Skip to content

Installation

ODAM: Installation


The ODAM software can be deployed at multiple scales (local, intranet, internet), depending on the need and the target community.


Deployment possibilities depending on the data storage medium




Local installation : Deployment of ODAM with full functionality

This type of deployment is reserved more specifically for computers running under MS Windows 10 and Apple MacOS 10.x.

All the steps involved in the installation of the virtual machine have been published and are available online on the INRAE Data website.

ODAM Virtual Disk Image (VDI) for Oracle VM VirtualBox - Zipped with 7zip.

https://doi.org/10.15454/C9LAEF, Portail Data INRAE, V1

The three main steps are shown below:


Overview of the ODAM software deployment on your PC (Windows 10, MacOS 10.x)



HOWTO

  1. Download the virtual machine file, to be unzipped with 7zip. Then follow the installation guide mentioned above. This requires prior installation of Oracle VM VirtualBox software
  2. Data collection and preparation
  3. Using ODAM

Local installation : ODAM API on Windows

Deploy the ODAM API with a very lightweight local web server for Windows.

For Windows users, ODAMwebserver offers a simple, fast and efficient way to deploy the ODAM API layer locally in (almost) one click.

ODAM API using Apache Web Server

ODAMwebserver.zip


Overview of the ODAM API deployment (Windows 10)



HOWTO

  • Download and Unzip ODAMwebserver.zip to the location of your choice on your hard drive

  • Go to the directory ODAMwebserver\settings, then with the notepad application for example, edit the getdata.conf file.

  • In the line starting with DATAROOT, specify the full path of the root of the data directory (be careful, it is necessary to put slashes "/" and not backslashes "\")
DATAROOT = C:/PATH/TO/YOUR/DATA/
  • Save the configuration file, then exit

  • Run WebServer.bat. A small window opens, indicating whether the Apache server is running or not.

  • At the first run, Windows should ask you for permission to pass through the firewall. Validate the authorization so that the web server can function properly.
  • Also, if the system tells you that the libssh2.dll file is missing, don't worry, it will still work.

  • In your browser, go To http://localhost/ to view your site root and to http://localhost/getdata/xml/<dataset> to view the <dataset> where <dataset> correspondings to your dataset put under the root of the data directory.

  • For testing, you could download the FRIM1 dataset (choose frim1.zip), then unzip the download file under the DATAROOT folder. In your browser, go To http://localhost/getdata/xml/frim1

Note: For MacOS (10.13 or newer), you can easily install Docker, then proceed as described in the next section "Intranet/Internet installation".


Intranet/Internet installation

This type of deployment is reserved more specifically for computers or IT infrastructures running Linux (including MacOS 10.13 or newer) which allow the deployment of Docker containers. See for more details on “What is Docker” on opensource.com.

Note: Depending on the scale you want to deploy, your intranet can be local or via a VPN (Virtual Private Network) and for the internet, a machine on an institutional data center (e.g. university) is better but a machine on a cloud computing is a perfectly feasible solution.

All the steps to deploy the Docker containers corresponding to the ODAM software have been put online :



Remark : Installing one's own ODAM instance on an institutional server for example allows in this way experimental data tables to be widely accessible and fully reusable including through scripting languages such as R or Python, and this with minimal effort on the part of the data provider. Thus, the web of data could be seen as a data network on the web, based on appropriate technologies (Web API), and using standard data formats (TSV, JSON). Web applications, each with a clearly defined objective, then operate this network. A data can therefore be used for several applications and vice versa. The data management system becomes completely independent of its operation. The data is thus “decompartmentalized", a sine qua non condition for the Web of Data.