feat: 添加 Dahuagov Spider（大化县政府网爬虫）+ 测试

2026-05-09 13:47:48 +08:00
parent b1bbf64287
commit 5e0582d499
2 changed files with 134 additions and 0 deletions
@@ -0,0 +1,62 @@
+import pytest
+from unittest.mock import AsyncMock, patch, MagicMock
+from app.crawler.dahuagov_spider import DahuagovSpider
+from app.crawler.base import PipelineConfig
+
+
+@pytest.mark.asyncio
+async def test_dahuagov_spider_attributes():
+    spider = DahuagovSpider()
+    assert spider.name == "dahuagov"
+    assert spider.source_code == "dahuagov"
+    assert spider.source_name == "大化县政府网采购公告"
+
+
+@pytest.mark.asyncio
+async def test_dahuagov_spider_pipeline_config():
+    spider = DahuagovSpider()
+    config = spider.get_pipeline_config()
+    assert isinstance(config, PipelineConfig)
+    assert config.filter_enabled is False
+    assert config.notify_mode == "all"
+    assert config.mark_sent is True
+
+
+@pytest.mark.asyncio
+async def test_dahuagov_spider_crawl():
+    spider = DahuagovSpider()
+
+    html = """
+    <html><body>
+      <ul class="more-list">
+        <li><span>2026-05-08</span>
+            <a href="./detail/1.html" title="公告1">公告1</a></li>
+        <li><span>2026-05-07</span>
+            <a href="./detail/2.html" title="公告2">公告2</a></li>
+      </ul>
+    </body></html>
+    """
+
+    mock_response = MagicMock()
+    mock_response.status_code = 200
+    mock_response.text = html
+
+    with patch("httpx.AsyncClient.get", AsyncMock(return_value=mock_response)):
+        result = await spider.crawl()
+        assert result.total_count == 2
+        assert len(result.announcements) == 2
+        assert result.success is True
+
+
+@pytest.mark.asyncio
+async def test_dahuagov_spider_crawl_empty():
+    spider = DahuagovSpider()
+
+    mock_response = MagicMock()
+    mock_response.status_code = 200
+    mock_response.text = "<html><body></body></html>"
+
+    with patch("httpx.AsyncClient.get", AsyncMock(return_value=mock_response)):
+        result = await spider.crawl()
+        assert result.total_count == 0
+        assert result.success is True