fix(audio): 兼容百炼 OpenAI 音频输入

为阿里云百炼 OpenAI-compatible Chat 统一保留标准 input_audio，并将旧版 audio_url 转换为 input_audio，同时根据 MIME 或扩展名补齐音频格式。补充请求资产水合测试，确认 input_audio.data 会转换为裸 Base64 且保留 format。验证：在 apps/api 执行 env -u AI_GATEWAY_TEST_DATABASE_URL go test ./... -count=1 全部通过。
2026-07-27 23:50:17 +08:00
parent 95776c84f6
commit 046c16fc69
4 changed files with 297 additions and 0 deletions
@@ -344,6 +344,79 @@ func TestOpenAIClientChatContract(t *testing.T) {
 	}
 }

+func TestOpenAIClientAliyunChatSendsStandardInputAudioAndNormalizesLegacyAudioURL(t *testing.T) {
+	var captured map[string]any
+	server := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		if err := json.NewDecoder(r.Body).Decode(&captured); err != nil {
+			t.Fatalf("decode request: %v", err)
+		}
+		_ = json.NewEncoder(w).Encode(map[string]any{
+			"id":     "chatcmpl-audio",
+			"object": "chat.completion",
+			"model":  captured["model"],
+			"choices": []any{map[string]any{
+				"message": map[string]any{"role": "assistant", "content": "ok"},
+			}},
+		})
+	}))
+	defer server.Close()
+
+	_, err := (OpenAIClient{HTTPClient: server.Client()}).Run(context.Background(), Request{
+		Kind:  "chat.completions",
+		Model: "qwen-omni",
+		Body: map[string]any{
+			"model": "qwen-omni",
+			"messages": []any{
+				map[string]any{
+					"role": "user",
+					"content": []any{
+						map[string]any{
+							"type": "input_audio",
+							"input_audio": map[string]any{
+								"data":   "https://cdn.example.com/audio/standard.wav",
+								"format": "wav",
+							},
+						},
+						map[string]any{
+							"type":      "audio_url",
+							"audio_url": map[string]any{"url": "https://cdn.example.com/audio/legacy.m4a"},
+						},
+					},
+				},
+			},
+		},
+		Candidate: store.RuntimeModelCandidate{
+			Provider:          "aliyun-bailian-openai",
+			BaseURL:           server.URL,
+			ProviderModelName: "qwen-omni",
+			Credentials:       map[string]any{"apiKey": "test-key"},
+		},
+	})
+	if err != nil {
+		t.Fatalf("run aliyun openai-compatible client: %v", err)
+	}
+
+	messages, _ := captured["messages"].([]any)
+	message, _ := messages[0].(map[string]any)
+	content, _ := message["content"].([]any)
+	standard, _ := content[0].(map[string]any)
+	if standard["type"] != "input_audio" {
+		t.Fatalf("standard input_audio was not preserved: %+v", standard)
+	}
+	standardAudio, _ := standard["input_audio"].(map[string]any)
+	if standardAudio["data"] != "https://cdn.example.com/audio/standard.wav" || standardAudio["format"] != "wav" {
+		t.Fatalf("unexpected standard input_audio: %+v", standardAudio)
+	}
+	legacy, _ := content[1].(map[string]any)
+	if legacy["type"] != "input_audio" {
+		t.Fatalf("legacy audio_url was not normalized: %+v", legacy)
+	}
+	legacyAudio, _ := legacy["input_audio"].(map[string]any)
+	if legacyAudio["data"] != "https://cdn.example.com/audio/legacy.m4a" || legacyAudio["format"] != "m4a" {
+		t.Fatalf("unexpected normalized legacy audio: %+v", legacyAudio)
+	}
+}
+
 func TestOpenAIClientImageEditUsesMultipartWithMultipleImages(t *testing.T) {
 	var contentType string
 	var receivedFields map[string][]string