Implementation feedback of the IVOA Provenance data model

TL;DR

This paper discusses the implementation experiences of the IVOA Provenance data model in distributing provenance metadata for HST data collections, highlighting lessons learned and strategies for complex query handling.

Contribution

It provides practical insights and lessons learned from implementing the IVOA Provenance data model in a real-world astronomical data service.

Findings

Effective graph representation strategies for complex provenance queries

Implementation challenges and solutions in ProvHiPS service

Insights into the use of ProvTAP protocol for provenance metadata distribution

Abstract

The IVOA Provenance Data model defines entities, agents and activities as container classes to describe the provenance of datasets, with the executed tasks and responsibilities attached to agents. It also provides a set of classes to describe the activities type and their configuration template, as well as the configuration applied effectively during the execution of a task. Here we highlight lessons learned in the implementation of the CDS ProvHiPS service distributing provenance metadata for the HST HiPS data collections, and for the HST archive original images used to produce the HiPS tiles. ProvHiPS is based on the ProvTAP protocol, the emerging TAP standard for distributing provenance metadata. ProvTAP queries may rapidly become very complex. Various graph representation strategies, including ad hoc solutions, triplestore and SQL CTE have been considered and are discussed shortly.