Campaign 26.04.1-stable

Earlier campaign

This page records an earlier campaign. Newer background-mixed campaigns (26.07.0, 26.07.1) and our local productions are listed on Available Datasets, the current catalog.

CSV conversion of the 26.04.1-stable ePIC background-mixed campaign.

Naming note

The campaign is labelled 26.04 (software release 26.04.1-stable), but the DIDs were registered in May 2026 (created_at: 2026-05-05). The 26.04 in the name refers to the software/release line, not the production date.

What this campaign contains

Background-mixed simulation with exactly one signal event per 2 µs frame (Bkg_Exact1S_2us) and the gold-coated (10 µm) beampipe configuration: DIS/NC at 10x100 and 10x275 GeV, minQ2 = 1, 10, 100, 1000. Each dataset has a FULL (npsim, edm4hep) and a RECO (EICrecon, *.eicrecon.edm4eic.root) partition in Rucio, under DIDs like:

epic:/RECO/26.04.1/epic_craterlake/Bkg_Exact1S_2us/GoldCt/10um/DIS/NC/10x100/minQ2=1

Key parameters (from each dataset's Rucio metadata):

Field	Value
software_release	26.04.1-stable
data_level	reconstruction
geometry_config	craterlake_<energy>
generator	pythia8
is_background_mixed	true
ion_species	p
beams	electron × ion GeV (e.g. 10×100)

How the CSVs are produced

The simulation pipeline handles official campaigns with Rucio dataset cards: the card's files: are the Rucio root:// PFNs, and each job runs the configured converter (edm4eic_* ROOT macros or eic2ai) on them. See CSV Convert for the converters and Data Format for the CSV columns.

1. generate_datasets csv_eicrecon --rucio -c <config> — lists the DIDs matching the config's rucio_did_pattern (and rucio_did_filters, e.g. is_background_mixed=true) and writes one dataset card YAML per DID into <datasets_dir>/csv_eicrecon/.
2. 40_csv_convert.py csv_eicrecon -c <config> — reads those cards and generates the conversion jobs, one output directory per dataset.

No config for 26.04.1 is kept in simulation-pipeline/configs/; config-off-26-06.yaml is the current example of a Rucio-driven conversion.

Output layout

The Rucio path tree is not mirrored. Each dataset collapses to one flat directory (slug), and each ROOT file gives one CSV per role, named after the input file:

<csv_eicrecon.output>/
  Bkg_Exact1S_2us_GoldCt_10um_DIS_NC_10x100_q2-gt-1/
    <stem>.trk_hits.csv.zip
    <stem>.calo_clusters.csv.zip
    ...
  Bkg_Exact1S_2us_GoldCt_10um_DIS_NC_10x275_q2-gt-1/
    ...

• Directory slug joins the DID path below epic_craterlake/ with _, with minQ2=1 rewritten to q2-gt-1 (no = in names).
• <stem> is the ROOT file name without .eicrecon.edm4eic.root.
• The ROOT-macro converter also keeps the unzipped <stem>.<role>.csv; eic2ai writes the .csv.zip only.

Dataset cards

Each dataset has a YAML card written by generate_datasets --rucio carrying the DID, slug, beam energy, output directory, the Rucio metadata, and the full file list:

dataset: epic:/RECO/26.04.1/epic_craterlake/Bkg_Exact1S_2us/GoldCt/10um/DIS/NC/10x100/minQ2=1
slug: Bkg_Exact1S_2us_GoldCt_10um_DIS_NC_10x100_q2-gt-1
stage: csv_eicrecon
energy: 10x100
output: <csv_eicrecon.output>/Bkg_Exact1S_2us_GoldCt_10um_DIS_NC_10x100_q2-gt-1
metadata:
  source: rucio
  rucio_metadata:
    software_release: 26.04.1-stable
    electron_beam_energy_gev: 10
    ion_beam_energy_gev: 100
    q2_min_gev2: 1
    is_background_mixed: true
n_files: 123
files:
  - root://.../<stem>.eicrecon.edm4eic.root