Release Management > Generate and publish database ER diagram with every release

[Candice-Louw]

Problem Description

SORMAS administrators are dependent on viewing the SORMAS database structure after each release in order to correctly interpret data received from the application and export it to other applications (MED export for CH, for example) after a database structural change has happened.

Proposed Solution

Generate a new ER Diagram of the database structure of each release and publish it with the release notes on GitHub (https://github.com/hzi-braunschweig/SORMAS-Project/releases) as a direct download or asset of the release. Please incorporate this into the release process going forward for every version.

Possible Alternatives

Alternative suggestions on locations where to publish the diagram (for example on the GitHub Wiki? Elsewhere?) after each release, are most welcome.

Additional Information

Acceptance criteria - ER Diagram:

• [ ] Generated and published automatically with every new release (on GitHub)
• [ ] Published in a shareable format (i.e. PNG, PDF, HTML format) that can be embedded on a website
• [ ] Data table names clearly visible
• [ ] Data table entries clearly visible
• [ ] Data table entries types (int, char, string etc.) clearly visible (or explained in a legend on the diagram)
• [ ] Data table relationships clearly visible

Current solution: Generating a SchemaSpy scheme: https://github.com/hzi-braunschweig/SORMAS-Glossary#schemaspy (preferred solution) Netzlink solution: PNG image

[Candice-Louw]

@VitaliHZI - please add any additional acceptance criteria that I may have missed.

[VitaliHZI]

@Candice-Louw for me it looks fine.

[Candice-Louw]

@Jbartsch - please feel free to add additional requirements from your side. :)

[JonasCir]

Ok I think we can do this as an automated step in our release:

1. Create a dedicated action and trigger it on a GitHub release
2. ~Take the SORMAS artifacts and deploy it into a container (like we do it in the Testbed)~ We can just use sormas_schema.sql in apply it to postgres container. No need to spin up a complete SORMAS isntance. This is much more lightweight.
3. Run schema-spy
4. Upload the result as release artifact or deploy to Github pages. Alternative, GitHub has some scratch storage space accessible from the CI which gets deleted after 90 days.

[Candice-Louw]

@MallerBeach - please see if this would be a solution going forward and if you have any additional requirements to add.

[MallerBeach]

@Candice-Louw thanks for your proposed solution. I would appreciate it.