github/sub2api
4
0
Fork 0
mirror of https://github.com/Wei-Shaw/sub2api.git synced 2026-10-22 07:30:51 +00:00
Code Issues Packages Projects Releases Wiki Activity

merge: sync upstream/main into release/custom-0.1.92

Browse Source 
Merge upstream main branch changes including:
- OpenAI Codex support and websearch adaptation
- API key rate limit auto-reset fix
- golangci-lint v2.11 upgrade for Go 1.26 compatibility
- Go 1.26.1 stdlib security fixes
- Empty stream failover fix
- CCSwitch import improvements
- API key credentials preserve existing fields
- Usage query rate limit optimization
- Setup Token real utilization fix

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
This commit is contained in:
erio
2026-03-07 12:53:12 +08:00
co-authored by Claude Opus 4.6
parent fcdd34ddb0 6411645ffc
commit 1864f41f33
52 changed files with 5260 additions and 373 deletions
Download Patch File Download Diff File Expand all files Collapse all files
.github/workflows/backend-ci.yml
+3 -3
View File
@@ -20,7 +20,7 @@ jobs:
cache-dependency-path: backend/go.sum
- name: Verify Go version
run: |
go version | grep -q 'go1.25.7'
go version | grep -q 'go1.26.1'
- name: Unit tests
working-directory: backend
run: make test-unit
@@ -40,10 +40,10 @@ jobs:
cache-dependency-path: backend/go.sum
- name: Verify Go version
run: |
go version | grep -q 'go1.25.7'
go version | grep -q 'go1.26.1'
- name: golangci-lint
uses: golangci/golangci-lint-action@v9
with:
version: v2.7
version: v2.11
args: --timeout=30m
working-directory: backend
.github/workflows/security-scan.yml
+1 -1
View File
@@ -23,7 +23,7 @@ jobs:
cache-dependency-path: backend/go.sum
- name: Verify Go version
run: |
go version | grep -q 'go1.25.7'
go version | grep -q 'go1.26.1'
- name: Run govulncheck
working-directory: backend
run: |
Dockerfile
+1 -1
View File
@@ -7,7 +7,7 @@
# =============================================================================
ARG NODE_IMAGE=node:24-alpine
ARG GOLANG_IMAGE=golang:1.25.7-alpine
ARG GOLANG_IMAGE=golang:1.26.1-alpine
ARG ALPINE_IMAGE=alpine:3.21
ARG GOPROXY=https://goproxy.cn,direct
ARG GOSUMDB=sum.golang.google.cn
backend/cmd/server/wire_gen.go
+1 -1
View File
@@ -164,7 +164,7 @@ func initializeApplication(buildInfo handler.BuildInfo) (*Application, error) {
digestSessionStore := service.NewDigestSessionStore()
gatewayService := service.NewGatewayService(accountRepository, groupRepository, usageLogRepository, userRepository, userSubscriptionRepository, userGroupRateRepository, gatewayCache, configConfig, schedulerSnapshotService, concurrencyService, billingService, rateLimitService, billingCacheService, identityService, httpUpstream, deferredService, claudeTokenProvider, sessionLimitCache, rpmCache, digestSessionStore)
openAITokenProvider := service.NewOpenAITokenProvider(accountRepository, geminiTokenCache, openAIOAuthService)
