test(02-03): add failing tests for ZhihuAdapter

- Test answer content extraction - Test zhuanlan (article) content extraction - Test cookie-based authentication - Test login page detection - Test AuthError handling for missing/expired cookies Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-03-13 18:24:37 +08:00
parent d17fbd8118
commit a6f4d34625
2 changed files with 447 additions and 0 deletions
@@ -0,0 +1,183 @@
+"""Tests for ZhihuAdapter
+
+Tests cover:
+- Answer content extraction
+- Zhuanlan (article) content extraction
+- Cookie-based authentication
+- Login page detection
+- AuthError handling
+"""
+
+import pytest
+from unittest.mock import MagicMock, patch, PropertyMock
+import asyncio
+
+from kb_pipeline.platforms.zhihu import ZhihuAdapter
+from kb_pipeline.errors import AuthError
+
+
+class TestZhihuAdapterAnswerContent:
+    """Tests for ZhihuAdapter answer extraction"""
+
+    def test_get_content_type_answer(self):
+        """Test content type detection for answer URLs"""
+        adapter = ZhihuAdapter()
+        adapter._cookie = "test"  # Avoid config loading
+
+        assert adapter.get_content_type("https://www.zhihu.com/question/123456/answer/789012") == "answer"
+        assert adapter.get_content_type("https://zhihu.com/question/123/answer/456") == "answer"
+
+    def test_get_content_type_zhuanlan(self):
+        """Test content type detection for zhuanlan URLs"""
+        adapter = ZhihuAdapter()
+        adapter._cookie = "test"  # Avoid config loading
+
+        assert adapter.get_content_type("https://zhuanlan.zhihu.com/p/123456") == "zhuanlan"
+        assert adapter.get_content_type("https://zhuanlan.zhihu.com/p/abc123") == "zhuanlan"
+
+    def test_get_content_type_unknown(self):
+        """Test content type detection for unknown URLs"""
+        adapter = ZhihuAdapter()
+        adapter._cookie = "test"  # Avoid config loading
+
+        assert adapter.get_content_type("https://www.zhihu.com/question/123456") == "unknown"
+        assert adapter.get_content_type("https://www.zhihu.com/") == "unknown"
+        assert adapter.get_content_type("https://example.com") == "unknown"
+
+    def test_get_output_filename(self):
+        """Test output filename generation"""
+        adapter = ZhihuAdapter()
+        adapter._cookie = "test"  # Avoid config loading
+
+        assert adapter.get_output_filename("789012", "answer", "metadata.json") == "zhihu_789012_metadata.json"
+        assert adapter.get_output_filename("123456", "zhuanlan", "content.md") == "zhihu_123456_content.md"
+
+    @pytest.mark.asyncio
+    async def test_fetch_metadata_answer(self, mock_zhihu_answer_result):
+        """Test fetching answer metadata"""
+        adapter = ZhihuAdapter()
+        adapter._cookie = "test_cookie"  # Set cookie directly
+
+        with patch.object(adapter, '_extract_with_playwright', return_value=mock_zhihu_answer_result):
+            metadata = await adapter.fetch_metadata("789012", "answer")
+
+        assert metadata["title"] == "Test Question Title"
+        assert metadata["author"] == "Test Author"
+        assert metadata["type"] == "answer"
+
+    @pytest.mark.asyncio
+    async def test_fetch_content_answer(self, mock_zhihu_answer_result):
+        """Test fetching answer content as Markdown"""
+        adapter = ZhihuAdapter()
+        adapter._cookie = "test_cookie"  # Set cookie directly
+
+        with patch.object(adapter, '_extract_with_playwright', return_value=mock_zhihu_answer_result):
+            content = await adapter.fetch_content("789012", "answer")
+
+        assert "---" in content  # YAML frontmatter
+        assert "title:" in content
+        assert "Test Question Title" in content
+        assert "This is the answer content" in content
+
+    @pytest.mark.asyncio
+    async def test_fetch_metadata_no_cookie_raises_auth_error(self):
+        """Test that missing cookie raises AuthError"""
+        adapter = ZhihuAdapter()
