Files
new-api-custom/relay/channel/openai/relay_responses_billing_test.go
T

534 lines
21 KiB
Go

package openai
import (
"bytes"
"io"
"net/http"
"net/http/httptest"
"slices"
"strings"
"testing"
"github.com/QuantumNous/new-api/common"
"github.com/QuantumNous/new-api/constant"
relaycommon "github.com/QuantumNous/new-api/relay/common"
relayconstant "github.com/QuantumNous/new-api/relay/constant"
"github.com/QuantumNous/new-api/relaykit/dto"
"github.com/QuantumNous/new-api/relaykit/types"
"github.com/QuantumNous/new-api/setting/operation_setting"
"github.com/gin-gonic/gin"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
)
func TestOaiResponsesHandlerBuffersUnexpectedSSEAndPreservesUsage(t *testing.T) {
terminal := `{"id":"resp_test","object":"response","model":"deepseek-v4.1-flash","status":"completed","extra_provider_field":true,"output":[{"type":"message","role":"assistant","content":[{"type":"output_text","text":"OK"}]}],"usage":{"input_tokens":86926,"output_tokens":163,"total_tokens":87089,"input_tokens_details":{"cached_tokens":86784}}}`
for _, tc := range []struct {
name string
body string
wantError bool
}{
{"completed", ": ping\n\nevent: response.completed\ndata: {\"type\":\"response.completed\",\"response\":" + terminal + "}\n\ndata: [DONE]\n\n", false},
{"incomplete", "data: {\"type\":\"response.incomplete\",\"response\":" + strings.Replace(terminal, `"completed"`, `"incomplete"`, 1) + "}\n\n", false},
{"truncated", "data: {\"type\":\"response.created\",\"response\":{\"model\":\"deepseek-v4.1-flash\"}}\n\ndata: [DONE]\n\n", true},
{"failed", "data: {\"type\":\"response.failed\",\"response\":{\"status\":\"failed\"}}\n\n", true},
{"invalid", "data: not-json\n\n", true},
} {
t.Run(tc.name, func(t *testing.T) {
w := httptest.NewRecorder()
c, _ := gin.CreateTestContext(w)
c.Request = httptest.NewRequest(http.MethodPost, "/v1/responses", nil)
info := &relaycommon.RelayInfo{OriginModelName: "deepseek-v4.1-flash", ChannelMeta: &relaycommon.ChannelMeta{}}
resp := &http.Response{StatusCode: 200, Header: http.Header{"Content-Type": {"text/event-stream; charset=utf-8"}}, Body: io.NopCloser(strings.NewReader(tc.body))}
usage, apiErr := OaiResponsesHandler(c, info, resp)
if tc.wantError {
require.NotNil(t, apiErr)
assert.Nil(t, usage)
assert.False(t, c.Writer.Written())
return
}
require.Nil(t, apiErr)
require.NotNil(t, usage)
assert.Equal(t, 86926, usage.PromptTokens)
assert.Equal(t, 163, usage.CompletionTokens)
assert.Equal(t, 86784, usage.PromptTokensDetails.CachedTokens)
assert.Equal(t, "application/json", w.Header().Get("Content-Type"))
assert.Contains(t, w.Body.String(), `"extra_provider_field":true`)
assert.Contains(t, w.Body.String(), `"text":"OK"`)
})
}
}
func TestOaiResponsesHandlerCountsOutputCallsNotDeclarations(t *testing.T) {
gin.SetMode(gin.TestMode)
operation_setting.SetToolPriceForTest("priced_fn", 5.0)
t.Cleanup(func() {
operation_setting.DeleteToolPriceForTest("priced_fn")
})
body, err := common.Marshal(dto.OpenAIResponsesResponse{
Tools: []map[string]any{
{"type": "web_search_preview"},
{"type": "file_search"},
},
Output: []dto.ResponsesOutput{
{Type: dto.BuildInCallWebSearchCall},
{Type: dto.BuildInCallWebSearchCall},
{Type: dto.BuildInCallFunctionCall, Name: "priced_fn"},
