LoginSign Up
← Back to all sparks
I

IsoriX

INFRA · APIS
Velocity0.0

Isoscape Computation and Inference of Spatial Origins using Mixed Models

github.com/courtiol/isorix ↗

An isotope geolocation package still recovering from the r-spatial retirement

stable-isotopesgeolocationr-packagespatial-modellingecology
◆Current state
IsoriX builds isoscapes and assigns organisms to geographic origin from stable isotope data. Its recent history is defined by the October 2023 retirement of the old R spatial stack: version 0.9.1 recoded substantial parts of the package to drop raster and sp, 0.9.2 restored capabilities that migration had broken - saving and reloading IsoriX objects, and plotting ocean masks containing holes - and documented the new plotting calls that replace the old sp.polygons and sp.points idioms. Since then 0.9.3 revised the bundled bat calibration datasets and 0.9.4 moved the documentation off bookdown.org.
◆Where it's heading
Two years of the visible history have gone on infrastructure the science did not ask for. The 0.9.1 notes are unusually candid, warning users that the surrounding package landscape was still in flux and that incompatibility bugs were likely to keep surfacing, which 0.9.2 then confirmed. Only in 0.9.3 does domain work reappear: the bat fur isotope values in the calibration and assignment datasets were corrected to align with current keratin reference material normalisation, so results computed against the old datasets are not comparable with new ones.
◆Prediction
With the spatial migration settled and the documentation rehomed, the next release is more likely to return to isoscape functionality than to infrastructure, though the entries do not name a specific feature.

◆Recent moves

  1. 7mo ago

    Documentation rehomed after bookdown.org shutdown

    Migrates the book-length documentation to a Posit Connect Cloud address as bookdown.org is discontinued, and updates every link accordingly. The R version requirement rises to 4.1.0 for native pipe usage. Infrastructure only, though it is the third consecutive release driven by something outside the project.

    View source ↗
  2. 1y ago

    Calibration datasets corrected to current keratin reference standards

    Ships revised versions of the bat calibration and assignment datasets with fur isotope values realigned to current delta values for keratin reference materials, so formerly and recently normalised data are comparable. Elevation values and site identifiers changed too, which the notes flag as breaking. A new title_delta_notation option lets the delta notation be set globally, useful for isotopes other than deuterium.

    View source ↗
  3. 2y ago

    Saving objects and masked plotting restored after the spatial migration

    Recovers two capabilities the previous release had cost: persisting IsoriX objects, now through saveRDS and readRDS, and plotting oceans and masks containing holes. The notes spell out the replacement plotting idiom in full, since the polygon and point functions and their colour arguments all changed shape.

    View source ↗
  4. 2y ago

    v0.9.1.9999

    A development tag carrying the rasterVis and lattice plotting rework that shipped properly in 0.9.2 a month later. A pre-release snapshot rather than a distinct release.

    View source ↗
  5. 2y ago

    Package recoded to drop the retiring spatial dependencies

    The forced migration release: substantial internal recoding to replace spatial packages retiring in October 2023, with plot-building workflows the most likely thing to break. The notes are unusually honest about the situation, warning that dependent packages were migrating simultaneously and that incompatibility bugs would surface - which the next two releases spent their time confirming and repairing.

    View source ↗
  6. 5y ago

    CRAN release version 0.5

    A one-line CRAN submission announcement with no description of what changed. The feed then goes quiet for three years before the spatial migration releases begin.

    View source ↗