"""PMC 全文转换测试 - TDD 驱动开发
✅ v0.2.1 更新：所有测试已重写为异步版本，使用 aiohttp mock
设计原则：
1. 必须有 PMCID 才能获取全文
2. 无 PMCID 直接报错，不降级
3. 只返回全文格式（XML/Markdown/Text），不返回元数据
4. 职责单一，与其他工具配合
5. 支持按章节提取内容
"""
import sys
from pathlib import Path
from unittest.mock import AsyncMock, Mock, patch
import pytest
# 添加 src 目录到路径
project_root = Path(__file__).parent.parent.parent.parent
src_path = project_root / "src"
if str(src_path) not in sys.path:
sys.path.insert(0, str(src_path))
# ============================================================================
# 测试数据 - 模拟 PMC XML 响应
# ============================================================================
# 基础 XML（用于原有测试）
SAMPLE_PMC_XML = """<?xml version="1.0" encoding="UTF-8"?>
<pmc-articleset>
<article>
<article-meta>
<article-title>Machine Learning in Healthcare: A Comprehensive Review</article-title>
<contrib-group>
<contrib contrib-type="author">
<name>
<surname>Smith</surname>
<given-names>John</given-names>
</name>
</contrib>
</contrib-group>
</article-meta>
<abstract>
<title>Abstract</title>
<p>This study explores machine learning in healthcare.</p>
</abstract>
<body>
<sec sec-type="intro">
<title>Introduction</title>
<p>Machine learning is transforming healthcare.</p>
</sec>
<sec sec-type="methods">
<title>Methods</title>
<p>We collected data from 1000 patients.</p>
</sec>
</body>
</article>
</pmc-articleset>
"""
# 完整 XML（用于章节提取测试）
SAMPLE_PMC_XML_WITH_SECTIONS = """<?xml version="1.0" encoding="UTF-8"?>
<pmc-articleset>
<article>
<article-meta>
<article-title>Machine Learning in Healthcare</article-title>
</article-meta>
<body>
<sec sec-type="intro">
<title>Introduction</title>
<p>This is the introduction section.</p>
<p>Machine learning is transforming healthcare.</p>
</sec>
<sec sec-type="methods">
<title>Methods</title>
<p>We collected data from 1000 patients.</p>
<p>The study was approved by the ethics committee.</p>
</sec>
<sec sec-type="results">
<title>Results</title>
<p>Our model achieved 95% accuracy.</p>
</sec>
<sec sec-type="discussion">
<title>Discussion</title>
<p>The results demonstrate the potential of ML in healthcare.</p>
</sec>
<sec sec-type="conclusion">
<title>Conclusion</title>
<p>Further research is needed to validate these findings.</p>
</sec>
</body>
</article>
</pmc-articleset>
"""
# ============================================================================
# aiohttp Mock 辅助函数
# ============================================================================
def create_mock_aiohttp_response(text_content: str, status: int = 200) -> Mock:
"""创建模拟的 aiohttp 响应对象"""
mock_response = Mock()
mock_response.status = status
mock_response.text = AsyncMock(return_value=text_content)
return mock_response
def create_mock_aiohttp_session(xml_content: str) -> Mock:
"""创建模拟的 aiohttp.ClientSession
返回一个可以用于 patch('aiohttp.ClientSession') 的 mock 对象
"""
# 创建 mock 响应
mock_response = create_mock_aiohttp_response(xml_content, 200)
# 创建模拟的 async context manager (用于 session.get() 的返回值)
# 关键：不能使用 AsyncMock，因为它本身就是一个 coroutine
# 我们需要一个普通对象，但有 __aenter__ 和 __aexit__ 异步方法
class MockGetContextManager:
async def __aenter__(self):
return mock_response
async def __aexit__(self, *args):
pass
# 创建 mock session
mock_session = Mock()
mock_session.get = Mock(return_value=MockGetContextManager())
# session 本身也需要是 async context manager
class MockSessionContextManager:
async def __aenter__(self):
return mock_session
async def __aexit__(self, *args):
pass
return MockSessionContextManager()