openAIGatewayService := service.NewOpenAIGatewayService(accountRepository, usageLogRepository, userRepository, userSubscriptionRepository, gatewayCache, configConfig, schedulerSnapshotService, concurrencyService, billingService, rateLimitService, billingCacheService, httpUpstream, deferredService, openAITokenProvider)
openAIGatewayService := service.NewOpenAIGatewayService(accountRepository, usageLogRepository, userRepository, userSubscriptionRepository, userGroupRateRepository, gatewayCache, configConfig, schedulerSnapshotService, concurrencyService, billingService, rateLimitService, billingCacheService, httpUpstream, deferredService, openAITokenProvider)
geminiMessagesCompatService := service.NewGeminiMessagesCompatService(accountRepository, groupRepository, gatewayCache, schedulerSnapshotService, geminiTokenProvider, rateLimitService, httpUpstream, antigravityGatewayService, configConfig)
opsSystemLogSink := service.ProvideOpsSystemLogSink(opsRepository)
opsService := service.NewOpsService(opsRepository, settingRepository, configConfig, accountRepository, userRepository, concurrencyService, gatewayService, openAIGatewayService, geminiMessagesCompatService, antigravityGatewayService, opsSystemLogSink)
backend/go.mod
+20 -18
View File
@@ -1,6 +1,6 @@
module github.com/Wei-Shaw/sub2api
go 1.25.7
go 1.26.1
require (
entgo.io/ent v0.14.5
@@ -38,11 +38,11 @@ require (
github.com/zeromicro/go-zero v1.9.4
go.uber.org/zap v1.24.0
golang.org/x/crypto v0.48.0
golang.org/x/net v0.51.0
golang.org/x/oauth2 v0.35.0
golang.org/x/net v0.49.0
golang.org/x/oauth2 v0.30.0
golang.org/x/sync v0.19.0
golang.org/x/term v0.40.0
google.golang.org/api v0.269.0
google.golang.org/api v0.153.0
gopkg.in/natefinch/lumberjack.v2 v2.2.1
gopkg.in/yaml.v3 v3.0.1
modernc.org/sqlite v1.44.3
@@ -50,9 +50,7 @@ require (
require (
ariga.io/atlas v0.32.1-0.20250325101103-175b25e1c1b9 // indirect
cloud.google.com/go/auth v0.18.2 // indirect
cloud.google.com/go/auth/oauth2adapt v0.2.8 // indirect
cloud.google.com/go/compute/metadata v0.9.0 // indirect
cloud.google.com/go/compute/metadata v0.7.0 // indirect
dario.cat/mergo v1.0.2 // indirect
github.com/Azure/go-ansiterm v0.0.0-20210617225240-d185dfc1b5a1 // indirect
github.com/Microsoft/go-winio v0.6.2 // indirect
@@ -113,11 +111,13 @@ require (
github.com/go-playground/universal-translator v0.18.1 // indirect
github.com/go-playground/validator/v10 v10.14.0 // indirect
github.com/goccy/go-json v0.10.2 // indirect
github.com/golang/groupcache v0.0.0-20210331224755-41bb18bfe9da // indirect
github.com/golang/protobuf v1.5.4 // indirect
github.com/google/go-cmp v0.7.0 // indirect
github.com/google/go-querystring v1.1.0 // indirect
github.com/google/s2a-go v0.1.9 // indirect
github.com/googleapis/enterprise-certificate-proxy v0.3.12 // indirect
github.com/googleapis/gax-go/v2 v2.17.0 // indirect
