Skip to content

Get help or report a problem

Choose the route that matches what you need. Keeping questions separate from trackable defects makes both easier to find later.

Route Best for
GitHub Issues Reproducible defects, incorrect documentation, and focused feature requests. Search existing issues first.
GitHub Discussions Installation help, workflow design questions, ideas, and examples that may help other VIPP users.
image.sc — napari Broader bioimage-analysis and scientific-method questions that benefit from the wider community. Put napari-vipp and its version in the title or opening paragraph.
Private security reporting A credible vulnerability or sensitive security detail that should not be public.

If a discussion reveals a reproducible software defect, link it from a focused GitHub issue. A small reproducible case is more useful than a large private dataset or an unexplained screenshot.

Include

  • napari-vipp, napari, Python, and operating-system versions;
  • whether the build came from PyPI, a tag, or a commit;
  • the smallest workflow JSON that reproduces the problem;
  • exact steps from a new napari session;
  • expected and observed behavior;
  • complete traceback as text;
  • source shape, axes, dtype, scale/unit, and format;
  • whether a bundled sample reproduces it;
  • for a numerical problem, a tiny synthetic array and expected invariant.
  • for compute/GPU behavior: requested mode, node badge/decision reason, fallback policy, provider/device, vipp-compute-doctor output, and the relevant execution report or redacted provenance sidecar;
  • for optimizer behavior: node/provider stage, both progress values, time limit, completed/remaining comparisons, and whether exact evidence was reused;
  • for ND2 T/Z/C behavior: ordered shape/axes reported by the reader, which slider is missing or wrong, and a minimal non-sensitive file when possible.

Get the core versions with:

python -c "import importlib.metadata as m, sys; print(sys.version); print('napari-vipp', m.version('napari-vipp')); print('napari', m.version('napari'))"

For an optional CUDA environment also run:

vipp-compute-doctor --track cuda13

Review its output before posting: local paths, host/device names, and environment details can be identifying. Do not attach a full batch manifest or workflow until source paths and metadata have been redacted.

Protect data and identities

Do not attach patient/research-participant data, credentials, restricted image metadata, identifiable paths, or unpublished results. Reproduce with a bundled sample, crop/synthetic phantom, or a privately shared artifact only after an appropriate route is agreed.

Review workflow JSON and screenshots: paths, source names, graph notes, metadata columns, window titles, and recent-file dialogs can reveal sensitive context.

Security problems

Do not disclose a credible vulnerability in an Issue, Discussion, or image.sc topic. Follow the application repository's security policy and use its private reporting route.