feat: OpenAI 兼容端点完整实现 + reasoning_content 输出通道

- StageContext 新增 ReasoningContent + TokenUsage 字段
- emitResponse 将 reasoning_content / usage 传入 Payload
- /v1/chat/completions:
  - 非流式返回 reasoning_content + token_usage
  - 流式 (stream=true) SSE 分块返回 reasoning/content/finish chunk
  - token_usage 来自精确 LLM 回报而非估算
This commit is contained in:
root
2026-07-03 20:49:43 +08:00
parent 239a22899b
commit d4956c23f0
3 changed files with 129 additions and 22 deletions

View File

@ -690,6 +690,15 @@ func (h *Handler) handleOpenAICompletions(w http.ResponseWriter, r *http.Request
return
}
content, _ := response.Payload["content"].(string)
reasoningContent, _ := response.Payload["reasoning_content"].(string)
usage, _ := response.Payload["usage"].(map[string]interface{})
if req.Stream {
h.writeOpenAIStream(w, req.Model, content, reasoningContent, usage)
return
}
resp := map[string]interface{}{
"id": fmt.Sprintf("chatcmpl-%d", time.Now().UnixNano()),
"object": "chat.completion",
@ -700,22 +709,104 @@ func (h *Handler) handleOpenAICompletions(w http.ResponseWriter, r *http.Request
"index": 0,
"message": map[string]interface{}{
"role": "assistant",
"content": response.Payload["content"],
"content": content,
},
"finish_reason": "stop",
},
},
"usage": map[string]interface{}{
"prompt_tokens": len(lastMsg.Content) / 2,
"completion_tokens": len(fmt.Sprint(response.Payload["content"])) / 2,
"total_tokens": (len(lastMsg.Content) + len(fmt.Sprint(response.Payload["content"]))) / 2,
},
}
if reasoningContent != "" {
resp["choices"].([]map[string]interface{})[0]["message"].(map[string]interface{})["reasoning_content"] = reasoningContent
}
if usage != nil {
resp["usage"] = usage
}
w.Header().Set("Content-Type", "application/json; charset=utf-8")
json.NewEncoder(w).Encode(resp)
}
func (h *Handler) writeOpenAIStream(w http.ResponseWriter, model, content, reasoningContent string, usage map[string]interface{}) {
flusher, ok := w.(http.Flusher)
if !ok {
writeJSON(w, http.StatusInternalServerError, map[string]string{"error": "streaming not supported"})
return
}
w.Header().Set("Content-Type", "text/event-stream")
w.Header().Set("Cache-Control", "no-cache")
w.Header().Set("Connection", "keep-alive")
w.WriteHeader(http.StatusOK)
flusher.Flush()
// 如果有 reasoning_content先发送一个 reasoning chunk
if reasoningContent != "" {
reasoningChunk := map[string]interface{}{
"id": fmt.Sprintf("chatcmpl-%d", time.Now().UnixNano()),
"object": "chat.completion.chunk",
"created": time.Now().Unix(),
"model": model,
"choices": []map[string]interface{}{
{
"index": 0,
"delta": map[string]interface{}{
"content": "",
"reasoning_content": reasoningContent,
},
"finish_reason": nil,
},
},
}
data, _ := json.Marshal(reasoningChunk)
fmt.Fprintf(w, "data: %s\n\n", data)
flusher.Flush()
}
// content chunk
contentChunk := map[string]interface{}{
"id": fmt.Sprintf("chatcmpl-%d", time.Now().UnixNano()),
"object": "chat.completion.chunk",
"created": time.Now().Unix(),
"model": model,
"choices": []map[string]interface{}{
{
"index": 0,
"delta": map[string]interface{}{
"content": content,
},
"finish_reason": nil,
},
},
}
data, _ := json.Marshal(contentChunk)
fmt.Fprintf(w, "data: %s\n\n", data)
flusher.Flush()
// finish chunk
finishChunk := map[string]interface{}{
"id": fmt.Sprintf("chatcmpl-%d", time.Now().UnixNano()),
"object": "chat.completion.chunk",
"created": time.Now().Unix(),
"model": model,
"choices": []map[string]interface{}{
{
"index": 0,
"delta": map[string]interface{}{},
"finish_reason": "stop",
},
},
}
if usage != nil {
finishChunk["usage"] = usage
}
data, _ = json.Marshal(finishChunk)
fmt.Fprintf(w, "data: %s\n\n", data)
flusher.Flush()
fmt.Fprintf(w, "data: [DONE]\n\n")
flusher.Flush()
}
func (h *Handler) handleTracker(w http.ResponseWriter, r *http.Request) {
if h.tracker == nil {
writeJSON(w, http.StatusServiceUnavailable, map[string]string{"error": "tracker not available"})