README.md
In wehoop: Access Women's Basketball Play by Play Data

wehoop test suite

Tests live in this directory and are organized by data source. Each external-data domain is gated behind a source-specific env var so routine local runs (and CI) don't hit upstream APIs or download large release assets unless the contributor opts in.

| Variable | Helper | What it gates | |----------------------|---------------------------|------------------------------------------------| | WNBA_STATS_TESTS=1 | skip_wnba_stats_test() | test-wnba_*.R (WNBA Stats API endpoints) | | ESPN_TESTS=1 | skip_espn_test() | test-espn_wbb_*.R, test-espn_wnba_*.R | | NCAA_WBB_TESTS=1 | skip_ncaa_wbb_test() | test-ncaa_wbb_*.R | | WEHOOP_LOAD_TESTS=1| skip_load_test() | test-load_*.R (data-repo release downloads) |

test-load_*.R files exercise the load_*() functions, which pull season files from https://github.com/sportsdataverse/sportsdataverse-data/releases/.... Each season asset is multiple megabytes, and a full multi-season run can be hundreds of megabytes plus several minutes of network I/O.

These are kept separate from WNBA_STATS_TESTS / ESPN_TESTS on purpose: the API tests issue small JSON requests, whereas the load tests download large release assets. Bundling them under the same env var would force every contributor running API tests to also pull down the full data repo, which is undesirable for routine local checks.

To run only the load tests locally:

WEHOOP_LOAD_TESTS=1 Rscript -e 'devtools::test(filter = "load_")'

Even with WEHOOP_LOAD_TESTS=1 set, every load test also calls skip_on_ci(), so the suite stays a no-op under GitHub Actions by default. This matches the pattern called out in CLAUDE.md:

Note: in CI, many live API tests still include skip_on_ci() guards. Env vars alone do not override those guards unless tests are intentionally changed.

If you want a specific load test to run in CI as well, remove the skip_on_ci() line from that individual test file (and weigh the bandwidth cost on the runner before doing so).