mirror of
https://github.com/farion1231/cc-switch.git
synced 2026-07-24 12:44:18 +08:00
Jason
a30e2096bb
Support Anthropic ↔ OpenAI Responses API format conversion alongside existing Chat Completions conversion. The Responses API uses a flat input/output structure with lifted function_call/function_call_output items and named SSE lifecycle events.
633 lines
23 KiB
Rust
633 lines
23 KiB
Rust
//! OpenAI Responses API 格式转换模块
|
|
//!
|
|
//! 实现 Anthropic Messages ↔ OpenAI Responses API 格式转换。
|
|
//! Responses API 是 OpenAI 2025 年推出的新一代 API，采用扁平化的 input/output 结构。
|
|
//!
|
|
//! 与 Chat Completions 的主要差异：
|
|
//! - tool_use/tool_result 从 message content 中"提升"为顶层 input item
|
|
//! - system prompt 使用 `instructions` 字段而非 system role message
|
|
//! - usage 字段命名与 Anthropic 一致 (input_tokens/output_tokens)
|
|
|
|
use crate::proxy::error::ProxyError;
|
|
use serde_json::{json, Value};
|
|
|
|
/// Anthropic 请求 → OpenAI Responses 请求
|
|
pub fn anthropic_to_responses(body: Value) -> Result<Value, ProxyError> {
|
|
let mut result = json!({});
|
|
|
|
// NOTE: 模型映射由上游统一处理（proxy::model_mapper），格式转换层只做结构转换。
|
|
if let Some(model) = body.get("model").and_then(|m| m.as_str()) {
|
|
result["model"] = json!(model);
|
|
}
|
|
|
|
// system → instructions (Responses API 使用 instructions 字段)
|
|
if let Some(system) = body.get("system") {
|
|
let instructions = if let Some(text) = system.as_str() {
|
|
text.to_string()
|
|
} else if let Some(arr) = system.as_array() {
|
|
arr.iter()
|
|
.filter_map(|msg| msg.get("text").and_then(|t| t.as_str()))
|
|
.collect::<Vec<_>>()
|
|
.join("\n\n")
|
|
} else {
|
|
String::new()
|
|
};
|
|
if !instructions.is_empty() {
|
|
result["instructions"] = json!(instructions);
|
|
}
|
|
}
|
|
|
|
// messages → input
|
|
if let Some(msgs) = body.get("messages").and_then(|m| m.as_array()) {
|
|
let input = convert_messages_to_input(msgs)?;
|
|
result["input"] = json!(input);
|
|
}
|
|
|
|
// max_tokens → max_output_tokens
|
|
if let Some(v) = body.get("max_tokens") {
|
|
result["max_output_tokens"] = v.clone();
|
|
}
|
|
|
|
// 直接透传的参数
|
|
if let Some(v) = body.get("temperature") {
|
|
result["temperature"] = v.clone();
|
|
}
|
|
if let Some(v) = body.get("top_p") {
|
|
result["top_p"] = v.clone();
|
|
}
|
|
if let Some(v) = body.get("stream") {
|
|
result["stream"] = v.clone();
|
|
}
|
|
|
|
// stop_sequences → 丢弃 (Responses API 不支持)
|
|
|
|
// 转换 tools (过滤 BatchTool)
|
|
if let Some(tools) = body.get("tools").and_then(|t| t.as_array()) {
|
|
let response_tools: Vec<Value> = tools
|
|
.iter()
|
|
.filter(|t| t.get("type").and_then(|v| v.as_str()) != Some("BatchTool"))
|
|
.map(|t| {
|
|
json!({
|
|
"type": "function",
|
|
"name": t.get("name").and_then(|n| n.as_str()).unwrap_or(""),
|
|
"description": t.get("description"),
|
|
"parameters": super::transform::clean_schema(
|
|
t.get("input_schema").cloned().unwrap_or(json!({}))
|
|
)
|
|
})
|
|
})
|
|
.collect();
|
|
|
|
if !response_tools.is_empty() {
|
|
result["tools"] = json!(response_tools);
|
|
}
|
|
}
|
|
|
|
if let Some(v) = body.get("tool_choice") {
