# Constructed data: a cytokine in 12 control and 12 treated mice; LOD 5 pg/mL. True model (model.py): log-normal,
# control median 8 pg/mL, treated median 16 pg/mL (twice as high), log-SD 0.8.
# The seed (3) is the first from 1 upward with three or four control values and at most one treated value below
# the LOD, the typical pattern for this model (about 28% of control values expected below the LOD).
import numpy as np, pandas as pd
from model import simulate
rng = np.random.default_rng(3)
a, ea = simulate(rng, 12, 8); b, eb = simulate(rng, 12, 16)
pd.DataFrame({"mouse": [f"c{i+1}" for i in range(12)] + [f"t{i+1}" for i in range(12)],
              "group": ["control"] * 12 + ["treated"] * 12,
              "conc_pg_ml": np.round(np.r_[a, b], 2), "detected": np.r_[ea, eb]}).to_csv("cytokine.csv", index=False)