Every upstream source retains its own canonical URL, release context, retrieval date, attribution, and license. Snapshot paths identify retained corpus build inputs; those source files are not packaged in the served binary. The mixed corpus does not receive a false blanket license.
uniprot
UniProtKB
UniProt-reviewed protein identity, sequence, catalytic activity and cross-reference records.
Release
2026_02
Release date
2026-06-10
Retrieved
2026-07-14
License
CC-BY-4.0
Attribution
UniProt Consortium
Attribution required
Yes
Share-alike
No
Retained build-input license provenance
Retrieved 2026-07-14T09:21:21Z · raw/licenses/uniprot-license.txt SHA-256 64e362ea88d408d318b9aa3fa7842e5fde78a2c8187b0070641982019bb16a32 Open exact license source
No blanket license is asserted for this mixed-source aggregation. Reusers must comply with each source license and preserve record provenance. Site code is Apache-2.0. ChEMBL-derived assay material remains share-alike, attribution-bearing sources retain their credits, and public-domain partitions remain identified as such.