VERAView is a GUI for visualization and engineering analyses of output data from VERA (Virtual Environment for Reactor Applications). Implemented in Python, it provides instantaneous 2D and 3D images, 1D plots, and alpha-numeric data from VERA multi-physics simulations.
VERAView is under rapid development, and significant changes in the software are expected. However, the snapshots in the repository represent comparatively stable versions intended for use. Bugs and feature requests can be entered using the Issues tab on this page or reported via email to
There are two primary ways to access and install VERAView: running an application installer, and updating from the repository. Installers for the latest stable (relatively speaking) version are available for download as follows:
Note installers will not be updated as frequently as the repository. So, if you want to track development versions, you'll need to install from the repository.
How to get VERAView from the repository and set it up for execution are described below.
The executable (link above) is an installer that does not require elevated privileges. By default it will install in your home folder, but you can install to any folder in which you have permissions. Execute the installer and follow the prompts.
The link above is for a disk image that can be mounted by opening the file in Finder. The resulting mount window will have an Applications folder as well as the VERAView application icon. Install VERAView by dragging the icon onto the Applications folder icon.
If you are a Admin user, VERAView will be installed in the root /Applications
folder. Otherwise, it will be installed in the $HOME/Applications
.
Linked above is a shell script that must be executed as follows:
$ bash VERAView-2.4.3-Linux-x86_64.sh
You can execute as root for a system wide install to any location. Otherwise, the default is to install in your home directory.
The easiest path for tracking repository updates is to install the pre-built Miniconda2 environment for your platform and extracting it in you home directory/folder. The launcher script files will look for Python there.
Download the Miniconda2 environment
zip file. Extract it in
your folder using File Explorer. Clone or pull from this repository to obtain
the latest VERAView version. Launch using the batch file vvconda.run.bat
in
the root VERAView folder.
Download the Miniconda2 environment gzipped tarball. Extract it in your home directory:
$ tar xvfz miniconda2-vview-macos.tar.gz
Clone or pull from this repository to obtain the latest VERAView version.
Launch using the script file vvconda.run.sh
in the root VERAView directory.
Download the Miniconda2 environment gzipped tarball. Extract it in your home directory:
$ tar xvfz miniconda2-vview-linux.tar.gz
Clone or pull from this repository to obtain the latest VERAView version.
Launch using the script file vvconda.run.sh
in the root VERAView directory.
You must have the following packages required by VERAView (minimum version numbers in parens):
- numpy (1.9.3)
- hp5y (2.5.0)
- scipy (0.16.0)
- matplotlib (1.4.3)
- pillow
- wxPython (3.0)
- mayavi (4.4.0)
Use the Conda command-line package manager or the Anaconda Navigator to install
any missing packages. For Windows, conda.exe
is in the scripts\
subfolder
under your Anaconda2 or Miniconda2 installation. For Mac and Linux it is the
conda
script in the bin/
subdir. The following command illustrates
installation of all required packages via the command line (Command Prompt for Windows)
conda install -y numpy=1.9.3 h5py=2.5.0 scipy=0.16.0 matplotlib pillow wxpython=3.0 mayavi=4.4.0
If you have installed Anaconda2 or Miniconda2 to a custom location, it might be necessary to edit the VERAView launcher script. This is described in platform-specific sections below.
If you have already installed Anaconda2 or Miniconda2 and want to use that Python environment, refer to the Installation to an Existing Anaconda2 or Miniconda2 Environment section above. Otherwise, you can choose to use the single-click application installer or download a pre-built Miniconda2 environment containing all packages needed by VERAView.
A Windows application installer that will install VERAView and the required
Python environment (Continuum's Miniconda2) is available
here.
The installer does not require elevated privileges and will install by default
in a VERAView
subfolder in your home folder.
With this option no other steps are required to install VERAView. After
completing the installation, you launch VERAView by executing (i.e.,
double-clicking in File Explorer or Windows Explorer) the file
vvconda.run.exe
in the VERAView
folder. Also, Windows 10 users should see
a CASL VERAView Start menu folder containing a VERAView launcher shortcut.
If you prefer not to use the application installer or plan on frequently updating VERAView from the repository, you will save time by using an existing Anaconda2 or Miniconda2 environment or installing one.
A zip file containing the Miniconda2 environment with packages required by
VERAView can be downloaded from
here. Extract it
in your home folder (usually c:\Users\username
but determined by the
%userprofile%
environment variable). This should result in a Miniconda2\
subfolder in your home folder. If you extract it to a location other than your
home folder, it will be necessary to edit the launch script as described in
Launch VERAView below.