|
|
result["tool_choice"] = v.clone();
|
|
}
|
|
|
|
Ok(result)
|
|
}
|
|
|
|
/// 将 Anthropic messages 数组转换为 Responses API input 数组
|
|
///
|
|
/// 核心转换逻辑：
|
|
/// - user/assistant 的 text 内容 → 对应 role 的 message item
|
|
/// - tool_use 从 assistant message 中"提升"为独立的 function_call item
|
|
/// - tool_result 从 user message 中"提升"为独立的 function_call_output item
|
|
/// - thinking blocks → 丢弃
|
|
fn convert_messages_to_input(messages: &[Value]) -> Result<Vec<Value>, ProxyError> {
|
|
let mut input = Vec::new();
|
|
|
|
for msg in messages {
|
|
let role = msg.get("role").and_then(|r| r.as_str()).unwrap_or("user");
|
|
let content = msg.get("content");
|
|
|
|
match content {
|
|
// 字符串内容
|
|
Some(Value::String(text)) => {
|
|
let content_type = if role == "assistant" {
|
|
"output_text"
|
|
} else {
|
|
"input_text"
|
|
};
|
|
input.push(json!({
|
|
"role": role,
|
|
"content": [{ "type": content_type, "text": text }]
|
|
}));
|
|
}
|
|
|
|
// 数组内容（多模态/工具调用）
|
|
Some(Value::Array(blocks)) => {
|
|
let mut message_content = Vec::new();
|
|
|
|
for block in blocks {
|
|
let block_type = block.get("type").and_then(|t| t.as_str()).unwrap_or("");
|
|
|
|
match block_type {
|
|
"text" => {
|
|
if let Some(text) = block.get("text").and_then(|t| t.as_str()) {
|
|
let content_type = if role == "assistant" {
|
|
"output_text"
|
|
} else {
|
|
"input_text"
|
|
};
|
|
message_content.push(json!({ "type": content_type, "text": text }));
|
|
}
|
|
}
|
|
|
|
"image" => {
|
|
if let Some(source) = block.get("source") {
|
|
let media_type = source
|
|
.get("media_type")
|
|
.and_then(|m| m.as_str())
|
|
.unwrap_or("image/png");
|
|
let data =
|
|
source.get("data").and_then(|d| d.as_str()).unwrap_or("");
|
|
message_content.push(json!({
|
|
"type": "input_image",
|
|
"image_url": format!("data:{media_type};base64,{data}")
|
|
}));
|
|
}
|
|
}
|
|
|
|
"tool_use" => {
|
|
// 先刷新已累积的消息内容
|
|
if !message_content.is_empty() {
|
|
input.push(json!({
|
|
"role": role,
|
|
"content": message_content.clone()
|
|
}));
|
|
message_content.clear();
|
|
}
|
|
|
|
// 提升为独立的 function_call item
|
|
let id = block.get("id").and_then(|i| i.as_str()).unwrap_or("");
|
|
let name = block.get("name").and_then(|n| n.as_str()).unwrap_or("");
|
|
let arguments = block.get("input").cloned().unwrap_or(json!({}));
|
|
|
|
input.push(json!({
|
|
"type": "function_call",
|
|
"call_id": id,
|
|
"name": name,
|
|
"arguments": serde_json::to_string(&arguments).unwrap_or_default()
|
|
}));
|
|
}
|
|
|
|
"tool_result" => {
|
|
// 先刷新已累积的消息内容
|
|
if !message_content.is_empty() {
|
|
input.push(json!({
|
|
"role": role,
|
|
"content": message_content.clone()
|
|
}));
|
|
message_content.clear();
|
|
}
|
|
|
|
// 提升为独立的 function_call_output item
|
|
let call_id = block
|
|
.get("tool_use_id")
|
|
.and_then(|i| i.as_str())
|
|
.unwrap_or("");
|
|
let output = match block.get("content") {
|
|
Some(Value::String(s)) => s.clone(),
|
|
Some(v) => serde_json::to_string(v).unwrap_or_default(),
|
|
None => String::new(),
|
|
};
|
|
|
|
input.push(json!({
|