{Type: dto.BuildInCallFunctionCall, Name: "unpriced_fn"},
},
Usage: &dto.Usage{InputTokens: 1, OutputTokens: 1, TotalTokens: 2},
})
require.NoError(t, err)
w := httptest.NewRecorder()
c, _ := gin.CreateTestContext(w)
c.Request = httptest.NewRequest(http.MethodPost, "/v1/responses", nil)
info := &relaycommon.RelayInfo{
OriginModelName: "gpt-5.1",
ResponsesUsageInfo: &relaycommon.ResponsesUsageInfo{
BuiltInTools: map[string]*relaycommon.BuildInToolInfo{
dto.BuildInToolWebSearchPreview: {ToolName: dto.BuildInToolWebSearchPreview, CallCount: 0},
dto.BuildInToolFileSearch: {ToolName: dto.BuildInToolFileSearch, CallCount: 0},
},
},
}
resp := &http.Response{
StatusCode: http.StatusOK,
Body: io.NopCloser(bytes.NewReader(body)),
Header: http.Header{"Content-Type": []string{"application/json"}},
}
usage, apiErr := OaiResponsesHandler(c, info, resp)
require.Nil(t, apiErr)
require.NotNil(t, usage)
assert.Equal(t, 2, info.ResponsesUsageInfo.BuiltInTools[dto.BuildInToolWebSearchPreview].CallCount)
assert.Equal(t, 0, info.ResponsesUsageInfo.BuiltInTools[dto.BuildInToolFileSearch].CallCount)
require.Contains(t, info.ResponsesUsageInfo.BuiltInTools, "priced_fn")
assert.Equal(t, 1, info.ResponsesUsageInfo.BuiltInTools["priced_fn"].CallCount)
assert.NotContains(t, info.ResponsesUsageInfo.BuiltInTools, "unpriced_fn")
}
func TestOaiResponsesHandlerDeclaredToolsWithoutOutputCountZero(t *testing.T) {
gin.SetMode(gin.TestMode)
body, err := common.Marshal(dto.OpenAIResponsesResponse{
Tools: []map[string]any{
{"type": "web_search_preview"},
{"type": "file_search"},
},
Output: []dto.ResponsesOutput{
{Type: "message", Role: "assistant"},
},
Usage: &dto.Usage{InputTokens: 1, OutputTokens: 1, TotalTokens: 2},
})
require.NoError(t, err)
w := httptest.NewRecorder()
c, _ := gin.CreateTestContext(w)
c.Request = httptest.NewRequest(http.MethodPost, "/v1/responses", nil)
info := &relaycommon.RelayInfo{
OriginModelName: "gpt-5.1",
ResponsesUsageInfo: &relaycommon.ResponsesUsageInfo{
BuiltInTools: map[string]*relaycommon.BuildInToolInfo{
dto.BuildInToolWebSearchPreview: {ToolName: dto.BuildInToolWebSearchPreview, CallCount: 0},
dto.BuildInToolFileSearch: {ToolName: dto.BuildInToolFileSearch, CallCount: 0},
},
},
}
resp := &http.Response{
StatusCode: http.StatusOK,
Body: io.NopCloser(bytes.NewReader(body)),
Header: http.Header{"Content-Type": []string{"application/json"}},
}
_, apiErr := OaiResponsesHandler(c, info, resp)
require.Nil(t, apiErr)
assert.Equal(t, 0, info.ResponsesUsageInfo.BuiltInTools[dto.BuildInToolWebSearchPreview].CallCount)
assert.Equal(t, 0, info.ResponsesUsageInfo.BuiltInTools[dto.BuildInToolFileSearch].CallCount)
}
func TestOaiResponsesHandlerCountsCompletedImageGenerationOutputs(t *testing.T) {
gin.SetMode(gin.TestMode)
body, err := common.Marshal(dto.OpenAIResponsesResponse{
Status: []byte(`"completed"`),
Output: []dto.ResponsesOutput{
{
Type: dto.ResponsesOutputTypeImageGenerationCall,
ID: "img_1",
Status: "completed",
Result: "base64-a",
},
{
Type: dto.ResponsesOutputTypeImageGenerationCall,
ID: "img_2",
Status: "completed",
Result: "base64-b",
},
{
Type: dto.ResponsesOutputTypeImageGenerationCall,
ID: "img_empty",
Status: "completed",
Result: "",
},
},
Usage: &dto.Usage{InputTokens: 1, OutputTokens: 1, TotalTokens: 2},
