"""test_humanizer.py —— Humanizer 单测（S4）。

覆盖：规则命中 / 块级保护（引用/表格/代码/图片逐字符一致）/ 幂等 / 空串不抛。
"""
from __future__ import annotations

import sys
from pathlib import Path

sys.path.insert(0, str(Path(__file__).resolve().parent.parent))

from app.services.humanizer import humanize_answer  # noqa: E402


class TestRules:
    def test_strikes_ai_phrases(self):
        text = "值得注意的是，水稻SPL基因家族在产量调控中发挥作用。此外，它影响穗发育。"
        out, stats = humanize_answer(text)
        assert "值得注意的是" not in out
        assert stats["rules_applied"] >= 1

    def test_idempotent(self):
        text = "此外，该基因影响发育。"
        out1, _ = humanize_answer(text)
        out2, _ = humanize_answer(out1)
        assert out1 == out2

    def test_empty_ok(self):
        assert humanize_answer("") == ("", {"rules_applied": 0})
        assert humanize_answer(None)[0] is None


class TestBlockProtection:
    def test_citation_spans_preserved(self):
        text = "该基因调控分蘖 [PMID: 12345678]。此外，另一研究 [DOI: 10.1000/x] 表明……"
        out, _ = humanize_answer(text)
        assert "[PMID: 12345678]" in out
        assert "[DOI: 10.1000/x]" in out

    def test_code_fence_preserved(self):
        code = "```python\nprint('值得注意的是')\n```"
        text = f"正文。{code}"
        out, _ = humanize_answer(text)
        assert code in out
        assert "值得注意的是" in out  # 代码块内的不被替换

    def test_table_preserved(self):
        table = "| 基因 | 功能 |\n|---|---|\n| OsSPL14 | 分蘖 |"
        text = f"值得注意的是，见下表。\n{table}"
        out, _ = humanize_answer(text)
        assert table in out

    def test_fig_marker_preserved(self):
        marker = "[[FIG:PMC12345]]"
        text = f"正文。\n{marker}"
        out, _ = humanize_answer(text)
        assert marker in out

    def test_image_md_preserved(self):
        img = "![](/api/figures/proxy?pmcid=PMC7820984&url=https%3A%2F%2Fexample)"
        text = f"正文 {img}"
        out, _ = humanize_answer(text)
        assert img in out