github.com/google/s2a-go v0.1.7 // indirect
github.com/googleapis/enterprise-certificate-proxy v0.3.2 // indirect
github.com/googleapis/gax-go/v2 v2.12.0 // indirect
github.com/grpc-ecosystem/grpc-gateway/v2 v2.27.3 // indirect
github.com/hashicorp/hcl v1.0.0 // indirect
github.com/hashicorp/hcl/v2 v2.18.1 // indirect
@@ -173,11 +173,12 @@ require (
github.com/yusufpapurcu/wmi v1.2.4 // indirect
github.com/zclconf/go-cty v1.14.4 // indirect
github.com/zclconf/go-cty-yaml v1.1.0 // indirect
go.opentelemetry.io/auto/sdk v1.2.1 // indirect
go.opentelemetry.io/contrib/instrumentation/net/http/otelhttp v0.61.0 // indirect
go.opentelemetry.io/otel v1.39.0 // indirect
go.opentelemetry.io/otel/metric v1.39.0 // indirect
go.opentelemetry.io/otel/trace v1.39.0 // indirect
go.opencensus.io v0.24.0 // indirect
go.opentelemetry.io/auto/sdk v1.1.0 // indirect
go.opentelemetry.io/contrib/instrumentation/net/http/otelhttp v0.49.0 // indirect
go.opentelemetry.io/otel v1.37.0 // indirect
go.opentelemetry.io/otel/metric v1.37.0 // indirect
go.opentelemetry.io/otel/trace v1.37.0 // indirect
go.uber.org/atomic v1.10.0 // indirect
go.uber.org/automaxprocs v1.6.0 // indirect
go.uber.org/multierr v1.9.0 // indirect
@@ -186,9 +187,10 @@ require (
golang.org/x/mod v0.32.0 // indirect
golang.org/x/sys v0.41.0 // indirect
golang.org/x/text v0.34.0 // indirect
google.golang.org/genproto/googleapis/rpc v0.0.0-20260217215200-42d3e9bedb6d // indirect
google.golang.org/grpc v1.79.1 // indirect
google.golang.org/protobuf v1.36.11 // indirect
google.golang.org/genproto/googleapis/api v0.0.0-20250929231259-57b25ae835d4 // indirect
google.golang.org/genproto/googleapis/rpc v0.0.0-20250929231259-57b25ae835d4 // indirect
google.golang.org/grpc v1.75.1 // indirect
google.golang.org/protobuf v1.36.10 // indirect
gopkg.in/ini.v1 v1.67.0 // indirect
modernc.org/libc v1.67.6 // indirect
modernc.org/mathutil v1.7.1 // indirect
backend/go.sum
+120 -46
View File
@@ -1,11 +1,8 @@
ariga.io/atlas v0.32.1-0.20250325101103-175b25e1c1b9 h1:E0wvcUXTkgyN4wy4LGtNzMNGMytJN8afmIWXJVMi4cc=
ariga.io/atlas v0.32.1-0.20250325101103-175b25e1c1b9/go.mod h1:Oe1xWPuu5q9LzyrWfbZmEZxFYeu4BHTyzfjeW2aZp/w=
cloud.google.com/go/auth v0.18.2 h1:+Nbt5Ev0xEqxlNjd6c+yYUeosQ5TtEUaNcN/3FozlaM=
cloud.google.com/go/auth v0.18.2/go.mod h1:xD+oY7gcahcu7G2SG2DsBerfFxgPAJz17zz2joOFF3M=
cloud.google.com/go/auth/oauth2adapt v0.2.8 h1:keo8NaayQZ6wimpNSmW5OPc283g65QNIiLpZnkHRbnc=
cloud.google.com/go/auth/oauth2adapt v0.2.8/go.mod h1:XQ9y31RkqZCcwJWNSx2Xvric3RrU88hAYYbjDWYDL+c=
cloud.google.com/go/compute/metadata v0.9.0 h1:pDUj4QMoPejqq20dK0Pg2N4yG9zIkYGdBtwLoEkH9Zs=
cloud.google.com/go/compute/metadata v0.9.0/go.mod h1:E0bWwX5wTnLPedCKqk3pJmVgCBSM6qQI1yTBdEb3C10=
cloud.google.com/go v0.26.0/go.mod h1:aQUYkXzVsufM+DwF1aE+0xfcU+56JwCaLick0ClmMTw=
