Changelog

Source: NEWS.md

gsm.datasim (development version)

gsm.datasim v2.0.0

This major release introduces a composable study builder API, longitudinal multi-snapshot study support, workflow-driven data generation, a domain registry, and parquet export. Analytics and reporting are now powered by the workr workflow engine.

Breaking Changes

• generate_rawdata_for_single_study() and raw_data_generator() are deprecated in favour of the study builder API (create_study_config() + generate_study_data()). Existing calls still work but emit a deprecation warning.
• gsm.core, gsm.kri, gsm.mapping, and gsm.reporting moved from hard dependencies to Suggests, as pipelines now run on the workr engine. Install those packages only if you run the analytics/reporting pipelines.

New Features

• Study builder API: build study configs with create_study_config() (or create_standard_study_config()), refine them with set_outlier_config(), set_temporal_config(), and add_dataset_config() / remove_dataset_config(), validate with validate_study_config(), and generate data via generate_study_data().
• Longitudinal studies: create_longitudinal_study() and quick_longitudinal_study() generate complete multi-snapshot studies with optional analytics/reporting; create_multiple_longitudinal_studies() and export_multiple_studies() handle portfolios. Extract results with get_snapshot_data(), get_domain_timeline(), and get_available_domains() (the latter also serves as a standalone registry lookup when called with no arguments), and run pipelines on existing studies with run_longitudinal_analytics() / run_longitudinal_reporting().
• Domain registry & workflow-driven generation: get_domain_registry() exposes known domains; generate_data_from_workflows() and generate_raw_data_for_endpoints() generate data directly from gsm.mapping workflow specs; generate_unknown_domain() handles domains outside the registry via inferred column types.
• Parquet export: export_study_data() now supports "csv" (default), "parquet", or "both" via the format argument (parquet requires the arrow package).
• New domains: added Raw_VISIT, Raw_Death (including a deathcls classification column), Raw_Consents, Raw_Baseline, Raw_Randomization, Raw_OverallResponse, and Raw_GENERAL, plus gradual enrollment simulation and timestamp support across existing domains.

gsm.datasim v1.1.3

This patch release makes the following updates:

• Added strStartDate to generate_rawdata_for_single_study() and raw_data_generator() to allow arbitrary snapshot start dates.
• Added db_lock_dt generation in Raw_STUDY based on the snapshot global max date.
• Updated default count scaling for SDRGCOMP and AntiCancer records in generate_rawdata_for_single_study().

gsm.datasim v1.1.2

This release resolves a bug involving duplicate records in the randomization domain.

gsm.datasim v1.1.1

This release adds new contributor guidelines and standardized issue templates.

gsm.datasim v1.1.0

Notable Changes:

New Mapping Workflows Support: - Additional workflow support for “IE” domain - dates/timestamp support for many domains now available.

gsm.datasim v1.0.0

We are excited to announce the first major release of the gsm.datasim package, a collection of functions to generate synthetic test data for the RBQM of Clinical Trials based on several parameters.

Notable Changes:

User-facing functions - generate_rawdata_for_single_study() can create snapshot(s) for a single study when provided with the proper parameters and appropriate mapping specifications. - raw_data_generator() is a wrapper to run generate_rawdata_for_single_study() to create multiple studies if providing a template/dataset containing a variety of these parameters.

New Mapping Workflows Support: - All workflows that exist in gsm.mapping’s inst/workflow/1_mappings are now supported, with “STUDY”, “SUBJ”, “SITE” and “ENROLL” always being required. - There are many gsm.endpoints specific domains, such as Anticancer and OverallResponse that now have support as well.

Replacing clindata with gsm.datasim: - The object gsm.core::lSource was created using gsm.datasim for examples, tests, and vignettes across the gsm ecosystem. This object is based on “core mappings” which include: “AE”, “COUNTRY”, “DATACHG”, “DATAENT”, “ENROLL”, “LB”, “PD”, “PK”, “QUERY”, “STUDY”, “STUDCOMP”, “SDRGCOMP”, “SITE”, “SUBJ”