DELPHI event display¶
One DELPHI event drawn straight from an edm4hep file — no DELGRA, no X11, no ROOT, no key4hep. A port of the matplotlib display written for the September 2026 talk, rebuilt on Plotly so the figure zooms, hovers, and runs in a browser.
In one minute¶
That writes a self-contained event.html, a event.json figure, a event.schema.json
describing the collections, and — with --view both — the same again for the 3-D scene.
Or open the app and drop a .edm4hep.root file on it: uproot runs in the browser under
Pyodide and reads the file locally. Nothing is uploaded.
What makes it different from a plotting script¶
Nothing is pinned by collection name. A DELPHI edm4hep file holds far more than one
collection of each kind — the 1994 data file carries 3 Track collections, 7 Vertex, 4
Cluster, 2 TrackerHit3D and 24 ParticleID. The display discovers them from podio's own
metadata and binds each part of the picture to one by role, which you can change.
See Collections and roles.
Only measured quantities are drawn. Tracks are helices from their own perigee parameters, started at the vertex they belong to, with the back-extrapolation to the IP dotted so the picture never claims a path the particle did not travel. In 3-D this matters more, not less: a vertex-detector hit is a strip, so it is a line in space, not a point. See Decoding rules and self-checks.
The numbers are checked on every run, not assumed. The curvature sign and the VD strip decoding are re-derived per event and printed. Those printed numbers are the test suite.
Credits¶
Ported from event_display.py by J. Zhang
(/eos/experiment/eealliance/Users/zhangj/event_display/), the reference for the physics and for
every decoding rule. Files are produced by
delphi-edm4hep; the collection
vocabulary matches its
collection map.
Styling via plotlyhep.