OMOP test fixtures

These fixtures are derived from the OHDSI Eunomia example dataset:

• https://github.com/OHDSI/Eunomia

This directory intentionally contains two different kinds of OMOP input fixtures used by the test suite.

1. Full SQL fixture

• omop_cdm_eunomia.sql

This is the larger PostgreSQL dump fixture used for:

• omop2bff
• omop2pxf
• streamed SQL-based OMOP regressions

Note:

• concept_id = 5001 was added to the fixture CONCEPT data so this SQL export remains self-contained without --ohdsi-db.

2. Reduced plain CSV fixtures

• PERSON.csv
• CONCEPT.csv
• DRUG_EXPOSURE.csv

These are lightweight plain-CSV fixtures used to exercise the OMOP CLI/API paths without relying on a full OMOP export.

Important notes:

• CONCEPT.csv is intentionally reduced and is not a complete OMOP vocabulary table.
• These files are useful for regression-testing mechanics of the conversion path, but they are not a gold-standard fixture for complete OHDSI concept resolution.
• When required OMOP concept_id values are missing from this reduced CONCEPT.csv, the conversion fails unless --ohdsi-db is enabled for Athena-OHDSI fallback.

These reduced CSV fixtures are currently covered by the CLI regression test:

• t/19-cli-regression.t

with a non-stream command equivalent to:

../../../bin/convert-pheno -iomop PERSON.csv CONCEPT.csv DRUG_EXPOSURE.csv \
  -obff individuals_csv.json --test

3. Reduced MIMIC specimen fixture

• mimic_specimen/PERSON.csv
• mimic_specimen/CONCEPT.csv
• mimic_specimen/SPECIMEN.csv

These files are a reduced fixture derived from the public MIMIC-IV demo OMOP CSV export hosted by PhysioNet:

• https://physionet.org/content/mimic-iv-demo-omop/0.9/

Source files used:

• 1_omop_data_csv/person.csv
• 1_omop_data_csv/specimen.csv

Important notes:

• This fixture is intentionally reduced to only the rows needed for OMOP SPECIMEN -> Beacon biosamples regression tests.
• PERSON.csv keeps the original MIMIC person_id values for the selected rows.
• The demo OMOP person.csv provides year-level birth data; this fixture synthesizes birth_datetime, month_of_birth, and day_of_birth as January 1 of year_of_birth so collectionMoment tests remain deterministic.
• CONCEPT.csv is reduced to the concept ids referenced by the selected PERSON and SPECIMEN rows.

These reduced MIMIC specimen CSV fixtures are covered by:

• t/26-omop-biosamples.t

Provenance details for mimic_specimen/:

• Download source: PhysioNet MIMIC-IV demo OMOP dataset, version 0.9
• Source URLs:
  • https://physionet.org/files/mimic-iv-demo-omop/0.9/1_omop_data_csv/person.csv
  • https://physionet.org/files/mimic-iv-demo-omop/0.9/1_omop_data_csv/specimen.csv
• Retrieved for this fixture on 2026-04-15
• Selected original MIMIC person_id values kept in the fixture:
  • 4668337230155062633
  • 2288881942133868955
  • 3192038106523208432
  • 7131048714591189903
• SPECIMEN.csv rows are reduced from those public MIMIC rows for biosample regression coverage only
• CONCEPT.csv is not copied from the MIMIC download; it is a reduced local companion fixture containing only the OMOP concept rows needed by the selected PERSON and SPECIMEN rows

4. Gzipped CSV fixtures

• gz/PERSON.csv.gz
• gz/CONCEPT.csv.gz
• gz/DRUG_EXPOSURE.csv.gz

These are the fixtures used for streamed OMOP CSV regressions.

They are covered by:

• t/04-api-stream-omop.t
• t/20-cli-stream-omop.t

with a command equivalent to:

../../../bin/convert-pheno -iomop gz/PERSON.csv.gz gz/CONCEPT.csv.gz gz/DRUG_EXPOSURE.csv.gz \
  -obff individuals_csv.json.gz --stream --ohdsi-db --sep $'\t' --max-lines-sql 2700 --test