✨ tests(model): 添加多模态功能测试用例，验证模型对图像和文本的处理能力

tests/test_multimodal_quick.py

@@ -0,0 +1,62 @@
+"""
+glm-4.5v 多模态功能测试
+"""
+import requests
+import json
+
+# 创建一个1x1像素的红色图片作为测试
+tiny_red_image = (
+    "data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAADUlEQVR42mP8z8DwHwAFBQIAX8jx0gAAAABJRU5ErkJggg=="
+)
+
+# API配置
+api_url = "http://localhost:8080/v1/chat/completions"
+api_key = "sk-your-api-key"
+
+# 构建正确的多模态请求
+request_data = {
+    "model": "glm-4.5v",  # 使用多模态模型
+    "messages": [
+        {
+            "role": "user",
+            "content": [  # content必须是数组
+                {
+                    "type": "text",
+                    "text": "这是什么颜色的图片？"
+                },
+                {
+                    "type": "image_url",
+                    "image_url": {
+                        "url": tiny_red_image
+                    }
+                }
+            ]
+        }
+    ],
+    "stream": False
+}
+
+print("发送的请求:")
+print(json.dumps(request_data, indent=2, ensure_ascii=False))
+print("\n" + "="*60)
+
+# 发送请求
+headers = {
+    "Authorization": f"Bearer {api_key}",
+    "Content-Type": "application/json"
+}
+
+try:
+    response = requests.post(api_url, json=request_data, headers=headers)
+    print(f"响应状态码: {response.status_code}")
+    
+    if response.status_code == 200:
+        result = response.json()
+        print("\n模型回复:")
+        print(result["choices"][0]["message"]["content"])
+    else:
+        print("\n错误响应:")
+        print(response.text)
+        
+except Exception as e:
+    print(f"\n发生错误: {e}")