e2e: Enhance the e2e testing of the ai-proxy plugin based on the LLM mock server

plugins/wasm-go/extensions/ai-proxy/README.md

@@ -130,10 +130,11 @@ Azure OpenAI 所对应的 `type` 为 `azure`。它特有的配置字段如下：
 
 通义千问所对应的 `type` 为 `qwen`。它特有的配置字段如下：
 
-| 名称                 | 数据类型            | 填写要求 | 默认值 | 描述                                                               |
-|--------------------|-----------------|------|-----|------------------------------------------------------------------|
-| `qwenEnableSearch` | boolean         | 非必填  | -   | 是否启用通义千问内置的互联网搜索功能。                          |
-| `qwenFileIds`      | array of string | 非必填  | -   | 通过文件接口上传至Dashscope的文件 ID，其内容将被用做 AI 对话的上下文。不可与 `context` 字段同时配置。 |
+| 名称                   | 数据类型        | 填写要求 | 默认值 | 描述                                                         |
+| ---------------------- | --------------- | -------- | ------ | ------------------------------------------------------------ |
+| `qwenEnableSearch`     | boolean         | 非必填   | -      | 是否启用通义千问内置的互联网搜索功能。                       |
+| `qwenFileIds`          | array of string | 非必填   | -      | 通过文件接口上传至Dashscope的文件 ID，其内容将被用做 AI 对话的上下文。不可与 `context` 字段同时配置。 |
+| `qwenEnableCompatible` | boolean         | 非必填   | false  | 开启通义千问兼容模式。启用通义千问兼容模式后，将调用千问的兼容模式接口，同时对请求/响应不做修改。 |
 
 #### 百川智能 (Baichuan AI)
 

plugins/wasm-go/extensions/ai-proxy/README_EN.md

@@ -106,6 +106,7 @@ For Qwen (Tongyi Qwen), the corresponding `type` is `qwen`. Its unique configura
 |--------------------|-----------------|----------------------|---------------|------------------------------------------------------------------------------------------------------------------------|
 | `qwenEnableSearch`  | boolean          | Optional             | -             | Whether to enable the built-in Internet search function provided by Qwen.                                             |
 | `qwenFileIds`       | array of string   | Optional             | -             | The file IDs uploaded via the Dashscope file interface, whose content will be used as context for AI conversations. Cannot be configured with the `context` field. |
+| `qwenEnableCompatible` | boolean          | Optional | false         | Enable Qwen compatibility mode. When Qwen compatibility mode is enabled, the compatible mode interface of Qwen will be called, and the request/response will not be modified. |
 
 #### Baichuan AI
 

test/e2e/conformance/base/llm-mock.yaml

@@ -0,0 +1,46 @@
+# Copyright (c) 2025 Alibaba Group Holding Ltd.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#      http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+apiVersion: v1
+kind: Namespace
+metadata:
+  name: higress-conformance-ai-backend
+  labels:
+    higress-conformance: infra
+---
+apiVersion: v1
+kind: Pod
+metadata:
+  name: llm-mock
+  namespace: higress-conformance-ai-backend
+  labels:
+    name: llm-mock
+spec:
+  containers:
+    - name: llm-mock
+      image: registry.cn-hangzhou.aliyuncs.com/hxt/llm-mock:latest
+      ports:
+        - containerPort: 3000
+---
+apiVersion: v1
+kind: Service
+metadata:
+  name: llm-mock-service
+  namespace: higress-conformance-ai-backend
+spec:
+  selector:
+    name: llm-mock
+  clusterIP: None
+  ports:
+    - port: 3000

