#!/usr/bin/env python3
"""Export flat CSV versions of the dataset and the summary tables."""
from __future__ import annotations

import csv
import json
import pathlib

ROOT = pathlib.Path(__file__).resolve().parent.parent
DATASET = ROOT / "dataset"
RESULTS = ROOT / "results"


def export_dataset() -> None:
    items = [json.loads(l) for l in (DATASET / "msq49.jsonl").open(encoding="utf-8")]
    path = DATASET / "msq49.csv"
    with path.open("w", newline="", encoding="utf-8") as fh:
        w = csv.writer(fh)
        w.writerow(["id", "cls", "domain", "question", "stem",
                    "n_positions", "position_texts", "position_schools",
                    "majority_schools", "minority_school", "sources", "notes"])
        for it in items:
            w.writerow([
                it["id"], it["cls"], it["domain"], it["question"], it["stem"],
                len(it["positions"]),
                " | ".join(p["txt"] for p in it["positions"]),
                " | ".join("/".join(p["schools"]) for p in it["positions"]),
                "/".join(it["majority_schools"]), it["minority_school"] or "",
                " ; ".join(it["sources"]), it.get("notes", ""),
            ])
    print(f"wrote {path} ({len(items)} rows)")


def export_tables() -> None:
    p = RESULTS / "metrics.json"
    if not p.exists():
        print("no metrics.json yet")
        return
    metrics = json.loads(p.read_text())
    rows = []
    for model, rep in metrics.items():
        h = rep.get("headline", {})
        n = rep.get("mcq", {}).get("mcq_neutral", {})
        s = rep.get("mcq", {}).get("mcq_specified", {})
        ft = rep.get("free_text", {})
        rows.append(dict(
            model=model,
            PRR=h.get("PRR"), PRR_A=h.get("PRR_A"), PRR_B=h.get("PRR_B"),
            PRR_C=h.get("PRR_C"), PRR_mass=h.get("PRR_mass"),
            MSQ=h.get("MSQ"), blind_spot_gap=h.get("blind_spot_gap"),
            acc_control_D=h.get("acc_D"),
            conf_contested=h.get("conf_contested"),
            conf_control=h.get("conf_control"),
            disclosure_default=ft.get("ft_default", {}).get("ALL", {}).get("disclosure_rate"),
            disclosure_technical=ft.get("ft_technical", {}).get("ALL", {}).get("disclosure_rate"),
            disclosure_two_schools=ft.get("ft_two_schools", {}).get("ALL", {}).get("disclosure_rate"),
            attr_two_schools=ft.get("ft_two_schools", {}).get("ALL", {}).get("plural_attribution_rate"),
            clarification_default=ft.get("ft_default", {}).get("ALL", {}).get("clarification_rate"),
            answer_changed_rate=rep.get("answer_change", {}).get("changed_rate"),
            mean_max_prob=n.get("mean_max_prob"),
            mean_entropy=n.get("mean_entropy"),
        ))
    path = RESULTS / "summary_tables.csv"
    with path.open("w", newline="", encoding="utf-8") as fh:
        w = csv.DictWriter(fh, fieldnames=list(rows[0].keys()))
        w.writeheader()
        w.writerows(rows)
    print(f"wrote {path} ({len(rows)} rows)")


if __name__ == "__main__":
    export_dataset()
    export_tables()