# ============================================================================
# 测试类
# ============================================================================
class TestPMCFulltextCore:
"""核心功能测试：有 PMCID 返回全文"""
@pytest.fixture
def pubmed_service(self):
"""创建 PubMed 服务实例"""
from article_mcp.services.pubmed_search import PubMedService
return PubMedService(logger=Mock())
@pytest.mark.asyncio
async def test_with_pmcid_returns_three_formats(self, pubmed_service):
"""测试：有 PMCID 时返回 XML、Markdown、Text 三种格式"""
with patch("aiohttp.ClientSession") as mock_session_class:
mock_session_class.return_value = create_mock_aiohttp_session(SAMPLE_PMC_XML)
result = await pubmed_service.get_pmc_fulltext_html_async("PMC1234567")
# 核心验证：三种格式都存在
assert result["pmc_id"] == "PMC1234567"
assert result["fulltext_available"] is True
assert result["error"] is None
assert "fulltext_xml" in result
assert "fulltext_markdown" in result
assert "fulltext_text" in result
# 内容非空
assert result["fulltext_xml"] is not None
assert result["fulltext_markdown"] is not None
assert result["fulltext_text"] is not None
# 验证不包含冗余的元数据字段（其他工具负责）
assert "title" not in result or result.get("title") is None
assert (
"authors" not in result
or result.get("authors") is None
or result.get("authors") == []
)
assert "journal_name" not in result or result.get("journal_name") is None
assert "abstract" not in result or result.get("abstract") is None
assert "publication_date" not in result or result.get("publication_date") is None
assert "pmc_link" not in result or result.get("pmc_link") is None
assert "fulltext_html" not in result or result.get("fulltext_html") is None
@pytest.mark.asyncio
async def test_markdown_format_is_valid(self, pubmed_service):
"""测试：Markdown 格式只包含正文，不包含标题、作者、摘要"""
with patch("aiohttp.ClientSession") as mock_session_class:
mock_session_class.return_value = create_mock_aiohttp_session(SAMPLE_PMC_XML)
result = await pubmed_service.get_pmc_fulltext_html_async("PMC1234567")
markdown = result["fulltext_markdown"]
# Markdown 应该有结构
assert len(markdown) > 100
# 不应该有 XML 标签
assert "<article-title>" not in markdown
assert "</article-title>" not in markdown
assert "<sec" not in markdown
# 不应该包含元数据（标题、作者、摘要）
# Markdown 不应该以文章标题开头（那是元数据）
assert not markdown.startswith("# Machine Learning")
# 不应该包含 "Authors:" 标记
assert "Authors:" not in markdown
# 不应该包含 "Abstract" 标题
assert markdown.find("Abstract") == -1 or markdown.find("Abstract") > markdown.find(
"Introduction"
)
@pytest.mark.asyncio
async def test_text_format_is_clean(self, pubmed_service):
"""测试：纯文本格式干净，无标签"""
with patch("aiohttp.ClientSession") as mock_session_class:
mock_session_class.return_value = create_mock_aiohttp_session(SAMPLE_PMC_XML)
result = await pubmed_service.get_pmc_fulltext_html_async("PMC1234567")
text = result["fulltext_text"]
# 纯文本应该有内容
assert len(text) > 100
# 不应该有任何 XML 标签
assert "<article-title>" not in text
assert "</article-title>" not in text
assert "<sec" not in text
assert "<p>" not in text
assert "</p>" not in text
assert "<?xml" not in text
class TestPMCFulltextErrors:
"""错误处理测试：无 PMCID 或请求失败"""
@pytest.fixture
def pubmed_service(self):
"""创建 PubMed 服务实例"""
from article_mcp.services.pubmed_search import PubMedService
return PubMedService(logger=Mock())
def test_empty_pmcid_returns_error(self, pubmed_service):
"""测试：空 PMCID 返回错误（不需要网络调用，保持同步测试）"""
result = pubmed_service.get_pmc_fulltext_html("")
# 应该返回错误
assert result["error"] is not None
assert result["fulltext_available"] is False
assert result["pmc_id"] is None