cloud.google.com/go/compute/metadata v0.7.0 h1:PBWF+iiAerVNe8UCHxdOt6eHLVc3ydFeOCw78U8ytSU=
cloud.google.com/go/compute/metadata v0.7.0/go.mod h1:j5MvL9PprKL39t166CoB1uVHfQMs4tFQZZcKwksXUjo=
dario.cat/mergo v1.0.2 h1:85+piFYR1tMbRrLcDwR18y4UKJ3aH1Tbzi24VRW1TK8=
dario.cat/mergo v1.0.2/go.mod h1:E/hbnu0NxMFBjpMIE34DRGLWqDy0g5FuKDhCb31ngxA=
entgo.io/ent v0.14.5 h1:Rj2WOYJtCkWyFo6a+5wB3EfBRP0rnx1fMk6gGA0UUe4=
@@ -14,6 +11,7 @@ github.com/AdaLogics/go-fuzz-headers v0.0.0-20240806141605-e8a1dd7889d6 h1:He8af
github.com/AdaLogics/go-fuzz-headers v0.0.0-20240806141605-e8a1dd7889d6/go.mod h1:8o94RPi1/7XTJvwPpRSzSUedZrtlirdB3r9Z20bi2f8=
github.com/Azure/go-ansiterm v0.0.0-20210617225240-d185dfc1b5a1 h1:UQHMgLO+TxOElx5B5HZ4hJQsoJ/PvUvKRhJHDQXO8P8=
github.com/Azure/go-ansiterm v0.0.0-20210617225240-d185dfc1b5a1/go.mod h1:xomTg63KZ2rFqZQzSB4Vz2SUXa1BpHTVz9L5PTmPC4E=
github.com/BurntSushi/toml v0.3.1/go.mod h1:xHWCNGjB5oqiDr8zfno3MHue2Ht5sIBksp03qcyfWMU=
github.com/DATA-DOG/go-sqlmock v1.5.2 h1:OcvFkGmslmlZibjAjaHm3L//6LiuBgolP7OputlJIzU=
github.com/DATA-DOG/go-sqlmock v1.5.2/go.mod h1:88MAG/4G7SMwSE3CeA0ZKzrT5CiOU3OJ+JlNzwDqpNU=
github.com/DouDOU-start/go-sora2api v1.1.0 h1:PxWiukK77StiHxEngOFwT1rKUn9oTAJJTl07wQUXwiU=
@@ -95,11 +93,14 @@ github.com/bytedance/sonic v1.9.1 h1:6iJ6NqdoxCDr6mbY8h18oSO+cShGSMRGCEo7F2h0x8s
github.com/bytedance/sonic v1.9.1/go.mod h1:i736AoUSYt75HyZLoJW9ERYxcy6eaN6h4BZXU064P/U=
github.com/cenkalti/backoff/v4 v4.3.0 h1:MyRJ/UdXutAwSAT+s3wNd7MfTIcy71VQueUuFK343L8=
github.com/cenkalti/backoff/v4 v4.3.0/go.mod h1:Y3VNntkOUPxTVeUxJ/G5vcM//AlwfmyYozVcomhLiZE=
github.com/census-instrumentation/opencensus-proto v0.2.1/go.mod h1:f6KPmirojxKA12rnyqOA5BBL4O983OfeGPqjHWSTneU=
github.com/cespare/xxhash/v2 v2.3.0 h1:UL815xU9SqsFlibzuggzjXhog7bL6oX9BbNZnL2UFvs=
github.com/cespare/xxhash/v2 v2.3.0/go.mod h1:VGX0DQ3Q6kWi7AoAeZDth3/j3BFtOZR5XLFGgcrjCOs=
github.com/chenzhuoyu/base64x v0.0.0-20211019084208-fb5309c8db06/go.mod h1:DH46F32mSOjUmXrMHnKwZdA8wcEefY7UVqBKYGjpdQY=
github.com/chenzhuoyu/base64x v0.0.0-20221115062448-fe3a3abad311 h1:qSGYFH7+jGhDF8vLC+iwCD4WpbV1EBDSzWkJODFLams=
github.com/chenzhuoyu/base64x v0.0.0-20221115062448-fe3a3abad311/go.mod h1:b583jCggY9gE99b6G5LEC39OIiVsWj+R97kbl5odCEk=
github.com/client9/misspell v0.3.4/go.mod h1:qj6jICC3Q7zFZvVWo7KLAzC3yx5G7kyvSDkc90ppPyw=
github.com/cncf/udpa/go v0.0.0-20191209042840-269d4d468f6f/go.mod h1:M8M6+tZqaGXZJjfX53e64911xZQV5JYwmTeXPW+k8Sc=
github.com/coder/websocket v1.8.14 h1:9L0p0iKiNOibykf283eHkKUHHrpG7f65OE3BhhO7v9g=
github.com/coder/websocket v1.8.14/go.mod h1:NX3SzP+inril6yawo5CQXx8+fk145lPDC6pumgx0mVg=
github.com/containerd/errdefs v1.0.0 h1:tg5yIfIlQIrxYtu9ajqY42W3lpS19XqdxRQeEwYG8PI=
@@ -138,6 +139,10 @@ github.com/dustin/go-humanize v1.0.1 h1:GzkhY7T5VNhEkwH0PVJgjz+fX1rhBrR7pRT3mDkp
github.com/dustin/go-humanize v1.0.1/go.mod h1:Mu1zIs6XwVuF/gI1OepvI0qD18qycQx+mFykh5fBlto=