test/e2e/conformance/tests/go-wasm-ai-proxy.yaml

@@ -1,4 +1,4 @@
-# Copyright (c) 2022 Alibaba Group Holding Ltd.
+# Copyright (c) 2025 Alibaba Group Holding Ltd.
 #
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.
@@ -14,74 +14,193 @@
 apiVersion: networking.k8s.io/v1
 kind: Ingress
 metadata:
-  annotations:
-  name: wasmplugin-ai-proxy-openai
-  namespace: higress-conformance-infra
+  name: wasmplugin-ai-proxy-baidu
+  namespace: higress-conformance-ai-backend
 spec:
   ingressClassName: higress
   rules:
-    - host: "openai.ai.com"
+    - host: "qianfan.baidubce.com"
       http:
         paths:
           - pathType: Prefix
             path: "/"
             backend:
               service:
-                name: infra-backend-v1
+                name: llm-mock-service
                 port:
-                  number: 8080
+                  number: 3000
+---
+apiVersion: networking.k8s.io/v1
+kind: Ingress
+metadata:
+  name: wasmplugin-ai-proxy-doubao
+  namespace: higress-conformance-ai-backend
+spec:
+  ingressClassName: higress
+  rules:
+    - host: "ark.cn-beijing.volces.com"
+      http:
+        paths:
+          - pathType: Prefix
+            path: "/"
+            backend:
+              service:
+                name: llm-mock-service
+                port:
+                  number: 3000
+---
+apiVersion: networking.k8s.io/v1
+kind: Ingress
+metadata:
+  name: wasmplugin-ai-proxy-minimax-v2-api
+  namespace: higress-conformance-ai-backend
+spec:
+  ingressClassName: higress
+  rules:
+    - host: "api.minimax.chat-v2-api"
+      http:
+        paths:
+          - pathType: Prefix
+            path: "/"
+            backend:
+              service:
+                name: llm-mock-service
+                port:
+                  number: 3000
+---
+apiVersion: networking.k8s.io/v1
+kind: Ingress
+metadata:
+  name: wasmplugin-ai-proxy-minimax-pro-api
+  namespace: higress-conformance-ai-backend
+spec:
+  ingressClassName: higress
+  rules:
+    - host: "api.minimax.chat-pro-api"
+      http:
+        paths:
+          - pathType: Prefix
+            path: "/"
+            backend:
+              service:
+                name: llm-mock-service
+                port:
+                  number: 3000
+---
+apiVersion: networking.k8s.io/v1
+kind: Ingress
+metadata:
+  name: wasmplugin-ai-proxy-qwen-compatible-mode
+  namespace: higress-conformance-ai-backend
+spec:
+  ingressClassName: higress
+  rules:
+    - host: "dashscope.aliyuncs.com-compatible-mode"
+      http:
+        paths:
+          - pathType: Prefix
+            path: "/"
+            backend:
+              service:
+                name: llm-mock-service
+                port:
+                  number: 3000
 ---
 apiVersion: networking.k8s.io/v1
 kind: Ingress
 metadata:
-  annotations:
   name: wasmplugin-ai-proxy-qwen
-  namespace: higress-conformance-infra
+  namespace: higress-conformance-ai-backend
 spec:
   ingressClassName: higress
   rules:
-    - host: "qwen.ai.com"
+    - host: "dashscope.aliyuncs.com"
       http:
         paths:
           - pathType: Prefix
             path: "/"
             backend:
               service:
-                name: infra-backend-v1
+                name: llm-mock-service
                 port:
-                  number: 8080
+                  number: 3000
 ---
 apiVersion: extensions.higress.io/v1alpha1
 kind: WasmPlugin
 metadata:
   name: ai-proxy
   namespace: higress-system
 spec:
-  priority: 200
+  defaultConfigDisable: true
+  phase: UNSPECIFIED_PHASE
+  priority: 100
   matchRules:
     - config:
         provider:
-          type: "openai"
-          customSettings:
-            - name: "max_tokens"
-              value: 123
-              overwrite: false
-            - name: "temperature"
-              value: 0.66
-              overwrite: true
+          apiTokens:
+            - fake_token
+          modelMapping:
+            'gpt-3': ernie-3.5-8k
+            '*': ernie-3.5-8k
+          type: baidu
+      ingress:
+        - higress-conformance-ai-backend/wasmplugin-ai-proxy-baidu
+    - config:
+        provider:
+          apiTokens:
+            - fake_token
+          modelMapping:
+            '*': fake_doubao_endpoint
+          type: doubao
+      ingress:
+        - higress-conformance-ai-backend/wasmplugin-ai-proxy-doubao
+    - config:
+        provider:
+          apiTokens:
+            - fake_token
+          modelMapping:
+            'gpt-3': abab6.5s-chat
+            'gpt-4': abab6.5g-chat
+            '*': abab6.5t-chat
+          type: minimax
+      ingress:
+        - higress-conformance-ai-backend/wasmplugin-ai-proxy-minimax-v2-api
+    - config:
+        provider:
+          apiTokens:
+            - fake_token
+          modelMapping:
+            'gpt-3': abab6.5s-chat
+            'gpt-4': abab6.5g-chat
+            '*': abab6.5t-chat
+          type: minimax
+          minimaxApiType: pro
+          minimaxGroupId: 1
+      ingress:
+        - higress-conformance-ai-backend/wasmplugin-ai-proxy-minimax-pro-api
+    - config:
+        provider:
+          apiTokens:
+            - fake_token
+          modelMapping:
+            'gpt-3': qwen-turbo
+            'gpt-35-turbo': qwen-plus
+            'gpt-4-*': qwen-max
+            '*': qwen-turbo
+          type: qwen
+          qwenEnableCompatible: true
       ingress:
-        - higress-conformance-infra/wasmplugin-ai-proxy-openai
+        - higress-conformance-ai-backend/wasmplugin-ai-proxy-qwen-compatible-mode
     - config:
         provider:
-          type: "qwen"
-          apiTokens: "fake-token"
-          customSettings:
-            - name: "max_tokens"
-              value: 123
-              overwrite: false
-            - name: "temperature"
-              value: 0.66
-              overwrite: true
+          apiTokens:
+            - fake_token
+          modelMapping:
+            'gpt-3': qwen-turbo
+            'gpt-35-turbo': qwen-plus
+            'gpt-4-*': qwen-max
+            '*': qwen-turbo
+          type: qwen
       ingress:
-        - higress-conformance-infra/wasmplugin-ai-proxy-qwen
-  url: oci://higress-registry.cn-hangzhou.cr.aliyuncs.com/plugins/ai-proxy:1.0.0
+        - higress-conformance-ai-backend/wasmplugin-ai-proxy-qwen
+  url: oci://higress-registry.cn-hangzhou.cr.aliyuncs.com/plugins/ai-proxy:1.0.0

