← Back to all sparks
A

admiral

ANALYTICS
Velocity0.0

ADaM in R Asset Library

admiral spent its last cycle making complex derivations documentable rather than adding more of them.

pharmaverseadamclinical-programmingdocumentationderivations
Current state
admiral is the pharmaverse's ADaM dataset derivation toolkit — a large library of derive_* functions for building analysis datasets from SDTM. The 1.2.0 and 1.3.0 releases added new derivation functions and then a substantial documentation overhaul: argument defaults now appear in the help text and complex functions carry titled, described examples. 1.3.1 was a hotfix for links and the search index.
Where it's heading
The direction is toward usability of an already-large API rather than growth of it. 1.1.0 rewrote error messaging to be user-facing, 1.2.0 added flexibility to duplicate handling across several functions via check_type, and 1.3.0 invested in the structured-example documentation that the rdx_roclet work in {admiraldev} made possible. The two packages move together.
Prediction
With the documentation infrastructure in place, expect the remaining complex derivation functions to be brought up to the structured-example standard, and new therapeutic-area work to keep landing in the extension packages rather than the core.

Recent moves

  1. 1y ago

    Search index and footer link fixes

    A hotfix restoring search index generation and correcting sidebar logos, footer links and vignette cross-references.

    View source ↗
  2. 1y ago

    Documentation overhaul: defaults shown, complex functions get structured examples

    Argument defaults now appear in the argument description, and a dozen of the most complex derivation functions gain titled, described examples that render in the RStudio help pane. A template and vignette fix makes SAFFL CDISC-compliant.

    View source ↗
  3. 1y ago

    New derivations for categorical pairs, criterion flags and range transforms

    derive_vars_cat(), derive_vars_crit_flag() and transform_range() are added, and check_type gains a "message" option across several functions so duplicate handling can be tuned rather than always erroring.

    View source ↗
  4. 2y ago

    User-facing error messages and merge helpers exposed

    Error messaging is rewritten for clarity, the internal helpers behind derive_vars_merged() and derive_vars_query() are exported for direct use, and a relationship argument brings dplyr-style join cardinality checks to the merge functions.

    View source ↗