Question: Embedding IMMARKUS as an annotation component in a scientific image repository

[leilaicruz]

Dear IMMARKUS maintainers,

First of all, thank you for developing and maintaining IMMARKUS. I am exploring its potential use in the context of a scientific image repository that exposes images through IIIF.

I attended your presentation on Day 1 of the IIIF Conference in Amsterdam, and I was very impressed by your work and by the range of functionality that IMMARKUS offers. I can see strong potential for this type of tool not only in the humanities and cultural heritage domains, but also for researchers working with scientific images in the natural and engineering sciences.

I would like to ask whether IMMARKUS can be embedded or integrated into another codebase, rather than used only as a standalone annotation environment.

I work at the 4TU.ResearchData repository, which is currently exposing images using IIIF standards. In this context, we are interested in understanding whether IMMARKUS could potentially be integrated as an annotation layer for images hosted in the repository, allowing users to annotate repository images while keeping the annotations connected to the corresponding datasets, images, canvases, or manifests.

Use case

We are interested in supporting scientific image datasets stored in a research data repository. These images may include, for example, microscopy images or other high-resolution scientific images. The repository already stores the images and associated metadata, and the goal would be to provide users with annotation functionality directly from the repository interface.

The intended workflow would be something like:

1. A user opens an image or image dataset in the repository.
2. The repository provides the image through IIIF Image API and/or IIIF Presentation API.
3. The user can create semantic or region-based annotations using IMMARKUS functionality.
4. The resulting annotations can be stored, exported, or linked back to the repository object using IIIF/Web Annotation-compatible structures.
5. The annotations remain associated with the image, dataset, canvas, or manifest in the repository.

Main question

Is IMMARKUS designed in a way that would allow it to be embedded into an existing repository platform, for example as:

• a reusable frontend component,
• an iframe-based integration,
• a separately deployed service connected to repository images,
• or a library/module that can be integrated into another application?

Specific integration questions

I would be grateful for guidance on the following points:

1. Input from an external repository
   Can IMMARKUS open and annotate images directly from externally hosted IIIF manifests or canvases?

2. Annotation persistence
   Does IMMARKUS provide a way to save annotations to an external backend, or would this require custom development?

3. Export format
   Are annotations exported in a IIIF/Web Annotation-compatible JSON-LD format that could be stored and re-used by another repository system?

4. Authentication and user management
   If embedded in another platform, would IMMARKUS rely on its own user/session model, or can authentication be delegated to the host repository?

5. Licensing and embedding
   Since IMMARKUS is licensed under AGPL-3.0, are there specific considerations or recommended practices for integrating it into another open-source repository codebase?

6. Recommended integration pattern
   From your perspective, what would be the most realistic way to integrate IMMARKUS functionality into an existing scientific image repository?

Any advice on whether this type of integration is feasible, and what architectural approach you would recommend, would be very helpful.

Thank you very much.

[rsimon]

Dear @leilaicruz,

thanks for getting in touch – it was nice meeting you at the IIIF Annual Conference!

We should definitely chat about your use case. But the bad new first: IMMARKUS' architecture is very much at odds with what you're looking for.

What's definitely not going to work:

• IMMARKUS does not have the concept of an external backend, since everything works via a local folder and in-memory storage. You can export data manually (incl. in IIIF-compatible format), and potentially upload elsewhere. But a seamless integration into an online repository system, where users can open images from inside the repository and then annotate them directly in the website, with annotation data getting stored back into a database backend would be a significant undertaking – pretty much equal to a full redesign.
• IMMARKUS would not (easily) support any collaborative annotation activity.
• There is no concept of authentication or user management in IMMARKUS, because it works entirely in isolation, in your browser, on a single folder on your computer.

So, in other words: IMMARKUS simply isn't designed for embedding in an online system. It's "delivered" via the browser, but other than that it's as much a local, single-user app as any installable desktop application.

As far as possible alternatives are concerned:

These would depend on which features from IMMARKUS you're most interested in integrating into your system, and we should definitely talk about it!

• First and foremost: IMMARKUS is built on top of a number of open source libraries internally. Potentially, these could be used to implement a custom "IMMARKUS-like" online annotation environment, embedded directly in your repository system, fairly easily. Depending on what features you are looking for specifically, "fairly easily" might be a bit of an understatement I guess; and bespoke solutions always mean additional maintenance responsibilities – but it could be a solution nonetheless.
• Another project of mine is this: liiive.now. It's very different from IMMARKUS, and lacks the all of IMMARKUS' advanced scholarly features (semantic annotation, data models, knowledge graph, AI integration, multi-image workspace). But it's designed for more loose integration with IIIF repositories, supports multi-user collaboration natively, could be modified to work with institutional authentication, and is fully open source and self-hostable – so I just want to mention it, in case basic annotation (comments and replies) might be interesting also.
• Last but not least: we're also thinking about where to take IMMARKUS next, as part of follow-up funding grants. I don't think that "make IMMARKUS embeddable" would be a direct goal. But we're definitely interested in finding ways to a) make IMMARKUS outputs easier to publish online; and b) provide some form of online sharing and/or collaborative curation of resources.

[leilaicruz]

Dear @rsimon ,

Thank you for your extensive response and explanations. We will now explore which roadmap we would like to follow regarding the features we want to implement in the repository for managing scientific images.

Good luck with your next steps for IMMARKUS, it looks great! . What you are doing is really valuable and inspiring.

I will get back to you once we have a clearer idea of which functionalities we would like to reuse from IMMARKUS.

Thanks again!

[rsimon]

Perfect, thanks!