github/new-api
1
0
Fork 0
mirror of https://github.com/QuantumNous/new-api.git synced 2026-03-30 02:25:00 +00:00
Code Issues Packages Projects Releases Wiki Activity

feat: openai response /v1/response/compact (#2644)

Browse Source 
* feat: openai response /v1/response/compact

* feat: /v1/response/compact bill

* feat: /v1/response/compact

* feat: /v1/responses/compact -> codex channel

* feat: /v1/responses/compact -> codex channel

* feat: /v1/responses/compact -> codex channel

* feat: codex channel default models

* feat: compact model price

* feat: /v1/responses/comapct test
This commit is contained in:
Seefs
2026-01-26 20:20:16 +08:00
committed by GitHub
parent d7d3a2f763
commit cc1da72d10
25 changed files with 483 additions and 42 deletions
Download Patch File Download Diff File Expand all files Collapse all files

58
controller/channel-test.go
View File

@@ -26,6 +26,7 @@ import (
"github.com/QuantumNous/new-api/relay/helper"
"github.com/QuantumNous/new-api/service"
"github.com/QuantumNous/new-api/setting/operation_setting"
"github.com/QuantumNous/new-api/setting/ratio_setting"
"github.com/QuantumNous/new-api/types"
"github.com/bytedance/gopkg/util/gopool"
@@ -107,6 +108,14 @@ func testChannel(channel *model.Channel, testModel string, endpointType string)
if strings.Contains(strings.ToLower(testModel), "codex") {
requestPath = "/v1/responses"
}
// responses compaction models (must use /v1/responses/compact)
if strings.HasSuffix(testModel, ratio_setting.CompactModelSuffix) {
requestPath = "/v1/responses/compact"
}
}
if strings.HasPrefix(requestPath, "/v1/responses/compact") {
testModel = ratio_setting.WithCompactModelSuffix(testModel)
}
c.Request = &http.Request{
@@ -150,6 +159,8 @@ func testChannel(channel *model.Channel, testModel string, endpointType string)
relayFormat = types.RelayFormatOpenAI
case constant.EndpointTypeOpenAIResponse:
relayFormat = types.RelayFormatOpenAIResponses
case constant.EndpointTypeOpenAIResponseCompact:
relayFormat = types.RelayFormatOpenAIResponsesCompaction
case constant.EndpointTypeAnthropic:
relayFormat = types.RelayFormatClaude
case constant.EndpointTypeGemini:
@@ -184,6 +195,9 @@ func testChannel(channel *model.Channel, testModel string, endpointType string)
if c.Request.URL.Path == "/v1/responses" {
relayFormat = types.RelayFormatOpenAIResponses
}
if strings.HasPrefix(c.Request.URL.Path, "/v1/responses/compact") {
relayFormat = types.RelayFormatOpenAIResponsesCompaction
}
}
request := buildTestRequest(testModel, endpointType, channel)
@@ -215,6 +229,15 @@ func testChannel(channel *model.Channel, testModel string, endpointType string)
request.SetModelName(testModel)
apiType, _ := common.ChannelType2APIType(channel.Type)
if info.RelayMode == relayconstant.RelayModeResponsesCompact &&
apiType != constant.APITypeOpenAI &&
apiType != constant.APITypeCodex {
return testResult{
context: c,
localErr: fmt.Errorf("responses compaction test only supports openai/codex channels, got api type %d", apiType),
newAPIError: types.NewError(fmt.Errorf("unsupported api type: %d", apiType), types.ErrorCodeInvalidApiType),
}
}
adaptor := relay.GetAdaptor(apiType)
if adaptor == nil {
return testResult{
@@ -287,6 +310,25 @@ func testChannel(channel *model.Channel, testModel string, endpointType string)
newAPIError: types.NewError(errors.New("invalid response request type"), types.ErrorCodeConvertRequestFailed),
}
}
case relayconstant.RelayModeResponsesCompact:
// Response compaction request - convert to OpenAIResponsesRequest before adapting
switch req := request.(type) {
case *dto.OpenAIResponsesCompactionRequest:
convertedRequest, err = adaptor.ConvertOpenAIResponsesRequest(c, info, dto.OpenAIResponsesRequest{
Model: req.Model,
Input: req.Input,
Instructions: req.Instructions,
PreviousResponseID: req.PreviousResponseID,
})
case *dto.OpenAIResponsesRequest:
convertedRequest, err = adaptor.ConvertOpenAIResponsesRequest(c, info, *req)
default:
return testResult{
context: c,
localErr: errors.New("invalid response compaction request type"),
newAPIError: types.NewError(errors.New("invalid response compaction request type"), types.ErrorCodeConvertRequestFailed),
}
}
default:
// Chat/Completion 等其他请求类型
if generalReq, ok := request.(*dto.GeneralOpenAIRequest); ok {
@@ -432,6 +474,8 @@ func testChannel(channel *model.Channel, testModel string, endpointType string)
}
func buildTestRequest(model string, endpointType string, channel *model.Channel) dto.Request {
testResponsesInput := json.RawMessage(`[{"role":"user","content":"hi"}]`)
// 根据端点类型构建不同的测试请求
if endpointType != "" {
switch constant.EndpointType(endpointType) {
@@ -463,6 +507,12 @@ func buildTestRequest(model string, endpointType string, channel *model.Channel)
Model: model,
Input: json.RawMessage(`[{"role":"user","content":"hi"}]`),
}
case constant.EndpointTypeOpenAIResponseCompact:
// 返回 OpenAIResponsesCompactionRequest
return &dto.OpenAIResponsesCompactionRequest{
Model: model,
Input: testResponsesInput,
}
case constant.EndpointTypeAnthropic, constant.EndpointTypeGemini, constant.EndpointTypeOpenAI:
// 返回 GeneralOpenAIRequest
maxTokens := uint(16)
@@ -504,6 +554,14 @@ func buildTestRequest(model string, endpointType string, channel *model.Channel)
}
}
// Responses compaction models (must use /v1/responses/compact)
if strings.HasSuffix(model, ratio_setting.CompactModelSuffix) {
return &dto.OpenAIResponsesCompactionRequest{
Model: model,
Input: testResponsesInput,
}
}
// Responses-only models (e.g. codex series)
if strings.Contains(strings.ToLower(model), "codex") {
return &dto.OpenAIResponsesRequest{