AROMA RO-Crate Editor

As part of the HUN-REN ARP Data Repository Platform (https://researchdata.hu/en), we have implemented full RO-Crate (https://www.researchobject.org/ro-crate/) import and export support for Dataverse, and introduced native RO-Crate authoring through a dedicated tool called AROMA (ARP RO-Crate Manager).

This deep RO-Crate integration has been added to overcome some shortcomings of metadata authoring with Dataverse’s metadata tools. Specifically, Dataverse provides only a minimal set of file-level metadata - such as size, type, checksum, and a free-text “description” - many research workflows require far more detail. For example, users may wish to record the configuration of the instrument that generated a file, software parameters used during analysis, or any number of other structured attributes that Dataverse cannot represent natively.

Because extending Dataverse to support such deeply granular metadata at the file level is architecturally challenging, we approached the problem from another angle: RO-Crate. The RO-Crate standard already offers a flexible metadata model, making it a natural extension layer for capturing detailed descriptions that go beyond Dataverse metadata blocks.

To make creating and editing RO-Crate metadata seamless, we introduced a new tool within Dataverse: AROMA - the ARP RO-Crate Manager. AROMA enables users to edit dataset metadata directly in RO-Crate format while keeping it synchronized with Dataverse. Any metadata edited via Dataverse metadata blocks is reflected in AROMA, and changes made in AROMA flow back to Dataverse in return.

In addition, AROMA supports integration with CEDAR (https://metadatacenter.org/), allowing users to associate metadata schemas beyond those provided by Dataverse. Researchers, for example,  can design their own file-level metadata templates in CEDAR and apply them within AROMA to describe their data in richer detail.

Through this deep integration and two-way synchronization, AROMA effectively provides an alternative metadata authoring interface for Dataverse, one that extends its functionality while remaining fully interoperable with existing workflows.

AROMA 2 - Standalone RO-Crate Editor with Repository Systems Integrations

AROMA currently functions as a Dataverse extension and supports metadata editing for individual Dataverse datasets. With AROMA 2, we aim to expand its capabilities into a standalone application, an “RO-Crate IDE” that can generate RO-Crate packages from any data source including local files, remote storage systems, and also enables depositing such datasets to data repository systems like Zenodo or Dataverse.

A key goal of AROMA 2 is to integrate AI-assisted authoring throughout the entire RO-Crate creation process, based on our work on VibeARP (https://www.youtube.com/watch?v=ZMvsET-5S5Y). The system will guide users in building high-quality metadata from raw input, automatically where possible. Planned capabilities include:

• Automatic extraction of metadata directly from files

• Retrieval of additional contextual information from internal or external sources (e.g., the web, institutional knowledge stores)

• Assistance in identifying, importing, or creating suitable metadata schemas for describing datasets and files

• Mapping and alignment of gathered information to the selected schemas

• FAIR compliance analysis, scoring, and recommendations for improvement

• Automatic generation of documentation, reports, and other outputs derived from dataset metadata

With these features, AROMA 2 will not only support RO-Crate creation across diverse storage environments, but also help researchers produce richer, more FAIR-compliant metadata with significantly less manual effort.