lisone/ds2api
1
0
Fork 0
mirror of https://github.com/CJackHwang/ds2api.git synced 2026-05-23 01:17:44 +08:00
Code Issues Packages Projects Releases Wiki Activity

remove upstream token-usage plumbing and always estimate from content

Browse Source
This commit is contained in:
CJACK.
2026-04-07 20:12:18 +08:00
parent a14e5b0847
commit d9e65c9710
17 changed files with 24 additions and 306 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
internal/adapter/claude/stream_runtime_core.go
View File

@@ -24,10 +24,9 @@ type claudeStreamRuntime struct {
bufferToolContent bool
stripReferenceMarkers bool
messageID string
thinking strings.Builder
text strings.Builder
outputTokens int
messageID string
thinking strings.Builder
text strings.Builder
nextBlockIndex int
thinkingBlockOpen bool
@@ -70,9 +69,6 @@ func (s *claudeStreamRuntime) onParsed(parsed sse.LineResult) streamengine.Parse
if !parsed.Parsed {
return streamengine.ParsedDecision{}
}
if parsed.OutputTokens > 0 {
s.outputTokens = parsed.OutputTokens
}
if parsed.ErrorMessage != "" {
s.upstreamErr = parsed.ErrorMessage
return streamengine.ParsedDecision{Stop: true, StopReason: streamengine.StopReason("upstream_error")}

3
internal/adapter/claude/stream_runtime_finalize.go
View File

@@ -109,9 +109,6 @@ func (s *claudeStreamRuntime) finalize(stopReason string) {
}
outputTokens := util.EstimateTokens(finalThinking) + util.EstimateTokens(finalText)
if s.outputTokens > 0 {
outputTokens = s.outputTokens
}
s.send("message_delta", map[string]any{
"type": "message_delta",
"delta": map[string]any{

11
internal/adapter/gemini/handler_generate.go
View File

@@ -149,14 +149,13 @@ func (h *Handler) handleNonStreamGenerateContent(w http.ResponseWriter, resp *ht
cleanVisibleOutput(result.Thinking, stripReferenceMarkers),
cleanVisibleOutput(result.Text, stripReferenceMarkers),
toolNames,
result.OutputTokens,
))
}
//nolint:unused // retained for native Gemini non-stream handling path.
func buildGeminiGenerateContentResponse(model, finalPrompt, finalThinking, finalText string, toolNames []string, outputTokens int) map[string]any {
func buildGeminiGenerateContentResponse(model, finalPrompt, finalThinking, finalText string, toolNames []string) map[string]any {
parts := buildGeminiPartsFromFinal(finalText, finalThinking, toolNames)
usage := buildGeminiUsage(finalPrompt, finalThinking, finalText, outputTokens)
usage := buildGeminiUsage(finalPrompt, finalThinking, finalText)
return map[string]any{
"candidates": []map[string]any{
{
@@ -174,14 +173,10 @@ func buildGeminiGenerateContentResponse(model, finalPrompt, finalThinking, final
}
//nolint:unused // retained for native Gemini non-stream handling path.
func buildGeminiUsage(finalPrompt, finalThinking, finalText string, outputTokens int) map[string]any {
func buildGeminiUsage(finalPrompt, finalThinking, finalText string) map[string]any {
promptTokens := util.EstimateTokens(finalPrompt)
reasoningTokens := util.EstimateTokens(finalThinking)
completionTokens := util.EstimateTokens(finalText)
if outputTokens > 0 {
completionTokens = outputTokens
reasoningTokens = 0
}
return map[string]any{
"promptTokenCount": promptTokens,
"candidatesTokenCount": reasoningTokens + completionTokens,

10
internal/adapter/gemini/handler_stream_runtime.go
View File

@@ -65,9 +65,8 @@ type geminiStreamRuntime struct {
stripReferenceMarkers bool
toolNames []string
thinking strings.Builder
text strings.Builder
outputTokens int
thinking strings.Builder
text strings.Builder
}
//nolint:unused // retained for native Gemini stream handling path.
@@ -112,9 +111,6 @@ func (s *geminiStreamRuntime) onParsed(parsed sse.LineResult) streamengine.Parse
if !parsed.Parsed {
return streamengine.ParsedDecision{}
}
if parsed.OutputTokens > 0 {
s.outputTokens = parsed.OutputTokens
}
if parsed.ContentFilter || parsed.ErrorMessage != "" || parsed.Stop {
return streamengine.ParsedDecision{Stop: true}
}
@@ -198,6 +194,6 @@ func (s *geminiStreamRuntime) finalize() {
},
},
"modelVersion": s.model,
"usageMetadata": buildGeminiUsage(s.finalPrompt, finalThinking, finalText, s.outputTokens),
"usageMetadata": buildGeminiUsage(s.finalPrompt, finalThinking, finalText),
})
}

