"""E1.1 — Jev sur le banc de reference, compare aux baselines sans IA."""
import os, sys
from collections import defaultdict
sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
sys.path.insert(0, os.path.dirname(os.path.abspath(__file__)))
from common import Jev, run_parallel, load, outfile, med, pct_ci
from bench.reference_set import build, CRITERIA

REPS = 2

def build_trials():
    items = build()
    tr = []
    for it in items:
        for r in range(REPS):
            tr.append({"tid": f"{it['id']}_{r}", "state": it["text"],
                       "meta": {"id": it["id"], "truth": it["truth"], "diff": it["diff"],
                                "leurre": it["leurre"], "tpl": it["tpl"], "rep": r},
                       "questions": {"cat": {"type": "choice",
                            "instructions": "Dans quelle categorie classer ce message ?",
                            "criteria": CRITERIA}}})
    return tr

def analyze(path):
    obs = load(path)
    print(f"\n{'='*86}\nE1.1 — JEV SUR LE BANC DE REFERENCE\n{'='*86}")
    print(f"{len(obs)} appels, {len(set(o['meta']['id'] for o in obs))} items distincts\n")

    par = defaultdict(lambda: [0, 0])
    conf_ok, conf_ko = [], []
    for o in obs:
        m = o["meta"]; a = o["answers"]["cat"]
        juste = a["choice"] == m["truth"]
        par[m["diff"]][1] += 1; par["GLOBAL"][1] += 1
        if juste:
            par[m["diff"]][0] += 1; par["GLOBAL"][0] += 1; conf_ok.append(a["confidence"])
        else:
            conf_ko.append(a["confidence"])

    print(f"{'difficulte':14s} {'justesse':>24s}")
    for d in ("explicite", "implicite", "piege", "GLOBAL"):
        k, n = par[d]
        print(f"{d:14s} {pct_ci(k, n):>24s}")

    print(f"\nconfiance si juste : {med(conf_ok):.3f} (n={len(conf_ok)})")
    print(f"confiance si faux  : {med(conf_ko):.3f} (n={len(conf_ko)})" if conf_ko
          else "aucune erreur")

    # piege : tombe-t-il sur le leurre ?
    leurres = defaultdict(lambda: [0, 0])
    for o in obs:
        m = o["meta"]
        if m["diff"] != "piege": continue
        c = o["answers"]["cat"]["choice"]
        leurres["total"][1] += 1
        if c != m["truth"]:
            leurres["total"][0] += 1
            if c == m["leurre"]: leurres["sur_leurre"][0] += 1
    if leurres["total"][1]:
        e, n = leurres["total"][0], leurres["total"][1]
        print(f"\npieges : {e} erreurs sur {n}, dont {leurres['sur_leurre'][0]} "
              f"exactement sur le leurre")

    # stabilite entre repetitions
    par_id = defaultdict(list)
    for o in obs: par_id[o["meta"]["id"]].append(o["answers"]["cat"]["choice"])
    instables = [i for i, v in par_id.items() if len(set(v)) > 1]
    print(f"items dont la reponse change entre les 2 passages : {len(instables)}/{len(par_id)}")

    # comparaison finale
    print(f"\n{'='*86}\nECHELLE DE LECTURE\n{'='*86}")
    sys.path.insert(0, os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
    from baselines.no_ai import main as bl
    import io, contextlib
    buf = io.StringIO()
    with contextlib.redirect_stdout(buf):
        res = bl()
    k, n = par["GLOBAL"]
    lignes = [("Jev (zero exemple)", k, n, {d: tuple(par[d]) for d in
               ("explicite", "implicite", "piege")})] + \
             [(nom, kk, nn, pd_) for nom, kk, nn, pd_ in res if "naif" not in nom]
    print(f"{'methode':32s} {'global':>13s} {'explicite':>13s} {'implicite':>13s} {'piege':>13s}")
    print("-" * 86)
    for nom, kk, nn, pd_ in lignes:
        row = f"{nom:32s} {kk/nn*100:11.1f} % "
        for d in ("explicite", "implicite", "piege"):
            a, b = pd_.get(d, (0, 1))
            row += f"{a/b*100:11.1f} % "
        print(row)

if __name__ == "__main__":
    if len(sys.argv) > 1 and sys.argv[1] == "analyze":
        analyze(sys.argv[2]); sys.exit()
    j = Jev(); tr = build_trials()
    print(f"E1.1 : {len(tr)} appels (400 items x {REPS})")
    p = run_parallel(j, tr, outfile("t6_banc"), workers=8, label="banc ")
    analyze(p)