github.com/ebitengine/purego v0.8.4 h1:CF7LEKg5FFOsASUj0+QwaXf8Ht6TlFxg09+S9wz0omw=
github.com/ebitengine/purego v0.8.4/go.mod h1:iIjxzd6CiRiOG0UyXP+V1+jWqUXVjPKLAI0mRfJZTmQ=
github.com/envoyproxy/go-control-plane v0.9.0/go.mod h1:YTl/9mNaCwkRvm6d1a2C3ymFceY/DCBVvsKhRF0iEA4=
github.com/envoyproxy/go-control-plane v0.9.1-0.20191026205805-5f8ba28d4473/go.mod h1:YTl/9mNaCwkRvm6d1a2C3ymFceY/DCBVvsKhRF0iEA4=
github.com/envoyproxy/go-control-plane v0.9.4/go.mod h1:6rpuAdCZL397s3pYoYcLgu1mIlRU8Am5FuJP05cCM98=
github.com/envoyproxy/protoc-gen-validate v0.1.0/go.mod h1:iSmxcyjqTsJpI2R4NaDN7+kN2VEUnK/pcBlmesArF7c=
github.com/fatih/color v1.18.0 h1:S8gINlzdQ840/4pfAwic/ZE0djQEH3wM94VfqLTZcOM=
github.com/fatih/color v1.18.0/go.mod h1:4FelSpRwEGDpQ12mAdzqdOukCy4u8WUtOY6lkT/6HfU=
github.com/felixge/httpsnoop v1.0.4 h1:NFTV2Zj1bL4mc9sqWACXbQFVBBg2W3GPvqp8/ESS2Wg=
@@ -175,9 +180,29 @@ github.com/goccy/go-json v0.10.2 h1:CrxCmQqYDkv1z7lO7Wbh2HN93uovUHgrECaO5ZrCXAU=
github.com/goccy/go-json v0.10.2/go.mod h1:6MelG93GURQebXPDq3khkgXZkazVtN9CRI+MGFi0w8I=
github.com/golang-jwt/jwt/v5 v5.2.2 h1:Rl4B7itRWVtYIHFrSNd7vhTiz9UpLdi6gZhZ3wEeDy8=
github.com/golang-jwt/jwt/v5 v5.2.2/go.mod h1:pqrtFR0X4osieyHYxtmOUWsAWrfe1Q5UVIyoH402zdk=
github.com/golang/glog v0.0.0-20160126235308-23def4e6c14b/go.mod h1:SBH7ygxi8pfUlaOkMMuAQtPIUF8ecWP5IEl/CR7VP2Q=
github.com/golang/groupcache v0.0.0-20200121045136-8c9f03a8e57e/go.mod h1:cIg4eruTrX1D+g88fzRXU5OdNfaM+9IcxsU14FzY7Hc=
github.com/golang/groupcache v0.0.0-20210331224755-41bb18bfe9da h1:oI5xCqsCo564l8iNU+DwB5epxmsaqB+rhGL0m5jtYqE=
github.com/golang/groupcache v0.0.0-20210331224755-41bb18bfe9da/go.mod h1:cIg4eruTrX1D+g88fzRXU5OdNfaM+9IcxsU14FzY7Hc=
github.com/golang/mock v1.1.1/go.mod h1:oTYuIxOrZwtPieC+H1uAHpcLFnEyAGVDL/k47Jfbm0A=
github.com/golang/protobuf v1.2.0/go.mod h1:6lQm79b+lXiMfvg/cZm0SGofjICqVBUtrP5yJMmIC1U=
github.com/golang/protobuf v1.3.2/go.mod h1:6lQm79b+lXiMfvg/cZm0SGofjICqVBUtrP5yJMmIC1U=
github.com/golang/protobuf v1.4.0-rc.1/go.mod h1:ceaxUfeHdC40wWswd/P6IGgMaK3YpKi5j83Wpe3EHw8=
github.com/golang/protobuf v1.4.0-rc.1.0.20200221234624-67d41d38c208/go.mod h1:xKAWHe0F5eneWXFV3EuXVDTCmh+JuBKY0li0aMyXATA=
github.com/golang/protobuf v1.4.0-rc.2/go.mod h1:LlEzMj4AhA7rCAGe4KMBDvJI+AwstrUpVNzEA03Pprs=
github.com/golang/protobuf v1.4.0-rc.4.0.20200313231945-b860323f09d0/go.mod h1:WU3c8KckQ9AFe+yFwt9sWVRKCVIyN9cPHBJSNnbL67w=
github.com/golang/protobuf v1.4.0/go.mod h1:jodUvKwWbYaEsadDk5Fwe5c77LiNKVO9IDvqG2KuDX0=
github.com/golang/protobuf v1.4.1/go.mod h1:U8fpvMrcmy5pZrNK1lt4xCsGvpyWQ/VVv6QDs8UjoX8=
github.com/golang/protobuf v1.4.3/go.mod h1:oDoupMAO8OvCJWAcko0GGGIgR6R6ocIYbsSw735rRwI=
github.com/golang/protobuf v1.5.4 h1:i7eJL8qZTpSEXOPTxNKhASYpMn+8e5Q6AdndVa1dWek=
github.com/golang/protobuf v1.5.4/go.mod h1:lnTiLA8Wa4RWRcIUkrtSVa5nRhsEGBg48fD6rSs7xps=
