I believe all necessary packages to run this should be in the shared stack.
Clone this repo, then from the repo directory, run
panel serve dashboard.py --port 12345 # your favorite port number
Or, for the gen3 backend, run
panel serve dashboard_gen3.py --port 12345 # your favorite port number
Then tunnel to that port locally:
ssh -NfL localhost:12345:localhost:12345 user@lsst-devl # your ssh info
Then point a browser to localhost:12345 to open the dashboard.
In the "Repository" text entry box, enter the path to a pipe_analysis plots directory that has images in it; e.g., the plots directory inside a repo, such as /datasets/hsc/repo/rerun/RC/w_2020_38/DM-26820/plots.
Interaction should be straightforward. You can multi-select the "Plots" box. More than four will require you to scroll; it should also in principle be possible to change the display sizes of the plots if that would be helpful.
If any of the plots are miscategorized, let me know.
For example:
docker build -t dashboard .
docker run -it -p 12345:12345 dashboard # for port forwarding