Click the green Clone or download button above to see the git
command
that will clone the VERAView repostory. This should be run in a Command Prompt
window and will create a VERAView\
subfolder in the current folder.
Alternatively, download the zip file and extract it in a folder of choice,
renaming the VERAView-master
subfolder to VERAView
.
A vvconda.run.bat
script is provided in the VERAView\
folder for launching
VERAView. If you installed Anaconda or Miniconda in a non-default location you
will need to edit vvconda.run.bat
to change the line
set CondaDir=%VERAViewDir%Miniconda2
to point to your Anaconda2 or Miniconda2 installation.
If you have already installed Anaconda2 or Miniconda2 and want to use that Python environment, refer to the Installation to an Existing Anaconda2 or Miniconda2 Environment section above. Otherwise, you can choose to use an application disk image or download a pre-built Miniconda2 environment containing all packages required by VERAView.
An application disk image containing VERAView with the required Python
environment is available
here.
Once the image is mounted, you may copy the VERAView.app
file to
$HOME/Applications
, or if you are an administrator you may copy it to
/Applications.
With this option no other steps are required to install VERAView. After
completing the installation, you launch VERAView by double-clicking VERAView
under Applications in Finder or by executing the VERAView
script under
VERAView.app/Contents/MacOS
in the applications directory.
If you prefer not to use the application disk image or plan on frequently updating VERAView from the repository, you will save time by using an existing Anaconda2 or Miniconda2 environment or installing one.
A gzipped tar file containing the Miniconda2 environment with packages required
by VERAView can be downloaded from
here. Extract it
in your home directory (usually /Users/username
but determined by the $HOME
environment variable). This should result in a miniconda2/
subdir in your
home directory. If you extract it to a location other than your home
directory, it will be necessary to edit the launch script as described in
Launch VERAView below.
Click the green Clone or download button above to see the git
command
needed to clone from the repostory. This should be run in a Bash shell
and will create the VERAView/
subdir in the current directory.
Alternatively, download the zip file and extract it in a directory of choice,
renaming the VERAView-master
subdir to VERAView
.
A script is provided in the VERAView/
subdir for launching VERAView,
vvconda.run.sh
. If you installed Anaconda or Miniconda in a non-default
location you will need to edit vvconda.run.sh
to change the line
CondaBinDir="${VERAViewDir}/miniconda2/bin"
to point to your Anaconda2 or Miniconda2 installation.
If you have already installed Anaconda2 or Miniconda2 and want to use that Python environment, refer to the Installation to an Existing Anaconda2 or Miniconda2 Environment section above. Otherwise, you can choose to use an application installer script or download a pre-built Miniconda2 environment containing all packages required by VERAView.
An installer script containing VERAView with the required Python
environment is available
here.
The installer does not require root privileges but can be installed as root.
When run as root, the default installation path is /usr/local/veraview-2.3.0
.
Otherwise the default install path is $HOME/veraview-2.3.0
. Execute the
installer with
$ bash VERAView-2.3.0-Linux-x86_64.sh
With this option no other steps are required to install VERAView. After
completing the installation and depending on you desktop environment, VERAView
might appear in your desktop Applications
menu (usually in the Other
folder), in which case you can launch by choosing it from the menu.
Alternatively, execute the vvconda.run.sh
script in the veraview-2.3.0
installation folder.
If you prefer not to use the application installer script or plan on frequently updating VERAView from the repository, you will save time by using an existing Anaconda2 or Miniconda2 environment or installing one.
A gzipped tar file containing the Miniconda2 environment with packages required
by VERAView can be downloaded from
here. Extract it
in your home directory (usually /home/username
but determined by the $HOME
environment variable). This should result in a miniconda2/
subdir in your
home directory. If you extract it to a location other than your home
directory, it will be necessary to edit the launch script as described in
Launch VERAView below.
Click the green Clone or download button above to see the git
command
needed to clone from the repostory. This should be run in a Bash shell
and will create the VERAView/
subdir in the current directory.
Alternatively, download the zip file and extract it in a directory of choice,
renaming the VERAView-master
subdir to VERAView
.
A script is provided in the VERAView/
subdir for launching VERAView,
vvconda.run.sh
. If you installed Anaconda or Miniconda in a non-default
location you will need to edit vvconda.run.sh
to change the line
CondaBinDir="${VERAViewDir}/miniconda2/bin"
to point to your Anaconda2 or Miniconda2 installation.
After an initial clone from the repository, execute
git pull
in your VERAView repository directory to update with the latest changes.
