Changes in version 1.50.0                        

SIGNIFICANT USER-VISIBLE CHANGES

  - The 12 datasets are no longer stored inside the package. They are
    hosted on Zenodo and downloaded individually on first use, then
    cached locally with BiocFileCache. This shrinks the installed
    package from about 75 MB to a few MB.
  - New exported function curatedBladderData(): call it with no
    arguments to list the available datasets, with one dataset name to
    get an ExpressionSet, or with several names to get a named list.
    test = TRUE loads small offline subsets bundled with the package,
    used by examples, tests and the vignette.
  - Legacy data(GSE89_eset) access still works. It downloads through the
    same cache, deferred until the object is first used, but is
    deprecated and will be removed in a future release.
  - inst/extdata/createEsetList.R now loads datasets through the getter.
    A new test.mode option in the patientselection config selects the
    offline subsets.
  - The cache is package-specific, at
    tools::R_user_dir("curatedBladderData", "cache"). Downloads are
    verified against md5 checksums recorded in
    inst/extdata/zenodo-manifest.csv.

BUG FIXES

  - biocViews listed OvarianCancerData; removed, and GEO added.
  - The package-level help page described ovarian cancer, listed the
    curatedOvarianData author list, and named a different maintainer
    than DESCRIPTION. Corrected to match DESCRIPTION.
  - Replaced the unused affy dependency with Biobase, which is where
    ExpressionSet, exprs() and pData() actually come from.
  - createEsetList.R no longer errors when the filters exclude every
    dataset (which() on an empty list).

INTERNAL

  - Added testthat tests and inst/scripts/ documenting how the Zenodo
    files, the offline fixtures and the data stubs are generated.