dabt

Dabt is an open-source data management library that implements a data-addressable-by-tag model. In this approach, each data item is associated with a set of metadata tags, and storage and retrieval are performed by querying these tags rather than by fixed keys. The design aims to improve flexibility and scalability in heterogeneous data environments, such as digital libraries, scientific repositories, and IoT data stores.

At its core, dabt provides a tag index, a query layer, and pluggable backends for storage. The tag index supports multi-dimensional queries, including conjunctions, disjunctions, and range filters on metadata fields. The API offers bindings for common programming languages and supports transactional guarantees for batch operations, enabling consistent updates across multiple items.

Key features include flexible schema handling, multi-tenant access control, data versioning, and built-in support for provenance and data lineage. Dabt emphasizes interoperability by exposing standard data formats for import and export and by offering connectors to popular storage systems. It is designed to work in environments ranging from local development machines to distributed clusters.

Common use cases involve data catalogs where users search by metadata such as author, instrument, location, or date, as well as scientific data repositories that require robust tagging and traceability. The library is typically integrated into applications that require dynamic data discovery and collaborative curation, and it often serves as a backend for research data management workflows.

The project is maintained by a community of contributors and has evolved through multiple releases. While dabt offers strong flexibility for tag-based querying, it may require careful index tuning and metadata governance to maintain performance at scale. See also tag-based databases and data provenance.