|
"type": "function_call_output",
|
|
"call_id": call_id,
|
|
"output": output
|
|
}));
|
|
}
|
|
|
|
"thinking" => {
|
|
// 丢弃 thinking blocks（与 openai_chat 一致）
|
|
}
|
|
|
|
_ => {}
|
|
}
|
|
}
|
|
|
|
// 刷新剩余的消息内容
|
|
if !message_content.is_empty() {
|
|
input.push(json!({
|
|
"role": role,
|
|
"content": message_content
|
|
}));
|
|
}
|
|
}
|
|
|
|
_ => {
|
|
// 无内容或 null
|
|
input.push(json!({ "role": role }));
|
|
}
|
|
}
|
|
}
|
|
|
|
Ok(input)
|
|
}
|
|
|
|
/// OpenAI Responses 响应 → Anthropic 响应
|
|
pub fn responses_to_anthropic(body: Value) -> Result<Value, ProxyError> {
|
|
let output = body
|
|
.get("output")
|
|
.and_then(|o| o.as_array())
|
|
.ok_or_else(|| ProxyError::TransformError("No output in response".to_string()))?;
|
|
|
|
let mut content = Vec::new();
|
|
|
|
for item in output {
|
|
let item_type = item.get("type").and_then(|t| t.as_str()).unwrap_or("");
|
|
|
|
match item_type {
|
|
"message" => {
|
|
if let Some(msg_content) = item.get("content").and_then(|c| c.as_array()) {
|
|
for block in msg_content {
|
|
let block_type = block.get("type").and_then(|t| t.as_str()).unwrap_or("");
|
|
if block_type == "output_text" {
|
|
if let Some(text) = block.get("text").and_then(|t| t.as_str()) {
|
|
if !text.is_empty() {
|
|
content.push(json!({"type": "text", "text": text}));
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
}
|
|
|
|
"function_call" => {
|
|
let call_id = item.get("call_id").and_then(|i| i.as_str()).unwrap_or("");
|
|
let name = item.get("name").and_then(|n| n.as_str()).unwrap_or("");
|
|
let args_str = item
|
|
.get("arguments")
|
|
.and_then(|a| a.as_str())
|
|
.unwrap_or("{}");
|
|
let input: Value = serde_json::from_str(args_str).unwrap_or(json!({}));
|
|
|
|
content.push(json!({
|
|
"type": "tool_use",
|
|
"id": call_id,
|
|
"name": name,
|
|
"input": input
|
|
}));
|
|
}
|
|
|
|
"reasoning" => {
|
|
// 映射 reasoning summary → thinking block
|
|
if let Some(summary) = item.get("summary").and_then(|s| s.as_array()) {
|
|
let thinking_text: String = summary
|
|
.iter()
|
|
.filter_map(|s| {
|
|
if s.get("type").and_then(|t| t.as_str()) == Some("summary_text") {
|
|
s.get("text").and_then(|t| t.as_str())
|
|
} else {
|
|
None
|
|
}
|
|
})
|
|
.collect::<Vec<_>>()
|
|
.join("");
|
|
|
|
if !thinking_text.is_empty() {
|
|
content.push(json!({
|
|
"type": "thinking",
|
|
"thinking": thinking_text
|
|
}));
|
|
}
|
|
}
|
|
}
|
|
|
|
_ => {}
|
|
}
|
|
}
|
|
|
|
// status → stop_reason
|
|
let stop_reason = body
|
|
.get("status")
|
|
.and_then(|s| s.as_str())
|
|
.map(|s| match s {
|
|
"completed" => "end_turn",
|
|
"incomplete" => "max_tokens",
|
|
_ => "end_turn",
|
|
});
|
|
|
|
// Usage — Responses API 使用与 Anthropic 相同的字段名
|
|
let usage = body.get("usage").cloned().unwrap_or(json!({}));
|
|
let input_tokens = usage
|
|
.get("input_tokens")
|
|
.and_then(|v| v.as_u64())
|
|
.unwrap_or(0) as u32;
|
|
let output_tokens = usage
|
|
.get("output_tokens")
|
|
.and_then(|v| v.as_u64())
|
|
.unwrap_or(0) as u32;
|
|
|
|
let result = json!({
|
|
