使用 5001 的向量服务
This commit is contained in:
		@@ -114,7 +114,7 @@ func GetNetWorkEmbedding(id int) (embedding []float32) {
 | 
			
		||||
	host := viper.GetString("embedding.host")
 | 
			
		||||
	port := viper.GetInt("embedding.port")
 | 
			
		||||
	httpClient := &http.Client{}
 | 
			
		||||
	request, err := http.NewRequest("PUT", fmt.Sprintf("http://%s:%d/reverse/%d", host, port, id), nil)
 | 
			
		||||
	request, err := http.NewRequest("PUT", fmt.Sprintf("http://%s:%d/api/default/%d", host, port, id), nil)
 | 
			
		||||
	if err != nil {
 | 
			
		||||
		log.Println("请求失败:", err)
 | 
			
		||||
		return
 | 
			
		||||
 
 | 
			
		||||
@@ -46,9 +46,6 @@ class ResNetServer(BaseHTTPRequestHandler):
 | 
			
		||||
            self.end_headers()                                   # 完成服務器響應的標頭
 | 
			
		||||
            content_length = int(self.headers['Content-Length']) # 獲取請求的內容長度
 | 
			
		||||
            body = self.rfile.read(content_length)               # 獲取請求的內容
 | 
			
		||||
            #img_path = './data/' + str(uuid.uuid4())             # 生成一個隨機的圖像文件名
 | 
			
		||||
            #with open(img_path, 'wb') as f:                      # 將二進制圖像文件解碼為圖像數據保存在本地
 | 
			
		||||
            #    f.write(body)
 | 
			
		||||
            feat = towhee.blob(body).image_decode().image_embedding.timm(model_name='resnet50').tensor_normalize().to_list()
 | 
			
		||||
            results = json.dumps(feat[0].tolist())               # 將結果轉換為JSON格式
 | 
			
		||||
            self.wfile.write(bytes(results, 'utf-8'))            # 將結果發送給客戶端
 | 
			
		||||
 
 | 
			
		||||
		Reference in New Issue
	
	Block a user