Workbench
The Convert-Pheno workbench is a local browser application for interactive conversions. It reads the routes available in the local installation and makes the generated files available for preview and download.
The Workbench is available from Convert-Pheno 0.35.
This is the current workbench. The original Convert-Pheno Web App is a legacy demonstration and does not reflect current conversion support.
Select a route, add its input files, choose the output, and run the conversion.
Run from Dockerβ
Start the Workbench from the published Docker image:
docker run --rm \
--publish 127.0.0.1:8080:8080 \
manuelrueda/convert-pheno:latest
Open http://127.0.0.1:8080 in your browser. Keep the
terminal running while you use the Workbench; press Ctrl+C to stop it.
This command makes the Workbench available only from your computer. It has no authentication and should not be exposed as a remote multi-user service.
OMOP output needs the separately distributed ohdsi.db. Mount it at the
standard database bundle location:
docker run --rm \
--publish 127.0.0.1:8080:8080 \
--volume "$PWD/ohdsi.db:/usr/share/convert-pheno/share/db/v0/ohdsi.db:ro" \
manuelrueda/convert-pheno:latest
Routes that need this database remain in the route chooser when it is missing, but are disabled with an explanation.
Run from a source checkoutβ
Development uses two processes. Start Mojolicious from the repository root:
morbo -l http://127.0.0.1:3000 api/perl/main.pl
Then start Vite in another terminal:
cd app
npm ci
npm run dev
Open http://127.0.0.1:5173. Vite proxies /api to Mojolicious. Node.js 24 is
required for the frontend toolchain; it is not required in the runtime Docker
image.
Privacy and scopeβ
JSON input and output stay in browser and API process memory. Uploaded files are copied to a private per-request temporary directory and removed after success or failure. The workbench does not use local storage, analytics, arbitrary server paths, or payload logging. Changing a route or input clears stale results.
The workbench accepts JSON for Beacon v2, Phenopacket v2, FHIR, openEHR, and OMOP, plus role-based uploads for OMOP table files or ZIP packages, CSV, REDCap, CDISC-ODM, Dataset-JSON, Dataset-XML, cBioPortal study packages, and i2b2, PCORnet, or Sentinel table packages. Entity-aware BFF routes also accept the optional compact Mapping V2 metadata file advertised for that route. Each source has a synthetic example drawn from the regression fixtures. Upload requests are limited to 100 MiB and run synchronously. Use the command-line interface for streaming or larger datasets.
Terminology reviewβ
For mapping-based routes, the workbench enables a color-coded XLSX terminology report by default. After conversion it counts exact or configured terms, similarity matches, unresolved terms, and source fallbacks. Filter the preview by text, ontology, or review action.
Terminology decisions can be filtered in the browser and downloaded as a complete XLSX or TSV report.
The browser shows a limited preview; the XLSX or TSV download contains every decision. Review recommendations come from the Perl audit writer rather than being recalculated in the browser. See Terminology Search for the fields and suggested review order.