+        adapter._cookie = ""  # Empty cookie
+
+        with pytest.raises(AuthError) as exc_info:
+            await adapter.fetch_metadata("789012", "answer")
+
+        assert exc_info.value.auth_type == "zhihu_cookie"
+
+
+class TestZhihuAdapterZhuanlanContent:
+    """Tests for ZhihuAdapter zhuanlan (article) extraction"""
+
+    @pytest.mark.asyncio
+    async def test_fetch_metadata_zhuanlan(self, mock_zhihu_zhuanlan_result):
+        """Test fetching zhuanlan metadata"""
+        adapter = ZhihuAdapter()
+        adapter._cookie = "test_cookie"  # Set cookie directly
+
+        with patch.object(adapter, '_extract_with_playwright', return_value=mock_zhihu_zhuanlan_result):
+            metadata = await adapter.fetch_metadata("123456", "zhuanlan")
+
+        assert metadata["title"] == "Test Article Title"
+        assert metadata["author"] == "Test Author"
+        assert metadata["type"] == "zhuanlan"
+
+    @pytest.mark.asyncio
+    async def test_fetch_content_zhuanlan(self, mock_zhihu_zhuanlan_result):
+        """Test fetching zhuanlan content as Markdown"""
+        adapter = ZhihuAdapter()
+        adapter._cookie = "test_cookie"  # Set cookie directly
+
+        with patch.object(adapter, '_extract_with_playwright', return_value=mock_zhihu_zhuanlan_result):
+            content = await adapter.fetch_content("123456", "zhuanlan")
+
+        assert "---" in content  # YAML frontmatter
+        assert "title:" in content
+        assert "Test Article Title" in content
+        assert "This is the article content" in content
+
+    @pytest.mark.asyncio
+    async def test_detect_login_page_raises_auth_error(self):
+        """Test that login page detection raises AuthError"""
+        adapter = ZhihuAdapter()
+        adapter._cookie = "expired_cookie"  # Set cookie directly
+
+        # Simulate login page response (empty content or login title)
+        login_page_result = {
+            "type": "answer",
+            "title": "登录 - 知乎",
+            "content": "",
+            "content_markdown": "",
+        }
+
+        with patch.object(adapter, '_extract_with_playwright', return_value=login_page_result):
+            with pytest.raises(AuthError) as exc_info:
+                await adapter.fetch_metadata("789012", "answer")
+
+        assert exc_info.value.auth_type == "zhihu_cookie"
+        assert "过期" in str(exc_info.value.cause) or "expired" in str(exc_info.value.cause).lower()
+
+    @pytest.mark.asyncio
+    async def test_detect_short_content_raises_auth_error(self):
+        """Test that very short content (< 100 chars) raises AuthError"""
+        adapter = ZhihuAdapter()
+        adapter._cookie = "test_cookie"
+
+        # Simulate response with very short content
+        short_content_result = {
+            "type": "answer",
+            "title": "Some Title",
+            "content": "x",  # Very short content
+            "content_markdown": "",
+        }
+
+        with patch.object(adapter, '_extract_with_playwright', return_value=short_content_result):
+            with pytest.raises(AuthError) as exc_info:
+                await adapter.fetch_metadata("789012", "answer")
+
+        assert exc_info.value.auth_type == "zhihu_cookie"
+
+
+class TestZhihuAdapterCookieHandling:
+    """Tests for cookie handling"""
+
+    def test_cookie_from_settings(self):
+        """Test cookie is loaded from settings"""
+        with patch('kb_pipeline.platforms.zhihu.get_settings') as mock_settings:
+            mock_secret = MagicMock()
+            mock_secret.get_secret_value.return_value = "test_cookie_value"
+            mock_settings.return_value.zhihu_cookie = mock_secret
+
+            adapter = ZhihuAdapter()
+            assert adapter.cookie == "test_cookie_value"
+
+    def test_platform_constant(self):
+        """Test PLATFORM constant"""
+        adapter = ZhihuAdapter()
+        adapter._cookie = "test"  # Avoid config loading
+        assert adapter.PLATFORM == "zhihu"