})
require.NoError(t, err)
w := httptest.NewRecorder()
c, _ := gin.CreateTestContext(w)
c.Request = httptest.NewRequest(http.MethodPost, "/v1/responses", nil)
info := &relaycommon.RelayInfo{OriginModelName: "gpt-5.1"}
resp := &http.Response{
StatusCode: http.StatusOK,
Body: io.NopCloser(bytes.NewReader(body)),
Header: http.Header{"Content-Type": []string{"application/json"}},
}
_, apiErr := OaiResponsesHandler(c, info, resp)
require.Nil(t, apiErr)
require.Contains(t, info.ResponsesUsageInfo.BuiltInTools, dto.BuildInToolImageGeneration)
assert.Equal(t, 2, info.ResponsesUsageInfo.BuiltInTools[dto.BuildInToolImageGeneration].CallCount)
assert.False(t, c.GetBool("image_generation_call"))
}
func TestOaiResponsesHandlerCountsImageGenerationOnIncompleteStatus(t *testing.T) {
gin.SetMode(gin.TestMode)
body, err := common.Marshal(dto.OpenAIResponsesResponse{
Status: []byte(`"incomplete"`),
Output: []dto.ResponsesOutput{
{
Type: dto.ResponsesOutputTypeImageGenerationCall,
ID: "img_1",
Status: "completed",
Result: "base64-a",
},
},
Usage: &dto.Usage{InputTokens: 1, OutputTokens: 1, TotalTokens: 2},
})
require.NoError(t, err)
w := httptest.NewRecorder()
c, _ := gin.CreateTestContext(w)
c.Request = httptest.NewRequest(http.MethodPost, "/v1/responses", nil)
info := &relaycommon.RelayInfo{
OriginModelName: "gpt-5.1",
ResponsesUsageInfo: &relaycommon.ResponsesUsageInfo{
BuiltInTools: map[string]*relaycommon.BuildInToolInfo{
dto.BuildInToolImageGeneration: {ToolName: dto.BuildInToolImageGeneration, CallCount: 0},
},
},
}
resp := &http.Response{
StatusCode: http.StatusOK,
Body: io.NopCloser(bytes.NewReader(body)),
Header: http.Header{"Content-Type": []string{"application/json"}},
}
_, apiErr := OaiResponsesHandler(c, info, resp)
require.Nil(t, apiErr)
assert.Equal(t, 1, info.ResponsesUsageInfo.BuiltInTools[dto.BuildInToolImageGeneration].CallCount)
}
func runResponsesImageBillingStream(t *testing.T, events ...string) *relaycommon.RelayInfo {
t.Helper()
gin.SetMode(gin.TestMode)
oldTimeout := constant.StreamingTimeout
constant.StreamingTimeout = 30
t.Cleanup(func() {
constant.StreamingTimeout = oldTimeout
})
var body strings.Builder
for _, event := range events {
body.WriteString("data: ")
body.WriteString(event)
body.WriteString("\n\n")
}
body.WriteString("data: [DONE]\n\n")
w := httptest.NewRecorder()
c, _ := gin.CreateTestContext(w)
c.Request = httptest.NewRequest(http.MethodPost, "/v1/responses", nil)
c.Set(common.RequestIdKey, "responses-image-billing-test")
info := &relaycommon.RelayInfo{
OriginModelName: "gpt-5.1",
DisablePing: true,
ChannelMeta: &relaycommon.ChannelMeta{
UpstreamModelName: "gpt-5.1",
},
}
resp := &http.Response{
StatusCode: http.StatusOK,
Body: io.NopCloser(strings.NewReader(body.String())),
Header: http.Header{"Content-Type": []string{"text/event-stream"}},
}
_, apiErr := OaiResponsesStreamHandler(c, info, resp)
require.Nil(t, apiErr)
require.NotNil(t, info.ResponsesUsageInfo)
require.Contains(t, info.ResponsesUsageInfo.BuiltInTools, dto.BuildInToolImageGeneration)
return info
}
func TestOaiResponsesStreamHandlerDeduplicatesCompletedImageOutput(t *testing.T) {
item := `{"type":"image_generation_call","id":"img_1","call_id":"call_1","status":"completed","result":"base64-a"}`