19
internal/adapter/openai/chat_stream_runtime.go
View File

@@ -37,8 +37,6 @@ type chatStreamRuntime struct {
streamToolNames map[int]string
thinking strings.Builder
text strings.Builder
promptTokens int
outputTokens int
}
func newChatStreamRuntime(
@@ -171,17 +169,6 @@ func (s *chatStreamRuntime) finalize(finishReason string) {
finishReason = "tool_calls"
}
usage := openaifmt.BuildChatUsage(s.finalPrompt, finalThinking, finalText)
if s.promptTokens > 0 {
usage["prompt_tokens"] = s.promptTokens
}
if s.outputTokens > 0 {
usage["completion_tokens"] = s.outputTokens
}
if s.promptTokens > 0 || s.outputTokens > 0 {
p := usage["prompt_tokens"].(int)
c := usage["completion_tokens"].(int)
usage["total_tokens"] = p + c
}
s.sendChunk(openaifmt.BuildChatStreamChunk(
s.completionID,
s.created,
@@ -196,12 +183,6 @@ func (s *chatStreamRuntime) onParsed(parsed sse.LineResult) streamengine.ParsedD
if !parsed.Parsed {
return streamengine.ParsedDecision{}
}
if parsed.PromptTokens > 0 {
s.promptTokens = parsed.PromptTokens
}
if parsed.OutputTokens > 0 {
s.outputTokens = parsed.OutputTokens
}
if parsed.ContentFilter {
return streamengine.ParsedDecision{Stop: true, StopReason: streamengine.StopReasonHandlerRequested}
}

13
internal/adapter/openai/handler_chat.go
View File

@@ -131,19 +131,6 @@ func (h *Handler) handleNonStream(w http.ResponseWriter, ctx context.Context, re
return
}
respBody := openaifmt.BuildChatCompletion(completionID, model, finalPrompt, finalThinking, finalText, toolNames)
if result.PromptTokens > 0 || result.OutputTokens > 0 {
if usage, ok := respBody["usage"].(map[string]any); ok {
if result.PromptTokens > 0 {
usage["prompt_tokens"] = result.PromptTokens
}
if result.OutputTokens > 0 {
usage["completion_tokens"] = result.OutputTokens
}
p, _ := usage["prompt_tokens"].(int)
c, _ := usage["completion_tokens"].(int)
usage["total_tokens"] = p + c
}
}
writeJSON(w, http.StatusOK, respBody)
}

13
internal/adapter/openai/responses_handler.go
View File

@@ -130,19 +130,6 @@ func (h *Handler) handleResponsesNonStream(w http.ResponseWriter, resp *http.Res
}
responseObj := openaifmt.BuildResponseObject(responseID, model, finalPrompt, sanitizedThinking, sanitizedText, toolNames)
if result.PromptTokens > 0 || result.OutputTokens > 0 {
if usage, ok := responseObj["usage"].(map[string]any); ok {
if result.PromptTokens > 0 {
usage["input_tokens"] = result.PromptTokens
}
if result.OutputTokens > 0 {
usage["output_tokens"] = result.OutputTokens
}
input, _ := usage["input_tokens"].(int)
output, _ := usage["output_tokens"].(int)
usage["total_tokens"] = input + output
}
}
h.getResponseStore().put(owner, responseID, responseObj)
writeJSON(w, http.StatusOK, responseObj)
}

26
internal/adapter/openai/responses_stream_runtime_core.go
View File

@@ -51,8 +51,6 @@ type responsesStreamRuntime struct {
messagePartAdded bool
sequence int
failed bool
promptTokens int
outputTokens int
persistResponse func(obj map[string]any)
}
@@ -150,24 +148,6 @@ func (s *responsesStreamRuntime) finalize() {
s.closeIncompleteFunctionItems()
obj := s.buildCompletedResponseObject(finalThinking, finalText, detected)
if s.outputTokens > 0 {
if usage, ok := obj["usage"].(map[string]any); ok {
usage["output_tokens"] = s.outputTokens
}
}
if s.promptTokens > 0 || s.outputTokens > 0 {
if usage, ok := obj["usage"].(map[string]any); ok {
if s.promptTokens > 0 {
usage["input_tokens"] = s.promptTokens
}
if s.outputTokens > 0 {
usage["output_tokens"] = s.outputTokens
}
input, _ := usage["input_tokens"].(int)
output, _ := usage["output_tokens"].(int)
usage["total_tokens"] = input + output
}
}
if s.persistResponse != nil {
s.persistResponse(obj)
}
@@ -196,12 +176,6 @@ func (s *responsesStreamRuntime) onParsed(parsed sse.LineResult) streamengine.Pa
if !parsed.Parsed {
return streamengine.ParsedDecision{}
}
if parsed.PromptTokens > 0 {
s.promptTokens = parsed.PromptTokens
}
if parsed.OutputTokens > 0 {
s.outputTokens = parsed.OutputTokens
}
if parsed.ContentFilter || parsed.ErrorMessage != "" || parsed.Stop {
return streamengine.ParsedDecision{Stop: true}
}