An alignment vitalization app, for streamlit.
pip install -r alignviz/requirements.txt
streamlit run alignviz/alignviz_app.pyEvery push to main builds a linux/amd64 image and pushes it to
ghcr.io/statisticalbiotechnology/alignviz (see
.github/workflows/docker-image-ghcr.yml).
Each build gets a unique UTC timestamp tag, e.g. 20260820-160621, alongside
latest; the exact tag to deploy is printed in the workflow run summary.
Serve needs to be able to pull the package anonymously. Check under repository
→ Packages → alignviz → Package settings that visibility is Public, and keep
it that way — Serve re-fetches the image at regular intervals.
To deploy or update the app, create/edit a Streamlit app at
serve.scilifelab.se with port 8501 and the
image set to the timestamped tag. Serve caches by tag, so always point it at a
new timestamp rather than latest when updating.
Building and testing the image by hand:
docker build --platform linux/amd64 -t alignviz:test .
docker run --rm -it -p 8501:8501 alignviz:test # then open http://localhost:8501