"id": body.get("id").and_then(|i| i.as_str()).unwrap_or(""),
|
|
"type": "message",
|
|
"role": "assistant",
|
|
"content": content,
|
|
"model": body.get("model").and_then(|m| m.as_str()).unwrap_or(""),
|
|
"stop_reason": stop_reason,
|
|
"stop_sequence": null,
|
|
"usage": {
|
|
"input_tokens": input_tokens,
|
|
"output_tokens": output_tokens
|
|
}
|
|
});
|
|
|
|
Ok(result)
|
|
}
|
|
|
|
#[cfg(test)]
|
|
mod tests {
|
|
use super::*;
|
|
|
|
#[test]
|
|
fn test_anthropic_to_responses_simple() {
|
|
let input = json!({
|
|
"model": "gpt-4o",
|
|
"max_tokens": 1024,
|
|
"messages": [{"role": "user", "content": "Hello"}]
|
|
});
|
|
|
|
let result = anthropic_to_responses(input).unwrap();
|
|
assert_eq!(result["model"], "gpt-4o");
|
|
assert_eq!(result["max_output_tokens"], 1024);
|
|
assert_eq!(result["input"][0]["role"], "user");
|
|
assert_eq!(result["input"][0]["content"][0]["type"], "input_text");
|
|
assert_eq!(result["input"][0]["content"][0]["text"], "Hello");
|
|
// stop_sequences should not appear
|
|
assert!(result.get("stop_sequences").is_none());
|
|
}
|
|
|
|
#[test]
|
|
fn test_anthropic_to_responses_with_system_string() {
|
|
let input = json!({
|
|
"model": "gpt-4o",
|
|
"max_tokens": 1024,
|
|
"system": "You are a helpful assistant.",
|
|
"messages": [{"role": "user", "content": "Hello"}]
|
|
});
|
|
|
|
let result = anthropic_to_responses(input).unwrap();
|
|
assert_eq!(result["instructions"], "You are a helpful assistant.");
|
|
// system should not appear in input
|
|
assert_eq!(result["input"].as_array().unwrap().len(), 1);
|
|
}
|
|
|
|
#[test]
|
|
fn test_anthropic_to_responses_with_system_array() {
|
|
let input = json!({
|
|
"model": "gpt-4o",
|
|
"max_tokens": 1024,
|
|
"system": [
|
|
{"type": "text", "text": "Part 1"},
|
|
{"type": "text", "text": "Part 2"}
|
|
],
|
|
"messages": [{"role": "user", "content": "Hello"}]
|
|
});
|
|
|
|
let result = anthropic_to_responses(input).unwrap();
|
|
assert_eq!(result["instructions"], "Part 1\n\nPart 2");
|
|
}
|
|
|
|
#[test]
|
|
fn test_anthropic_to_responses_with_tools() {
|
|
let input = json!({
|
|
"model": "gpt-4o",
|
|
"max_tokens": 1024,
|
|
"messages": [{"role": "user", "content": "Weather?"}],
|
|
"tools": [{
|
|
"name": "get_weather",
|
|
"description": "Get weather info",
|
|
"input_schema": {"type": "object", "properties": {"location": {"type": "string"}}}
|
|
}]
|
|
});
|
|
|
|
let result = anthropic_to_responses(input).unwrap();
|
|
assert_eq!(result["tools"][0]["type"], "function");
|
|
assert_eq!(result["tools"][0]["name"], "get_weather");
|
|
assert!(result["tools"][0].get("parameters").is_some());
|
|
// input_schema should not appear
|
|
assert!(result["tools"][0].get("input_schema").is_none());
|
|
}
|
|
|
|
#[test]
|
|
fn test_anthropic_to_responses_tool_use_lifting() {
|
|
let input = json!({
|
|
"model": "gpt-4o",
|
|
"max_tokens": 1024,
|
|
"messages": [{
|
|
"role": "assistant",
|
|
"content": [
|
|
{"type": "text", "text": "Let me check"},
|
|
{"type": "tool_use", "id": "call_123", "name": "get_weather", "input": {"location": "Tokyo"}}