github.com/google/go-cmp v0.2.0/go.mod h1:oXzfMopK8JAjlY9xF4vHSVASa0yLyX7SntLO5aqRK0M=
github.com/google/go-cmp v0.3.0/go.mod h1:8QqcDgzrUqlUb/G2PQTWiueGozuR1884gddMywk6iLU=
github.com/google/go-cmp v0.3.1/go.mod h1:8QqcDgzrUqlUb/G2PQTWiueGozuR1884gddMywk6iLU=
github.com/google/go-cmp v0.4.0/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/gNBxE=
github.com/google/go-cmp v0.5.0/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/gNBxE=
github.com/google/go-cmp v0.5.2/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/gNBxE=
github.com/google/go-cmp v0.5.3/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/gNBxE=
github.com/google/go-cmp v0.5.6/go.mod h1:v8dTdLbMG2kIc/vJvl+f65V22dbkXbowE6jgT/gNBxE=
github.com/google/go-cmp v0.7.0 h1:wk8382ETsv4JYUZwIsn6YpYiWiBsYLSJiTsyBybVuN8=
github.com/google/go-cmp v0.7.0/go.mod h1:pXiqmnSA92OHEEa9HXL2W4E7lf9JzCmGVUdgjX3N/iU=
@@ -186,16 +211,17 @@ github.com/google/go-querystring v1.1.0/go.mod h1:Kcdr2DB4koayq7X8pmAG4sNG59So17
github.com/google/gofuzz v1.0.0/go.mod h1:dBl0BpW6vV/+mYPU4Po3pmUjxk6FQPldtuIdl/M65Eg=
github.com/google/pprof v0.0.0-20250317173921-a4b03ec1a45e h1:ijClszYn+mADRFY17kjQEVQ1XRhq2/JR1M3sGqeJoxs=
github.com/google/pprof v0.0.0-20250317173921-a4b03ec1a45e/go.mod h1:boTsfXsheKC2y+lKOCMpSfarhxDeIzfZG1jqGcPl3cA=
github.com/google/s2a-go v0.1.9 h1:LGD7gtMgezd8a/Xak7mEWL0PjoTQFvpRudN895yqKW0=
github.com/google/s2a-go v0.1.9/go.mod h1:YA0Ei2ZQL3acow2O62kdp9UlnvMmU7kA6Eutn0dXayM=
github.com/google/s2a-go v0.1.7 h1:60BLSyTrOV4/haCDW4zb1guZItoSq8foHCXrAnjBo/o=
github.com/google/s2a-go v0.1.7/go.mod h1:50CgR4k1jNlWBu4UfS4AcfhVe1r6pdZPygJ3R8F0Qdw=
github.com/google/uuid v1.1.2/go.mod h1:TIyPZe4MgqvfeYDBFedMoGGpEw/LqOeaOT+nhxU+yHo=
github.com/google/uuid v1.6.0 h1:NIvaJDMOsjHA8n1jAhLSgzrAzy1Hgr+hNrb57e+94F0=
github.com/google/uuid v1.6.0/go.mod h1:TIyPZe4MgqvfeYDBFedMoGGpEw/LqOeaOT+nhxU+yHo=
github.com/google/wire v0.7.0 h1:JxUKI6+CVBgCO2WToKy/nQk0sS+amI9z9EjVmdaocj4=
github.com/google/wire v0.7.0/go.mod h1:n6YbUQD9cPKTnHXEBN2DXlOp/mVADhVErcMFb0v3J18=
github.com/googleapis/enterprise-certificate-proxy v0.3.12 h1:Fg+zsqzYEs1ZnvmcztTYxhgCBsx3eEhEwQ1W/lHq/sQ=
github.com/googleapis/enterprise-certificate-proxy v0.3.12/go.mod h1:vqVt9yG9480NtzREnTlmGSBmFrA+bzb0yl0TxoBQXOg=
github.com/googleapis/gax-go/v2 v2.17.0 h1:RksgfBpxqff0EZkDWYuz9q/uWsTVz+kf43LsZ1J6SMc=
github.com/googleapis/gax-go/v2 v2.17.0/go.mod h1:mzaqghpQp4JDh3HvADwrat+6M3MOIDp5YKHhb9PAgDY=
github.com/googleapis/enterprise-certificate-proxy v0.3.2 h1:Vie5ybvEvT75RniqhfFxPRy3Bf7vr3h0cechB90XaQs=
github.com/googleapis/enterprise-certificate-proxy v0.3.2/go.mod h1:VLSiSSBs/ksPL8kq3OBOQ6WRI2QnaFynd1DCjZ62+V0=
github.com/googleapis/gax-go/v2 v2.12.0 h1:A+gCJKdRfqXkr+BIRGtZLibNXf0m1f9E4HG56etFpas=
github.com/googleapis/gax-go/v2 v2.12.0/go.mod h1:y+aIqrI5eb1YGMVJfuV3185Ts/D7qKpsEkdD5+I6QGU=
