LoginSign Up
← Back to all sparks
T

traits.build

INFRA · APIS
Velocity0.0

A Workflow for Harmonising Trait Data from Diverse Sources into a Documented Standard Structure

github.com/traitecoevo/traits.build ↗

The AusTraits engine, generalised for anyone's trait database, now links measurements to real specimens.

trait-databasesdata-harmonisationecologyprovenanceontologyr-package
◆Current state
traits.build is the harmonisation workflow extracted from AusTraits and generalised so other groups can assemble trait databases from heterogeneous sources. Its schema and ontology reached 1.0.0 in late 2024, and the package now leans on austraits itself for the functions that belong to database consumption rather than construction. The 2025 release extends the data model with identifiers and methods tables.
◆Where it's heading
The project's direction is toward provenance and interoperability rather than throughput. Value types grew to carry standard error and standard deviation, the methods table now records what kind of source each dataset came from, and the identifiers table lets a trait value point at a herbarium sheet, a museum accession or a GenBank record. Alongside that, responsibilities have been split with the sibling austraits package, with shared functions moved out under deprecation shims. A published paper and a versioned ontology mark it as infrastructure meant for outside adoption, not just for AusTraits.
◆Prediction
Expect further schema extensions in the same provenance direction, since the last two releases both added structure for describing where a measurement came from rather than new processing capability.

◆Recent moves

  1. 1y ago

    Trait values gain specimen and GenBank identifiers

    ⚡ SPARK

    Adds an identifiers table linking trait values to herbarium, museum, GenBank or arboretum records — and to the same individual plant measured across different datasets. Combined with a methods table recording source types, it extends what the schema can assert about a measurement's origin rather than how much it can process.

    View source ↗
  2. 1y ago

    Functions move to austraits as the ontology reaches 1.0.0

    Redraws the boundary with the sibling austraits package: shared functions move out, austraits becomes a dependency, and several functions are renamed with their old names kept as deprecated aliases. The ontology reaches 1.0.0 and standard error and deviation arrive as value types, while the published paper in Ecological Informatics marks the project as citable infrastructure.

    View source ↗
  3. 2y ago

    Fixes to dataset testing, reports and name standardisation

    A patch three days after the first major release, fixing dataset testing, report generation and taxonomic name standardisation. The usual shakeout of issues that only appear once outside groups run the workflow.

    View source ↗
  4. 2y ago

    AusTraits workflow generalised into a reusable package

    The first major release, generalising code originally written for AusTraits into a workflow other groups can use to build their own trait databases. The release note points mostly at the website and book rather than describing what shipped, so the substance has to be read from the releases that follow.

    View source ↗