feat: support huawei cloud vector database (#16141)

api/tests/integration_tests/vdb/__mock/huaweicloudvectordb.py

@@ -0,0 +1,88 @@
+import os
+
+import pytest
+from _pytest.monkeypatch import MonkeyPatch
+from api.core.rag.datasource.vdb.field import Field
+from elasticsearch import Elasticsearch
+
+
+class MockIndicesClient:
+    def __init__(self):
+        pass
+
+    def create(self, index, mappings, settings):
+        return {"acknowledge": True}
+
+    def refresh(self, index):
+        return {"acknowledge": True}
+
+    def delete(self, index):
+        return {"acknowledge": True}
+
+    def exists(self, index):
+        return True
+
+
+class MockClient:
+    def __init__(self, **kwargs):
+        self.indices = MockIndicesClient()
+
+    def index(self, **kwargs):
+        return {"acknowledge": True}
+
+    def exists(self, **kwargs):
+        return True
+
+    def delete(self, **kwargs):
+        return {"acknowledge": True}
+
+    def search(self, **kwargs):
+        return {
+            "took": 1,
+            "hits": {
+                "hits": [
+                    {
+                        "_source": {
+                            Field.CONTENT_KEY.value: "abcdef",
+                            Field.VECTOR.value: [1, 2],
+                            Field.METADATA_KEY.value: {},
+                        },
+                        "_score": 1.0,
+                    },
+                    {
+                        "_source": {
+                            Field.CONTENT_KEY.value: "123456",
+                            Field.VECTOR.value: [2, 2],
+                            Field.METADATA_KEY.value: {},
+                        },
+                        "_score": 0.9,
+                    },
+                    {
+                        "_source": {
+                            Field.CONTENT_KEY.value: "a1b2c3",
+                            Field.VECTOR.value: [3, 2],
+                            Field.METADATA_KEY.value: {},
+                        },
+                        "_score": 0.8,
+                    },
+                ]
+            },
+        }
+
+
+MOCK = os.getenv("MOCK_SWITCH", "false").lower() == "true"
+
+
+@pytest.fixture
+def setup_client_mock(request, monkeypatch: MonkeyPatch):
+    if MOCK:
+        monkeypatch.setattr(Elasticsearch, "__init__", MockClient.__init__)
+        monkeypatch.setattr(Elasticsearch, "index", MockClient.index)
+        monkeypatch.setattr(Elasticsearch, "exists", MockClient.exists)
+        monkeypatch.setattr(Elasticsearch, "delete", MockClient.delete)
+        monkeypatch.setattr(Elasticsearch, "search", MockClient.search)
+
+    yield
+
+    if MOCK:
+        monkeypatch.undo()

api/tests/integration_tests/vdb/huawei/test_huawei_cloud.py

@@ -0,0 +1,28 @@
+from core.rag.datasource.vdb.huawei.huawei_cloud_vector import HuaweiCloudVector, HuaweiCloudVectorConfig
+from tests.integration_tests.vdb.__mock.huaweicloudvectordb import setup_client_mock
+from tests.integration_tests.vdb.test_vector_store import AbstractVectorTest, get_example_text, setup_mock_redis
+
+
+class HuaweiCloudVectorTest(AbstractVectorTest):
+    def __init__(self):
+        super().__init__()
+        self.vector = HuaweiCloudVector(
+            "dify",
+            HuaweiCloudVectorConfig(
+                hosts="https://127.0.0.1:9200",
+                username="dify",
+                password="dify",
+            ),
+        )
+
+    def search_by_vector(self):
+        hits_by_vector = self.vector.search_by_vector(query_vector=self.example_embedding)
+        assert len(hits_by_vector) == 3
+
+    def search_by_full_text(self):
+        hits_by_full_text = self.vector.search_by_full_text(query=get_example_text())
+        assert len(hits_by_full_text) == 3
+
+
+def test_huawei_cloud_vector(setup_mock_redis, setup_client_mock):
+    HuaweiCloudVectorTest().run_all_tests()