ASReview

Description

ASReview LAB is an open-source machine learning tool designed to streamline the systematic screening and labeling of large textual datasets. It is widely used for tasks such as title and abstract screening in systematic reviews or meta-analyses, but its applications extend to any scenario requiring systematic text screening.

This workspace type will start ASReview as a webapplication.

Anyone who has access to your Collaboration will be able to easily login to ASReview using Single-Sign on. Since ASReview 2.0 supports multiple users, this workspace is ideal for crowd-based screening. You can even use it for a screenathon! Simply add the members of your screening group to your Collaboration in SRAM and start screening together.

Extensions (such as asreview-dory) are not installed on this workspace.

Please also read the tips below.

Creation

Create a storage volume

If desired, first create a storage volume before creating the workspace.

See the Getting started page for more info about how and why to create a storage volume.

If you select a storage unit when creating an ASReview workspace, all ASReview data will automatically be stored on it. Since ASReview can potentially generate a lot of data, using a storage unit may be advised (the default, non-permanent storage on the workspace amounts to 150GB).

Create a workspace

In the Research Cloud portal click the ‘Create a new workspace’ button and follow the steps in the wizzard.

See the workspace creation manual page for more guidance.

Choosing the right hardware: ASReview uses an AI model that runs in parallel to the ASReview webapplication. It is therefore recommended to choose a workspace with at least two CPU cores. Four CPUs is probably enough for most cases. The more users that will be using the workspace to do screening in parallel, the more hardware will be required.

Please also see the tips for responsible use.

Access

This workspace can be accessed via the yellow ‘Access’ button, or by opening the URL listed in the dashboard in your browser. Any member of the collaboration can login to the workspace using Single-Sign On.

When new members are added to the collaboration, you can select them in ASReview in the ‘Projects’ view, and invite them to a screening project. It may take up to 15 minutes after adding them to the CO for them to become visible.

Command line

Alternatively, only those users who are admins in the Collaborative Organisation (in the src_co_admin group) can login to the workspace via the commandline, using ssh. This may be useful for running ASReview in simulation mode, or e.g. for making backups.

After logging in, the asreview command will be on your path. The ASREVIEW_PATH environment variable will be set to reflect the location of ASReview’s data (default: /var/www/asreview/asreview_data if you do not create the workspace with a storage unit, /data/yourstoragename/asreview_data if you did), so you should be able to interact with it straight away.

Admin users will also be able to use sudo (after entering their time-based passwords).

Usage

When you log in to this workspace, the ASReview web app will open automatically.

Tips

• Remember that you are responsible for making backups of any data yourself! This also applies when using a storage unit. Although data on a storage unit will not be deleted when you delete the workspace, it can still become corrupt. So if it is important to your work, make sure to make periodical (e.g. daily or weekly) backups of your projects. You can make a backup by logging in to the workspace using the commandline, or by downloading your project from the webapplication.
• Remember that ASReview may potentially generate a lot of data, so think about how much storage you may need. For most cases, the standard amount of storage should be enough. But if you’re unsure, just ask for advise!
• Similarly, think about how many resources (CPUs) your workspace will need. Again, if you’re unsure, just ask for advise!