def test_none_pmcid_returns_error(self, pubmed_service):
"""测试：None PMCID 返回错误（不需要网络调用，保持同步测试）"""
result = pubmed_service.get_pmc_fulltext_html(None)
# 应该返回错误
assert result["error"] is not None
assert result["fulltext_available"] is False
@pytest.mark.asyncio
async def test_network_error_returns_error(self, pubmed_service):
"""测试：网络错误返回错误"""
with patch("aiohttp.ClientSession") as mock_session_class:
# 模拟网络错误 - 在 text() 调用时抛出异常
mock_session = AsyncMock()
mock_get_cm = AsyncMock()
mock_get_cm.__aenter__.side_effect = Exception("Network error")
mock_get_cm.__aexit__.return_value = None
mock_session.get.return_value = mock_get_cm
mock_session.__aenter__.return_value = mock_session
mock_session.__aexit__.return_value = None
mock_session_class.return_value = mock_session
result = await pubmed_service.get_pmc_fulltext_html_async("PMC1234567")
# 应该返回错误
assert result["error"] is not None
assert result["fulltext_available"] is False
assert result["pmc_id"] == "PMC1234567"
@pytest.mark.asyncio
async def test_empty_xml_returns_error(self, pubmed_service):
"""测试：空 XML 返回错误"""
with patch("aiohttp.ClientSession") as mock_session_class:
mock_session_class.return_value = create_mock_aiohttp_session("")
result = await pubmed_service.get_pmc_fulltext_html_async("PMC1234567")
# 应该返回错误
assert result.get("error") is not None or result.get("fulltext_available") is False
class TestPMCFulltextNormalization:
"""PMCID 标准化测试"""
@pytest.fixture
def pubmed_service(self):
"""创建 PubMed 服务实例"""
from article_mcp.services.pubmed_search import PubMedService
return PubMedService(logger=Mock())
@pytest.mark.asyncio
async def test_pmcid_without_prefix_normalized(self, pubmed_service):
"""测试：不带 PMC 前缀的 ID 被标准化"""
with patch("aiohttp.ClientSession") as mock_session_class:
mock_session_class.return_value = create_mock_aiohttp_session(SAMPLE_PMC_XML)
result = await pubmed_service.get_pmc_fulltext_html_async("1234567")
# 应该自动添加 PMC 前缀
assert result["pmc_id"] == "PMC1234567"
assert result["fulltext_available"] is True
# ============================================================================
# 新增：章节提取测试
# ============================================================================
class TestPMCSectionExtraction:
"""章节提取功能测试"""
@pytest.fixture
def pubmed_service(self):
"""创建 PubMed 服务实例"""
from article_mcp.services.pubmed_search import PubMedService
return PubMedService(logger=Mock())
@pytest.mark.asyncio
async def test_extract_single_section_methods(self, pubmed_service):
"""测试：提取单个章节（Methods）"""
with patch("aiohttp.ClientSession") as mock_session_class:
mock_session_class.return_value = create_mock_aiohttp_session(
SAMPLE_PMC_XML_WITH_SECTIONS
)
result = await pubmed_service.get_pmc_fulltext_html_async(
"PMC1234567", sections=["methods"]
)
# 验证返回值
assert result["fulltext_available"] is True
assert "fulltext_markdown" in result
markdown = result["fulltext_markdown"]
# 只包含 Methods 章节
assert "Methods" in markdown
assert "We collected data from 1000 patients" in markdown
# 不包含其他章节
assert "Introduction" not in markdown or markdown.index("Methods") < markdown.index(
"Introduction"
)
assert "Results" not in markdown
assert "Discussion" not in markdown
assert "Conclusion" not in markdown
# 验证章节信息
assert result.get("sections_requested") == ["methods"]
assert result.get("sections_found") == ["methods"]
assert result.get("sections_missing") == []
@pytest.mark.asyncio
async def test_extract_multiple_sections(self, pubmed_service):
"""测试：提取多个章节（Methods + Discussion）"""
with patch("aiohttp.ClientSession") as mock_session_class:
