fix: 补全 streaming message_delta 事件缺失的 input_tokens 和 cache 相关字段 (#2881)
当上游为 AWS Bedrock 时,message_delta 的 usage 可能缺少 input_tokens、 cache_creation_input_tokens、cache_read_input_tokens 等字段,导致与原生 Anthropic 格式不一致。从 message_start 积累的 claudeInfo 中补全这些字段后 重新序列化,确保客户端收到一致的 usage 格式。
This commit is contained in:
@@ -544,6 +544,30 @@ type ClaudeResponseInfo struct {
|
||||
Done bool
|
||||
}
|
||||
|
||||
// enrichMessageDeltaUsage 补全 message_delta 事件中缺失的 input_tokens 和 cache 相关字段
|
||||
// 当上游(如 AWS Bedrock)的 message_delta 不包含这些字段时,从 claudeInfo 中积累的数据补全
|
||||
func enrichMessageDeltaUsage(claudeResponse *dto.ClaudeResponse, claudeInfo *ClaudeResponseInfo) {
|
||||
if claudeResponse.Usage == nil {
|
||||
claudeResponse.Usage = &dto.ClaudeUsage{}
|
||||
}
|
||||
if claudeResponse.Usage.InputTokens == 0 && claudeInfo.Usage.PromptTokens > 0 {
|
||||
claudeResponse.Usage.InputTokens = claudeInfo.Usage.PromptTokens
|
||||
}
|
||||
if claudeResponse.Usage.CacheReadInputTokens == 0 && claudeInfo.Usage.PromptTokensDetails.CachedTokens > 0 {
|
||||
claudeResponse.Usage.CacheReadInputTokens = claudeInfo.Usage.PromptTokensDetails.CachedTokens
|
||||
}
|
||||
if claudeResponse.Usage.CacheCreationInputTokens == 0 && claudeInfo.Usage.PromptTokensDetails.CachedCreationTokens > 0 {
|
||||
claudeResponse.Usage.CacheCreationInputTokens = claudeInfo.Usage.PromptTokensDetails.CachedCreationTokens
|
||||
}
|
||||
if claudeResponse.Usage.CacheCreation == nil &&
|
||||
(claudeInfo.Usage.ClaudeCacheCreation5mTokens > 0 || claudeInfo.Usage.ClaudeCacheCreation1hTokens > 0) {
|
||||
claudeResponse.Usage.CacheCreation = &dto.ClaudeCacheCreationUsage{
|
||||
Ephemeral5mInputTokens: claudeInfo.Usage.ClaudeCacheCreation5mTokens,
|
||||
Ephemeral1hInputTokens: claudeInfo.Usage.ClaudeCacheCreation1hTokens,
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
func FormatClaudeResponseInfo(claudeResponse *dto.ClaudeResponse, oaiResponse *dto.ChatCompletionsStreamResponse, claudeInfo *ClaudeResponseInfo) bool {
|
||||
if claudeInfo == nil {
|
||||
return false
|
||||
@@ -638,6 +662,13 @@ func HandleStreamResponseData(c *gin.Context, info *relaycommon.RelayInfo, claud
|
||||
if claudeResponse.Message != nil {
|
||||
info.UpstreamModelName = claudeResponse.Message.Model
|
||||
}
|
||||
} else if claudeResponse.Type == "message_delta" {
|
||||
// 确保 message_delta 的 usage 包含完整的 input_tokens 和 cache 相关字段
|
||||
// 解决 AWS Bedrock 等上游返回的 message_delta 缺少这些字段的问题
|
||||
enrichMessageDeltaUsage(&claudeResponse, claudeInfo)
|
||||
if newData, err := json.Marshal(claudeResponse); err == nil {
|
||||
data = string(newData)
|
||||
}
|
||||
}
|
||||
helper.ClaudeChunkData(c, claudeResponse, data)
|
||||
} else if info.RelayFormat == types.RelayFormatOpenAI {
|
||||
|
||||
Reference in New Issue
Block a user