info := runResponsesImageBillingStream(
t,
`{"type":"response.output_item.done","output_index":0,"item":`+item+`}`,
`{"type":"response.completed","response":{"status":"completed","output":[`+item+`],"usage":{"input_tokens":1,"output_tokens":1,"total_tokens":2}}}`,
)
assert.Equal(t, 1, info.ResponsesUsageInfo.BuiltInTools[dto.BuildInToolImageGeneration].CallCount)
}
func TestOaiResponsesStreamHandlerCountsImageOutputOnIncomplete(t *testing.T) {
info := runResponsesImageBillingStream(
t,
`{"type":"response.output_item.done","output_index":0,"item":{"type":"image_generation_call","id":"img_1","status":"completed","result":"base64-a"}}`,
`{"type":"response.incomplete","response":{"status":"incomplete"}}`,
)
assert.Equal(t, 1, info.ResponsesUsageInfo.BuiltInTools[dto.BuildInToolImageGeneration].CallCount)
}
func TestOaiResponsesStreamHandlerDoesNotCountPartialImageEvent(t *testing.T) {
info := runResponsesImageBillingStream(
t,
`{"type":"response.image_generation_call.partial_image","output_index":0,"partial_image_b64":"partial-bytes"}`,
`{"type":"response.completed","response":{"status":"completed","output":[]}}`,
)
assert.Equal(t, 0, info.ResponsesUsageInfo.BuiltInTools[dto.BuildInToolImageGeneration].CallCount)
}
func TestOaiResponsesHandlerRewritesSGLangCreatedAtToInt(t *testing.T) {
gin.SetMode(gin.TestMode)
w := httptest.NewRecorder()
c, _ := gin.CreateTestContext(w)
c.Request = httptest.NewRequest(http.MethodPost, "/v1/responses", nil)
info := &relaycommon.RelayInfo{
ChannelMeta: &relaycommon.ChannelMeta{ChannelType: constant.ChannelTypeSGLang},
}
resp := &http.Response{
StatusCode: http.StatusOK,
Body: io.NopCloser(strings.NewReader(`{"id":"resp_1","object":"response","created_at":1786588600.0,"status":"completed","output":[]}`)),
Header: http.Header{"Content-Type": []string{"application/json"}},
}
_, apiErr := OaiResponsesHandler(c, info, resp)
require.Nil(t, apiErr)
assert.Contains(t, w.Body.String(), `"created_at":1786588600`)
assert.NotContains(t, w.Body.String(), `1786588600.0`)
}
func TestOaiResponsesStreamHandlerRewritesSGLangCreatedAtToInt(t *testing.T) {
gin.SetMode(gin.TestMode)
oldTimeout := constant.StreamingTimeout
constant.StreamingTimeout = 30
t.Cleanup(func() {
constant.StreamingTimeout = oldTimeout
})
w := httptest.NewRecorder()
c, _ := gin.CreateTestContext(w)
c.Request = httptest.NewRequest(http.MethodPost, "/v1/responses", nil)
c.Set(common.RequestIdKey, "sglang-created-at-test")
info := &relaycommon.RelayInfo{
DisablePing: true,
ChannelMeta: &relaycommon.ChannelMeta{
ChannelType: constant.ChannelTypeSGLang,
UpstreamModelName: "served-model",
},
}
resp := &http.Response{
StatusCode: http.StatusOK,
Body: io.NopCloser(strings.NewReader(
"data: {\"type\":\"response.completed\",\"response\":{\"id\":\"resp_1\",\"created_at\":1.7865886E9,\"status\":\"completed\",\"output\":[]}}\n\n" +
"data: [DONE]\n\n",
)),
Header: http.Header{"Content-Type": []string{"text/event-stream"}},
}
_, apiErr := OaiResponsesStreamHandler(c, info, resp)
require.Nil(t, apiErr)
assert.Contains(t, w.Body.String(), `"created_at":1786588600`)
assert.NotContains(t, w.Body.String(), `1.7865886E9`)
assert.NotContains(t, w.Body.String(), `1786588600.0`)
}
func TestOaiResponsesStreamHandlerKeepsNonSGLangCreatedAt(t *testing.T) {