mock_session_class.return_value = create_mock_aiohttp_session(
SAMPLE_PMC_XML_WITH_SECTIONS
)
result = await pubmed_service.get_pmc_fulltext_html_async(
"PMC1234567", sections=["methods", "discussion"]
)
markdown = result["fulltext_markdown"]
# 包含 Methods 和 Discussion
assert "Methods" in markdown
assert "Discussion" in markdown
assert "We collected data from 1000 patients" in markdown
assert "results demonstrate the potential" in markdown
# 不包含其他章节
assert "Results" not in markdown or markdown.index("Results") > markdown.index(
"Discussion"
)
assert "Conclusion" not in markdown
# 验证章节信息
assert set(result.get("sections_found", [])) == {"methods", "discussion"}
assert result.get("sections_missing") == []
@pytest.mark.asyncio
async def test_extract_nonexistent_section(self, pubmed_service):
"""测试：提取不存在的章节"""
with patch("aiohttp.ClientSession") as mock_session_class:
mock_session_class.return_value = create_mock_aiohttp_session(
SAMPLE_PMC_XML_WITH_SECTIONS
)
result = await pubmed_service.get_pmc_fulltext_html_async(
"PMC1234567",
sections=["appendix"], # 不存在的章节
)
# 仍然成功返回，但章节未找到
assert result["fulltext_available"] is True
assert result.get("sections_found") == []
assert result.get("sections_missing") == ["appendix"]
# Markdown 为空或只有占位符
markdown = result.get("fulltext_markdown", "")
assert len(markdown) == 0 or "appendix" not in markdown.lower()
@pytest.mark.asyncio
async def test_sections_none_returns_all(self, pubmed_service):
"""测试：sections=None 返回全部章节"""
with patch("aiohttp.ClientSession") as mock_session_class:
mock_session_class.return_value = create_mock_aiohttp_session(
SAMPLE_PMC_XML_WITH_SECTIONS
)
result = await pubmed_service.get_pmc_fulltext_html_async("PMC1234567", sections=None)
markdown = result["fulltext_markdown"]
# 包含所有章节
assert "Introduction" in markdown
assert "Methods" in markdown
assert "Results" in markdown
assert "Discussion" in markdown
assert "Conclusion" in markdown
# 不包含章节信息字段（全部章节时不标记）
assert "sections_requested" not in result
assert "sections_found" not in result
@pytest.mark.asyncio
async def test_partial_sections_found(self, pubmed_service):
"""测试：部分章节找到，部分未找到"""
with patch("aiohttp.ClientSession") as mock_session_class:
mock_session_class.return_value = create_mock_aiohttp_session(
SAMPLE_PMC_XML_WITH_SECTIONS
)
result = await pubmed_service.get_pmc_fulltext_html_async(
"PMC1234567", sections=["methods", "appendix"]
)
# Methods 找到了，Appendix 未找到
assert result.get("sections_found") == ["methods"]
assert result.get("sections_missing") == ["appendix"]
# Markdown 只包含 Methods
markdown = result.get("fulltext_markdown", "")
assert "Methods" in markdown
assert "appendix" not in markdown.lower()
@pytest.mark.asyncio
async def test_empty_sections_list(self, pubmed_service):
"""测试：空章节列表"""
with patch("aiohttp.ClientSession") as mock_session_class:
mock_session_class.return_value = create_mock_aiohttp_session(
SAMPLE_PMC_XML_WITH_SECTIONS
)
result = await pubmed_service.get_pmc_fulltext_html_async("PMC1234567", sections=[])
# 返回空内容
assert result["fulltext_available"] is True
markdown = result.get("fulltext_markdown", "")
assert len(markdown) == 0 or markdown.strip() == ""
assert result.get("sections_found") == []
assert result.get("sections_requested") == []
# ============================================================================
# 运行测试
# ============================================================================
if __name__ == "__main__":
pytest.main([__file__, "-v", "-s", "--tb=short"])
