← Back to all sparks
O

omock

ANALYTICS
Velocity0.0

Creation of Mock Observational Medical Outcomes Partnership Common Data Model

A mock-data generator for OMOP studies that keeps widening what it can fake.

omop-cdmsynthetic-datatest-fixtureshealth-dataohdsi
Current state
omock builds synthetic OMOP Common Data Model tables so packages in the darwin-eu and OHDSI ecosystem can be tested without touching patient data. The 0.7.0 release adds concept set subsetting with its own vignette, unit and value support in mockMeasurement, observation date validation and cohort attrition initialization, while deprecating mockConcepts. Releases arrive as PR digests rather than written notes.
Where it's heading
The package has been moving from generating tables to shipping and managing reference datasets — mockDatasets arrived in 0.4.0, mockCdmFromDataset gained a source argument in 0.5.0, and 0.6.1 added download retry handling plus an internal GiBleed dataset after the hosted files moved. The recent work is filling in CDM fidelity: type concepts, measurement units, attrition, and guards for degenerate cases like an empty person table.
Prediction
Expect coverage to keep extending table by table toward the parts of the CDM that omock still cannot mock, with mockConcepts removed outright once the concept set subsetting path settles.

Recent moves

  1. 1mo ago

    Concept set subsetting and richer measurement mocks

    Adds concept set subsetting with a dedicated vignette, unit and value support in mockMeasurement, observation date validation with type concepts, and cohort attrition initialization, while deprecating mockConcepts. The broadest release in this feed and squarely on the CDM-fidelity track.

    View source ↗
  2. 4mo ago

    Delphi dataset added and CDM version bumped

    Adds a delphi dataset, adjusts a timeout message and bumps the declared CDM version. Housekeeping between the two substantive releases on either side.

    View source ↗
  3. 6mo ago

    Dataset downloads survive the storage move

    Repoints the mock dataset URLs at new storage, adds retry and timeout handling to the download path, and bundles GiBleed internally so tests are not wholly dependent on a remote fetch. The hazard of shipping datasets rather than generating them, handled.

    View source ↗
  4. 11mo ago

    JOSS submission tag alongside the 0.5.0 release

    A tag carrying no release notes beyond a changelog link, cut roughly ninety minutes after 0.5.0 itself for the JOSS submission. Not a separate body of work.

    View source ↗
  5. 11mo ago

    mockCdmFromDataset gains a source argument

    Adds a source argument to mockCdmFromDataset() and a mock vocabulary set with accompanying vignettes, alongside the JOSS paper work. Continues the shift toward driving mock CDMs from named datasets rather than pure generation.

    View source ↗
  6. 1y ago

    mockDatasets introduces bundled reference data

    Adds mockDatasets with tests, plus faster start and end date generation and contributing guidelines. This is where the package stops being only a generator and starts distributing reference datasets — the decision that the download and storage work in 0.6.1 later paid for.

    View source ↗