gin.SetMode(gin.TestMode)
oldTimeout := constant.StreamingTimeout
constant.StreamingTimeout = 30
t.Cleanup(func() {
constant.StreamingTimeout = oldTimeout
})
w := httptest.NewRecorder()
c, _ := gin.CreateTestContext(w)
c.Request = httptest.NewRequest(http.MethodPost, "/v1/responses", nil)
c.Set(common.RequestIdKey, "openai-created-at-test")
info := &relaycommon.RelayInfo{
DisablePing: true,
ChannelMeta: &relaycommon.ChannelMeta{ChannelType: constant.ChannelTypeOpenAI, UpstreamModelName: "gpt-test"},
}
resp := &http.Response{
StatusCode: http.StatusOK,
Body: io.NopCloser(strings.NewReader(
"data: {\"type\":\"response.completed\",\"response\":{\"id\":\"resp_1\",\"created_at\":1786588600.0,\"status\":\"completed\",\"output\":[]}}\n\n" +
"data: [DONE]\n\n",
)),
Header: http.Header{"Content-Type": []string{"text/event-stream"}},
}
_, apiErr := OaiResponsesStreamHandler(c, info, resp)
require.Nil(t, apiErr)
assert.Contains(t, w.Body.String(), `1786588600.0`)
}
func TestResponsesStreamFailureHidesRoutingDetails(t *testing.T) {
for _, eventType := range []string{"error", "response.error", "response.failed"} {
t.Run(eventType, func(t *testing.T) {
secret := "智能路由流式响应中途失败。原因:个人池商家在流式响应开始后中断。 request_id: upstream-private-id"
original := map[string]any{"type": eventType, "sequence_number": 3, "code": "vendor_stream_error", "message": secret, "debug": secret, "error": map[string]any{"message": secret}, "response": map[string]any{"id": "resp_test", "status": "failed", "error": map[string]any{"type": "vendor_error", "code": "vendor_stream_error", "message": secret}, "usage": map[string]any{"input_tokens": 10, "output_tokens": 2}, "debug": secret}}
encoded, err := common.Marshal(original)
require.NoError(t, err)
var event dto.ResponsesStreamResponse
require.NoError(t, common.Unmarshal(encoded, &event))
w := httptest.NewRecorder()
c, _ := gin.CreateTestContext(w)
c.Request = httptest.NewRequest(http.MethodPost, "/v1/responses", nil)
c.Set(common.RequestIdKey, "local-request-id")
sendResponsesStreamData(c, event, string(encoded))
body := w.Body.String()
assert.Contains(t, body, "event: "+eventType)
assert.Contains(t, body, "model_service_error")
assert.Contains(t, body, "local-request-id")
assert.Contains(t, body, `"status":"failed"`)
assert.Contains(t, body, `"input_tokens":10`)
for _, detail := range []string{"个人池", "智能路由", "upstream-private-id", "vendor_stream_error", "debug"} {
assert.NotContains(t, body, detail)
}
// Projection must not change the event used for outcome and billing.
require.NotNil(t, event.Response)
assert.Equal(t, "vendor_stream_error", event.Response.GetOpenAIError().Code)
})
}
}
func TestResponsesSuccessfulStreamDataUnchanged(t *testing.T) {
body := `{"type":"response.output_text.delta","delta":"个人池 is user content"}`
var event dto.ResponsesStreamResponse
require.NoError(t, common.UnmarshalJsonStr(body, &event))
w := httptest.NewRecorder()
c, _ := gin.CreateTestContext(w)
c.Request = httptest.NewRequest(http.MethodPost, "/v1/responses", nil)
sendResponsesStreamData(c, event, body)
assert.Contains(t, w.Body.String(), body)
}
// A converted hosted web search reaches Azure and Xinference Chat Completions
// as web_search_options, like OpenAI; the upstream decides whether it applies.