|
|
]
|
|
}]
|
|
});
|
|
|
|
let result = anthropic_to_responses(input).unwrap();
|
|
let input_arr = result["input"].as_array().unwrap();
|
|
|
|
// Should produce: assistant message (text) + function_call item
|
|
assert_eq!(input_arr.len(), 2);
|
|
|
|
// First: assistant message with output_text
|
|
assert_eq!(input_arr[0]["role"], "assistant");
|
|
assert_eq!(input_arr[0]["content"][0]["type"], "output_text");
|
|
assert_eq!(input_arr[0]["content"][0]["text"], "Let me check");
|
|
|
|
// Second: function_call item (lifted from message)
|
|
assert_eq!(input_arr[1]["type"], "function_call");
|
|
assert_eq!(input_arr[1]["call_id"], "call_123");
|
|
assert_eq!(input_arr[1]["name"], "get_weather");
|
|
}
|
|
|
|
#[test]
|
|
fn test_anthropic_to_responses_tool_result_lifting() {
|
|
let input = json!({
|
|
"model": "gpt-4o",
|
|
"max_tokens": 1024,
|
|
"messages": [{
|
|
"role": "user",
|
|
"content": [
|
|
{"type": "tool_result", "tool_use_id": "call_123", "content": "Sunny, 25°C"}
|
|
]
|
|
}]
|
|
});
|
|
|
|
let result = anthropic_to_responses(input).unwrap();
|
|
let input_arr = result["input"].as_array().unwrap();
|
|
|
|
// Should produce: function_call_output item (lifted)
|
|
assert_eq!(input_arr.len(), 1);
|
|
assert_eq!(input_arr[0]["type"], "function_call_output");
|
|
assert_eq!(input_arr[0]["call_id"], "call_123");
|
|
assert_eq!(input_arr[0]["output"], "Sunny, 25°C");
|
|
}
|
|
|
|
#[test]
|
|
fn test_anthropic_to_responses_thinking_discarded() {
|
|
let input = json!({
|
|
"model": "gpt-4o",
|
|
"max_tokens": 1024,
|
|
"messages": [{
|
|
"role": "assistant",
|
|
"content": [
|
|
{"type": "thinking", "thinking": "Let me think..."},
|
|
{"type": "text", "text": "The answer is 42"}
|
|
]
|
|
}]
|
|
});
|
|
|
|
let result = anthropic_to_responses(input).unwrap();
|
|
let input_arr = result["input"].as_array().unwrap();
|
|
|
|
// thinking should be discarded, only text remains
|
|
assert_eq!(input_arr.len(), 1);
|
|
assert_eq!(input_arr[0]["content"][0]["type"], "output_text");
|
|
assert_eq!(input_arr[0]["content"][0]["text"], "The answer is 42");
|
|
}
|
|
|
|
#[test]
|
|
fn test_anthropic_to_responses_image() {
|
|
let input = json!({
|
|
"model": "gpt-4o",
|
|
"max_tokens": 1024,
|
|
"messages": [{
|
|
"role": "user",
|
|
"content": [
|
|
{"type": "text", "text": "What is this?"},
|
|
{"type": "image", "source": {"type": "base64", "media_type": "image/png", "data": "abc123"}}
|
|
]
|
|
}]
|
|
});
|
|
|
|
let result = anthropic_to_responses(input).unwrap();
|
|
let content = result["input"][0]["content"].as_array().unwrap();
|
|
|
|
assert_eq!(content[0]["type"], "input_text");
|
|
assert_eq!(content[1]["type"], "input_image");
|
|
assert_eq!(content[1]["image_url"], "data:image/png;base64,abc123");
|
|
}
|
|
|
|
#[test]
|
|
fn test_responses_to_anthropic_simple() {
|
|
let input = json!({
|
|
"id": "resp_123",
|
|
"object": "response",
|
|
"status": "completed",
|
|
"model": "gpt-4o",
|
|
"output": [{
|
|
"type": "message",
|
|
"id": "msg_123",
|
|
"role": "assistant",
|
|
