This repository contains scripts and configuration to create development environment for Bodylight tool chain. Base box is minimal CENTOS 9 stream (RHEL derivative) and Vagrant scripts prepare the rest of virtual machine with recent Bodylight.js, OpenModelica (nightly 1.20 build), Python 3.x, Jupyter, Anaconda, http server. See web apps after installation at http://localhost:8080.
A binary snapshots are sometimes created usually in yearly (or less) bases.
All the tools are provided as is - usually in best effort or beta version quality. If it works, than it works, if something is wrong, see logs, console logs, submit issue.
Vagrant tool automates configuration (port forwarding, secure ssh keys, shared folders) and provisioning of virtual machine and creates exemplar configuration. Thus preventing excuses: works on my machine
.
Requirement:
- HW: minimum 1 CPU, 4 GB RAM, min 5GB disk space.
- OS: Any OS supported by VirtualBox and Vagrant tool (succesfully tested on Windows 7,Windows 10, Ubuntu 16.04, ...)
- SW: Install VirtualBox, succesfully tested with version Virtualbox 6.1.38 (( Note we experienced issue VERR_NEM_VM_CREATE_FAILED - need to disable Windows features (V-Host) and see howto setup Windows 10 at https://forums.virtualbox.org/viewtopic.php?f=6&t=93712))
- SW: Install Vagrant tested version 2.2.6 and 2.2.9
- SW: GIT (EITHER Install GIT any version OR you may download master.zip and unzip instead of
git clone
from the URLs bellow)
Some OS has their own distribution of vagrant
, virtualbox
and git
, so you may try to use it: yum install vagrant virtualbox git
OR apt install vagrant virtualbox git
.
This is not required step, it adds demo Jupyter notebooks and Physiolibrary-models next to the Bodylight-VirtualMachine on host machine. They will appear as /vagrant_data in virtual machine and will be available for jupyter notebook after installation.
In command-line (Linux xterm
, bash
etc. for Windows Start-> type 'cmd' -> choos 'Command Prompt'
) do:
git clone https://github.com/creative-connections/Bodylight-notebooks.git
git clone https://github.com/creative-connections/Physiolibrary-models.git
Choose either installation from sources or binary VM installation.
- replace
Vagrantfile
withVagrantfile.preinstalled
where scripts are modified to download prepared image and configure the machine only. In linux BASH:
git clone https://github.com/creative-connections/Bodylight-VirtualMachine.git
cd Bodylight-VirtualMachine
cp Vagrantfile.preinstalled Vagrantfile
or in windows command
git clone https://github.com/creative-connections/Bodylight-VirtualMachine.git
cd Bodylight-VirtualMachine
copy /Y Vagrantfile.preinstalled Vagrantfile
- start "vagrant up" by
vagrant up
Clone repository with Virtual machine scripts and run vagrant up (In command-line (Linux xterm
, bash
etc. for Windows Start-> type 'cmd' -> choos 'Command Prompt'
) do)
git clone https://github.com/creative-connections/Bodylight-VirtualMachine.git
cd Bodylight-VirtualMachine
vagrant up
If you do not have git
then you may download and unzip manually the https://github.com/creative-connections/Bodylight-VirtualMachine/archive/master.zip and do vagrant up
in the unzipped directory.
The first vagrant up
takes 15-45 mins (or more depending on network speed) and The bootstrap scripts downloads, installs and configures all required software, ~500 MB of Anaconda (distribution of Python and Jupyter), ~700 MB OpenModelica, ~100 MB Julia and other packages. You may disable some bootstrap script by commenting them in Vagrantfile
. You should see success:
...
default: [yLpj] (webpack)/buildin/global.js 472 bytes {0} [built]
default: + exit 0
1.5 GB of depended packages (OpenModelica, Anaconda, Julia) are downloaded and persisted in host /cache
subdirectory during installation.
After several minutes the VM is installed and configured. Port forwarding is done from guest VM 80 to host 8080 by default, refer Vagrantfile for exact port number. Refer default page at http://localhost:8080
The default installation contains these applications, some available from web interface:
- Jupyter notebook with Modelica kernel, link http://localhost:8080/jupyter/ [1]
- OpenModelica (v 1.14) - use e.g.
OMEdit
[2] - Python v3 - including full Anaconda environment
- Bodylight.js-components - HTML web components - documentation and usage at http://localhost:8080/components/
- Bodylight.js-FMU-Compiler refer http://localhost:8080/compiler/
- Bodylight.js-Composer - refer http://localhost:8080/composer/
- Bodylight-Scenarios - Scenarios written in MD using web components http://localhost:8080/scenarios/
- Bodylight-VirtualBody - Virtualbody using WebGL http://localhost:8080/virtualbody/
- Visual Studio Code - Lightweight IDE for Linux - just type
code
within terminal window [7]
You may access virtual desktop using VirtualBox capabilities or you may access using vagrant ssh
.
If you don't need VM, you can halt it using Virtualbox UI or using
vagrant halt
If you need VM, you can boot it again using:
vagrant up
To restart VM
vagrant reload
The second and other vagrant up
is rapid and should take couple of seconds, because time consuming provisioning (bootstrap scripts) is already done.
If you have previously installed VM and would like to update or reinstall from scratch, do:
- Save any documents/data from VM to shared folder
/vagrant
folder. These are available and preserved in host. Other files and data will be erased. - Consider to delete
/cache
directory in host machine. It contains some installation packages, if it is not too old you may preserve it and go to next step. - Then do following:
vagrant destroy
vagrant box update
git pull
# if you made some local changes to Vagrantfile - git pull may fail,
# try: git stash;git pull;git stash apply
vagrant up
This will clean VM,check and update the base box and install the software again - if /cache
is present from previous installation it will use most packages from it rather to download again from Internet repositories.
Installation scripts are preserved in /scripts
directory, they are launched only when vagrant up
is made first or when provisioning is explicitly requested by vagrant up --provision
.
To stop VM.
vagrant halt
To start VM again - it wil start quickly as bootstrap is not needed
vagrant up
To restart VM
vagrant reload
To destroy VM and remove all VM files do the following. The files stored in shared folders /vagrant and /vagrant_data are preserved.
vagrant destroy
There are preserved installation binaries in /cache
directory. Delete cache if you don't need it anymore
rm -rf cache
References:
- [1] https://github.com/OpenModelica/jupyter-openmodelica
- [2] https://openmodelica.org/
- [3] vagrant https://vagrantup.com
- [4] virtualbox https://www.virtualbox.com
- [5] Anaconda https://anaconda.org/
- [6] GIT https://git-scm.com
- [7] Visual Studio Code https://code.visualstudio.com/