If you chose to download a zip file, you will need re-download the latest zip
file and extract it. Before renaming the extracted VERAView-master
subdir/subfolder to VERAView
, either delete your existing VERAView
folder
or rename it in order to preserve it.
The easiest way to migrate to using the repository after an installer is to follow these steps after downloading a zip.
- Move the
Miniconda2
subfolder orminiconda2
subdir in your installation to your home folder or directory.
Windows: The Miniconda2
subfolder is in the root of the VERAView
installation directory.
Mac: The miniconda2
subdir is in the VERAView.app/Contents/MacOS
directory in the applications directory where VERAView is installed, which
usually will be either /Applications
or $HOME/Applications
.
Linux: The miniconda2
subdir is the in the root of the VERAView
installation.
- Extract the zip file.
Windows: Extract in the folder above (containing) your VERAView installation directory.
Mac: Extract in the VERAView.app/Contents
directory in the applications
directory where VERAView is installed. Either rename or remove your
existing MacOS
directory. Rename the extracted veraview
subdirectory to
MacOS
.
Linux: Extract in the directory above (containing) your VERAView installation directory.
- Copy files.
Windows: Copy all the files in the extracted VERAView-master
folder into
your installed VERAView
folder. You can do this with File Explorer or
with the robocopy
in a Command Prompt window:
> robocopy VERAView-master VERAView * /e
Mac: Since application files are in the MacOS
subdir under
VERAView.app/Contents
the files in the newly extracted VERAView-master
subdirectory must copied into MacOS
. For example:
$ cd .../VERAView.app/Contents/VERAView-master
$ cp -r * ../MacOS
Linux: Copy all the files in the extracted VERAView-master
directory into
your installed veraview
folder. For example:
$ cd .../VERAView-master
$ cp -r * ../veraview
VERAView is specifically designed to interpret the output data from VERA codes. The VERAOut specification specifies how reactor data is structured in an HDF5 file. Understanding this structure, VERAView implicitly displays the data in the form of simplified pressurized water reactor (PWR) geometry.
-
Though VERAView can only process files in the VERAOut format, it is not connected directly to VERA codes and does not require any specific data from the physics software. It is currently limited to common physical geometries such as fuel rods, coolant channels, fuel assemblies, etc., and does not display any more specific information (though this may be an option in the future). The result is that data from any reactor methods can be displayed by first converting the data to the VERAOut format. Numerous converting codes have been created for this purpose.
-
VERAView is designed to be used with nearly no experience or training. While applications such as ParaView and VisIt are powerful, general-purpose data analysis and visualization tools, setting up data sources and rendering pipelines requires a fair amount of expertise and can be complex. VERAView provides an alternative, simplified interaction for engineers and students that is designed from the beginning for reactor analysis.
-
VERAView is to be a true multi-physics analysis tool, specifically for the simultaneous visualization of neutronics, thermal-hydraulics, and fuel performance data. Generally the fidelity of the displayed data is in the form of fuel rod or coolant channel quantities, but VERAView can also provide coarser quantities derived from rods and channels (such as assemblies or axial distributions), and can also display in-core detector data as well. Currently VERAView does not support methods-specific geometries, meshes, or intra-pin distributions, though these features could be added in the future.
-
The purpose of VERAView is to go beyond visualization and provide numeric data for engineering analyses. Features such as labels, plots, tool tips, image extraction, and comma separated values (CSV) file export allow VERAView to be extremely useful for interactions with Microsoft Excel and PowerPoint. Functions such as finding the maximum values have already been implemented, and there is an initial capability for displaying differences between files.
-
VERAView allows inspection of data in multiple dimensions, and visualization by the user in different perspectives, to accommodate various types of reactor data and quantities. It also incorporates time (or exposure) as a fourth dimension into this philosophy in a seamless manner, so that the user can discover critical aspects of data as a function of space and/or time, quickly able to observe trends and distributions.
-
The analysis capabilities are designed to be extensible to new codes, features, or data as needed. This is accomplished through the use of the custom widgets, which could be added by developers or even implemented and integrated by users.
-
VERAView is intended to be executed locally on the user’s personal computer, and is supported on Windows, Mac OS X, and Linux.
VERAView expects a Python-2.7 runtime environment with packages h5py, matplotlib, mayavi, numpy, pillow, scipy, and wxPython. Although it is possible to install each of these packages in a custom Python environment, VERAView is developed and tested with the Anaconda Python-2 environment. Our experience is that Anaconda provides a consistent environment on the three platforms, thus Anaconda is our recommended runtime environment, and we only provide instructions for it. The Anaconda2 or Miniconda2 environment provides all the necessary requirements as pre-built packages, but packages required by VERAView must be installed.