jiazhizhong/higress
1
0
Fork 0
mirror of https://github.com/alibaba/higress.git synced 2026-05-23 12:17:30 +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

9
plugins/wasm-go/extensions/ai-proxy/provider/vertex.go
View File

@@ -32,8 +32,7 @@ const (
vertexEmbeddingAction = "predict"
)
type vertexProviderInitializer struct {
}
type vertexProviderInitializer struct{}
func (v *vertexProviderInitializer) ValidateConfig(config *ProviderConfig) error {
if config.vertexAuthKey == "" {
@@ -245,7 +244,7 @@ func (v *vertexProvider) buildChatCompletionResponse(ctx wrapper.HttpContext, re
Created: time.Now().UnixMilli() / 1000,
Model: ctx.GetStringContext(ctxKeyFinalRequestModel, ""),
Choices: make([]chatCompletionChoice, 0, len(response.Candidates)),
Usage: usage{
Usage: &usage{
PromptTokens: response.UsageMetadata.PromptTokenCount,
CompletionTokens: response.UsageMetadata.CandidatesTokenCount,
TotalTokens: response.UsageMetadata.TotalTokenCount,
@@ -257,7 +256,7 @@ func (v *vertexProvider) buildChatCompletionResponse(ctx wrapper.HttpContext, re
Message: &chatMessage{
Role: roleAssistant,
},
FinishReason: candidate.FinishReason,
FinishReason: util.Ptr(candidate.FinishReason),
}
if len(candidate.Content.Parts) > 0 {
choice.Message.Content = candidate.Content.Parts[0].Text
@@ -310,7 +309,7 @@ func (v *vertexProvider) buildChatCompletionStreamResponse(ctx wrapper.HttpConte
Created: time.Now().UnixMilli() / 1000,
Model: ctx.GetStringContext(ctxKeyFinalRequestModel, ""),
Choices: []chatCompletionChoice{choice},
Usage: usage{
Usage: &usage{
PromptTokens: vertexResp.UsageMetadata.PromptTokenCount,
CompletionTokens: vertexResp.UsageMetadata.CandidatesTokenCount,
TotalTokens: vertexResp.UsageMetadata.TotalTokenCount,