Complete local hybrid search and improve agent reliability

SovereignAI-Starter/tests/test_web_search_api.py

@@ -0,0 +1,74 @@
+import asyncio
+import os
+import tempfile
+import unittest
+from types import SimpleNamespace
+from unittest.mock import AsyncMock, MagicMock, patch
+
+_TEST_DATA_DIR = None
+if "SOVEREIGNAI_DATA_DIR" not in os.environ:
+    _TEST_DATA_DIR = tempfile.TemporaryDirectory(prefix="sovereignai-web-search-tests-")
+    os.environ["SOVEREIGNAI_DATA_DIR"] = _TEST_DATA_DIR.name
+
+from fastapi.testclient import TestClient
+
+from app.main import app
+
+
+class _FakeSearchClient:
+    async def __aenter__(self):
+        return self
+
+    async def __aexit__(self, exc_type, exc, traceback):
+        return None
+
+    async def get(self, *args, **kwargs):
+        return SimpleNamespace(
+            content=b"<html>mock search</html>",
+            text="<html>mock search</html>",
+            raise_for_status=lambda: None,
+        )
+
+
+class WebSearchApiTests(unittest.TestCase):
+    @classmethod
+    def setUpClass(cls) -> None:
+        cls.client = TestClient(app)
+
+    def test_search_returns_per_source_and_total_fetch_durations(self) -> None:
+        candidates = [
+            {"title": "Source A", "url": "https://a.example/article", "snippet": "A excerpt"},
+            {"title": "Source B", "url": "https://b.example/article", "snippet": "B excerpt"},
+        ]
+
+        async def read_page(url: str) -> tuple[str, str, str]:
+            await asyncio.sleep(0.01)
+            title = "A article" if "a.example" in url else "B article"
+            return url, title, f"Content for {title}"
+
+        provider = MagicMock(default_model="test-model")
+        completion = {"choices": [{"message": {"content": "ملخص موثق."}}]}
+        with (
+            patch("app.main.httpx.AsyncClient", return_value=_FakeSearchClient()),
+            patch("app.main.parse_duckduckgo_results", return_value=candidates),
+            patch("app.main._validate_public_http_url", side_effect=lambda url: url),
+            patch("app.main._read_public_page", side_effect=read_page),
+            patch("app.main.get_model_provider", return_value=provider),
+            patch("app.main.get_completion", new=AsyncMock(return_value=completion)),
+        ):
+            response = self.client.post(
+                "/v1/web/search",
+                json={"query": "اختبار البحث", "max_results": 2},
+            )
+
+        self.assertEqual(response.status_code, 200, response.text)
+        body = response.json()
+        self.assertEqual(body["result"], "ملخص موثق.")
+        self.assertGreaterEqual(body["source_fetch_ms"], 0)
+        self.assertEqual(len(body["sources"]), 2)
+        self.assertTrue(all(source["fetch_ms"] >= 0 for source in body["sources"]))
+        self.assertTrue(all(source["status"] == "read" for source in body["sources"]))
+
+
+if __name__ == "__main__":
+    unittest.main()