citesync

Software citation metadata.

On this page

Contribution contract fixture

These files are a hand-written integration example. expected.json was written from the formats' field meanings, before running the converter; do not regenerate it from citesync output. The CLI tests copy only the four input files into a temporary directory, so the tracked inputs stay unchanged.

The historical names are test labels. The identities and affiliations are fictional combinations, not assertions about those people. The ORCID 0000-0002-1825-0097 is the published fictional Josiah Carberry example. The missing ORCIDs are deliberate and must stay missing.

  • codemeta.json supplies author and contributor Person records by hand. Its author Role follows the shape of CRediT Matrix's exporter at b39d3aa. No producer-generated Person export is claimed.
  • CITATION.cff and .zenodo.json start with an abbreviated author name. The shared ORCID permits a rename without dropping affiliations or name particles.
  • expected.json fixes the expected people in each target. The non-author stays a contributor; Zenodo's DataManager type stays unchanged. No Role becomes a Person or supplies a Zenodo type.
  • citesync.toml explicitly selects CodeMeta as the source and disables manifest discovery. Tests also select CFF to exercise writing a CodeMeta target.

The tests use the vendored mapping whose release, commit and artifact hashes are in MAPPING_SOURCE. Existing mapping tests verify those hashes. No remote fetch or producer installation is required.

See the contribution guide for supported shapes, format losses and remaining producer work.

Matrix producer inputs

matrix-roster.json and matrix-attribution.json copy the native roster and hand-checked CodeMeta expectation from Matrix commit a8a55c648bbb45862d60fbdbd6d51cd75eea3cc3, under src/core/__tests__/fixtures/export-codemeta{,.expected}.json. The names and reserved identifiers are the same as the existing contribution example. Their MIT licence is in MATRIX-LICENSE.

  • matrix-roster.json SHA-256: 65d930d8c62987f88d3751f9206cfb3a1541838befd90ae79b81d67a807045fb
  • matrix-attribution.json SHA-256: 5f7283afd7d9c6215c0776bacf2f57798ba9143075d8adba835ceec206bee4d0

The existing playground CI job installs the pinned [email protected] development dependency and runs tests/test_matrix_consumer.py. The producer expectation is compared before citesync runs; expected.json remains the independent oracle for CFF and Zenodo. Python-only runs skip these four tests when Node dependencies are not installed.