document-management/backend/tests/dms/test_q4a_search_highlight.py

"""搜索关键词高亮与正文命中片段的单元测试。"""

from __future__ import annotations

import sys
from pathlib import Path

import pytest

BACKEND_ROOT = Path(__file__).resolve().parents[2]
if str(BACKEND_ROOT) not in sys.path:
    sys.path.insert(0, str(BACKEND_ROOT))

from dms.services.document_query_service import _content_snippet


class TestContentSnippet:
    """_content_snippet 纯函数测试。"""

    def test_returns_none_when_no_content(self):
        assert _content_snippet(None, "关键词") is None

    def test_returns_none_when_no_keyword(self):
        assert _content_snippet("一些正文内容", None) is None
        assert _content_snippet("一些正文内容", "") is None
        assert _content_snippet("一些正文内容", "  ") is None

    def test_returns_none_when_keyword_not_in_content(self):
        assert _content_snippet("这是一段正文", "不存在") is None

    def test_keyword_at_start_no_prefix_ellipsis(self):
        content = "关键词在开头，后面有很多内容" * 10
        result = _content_snippet(content, "关键词")
        assert result is not None
        assert not result.startswith("……")
        assert result.endswith("……")
        assert "关键词" in result

    def test_keyword_at_end_no_suffix_ellipsis(self):
        content = "前面有很多内容" * 20 + "尾部关键词"
        result = _content_snippet(content, "尾部关键词")
        assert result is not None
        assert result.startswith("……")
        assert not result.endswith("……")
        assert "尾部关键词" in result

    def test_keyword_in_middle_both_ellipsis(self):
        content = "前面的内容。" * 20 + "目标词" + "后面的内容。" * 20
        result = _content_snippet(content, "目标词")
        assert result is not None
        assert result.startswith("……")
        assert result.endswith("……")
        assert "目标词" in result

    def test_short_content_no_ellipsis(self):
        content = "很短的关键词"
        result = _content_snippet(content, "关键词")
        assert result is not None
        assert not result.startswith("……")
        assert not result.endswith("……")
        assert result == content

    def test_case_insensitive_match(self):
        content = "Hello World Test"
        result = _content_snippet(content, "hello")
        assert result is not None
        assert "Hello" in result

    def test_chinese_case_insensitive(self):
        """中文无大小写，但函数不应崩溃。"""
        content = "这是一段包含关键词的正文内容"
        result = _content_snippet(content, "关键词")
        assert result is not None
        assert "关键词" in result

    def test_snippet_length_bounded(self):
        content = "A" * 500 + "目标" + "B" * 500
        result = _content_snippet(content, "目标")
        assert result is not None
        # 40 before + keyword + 80 after + 2 ellipsis chars
        assert len(result) <= 40 + 2 + 80 + 4  # 126

    def test_strips_keyword_whitespace(self):
        content = "前面的内容目标词后面的内容"
        result = _content_snippet(content, "  目标词  ")
        assert result is not None
        assert "目标词" in result