jiazhizhong/higress
1
0
Fork 0
mirror of https://github.com/alibaba/higress.git synced 2026-05-26 05:37:25 +08:00
Code Issues Packages Projects Releases Wiki Activity

feat(ai-proxy): Adjust the streaming response structure to keep it consistent with the openai (#2391)

Browse Source 
Signed-off-by: Xijun Dai <daixijun1990@gmail.com>
This commit is contained in:
Xijun Dai
2025-06-12 16:25:35 +08:00
committed by GitHub
parent 86934b3203
commit fb8dd819e9
12 changed files with 180 additions and 178 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
plugins/wasm-go/extensions/ai-proxy/provider/qwen.go
View File

@@ -302,7 +302,7 @@ func (m *qwenProvider) buildChatCompletionResponse(ctx wrapper.HttpContext, qwen
message := qwenMessageToChatMessage(qwenChoice.Message, m.config.reasoningContentMode)
choices = append(choices, chatCompletionChoice{
Message: &message,
FinishReason: qwenChoice.FinishReason,
FinishReason: util.Ptr(qwenChoice.FinishReason),
})
}
return &chatCompletionResponse{
@@ -312,7 +312,7 @@ func (m *qwenProvider) buildChatCompletionResponse(ctx wrapper.HttpContext, qwen
SystemFingerprint: "",
Object: objectChatCompletion,
Choices: choices,
Usage: usage{
Usage: &usage{
PromptTokens: qwenResponse.Usage.InputTokens,
CompletionTokens: qwenResponse.Usage.OutputTokens,
TotalTokens: qwenResponse.Usage.TotalTokens,
@@ -413,11 +413,11 @@ func (m *qwenProvider) buildChatCompletionStreamingResponse(ctx wrapper.HttpCont
if finished {
finishResponse := *&baseMessage
finishResponse.Choices = append(finishResponse.Choices, chatCompletionChoice{Delta: &chatMessage{}, FinishReason: qwenChoice.FinishReason})
finishResponse.Choices = append(finishResponse.Choices, chatCompletionChoice{Delta: &chatMessage{}, FinishReason: util.Ptr(qwenChoice.FinishReason)})
usageResponse := *&baseMessage
usageResponse.Choices = []chatCompletionChoice{{Delta: &chatMessage{}}}
usageResponse.Usage = usage{
usageResponse.Usage = &usage{
PromptTokens: qwenResponse.Usage.InputTokens,
CompletionTokens: qwenResponse.Usage.OutputTokens,
TotalTokens: qwenResponse.Usage.TotalTokens,