Skip to content

linkml-redcap

LinkML schemas modelling REDCap structures — vendor-neutral, fully typed, and backwards-compatible — for use as the canonical substrate in schema-driven conversion pipelines.

The package ships two schemas plus small loaders:

  • Data dictionary (linkml_redcap.data_dictionary) — a meta-schema for the REDCap data dictionary (the 18-column CSV that describes a project).
  • Record (linkml_redcap.record) — the REDCap record data itself, in both serializations REDCap's data uses: FlatRecord (the flat export row) and the generic, lossless StructuredRecord / RepeatedElement (the record-grouped form), plus a dependency-free grouping helper for the flat ⇄ structured step.

It models REDCap itself — nothing project- or RareLink-specific. The RareLink ontology conventions and the Phenopacket/FHIR engine live in rarelink, layered on top.

Where to go next

  • Schema reference — auto-generated, browsable pages for every class, slot, enum and type in each schema (regenerated from the schema on every release, so it never goes stale).
  • Design decisions — what the package is, the scope boundary, and why each choice was made.
  • Mapping workflow — how consumer repos use this: tabular → REDCap-flat (linkml-map + SSSOM), and REDCap-flat → Phenopackets.
  • Compatibility — the backwards-compatibility policy for downstream schemas that import this one.

Install

pip install linkml-redcap

from linkml_redcap.record import schema_view, group_flat_records
from linkml_redcap.data_dictionary import schema_view as dd_view

sv = schema_view()          # SchemaView over the record schema
print(sorted(sv.all_classes()))