Making clinical data ready for secondary use.
(ready42 = "ready for two[ndary use]". The 42 is a nod to Douglas Adams.)
ready42 is a federated toolkit for profiling, assessing, curating, and mapping clinical data toward standardized models for secondary use. Version 1 targets OMOP CDM; the architecture is model-agnostic and future versions can add other targets (i2b2, Sentinel, PCORnet, …).
The federated story: researchers run the tooling locally on privacy-sensitive data. Only enriched metadata (data dictionaries, mappings, readiness scores) ever leaves the institution.
Six Python engine packages + a thin product surface:
| Package | What it does |
|---|---|
common |
Shared LiteLLM client, OMOPhub client, OMOP CDM specs, LinkML schemas, the modular bundle system |
profiler |
White-Rabbit-compatible data profiler — schema + value-frequency summaries (with min-cell-count censoring) and a WR-compatible Excel writer, from REDCap dictionaries, generic CSVs, and SQLAlchemy/SQLite databases; houses the input-adapter layer |
assessor |
OMOP-readiness rubric scorer with per-source history tracking |
curator |
LLM-assisted, guideline-driven chat curation of data-dictionary metadata (resumable, audit-logged, replayable) |
mapper |
OMOP concept + structural mapping suggestions (vector-DB or OMOPhub candidates, LLM re-ranking), LinkML output |
athena_vectorizer |
Portable embedded vector DB (LanceDB + SQLite) over an Athena OMOP vocabulary download, for filtered semantic concept search |
Plus:
- A FastAPI backend (
backend/) exposing the engine over HTTP. - A SvelteKit frontend (
frontend/) for the chat curation UI. - Modular bundles (
bundles/) — composable rubric overlays + curator guidelines (OMOP core, oncology, …).
v0.1 — engines implemented. All six engine packages (profiler, assessor, curator, mapper, athena_vectorizer, common), the FastAPI backend (REST + curate WebSocket), and the SvelteKit frontend (five screens) are in place with passing tests. Active development toward a TEF Symposium 2026 milestone on 2026-06-07, then continued community development toward a contributor-ready v1 by ~September 2026.
See ARCHITECTURE.md for the design and CONTRIBUTING.md to get involved.
Requires Python 3.11+ and Poetry.
git clone https://github.com/<user>/ready42
cd ready42
poetry install # installs the workspace (5 engine packages + dev tools)Configuration goes in .env (copy from .env.example):
cp .env.example .env
# edit with your LLM provider key, OMOPhub key, etc.# FastAPI + SvelteKit on localhost:
bash scripts/dev.sh
# → backend at http://localhost:8000 · frontend at http://localhost:5173ARCHITECTURE.md— system design, the engine + backend + frontend split, the bundle system.CONTRIBUTING.md— how to get involved, the bundle-contribution pattern, the design-choices doc convention.docs/— MkDocs Material site (runpoetry run mkdocs serve).planning/(in a separate showcase repository) — detailed per-tool specs and design registers.
MIT. See LICENSE.