"content": [{"type": "output_text", "text": "Hello!"}]
|
|
}],
|
|
"usage": {"input_tokens": 10, "output_tokens": 5, "total_tokens": 15}
|
|
});
|
|
|
|
let result = responses_to_anthropic(input).unwrap();
|
|
assert_eq!(result["id"], "resp_123");
|
|
assert_eq!(result["type"], "message");
|
|
assert_eq!(result["content"][0]["type"], "text");
|
|
assert_eq!(result["content"][0]["text"], "Hello!");
|
|
assert_eq!(result["stop_reason"], "end_turn");
|
|
assert_eq!(result["usage"]["input_tokens"], 10);
|
|
assert_eq!(result["usage"]["output_tokens"], 5);
|
|
}
|
|
|
|
#[test]
|
|
fn test_responses_to_anthropic_with_function_call() {
|
|
let input = json!({
|
|
"id": "resp_123",
|
|
"object": "response",
|
|
"status": "completed",
|
|
"model": "gpt-4o",
|
|
"output": [{
|
|
"type": "function_call",
|
|
"id": "fc_123",
|
|
"call_id": "call_123",
|
|
"name": "get_weather",
|
|
"arguments": "{\"location\": \"Tokyo\"}",
|
|
"status": "completed"
|
|
}],
|
|
"usage": {"input_tokens": 10, "output_tokens": 15}
|
|
});
|
|
|
|
let result = responses_to_anthropic(input).unwrap();
|
|
assert_eq!(result["content"][0]["type"], "tool_use");
|
|
assert_eq!(result["content"][0]["id"], "call_123");
|
|
assert_eq!(result["content"][0]["name"], "get_weather");
|
|
assert_eq!(result["content"][0]["input"]["location"], "Tokyo");
|
|
}
|
|
|
|
#[test]
|
|
fn test_responses_to_anthropic_with_reasoning() {
|
|
let input = json!({
|
|
"id": "resp_123",
|
|
"object": "response",
|
|
"status": "completed",
|
|
"model": "gpt-4o",
|
|
"output": [
|
|
{
|
|
"type": "reasoning",
|
|
"id": "rs_123",
|
|
"summary": [
|
|
{"type": "summary_text", "text": "Thinking about the problem..."}
|
|
]
|
|
},
|
|
{
|
|
"type": "message",
|
|
"id": "msg_123",
|
|
"role": "assistant",
|
|
"content": [{"type": "output_text", "text": "The answer is 42"}]
|
|
}
|
|
],
|
|
"usage": {"input_tokens": 10, "output_tokens": 20}
|
|
});
|
|
|
|
let result = responses_to_anthropic(input).unwrap();
|
|
// Should have thinking + text
|
|
assert_eq!(result["content"][0]["type"], "thinking");
|
|
assert_eq!(
|
|
result["content"][0]["thinking"],
|
|
"Thinking about the problem..."
|
|
);
|
|
assert_eq!(result["content"][1]["type"], "text");
|
|
assert_eq!(result["content"][1]["text"], "The answer is 42");
|
|
}
|
|
|
|
#[test]
|
|
fn test_responses_to_anthropic_incomplete_status() {
|
|
let input = json!({
|
|
"id": "resp_123",
|
|
"status": "incomplete",
|
|
"model": "gpt-4o",
|
|
"output": [{
|
|
"type": "message",
|
|
"content": [{"type": "output_text", "text": "Partial..."}]
|
|
}],
|
|
"usage": {"input_tokens": 10, "output_tokens": 4096}
|
|
});
|
|
|
|
let result = responses_to_anthropic(input).unwrap();
|
|
assert_eq!(result["stop_reason"], "max_tokens");
|
|
}
|
|
|
|
#[test]
|
|
fn test_model_passthrough() {
|
|
let input = json!({
|
|
"model": "o3-mini",
|
|
"max_tokens": 1024,
|
|
"messages": [{"role": "user", "content": "Hello"}]
|
|
});
|
|
|
|
let result = anthropic_to_responses(input).unwrap();
|
|
assert_eq!(result["model"], "o3-mini");
|
|
}
|
|
}
|