"""partition_entry_text grounding: paraphrased percentages must survive (截图 GPA 丢失根因)."""

from __future__ import annotations

from app.claim_validator import partition_entry_text

FACTS = [{"id": "entry_description", "field": "description", "text": "学习数据结构、计算机视觉课程。GPA: 4.3/5.0，排名前百分之10。"}]


def test_percentage_paraphrase_is_not_quarantined() -> None:
    """LLM 把「前百分之10」改写成「前 10%」是同一事实，不得隔离。"""
    candidate = "主修数据结构、计算机视觉等课程。GPA 4.3/5.0，年级排名前 10%。"
    optimized, suggestions, _warnings = partition_entry_text(candidate, FACTS)

    assert "10%" in optimized
    assert "4.3" in optimized
    assert suggestions == []


def test_truly_new_numbers_remain_visible_and_require_confirmation() -> None:
    """用户没提过的数字（如「提升 37%」）必须继续被隔离。"""
    facts = [{"id": "entry_description", "field": "description", "text": "完成数据库课程项目。"}]
    optimized, suggestions, _warnings = partition_entry_text("完成数据库课程项目，性能提升 37%。", facts)

    assert "37" in optimized
    assert suggestions
    assert _warnings == ["candidate_requires_confirmation"]


def test_bullet_line_structure_is_preserved() -> None:
    """LLM 按行输出的 bullet 不得在防虚构分区时被拍平成一行（前端排版根因）。"""
    facts = [{"id": "entry_description", "field": "description", "text": "负责需求分析与全链路开发。使用 LangGraph 编排优化流程。完成部署上线。"}]
    candidate = "• 负责需求分析与全链路开发。\n• 使用 LangGraph 编排优化流程。\n• 完成部署上线。"
    optimized, _suggestions, _warnings = partition_entry_text(candidate, facts)

    assert optimized == candidate