# Constructed data: 2,000 genes; 1,700 unchanged and 300 changed, with true |log2 fold change| evenly spread from 0.25 to 2
# (1.19-fold to 4-fold), half up and half down. Python 3.13, NumPy, pandas. Seed 1, the first one tried.
import numpy as np, pandas as pd
from model import study, true_lfc, GENES, N
rng = np.random.default_rng(1)
ctl, trt = study(rng)
d = pd.DataFrame(np.vstack([ctl, trt]).T.round(3), columns=[f"control_{i + 1}" for i in range(N)] + [f"treated_{i + 1}" for i in range(N)])
d.insert(0, "gene", [f"g{g + 1:04d}" for g in range(GENES)])
d.to_csv("log2_expression.csv", index=False)
pd.DataFrame({"gene": d.gene, "true_log2fc": true_lfc()}).to_csv("truth.csv", index=False)