] install.sh
[--git]
[--install-prefix=<desired installation path>]
[--install-python-prefix=<desired python installation path>]
[options]

For Windows:

install.bat
[--install-prefix=<desired installation path>]
[--install-python-prefix=<desired python installation path>]
[options]

The --git (Linux only) option can be used to clone the required submodules as a prerequisite. Otherwise you should install the submodules manually with git commands, for instance:

$ git submodule init third-party/Processlib
$ git submodule init camera/basler
$ git submodule init applications/tango/python
$ git submodule update

Options are <camera-name> <saving-format> python pytango-server:

<camera-name> can be a combination of any of the following options:

andor|andor3|basler|prosilica|adsc|mythen3|ueye|xh|xspress3|ultra|
xpad|mythen|pco|marccd|pointgrey|imxpad|dexela|merlin|v4l2|
eiger|pixirad|hexitec|aviex|roperscientific|rayonixhs|espia|maxipix|frelon

<saving-format> can be a combination of any of the following options:

cbf|nxs|fits|edfgz|edflz4|tiff|hdf5|hdf5jp2k

python will install the python module

pytango-server will install the PyTango server

For example, to install the Basler camera, use the TIFF output format, the python binding and the TANGO server, one would run:

$ sudo install.sh --git --install-prefix=./install --install-python-prefix=./install/python tiff basler python pytango-server

Using CMake¶

Install first the project submodules:

git submodule init third-party/Processlib
git submodule init camera/basler
git submodule init applications/tango/python
git submodule update

Run cmake in the build directory:

mkdir build
cd build
cmake ..
   [-G "Visual Studio 15 2017 Win64" | -G "Visual Studio 15 2017" | -G "Unix Makefiles"]
   [-DCMAKE_INSTALL_PREFIX=<desired installation path>]
   -DLIMA_ENABLE_TIFF=true
   -DLIMACAMERA_BASLER=true
   -DLIMA_ENABLE_PYTANGO_SERVER=true
   -DLIMA_ENABLE_PYTHON=true

HDF5 JPEG 2000 saving can be enabled with:

cmake ..
   -DLIMA_ENABLE_HDF5=true
   -DLIMA_ENABLE_HDF5_JP2K=true

OpenJPEG is required when LIMA_ENABLE_HDF5_JP2K is enabled. If it is not installed in a standard location, point CMake to it with either OPENJPEG_ROOT or the explicit OPENJPEG_INCLUDE_DIR and OPENJPEG_LIBRARY variables:

cmake ..
   -DLIMA_ENABLE_HDF5=true
   -DLIMA_ENABLE_HDF5_JP2K=true
   -DOPENJPEG_ROOT=/path/to/openjpeg

Kakadu support is optional and must be explicitly enabled. When enabled and found, it can be selected at run time as the JPEG 2000 encoder. It can be discovered with KAKADU_ROOT or with the explicit Kakadu include and library variables:

cmake ..
   -DLIMA_ENABLE_HDF5=true
   -DLIMA_ENABLE_HDF5_JP2K=true
   -DLIMA_ENABLE_KAKADU_JP2K=true
   -DKAKADU_ROOT=/path/to/kakadu

The same paths can also be provided through environment variables, for example:

export OPENJPEG_ROOT=/path/to/openjpeg
export KAKADU_ROOT=/path/to/kakadu
cmake .. -DLIMA_ENABLE_HDF5=true -DLIMA_ENABLE_HDF5_JP2K=true -DLIMA_ENABLE_KAKADU_JP2K=true

Files written with HDF5JP2K use a custom HDF5 filter. Applications that read the files outside the Lima process, such as h5py or silx, must load a matching HDF5 filter plugin. The reader plugin is expected to be distributed outside Lima, for example through hdf5plugin.

Then compile and install:

cmake --build
sudo cmake --build --target install

Environment Setup¶

If the install path for libraries and python modules are not the default, you need to update your environment variables as follow:

For Linux:

export LD_LIBRARY_PATH=$LD_LIBRARY_PATH:<my-custom-install-dir>/Lima/lib
export PYTHONPATH=$PYTHONPATH:<my-custom-install-dir>

For Windows:

set PATH=%PATH%;<my-custom-install-dir>\Lima\lib
set PYTHONPATH=%PYTHONPATH%;<my-custom-install-dir>

or update the system wide variables PATH for the libraries and PYTHONPATH for python.