VSCode中C#调用OpenAI接口传递Payload遇BadRequest错误求助

我在VSCode中使用C#基于OpenAI Chat API开发代码翻译器，该功能在Postman中测试可正常完成代码翻译，但运行下方代码时收到BadRequest错误：

我怀疑是Payload格式问题，调整后依旧无法解决，恳请帮忙排查。

原始代码 #

using System;
using System.Net.Http;
using System.Text;
using System.Threading.Tasks;

namespace OpenAIProgram
{
    public class CodeTranslator
    {
        public static async Task<string> TranslateCodeAsync(string apiKey, string apiUrl, string jsonBody)
        {
            string result = string.Empty;

            try
            {
                if (string.IsNullOrEmpty(apiUrl) || string.IsNullOrEmpty(jsonBody))
                    throw new Exception("Missing parameter for execution");

                using var httpClient = new HttpClient();
                httpClient.DefaultRequestHeaders.Add("Authorization", $"Bearer {apiKey}");

                var request = new HttpRequestMessage(HttpMethod.Post, apiUrl);
                request.Headers.Add("Connection", "keep-alive");
                request.Content = new StringContent(jsonBody, Encoding.UTF8, "application/json");

                var response = await httpClient.SendAsync(request);

                if (response.IsSuccessStatusCode)
                {
                    result = await response.Content.ReadAsStringAsync();
                }
                else
                {
                    throw new Exception("Response code was " + response.StatusCode);
                }
            }
            catch (Exception e)
            {
                result = e.Message;
                result += "\n" + e.GetType() + ": " + e.Message;
                if (e.StackTrace != null)
                {
                    foreach (var trace in e.StackTrace.Split("\n"))
                    {
                        result += "\n\t" + trace.Trim();
                    }
                }
            }

            return result;
        }

        public static async Task Main()
        {
            string apiKey = "---";
            string apiUrl = "https://api.openai.com/v1/chat/completions";
            string jsonBody = @"
{
    ""model"": ""gpt-3.5-turbo"",
    ""messages"": [
        {
            ""role"": ""system"",
            ""content"": ""Ready to Translate .""
        },
        {
            ""role"": ""user"",
            ""content"": ""class FibonacciExample1{public static void main(String[] args){int n1=0,n2=1,n3,i,count=10;System.out.print(n1+"" + n2); for (i = 2; i < count; ++i) { n3 = n1 + n2; System.out.print(\"" + n3 + ""\"); n1 = n2; n2 = n3; }} }""
        },
        {
            ""role"": ""system"",
            ""content"": ""Translate the code from[Java] to[Python]."""
        }
    ],
    ""temperature"": 1.0,
    ""max_tokens"": 150
}";

            // Translate the code
            string translatedCode = await TranslateCodeAsync(apiKey, apiUrl, jsonBody);

            Console.WriteLine(translatedCode);
        }
    }
}

问题排查与解决方案 #

1. JSON格式转义错误

原始jsonBody中用户消息的Java代码部分存在转义混乱，导致JSON结构无效。比如System.out.print(\"" + n3 + ""\")的转义方式错误，正确的JSON字符串转义应该统一使用\"。

2. 消息角色与顺序问题

Chat API的消息序列需遵循合理逻辑：system角色用于设定上下文，user角色用于提交具体请求。原始代码中最后一条消息用system角色发送翻译指令，不符合API设计逻辑，应改为user角色，或合并到同一条用户消息中。

3. 多余的请求头

手动添加Connection: keep-alive头无必要，HttpClient会自动处理连接管理，多余的头可能引发请求格式问题。

修正后的代码

建议使用对象序列化生成JSON（需引入System.Text.Json），彻底避免手动拼接的格式错误：

using System;
using System.Net.Http;
using System.Text;
using System.Text.Json;
using System.Threading.Tasks;

namespace OpenAIProgram
{
    public class CodeTranslator
    {
        // 定义API请求模型
        private class ChatCompletionRequest
        {
            public string model { get; set; }
            public ChatMessage[] messages { get; set; }
            public double temperature { get; set; }
            public int max_tokens { get; set; }
        }

        private class ChatMessage
        {
            public string role { get; set; }
            public string content { get; set; }
        }

        public static async Task<string> TranslateCodeAsync(string apiKey, string apiUrl, ChatCompletionRequest requestBody)
        {
            string result = string.Empty;

            try
            {
                if (string.IsNullOrEmpty(apiUrl) || requestBody == null)
                    throw new Exception("Missing parameter for execution");

                using var httpClient = new HttpClient();
                httpClient.DefaultRequestHeaders.Add("Authorization", $"Bearer {apiKey}");

                // 序列化请求对象为JSON
                string jsonBody = JsonSerializer.Serialize(requestBody);
                var request = new HttpRequestMessage(HttpMethod.Post, apiUrl);
                request.Content = new StringContent(jsonBody, Encoding.UTF8, "application/json");

                var response = await httpClient.SendAsync(request);

                if (response.IsSuccessStatusCode)
                {
                    result = await response.Content.ReadAsStringAsync();
                }
                else
                {
                    // 获取错误详情，帮助排查
                    string errorContent = await response.Content.ReadAsStringAsync();
                    throw new Exception($"Response code was {response.StatusCode}. Error details: {errorContent}");
                }
            }
            catch (Exception e)
            {
                result = e.Message;
                result += "\n" + e.GetType() + ": " + e.Message;
                if (e.StackTrace != null)
                {
                    foreach (var trace in e.StackTrace.Split("\n"))
                    {
                        result += "\n\t" + trace.Trim();
                    }
                }
            }

            return result;
        }

        public static async Task Main()
        {
            string apiKey = "---";
            string apiUrl = "https://api.openai.com/v1/chat/completions";
            
            // 构造正确的请求体
            var requestBody = new ChatCompletionRequest
            {
                model = "gpt-3.5-turbo",
                temperature = 1.0,
                max_tokens = 150,
                messages = new[]
                {
                    new ChatMessage
                    {
                        role = "system",
                        content = "你是专业的代码翻译工具，将用户提供的代码准确翻译成指定语言。"
                    },
                    new ChatMessage
                    {
                        role = "user",
                        content = @"请将以下Java代码翻译成Python：
class FibonacciExample1{public static void main(String[] args){int n1=0,n2=1,n3,i,count=10;System.out.print(n1+"" ""+n2); for (i = 2; i < count; ++i) { n3 = n1 + n2; System.out.print("" ""+n3); n1 = n2; n2 = n3; }} }"
                    }
                }
            };

            // 执行翻译
            string translatedCode = await TranslateCodeAsync(apiKey, apiUrl, requestBody);

            Console.WriteLine(translatedCode);
        }
    }
}

额外建议 #

• 捕获错误时，读取响应内容获取OpenAI返回的具体错误信息，比仅返回状态码更有助于排查问题。
• 始终使用序列化工具生成JSON，避免手动拼接导致的格式疏漏。

内容的提问来源于stack exchange，提问作者Jes