This glossary was developed to enable transdisciplinary collaboration within the European Platform for Neurodegenerative Diseases (EPND) project, which requires heterogeneous partners (e.g., researchers, regulators, legal experts) and competencies (e.g., technical, academic, industry, etc.) in the field of neurodegenerative disorders.

The collaborative EPND glossary was created using the open source platform NocoDB and deployed as a Web App in the Azure cloud. The power of NocoDB, and reason for which it was chosen, is the fact that it can connect to any relational database and convert it into a collaborative spreadsheet with a user-friendly interface. Users can simultaneously work on tables, change views, and comment on the content, with an audit process that tracks all operations on the system. Furthermore, REST APIs are provided to interact with the database and perform task automation, which can help with the management of the glossary. NocoDB allowed to create a modular glossary organised by themes. Each module was structured as a table defining terms, their definitions, and references related to them.

The most challenge part of establishing a glossary is achieving convergence on the terms' definitions. In the EPND glossary, we took a Wiki-like approach, where certain participants were given an editor role and could freely define and revise terms related to their expertise. Another group of participants were given a commenter role and used to further peer-review the defined terms. This workflow enabled the development of a dynamic glossary that can quickly adapt. A public view of the EPND glossary is published online and can be freely consulted.

The glossary was deployed in the Azure cloud, following the user's documentation with the necessary changes for Azure deployment. Below is a summary of the process using the Azure client.

1. Login to Azure

az login

2. Create a resource group

az group create --name <group> \
                --location westeurope

where <group> is the desired name for your resource group, which is deployed in West Europe

3. Create an app service plan

az appservice plan create --resource-group <group> \
                          --name <service-plan> \
                          --location westeurope \
                          --sku B2 --is-linux

where <service-plan> is the name you wish to give to your service plan, with a pricing plan B2

4. Deploy NocoDB as a web app in the app service plan

az webapp create --resource-group <group> \
                 --plan <service-plan> \
                 --name <web-address> \
                 -i registry.hub.docker.com/nocodb/nocodb:latest

where <web-address> is the name you wish to give to your web app, if for instance <web-address> is nocodb-test, you will be able to access the web app at http://nocodb-test.azurewebsites.net/dashboard, here we also use the latest NocoDB image from Docker Hub

5. Configure the web app to send requests to port 8080

az webapp config appsettings set --resource-group <group> \
                                 --name <web-address> \
				 --settings WEBSITES_PORT=8080

6. Create a storage account for persistent storage to prevent data loss

az storage account create --name <storage-name> \
                          --resource-group <group> \
			  --location westeurope \
			  --sku Standard_RAGRS --https-only

where <storage-name> is the name you wish to give for the storage account, which must be between 3 and 24 characters in length and use numbers and lower-case letters only

7. Create an Azure file share for the web app container

az storage share-rm create --resource-group <group> \ 
                           --storage-account <storage-name> \
			   --name <share-name>

where <share-name> is the name id for the file share

8. Get the access key for your storage account, which you will need in step 9

az storage account keys list --resource-group <group> \
                             --account-name <storage-name>

9. Add a storage mount for the file share

az webapp config storage-account add --resource-group <group> \
                                     --name <web-address> \
				     --storage-type AzureFiles \
				     --account-name <storage-name> \
				     --share-name <share-name> \
				     --mount-path /usr/app/data \
				     --custom-id <webapp-storage> \
				     --access-key <key>

where <webapp-storage> is a custom name given for the share configured within the web app, --mount-path is the path with the web app data, and <key> is the key retrieved in step 8

10. Open the web app in your browser to test the deployment

az webapp browse --resource-group <group> \
                 --name <web-address>

if all went well you should see NocoDB's registration page, mind that the first person to sign up will become the super admin

11. If you wish to delete all resources run the following

az group delete --name <group>

The glossary was created using the web interface. Each module is a different table, where the columns are the necessary elements to properly define its modules' terms and each row is a different term.

The comment feature of NocoDB was used to help discussion about the terms and facilitate the review process of the definitions. Comment tracking was automated with a combination of comments' retrieval using the REST APIs and notification of the new comments via email using the comments_tracking.py script. A crontab job was scheduled to automatically run the script and moderators got notified of the comments with the desired frequency.

The config.json.example has an example of the necessary input to run the script. Below is a summary of how this script should be run.

1. Install dependencies (tested in Python 3.10), it is a good idea to create a virtual environment for the project

pip install -r requirements.txt

2. Create the config.json input

cp config.json.example config.json

3. Open the config.json file and edit it accordingly

4. Run the script to get comments and send them by email

python comments_tracking.py