test/e2e/conformance/utils/http/http.go

@@ -76,6 +76,7 @@ const (
 	ContentTypeFormUrlencoded         = "application/x-www-form-urlencoded"
 	ContentTypeMultipartForm          = "multipart/form-data"
 	ContentTypeTextPlain              = "text/plain"
+	ContentTypeTextEventStream        = "text/event-stream"
 )
 
 const (
@@ -140,11 +141,12 @@ type ExpectedRequest struct {
 
 // Response defines expected properties of a response from a backend.
 type Response struct {
-	StatusCode    int
-	Headers       map[string]string
-	Body          []byte
-	ContentType   string
-	AbsentHeaders []string
+	StatusCode           int
+	Headers              map[string]string
+	Body                 []byte
+	JsonBodyIgnoreFields []string
+	ContentType          string
+	AbsentHeaders        []string
 }
 
 // requiredConsecutiveSuccesses is the number of requests that must succeed in a row
@@ -601,6 +603,7 @@ func CompareResponse(cRes *roundtripper.CapturedResponse, expected Assertion) er
 
 			switch cTyp {
 			case ContentTypeTextPlain:
+			case ContentTypeTextEventStream:
 				if !bytes.Equal(expected.Response.ExpectedResponse.Body, cRes.Body) {
 					return fmt.Errorf("expected %s body to be %s, got %s", cTyp, string(expected.Response.ExpectedResponse.Body), string(cRes.Body))
 				}
@@ -616,7 +619,7 @@ func CompareResponse(cRes *roundtripper.CapturedResponse, expected Assertion) er
 					return fmt.Errorf("failed to unmarshall CapturedResponse body %s, %s", string(cRes.Body), err.Error())
 				}
 
-				if !reflect.DeepEqual(eResBody, cResBody) {
+				if err := CompareJSONWithIgnoreFields(eResBody, cResBody, expected.Response.ExpectedResponse.JsonBodyIgnoreFields); err != nil {
 					return fmt.Errorf("expected %s body to be %s, got %s", cTyp, string(expected.Response.ExpectedResponse.Body), string(cRes.Body))
 				}
 			case ContentTypeFormUrlencoded:
@@ -663,6 +666,47 @@ func CompareResponse(cRes *roundtripper.CapturedResponse, expected Assertion) er
 	}
 	return nil
 }
+
+// CompareJSONWithIgnoreFields compares two JSON objects, ignoring specified fields
+func CompareJSONWithIgnoreFields(eResBody, cResBody map[string]interface{}, ignoreFields []string) error {
+	for key, eVal := range eResBody {
+		if contains(ignoreFields, key) {
+			continue
+		}
+
+		cVal, exists := cResBody[key]
+		if !exists {
+			return fmt.Errorf("field %s exists in expected response but not in captured response", key)
+		}
+
+		if !reflect.DeepEqual(eVal, cVal) {
+			return fmt.Errorf("field %s mismatch: expected %v, got %v", key, eVal, cVal)
+		}
+	}
+
+	// Check if captured response has extra fields (excluding ignored fields)
+	for key := range cResBody {
+		if contains(ignoreFields, key) {
+			continue
+		}
+
+		if _, exists := eResBody[key]; !exists {
+			return fmt.Errorf("field %s exists in captured response but not in expected response", key)
+		}
+	}
+
+	return nil
+}
+
+func contains(slice []string, str string) bool {
+	for _, s := range slice {
+		if s == str {
+			return true
+		}
+	}
+	return false
+}
+
 func ParseFormUrlencodedBody(body []byte) (map[string][]string, error) {
 	ret := make(map[string][]string)
 	kvs, err := url.ParseQuery(string(body))

test/e2e/conformance/utils/suite/suite.go

@@ -136,6 +136,7 @@ func New(s Options) *ConformanceTestSuite {
 			"base/nacos.yaml",
 			"base/dubbo.yaml",
 			"base/opa.yaml",
+			"base/llm-mock.yaml",
 		}
 	}
 
@@ -173,6 +174,7 @@ func (suite *ConformanceTestSuite) Setup(t *testing.T) {
 		"higress-conformance-infra",
 		"higress-conformance-app-backend",
 		"higress-conformance-web-backend",
+		"higress-conformance-ai-backend",
 	}
 	kubernetes.NamespacesMustBeAccepted(t, suite.Client, suite.TimeoutConfig, namespaces)