github.com/gorilla/websocket v1.5.3 h1:saDtZ6Pbx/0u+bgYQ3q96pZgCzfhKXGPqt7kZ72aNNg=
github.com/gorilla/websocket v1.5.3/go.mod h1:YR8l580nyteQvAITg2hZ9XVh4b55+EU/adAjf1fMHhE=
github.com/grpc-ecosystem/grpc-gateway/v2 v2.27.3 h1:NmZ1PKzSTQbuGHw9DGPFomqkkLWMC+vZCkfs+FHv1Vg=
@@ -302,6 +328,7 @@ github.com/pquerna/otp v1.5.0 h1:NMMR+WrmaqXU4EzdGJEE1aUUI0AMRzsp96fFFWNPwxs=
github.com/pquerna/otp v1.5.0/go.mod h1:dkJfzwRKNiegxyNb54X/3fLwhCynbMspSyWKnvi1AEg=
github.com/prashantv/gostub v1.1.0 h1:BTyx3RfQjRHnUWaGF9oQos79AlQ5k8WNktv7VGvVH4g=
github.com/prashantv/gostub v1.1.0/go.mod h1:A5zLQHz7ieHGG7is6LLXLz7I8+3LZzsrV0P1IAHhP5U=
github.com/prometheus/client_model v0.0.0-20190812154241-14fe0d1b01d4/go.mod h1:xMI15A0UPsDsEKsMN9yxemIoYk6Tm2C1GtYGdfGttqA=
github.com/quic-go/qpack v0.6.0 h1:g7W+BMYynC1LbYLSqRt8PBg5Tgwxn214ZZR34VIOjz8=
github.com/quic-go/qpack v0.6.0/go.mod h1:lUpLKChi8njB4ty2bFLX2x4gzDqXwUpaO1DP9qMDZII=
github.com/quic-go/quic-go v0.57.1 h1:25KAAR9QR8KZrCZRThWMKVAwGoiHIrNbT72ULHTuI10=
@@ -314,8 +341,8 @@ github.com/remyoudompheng/bigfft v0.0.0-20230129092748-24d4a6f8daec h1:W09IVJc94
github.com/remyoudompheng/bigfft v0.0.0-20230129092748-24d4a6f8daec/go.mod h1:qqbHyh8v60DhA7CoWK5oRCqLrMHRGoxYCSS9EjAz6Eo=
github.com/robfig/cron/v3 v3.0.1 h1:WdRxkvbJztn8LMz/QEvLN5sBU+xKpSqwwUO1Pjr4qDs=
github.com/robfig/cron/v3 v3.0.1/go.mod h1:eQICP3HwyT7UooqI/z+Ov+PtYAWygg1TEWWzGIFLtro=
github.com/rogpeppe/go-internal v1.14.1 h1:UQB4HGPB6osV0SQTLymcB4TgvyWu6ZyliaW0tI/otEQ=
github.com/rogpeppe/go-internal v1.14.1/go.mod h1:MaRKkUm5W0goXpeCfT7UZI6fk/L7L7so1lCWt35ZSgc=
github.com/rogpeppe/go-internal v1.13.1 h1:KvO1DLK/DRN07sQ1LQKScxyZJuNnedQ5/wKSR38lUII=
github.com/rogpeppe/go-internal v1.13.1/go.mod h1:uMEvuHeurkdAXX61udpOXGD/AzZDWNMNyH2VO9fmH0o=
github.com/sagikazarmark/locafero v0.4.0 h1:HApY1R9zGo4DBgr7dqsTH/JJxLTTsOt7u6keLGt6kNQ=
github.com/sagikazarmark/locafero v0.4.0/go.mod h1:Pe1W6UlPYUk/+wc/6KFhbORCfqzgYEpgQ3O5fPuL3H4=
github.com/sagikazarmark/slog-shim v0.1.0 h1:diDBnUNK9N/354PgrxMywXnAwEr1QZcOr6gto+ugjYE=
@@ -390,24 +417,26 @@ github.com/zclconf/go-cty-yaml v1.1.0 h1:nP+jp0qPHv2IhUVqmQSzjvqAWcObN0KBkUl2rWB
github.com/zclconf/go-cty-yaml v1.1.0/go.mod h1:9YLUH4g7lOhVWqUbctnVlZ5KLpg7JAprQNgxSZ1Gyxs=
github.com/zeromicro/go-zero v1.9.4 h1:aRLFoISqAYijABtkbliQC5SsI5TbizJpQvoHc9xup8k=
github.com/zeromicro/go-zero v1.9.4/go.mod h1:a17JOTch25SWxBcUgJZYps60hygK3pIYdw7nGwlcS38=
go.opentelemetry.io/auto/sdk v1.2.1 h1:jXsnJ4Lmnqd11kwkBV2LgLoFMZKizbCi5fNZ/ipaZ64=
go.opentelemetry.io/auto/sdk v1.2.1/go.mod h1:KRTj+aOaElaLi+wW1kO/DZRXwkF4C5xPbEe3ZiIhN7Y=
go.opentelemetry.io/contrib/instrumentation/net/http/otelhttp v0.61.0 h1:F7Jx+6hwnZ41NSFTO5q4LYDtJRXBf2PD0rNBkeB/lus=
go.opentelemetry.io/contrib/instrumentation/net/http/otelhttp v0.61.0/go.mod h1:UHB22Z8QsdRDrnAtX4PntOl36ajSxcdUMt1sF7Y6E7Q=
go.opentelemetry.io/otel v1.39.0 h1:8yPrr/S0ND9QEfTfdP9V+SiwT4E0G7Y5MO7p85nis48=