// Azure Responses bills web search by tool_usage.web_search.num_requests (Bing
// transactions) under bing_web_search, and the explicit zeros silence the
// web_search_call items, which leaves one surcharge row.
func TestAzureHostedWebSearch(t *testing.T) {
gin.SetMode(gin.TestMode)
oldTimeout := constant.StreamingTimeout
constant.StreamingTimeout = 30
t.Cleanup(func() {
constant.StreamingTimeout = oldTimeout
})
t.Run("Claude to Chat", func(t *testing.T) {
const body = `{"model":"gpt-5.1","max_tokens":64,"messages":[{"role":"user","content":"Weather?"}],"tools":[{"type":"web_search_20250305","name":"web_search"}]}`
for _, channelType := range []int{constant.ChannelTypeAzure, constant.ChannelTypeXinference, constant.ChannelTypeOpenAI} {
c, _ := gin.CreateTestContext(httptest.NewRecorder())
info := &relaycommon.RelayInfo{
RelayFormat: types.RelayFormatClaude,
ChannelMeta: &relaycommon.ChannelMeta{ChannelType: channelType, UpstreamModelName: "gpt-5.1"},
}
adaptor := &Adaptor{}
adaptor.Init(info)
var request dto.ClaudeRequest
require.NoError(t, common.UnmarshalJsonStr(body, &request))
converted, err := adaptor.ConvertClaudeRequest(c, info, &request)
require.NoError(t, err)
wire, err := common.Marshal(converted)
require.NoError(t, err)
assert.True(t, strings.Contains(string(wire), `"web_search_options"`), channelType)
var codes []string
for _, diagnostic := range info.ConversionDiagnostics() {
codes = append(codes, diagnostic.Code)
}
assert.False(t, slices.Contains(codes, "hosted_web_search_omitted"), channelType)
}
})
const searchCall = `{"type":"web_search_call","status":"completed","action":{"type":"search","query":"go"}}`
const usage = `"usage":{"input_tokens":10,"output_tokens":3,"total_tokens":13},"tool_usage":{"web_search":{"num_requests":5}}`
output := strings.Repeat(searchCall+",", 3) + `{"type":"message","role":"assistant","content":[{"type":"output_text","text":"Go 1.25"}]}`
tests := []struct {
name string
contentType string
body string
}{
{
name: "json",
contentType: "application/json",
body: `{"id":"resp_1","object":"response","model":"gpt-5.1","status":"completed","output":[` + output + `],` + usage + `}`,
},
{
name: "stream",
contentType: "text/event-stream",
body: strings.Repeat(`data: {"type":"response.output_item.done","output_index":0,"item":`+searchCall+"}\n\n", 3) +
`data: {"type":"response.completed","response":{"id":"resp_1","object":"response","model":"gpt-5.1","status":"completed","output":[` + output + `],` + usage + "}}\n\ndata: [DONE]\n\n",
},
}
for _, tt := range tests {
t.Run("Responses "+tt.name, func(t *testing.T) {
c, _ := gin.CreateTestContext(httptest.NewRecorder())
c.Request = httptest.NewRequest(http.MethodPost, "/v1/responses", nil)
info := &relaycommon.RelayInfo{
RelayMode: relayconstant.RelayModeResponses,
RelayFormat: types.RelayFormatOpenAIResponses,
OriginModelName: "gpt-5.1",
IsStream: tt.name == "stream",
DisablePing: true,
ChannelMeta: &relaycommon.ChannelMeta{ChannelType: constant.ChannelTypeAzure, UpstreamModelName: "gpt-5.1"},
}
adaptor := &Adaptor{}
adaptor.Init(info)
resp := &http.Response{
StatusCode: http.StatusOK,
Body: io.NopCloser(strings.NewReader(tt.body)),
Header: http.Header{"Content-Type": []string{tt.contentType}},
}
var apiErr *types.NewAPIError
if info.IsStream {
_, apiErr = OaiResponsesStreamHandler(c, info, resp)
} else {
_, apiErr = OaiResponsesHandler(c, info, resp)
}
require.Nil(t, apiErr)
billed := map[string]int{}
for name, tool := range info.ResponsesUsageInfo.BuiltInTools {
if tool.CallCount > 0 {
billed[name] = tool.CallCount
}
}
assert.Equal(t, map[string]int{"bing_web_search": 5}, billed)
})
}
}