ENH: Add nightly health dashboard and fix Windows GPU runner labels#39

Merged

aylward merged 19 commits into

Project-MONAI:mainfrom

aylward:gpu_nightly_runner

Apr 2, 2026

Commits on Apr 1, 2026

Commits on Apr 2, 2026

MNT: Drop actions/setup-python from Windows GPU jobs; use pre-installed Python
aylward
committed
FIX: Remove Pages deploy from nightly-health; fix status badge for pytest failures
aylward
committed
MNT: Add actionlint config to allow custom gpu runner label
aylward
committed
FIX: Paginate GraphQL review threads and per-thread comments
aylward
committed
FIX: Make --no-resolve and --mark-resolved mutually exclusive
aylward
committed
ENH: Use per-project venv on Windows GPU runner via GITHUB_PATH
aylward
committed
ENH: Assert torch.cuda.is_available() before pytest on GPU runners
aylward
committed
FIX: Assert CuPy device count in nightly CUDA gate step
aylward
committed
ENH: Create throwaway venv in RUNNER_TEMP for GPU jobs; cache uv wheels
aylward
committed
FIX: Remove free-text ref input from nightly-health workflow_dispatch
aylward
committed
FIX: Correct Python path to C:\Program Files\Python310\python.exe
aylward
committed
FIX: Add CuPy device count assertion to ci.yml GPU gate step
aylward
committed
REF: Rename inline_comments to thread_comments in filter_since_cutoff and build_prompt
aylward
committed
ENH: Publish status.json to Pages via docs.yml artifact merge
aylward
committed
FIX: Invoke uv via python -m uv to target the active venv
aylward
committed
FIX: Guard --since-last-push auto-resolve against partial threads
aylward
committed
FIX: Transport status.json via nightly-status branch instead of artifact
aylward
committed