"""Replication script for F2-smoke-01. Run from workspace root."""
import json, os
import numpy as np, pandas as pd, statsmodels.api as sm
from scipy import stats

d = pd.read_csv("data/panel.csv")
n_in = len(d)
issues = []
na = d[["firm_id", "x", "y"]].isna().any(axis=1)
if na.any():
    issues.append({"issue": f"{int(na.sum())} rows with missing values", "handling": "dropped (complete-case)"})
    d = d[~na]
dup = d.duplicated()
if dup.any():
    issues.append({"issue": f"{int(dup.sum())} exact duplicate rows", "handling": "retained (cannot distinguish from genuine repeats); flagged"})
fin = np.isfinite(d[["x", "y"]]).all(axis=1)
if not fin.all():
    issues.append({"issue": f"{int((~fin).sum())} non-finite values", "handling": "dropped"})
    d = d[fin]

X = sm.add_constant(d["x"].to_numpy())
y = d["y"].to_numpy()
ols = sm.OLS(y, X).fit()
cl = sm.OLS(y, X).fit(cov_type="cluster", cov_kwds={"groups": d["firm_id"].to_numpy(), "use_correction": True})
G = int(d["firm_id"].nunique())
b, se_c = float(cl.params[1]), float(cl.bse[1])
tcrit = stats.t.ppf(0.975, G - 1)  # t with G-1 df for clustered inference

res = {
    "task_id": "F2-smoke-01",
    "beta_hat": b,
    "alpha_hat": float(cl.params[0]),
    "se_ols": float(ols.bse[1]),
    "se_cluster": se_c,
    "beta_ci_low": b - tcrit * se_c,
    "beta_ci_high": b + tcrit * se_c,
    "n_analysed": int(len(d)),
    "n_clusters": G,
    "feasibility": "complete",
    "feasibility_reason": None,
    "data_issues": issues,
    "n_input_rows": {"panel.csv": int(n_in)},
    "random_seed": None,
    # supplementary diagnostics
    "se_cluster_ratio_to_ols": se_c / float(ols.bse[1]),
    "ci_t_df": G - 1,
    "beta_ci_low_normal": b - 1.959963985 * se_c,
    "beta_ci_high_normal": b + 1.959963985 * se_c,
    "beta_ci_low_ols": float(ols.conf_int()[1, 0]),
    "beta_ci_high_ols": float(ols.conf_int()[1, 1]),
}
os.makedirs("out", exist_ok=True)
with open("out/results.json", "w") as f:
    json.dump(res, f, indent=2)
print(json.dumps(res, indent=2))