# -*- coding: utf-8 -*-
"""讀取 oov_llm_screening.csv，計算 LLM 專名篩選之混淆矩陣與 precision/recall/F1。
gold standard = human_accept(1/0)。用法：python eval_llm_screening.py"""
import csv, os, sys
try:
    sys.stdout.reconfigure(encoding="utf-8", errors="replace")
except Exception:
    pass

here = os.path.dirname(os.path.abspath(__file__))
rows = list(csv.DictReader(open(os.path.join(here, "..", "oov_llm_screening.csv"), encoding="utf-8-sig")))
def keep_llm(r): return r["llm_decision"].strip() == "收錄"
def keep_h(r): return str(r["human_accept(1/0)"]).strip() in ("1", "1.0")

TP = sum(1 for r in rows if keep_llm(r) and keep_h(r))
FP = sum(1 for r in rows if keep_llm(r) and not keep_h(r))
FN = sum(1 for r in rows if not keep_llm(r) and keep_h(r))
TN = sum(1 for r in rows if not keep_llm(r) and not keep_h(r))
prec = TP/(TP+FP) if TP+FP else float("nan")
rec = TP/(TP+FN) if TP+FN else float("nan")
f1 = 2*prec*rec/(prec+rec) if prec+rec else float("nan")
print(f"N={len(rows)}  TP={TP} FP={FP} FN={FN} TN={TN}")
print(f"Precision={prec:.4f}  Recall={rec:.4f}  F1={f1:.4f}")
print(f"LLM suggested-keep={TP+FP} -> human-accepted={TP} ({TP/(TP+FP):.1%}); human-accepted={TP+FN}, LLM-missed(FN)={FN}")