go.opentelemetry.io/otel v1.39.0/go.mod h1:kLlFTywNWrFyEdH0oj2xK0bFYZtHRYUdv1NklR/tgc8=
go.opencensus.io v0.24.0 h1:y73uSU6J157QMP2kn2r30vwW1A2W2WFwSCGnAVxeaD0=
go.opencensus.io v0.24.0/go.mod h1:vNK8G9p7aAivkbmorf4v+7Hgx+Zs0yY+0fOtgBfjQKo=
go.opentelemetry.io/auto/sdk v1.1.0 h1:cH53jehLUN6UFLY71z+NDOiNJqDdPRaXzTel0sJySYA=
go.opentelemetry.io/auto/sdk v1.1.0/go.mod h1:3wSPjt5PWp2RhlCcmmOial7AvC4DQqZb7a7wCow3W8A=
go.opentelemetry.io/contrib/instrumentation/net/http/otelhttp v0.49.0 h1:jq9TW8u3so/bN+JPT166wjOI6/vQPF6Xe7nMNIltagk=
go.opentelemetry.io/contrib/instrumentation/net/http/otelhttp v0.49.0/go.mod h1:p8pYQP+m5XfbZm9fxtSKAbM6oIllS7s2AfxrChvc7iw=
go.opentelemetry.io/otel v1.37.0 h1:9zhNfelUvx0KBfu/gb+ZgeAfAgtWrfHJZcAqFC228wQ=
go.opentelemetry.io/otel v1.37.0/go.mod h1:ehE/umFRLnuLa/vSccNq9oS1ErUlkkK71gMcN34UG8I=
go.opentelemetry.io/otel/exporters/otlp/otlptrace v1.24.0 h1:t6wl9SPayj+c7lEIFgm4ooDBZVb01IhLB4InpomhRw8=
go.opentelemetry.io/otel/exporters/otlp/otlptrace v1.24.0/go.mod h1:iSDOcsnSA5INXzZtwaBPrKp/lWu/V14Dd+llD0oI2EA=
go.opentelemetry.io/otel/exporters/otlp/otlptrace/otlptracehttp v1.24.0 h1:Xw8U6u2f8DK2XAkGRFV7BBLENgnTGX9i4rQRxJf+/vs=
go.opentelemetry.io/otel/exporters/otlp/otlptrace/otlptracehttp v1.24.0/go.mod h1:6KW1Fm6R/s6Z3PGXwSJN2K4eT6wQB3vXX6CVnYX9NmM=
go.opentelemetry.io/otel/metric v1.39.0 h1:d1UzonvEZriVfpNKEVmHXbdf909uGTOQjA0HF0Ls5Q0=
go.opentelemetry.io/otel/metric v1.39.0/go.mod h1:jrZSWL33sD7bBxg1xjrqyDjnuzTUB0x1nBERXd7Ftcs=
go.opentelemetry.io/otel/sdk v1.39.0 h1:nMLYcjVsvdui1B/4FRkwjzoRVsMK8uL/cj0OyhKzt18=
go.opentelemetry.io/otel/sdk v1.39.0/go.mod h1:vDojkC4/jsTJsE+kh+LXYQlbL8CgrEcwmt1ENZszdJE=
go.opentelemetry.io/otel/sdk/metric v1.39.0 h1:cXMVVFVgsIf2YL6QkRF4Urbr/aMInf+2WKg+sEJTtB8=
go.opentelemetry.io/otel/sdk/metric v1.39.0/go.mod h1:xq9HEVH7qeX69/JnwEfp6fVq5wosJsY1mt4lLfYdVew=
go.opentelemetry.io/otel/trace v1.39.0 h1:2d2vfpEDmCJ5zVYz7ijaJdOF59xLomrvj7bjt6/qCJI=
go.opentelemetry.io/otel/trace v1.39.0/go.mod h1:88w4/PnZSazkGzz/w84VHpQafiU4EtqqlVdxWy+rNOA=
go.opentelemetry.io/otel/metric v1.37.0 h1:mvwbQS5m0tbmqML4NqK+e3aDiO02vsf/WgbsdpcPoZE=
go.opentelemetry.io/otel/metric v1.37.0/go.mod h1:04wGrZurHYKOc+RKeye86GwKiTb9FKm1WHtO+4EVr2E=
go.opentelemetry.io/otel/sdk v1.37.0 h1:ItB0QUqnjesGRvNcmAcU0LyvkVyGJ2xftD29bWdDvKI=
go.opentelemetry.io/otel/sdk v1.37.0/go.mod h1:VredYzxUvuo2q3WRcDnKDjbdvmO0sCzOvVAiY+yUkAg=
go.opentelemetry.io/otel/sdk/metric v1.37.0 h1:90lI228XrB9jCMuSdA0673aubgRobVZFhbjxHHspCPc=
go.opentelemetry.io/otel/sdk/metric v1.37.0/go.mod h1:cNen4ZWfiD37l5NhS+Keb5RXVWZWpRE+9WyVCpbo5ps=
go.opentelemetry.io/otel/trace v1.37.0 h1:HLdcFNbRQBE2imdSEgm/kwqmQj1Or1l/7bW6mxVK7z4=
go.opentelemetry.io/otel/trace v1.37.0/go.mod h1:TlgrlQ+PtQO5XFerSPUYG0JSgGyryXewPGyayAWSBS0=
go.opentelemetry.io/proto/otlp v1.3.1 h1:TrMUixzpM0yuc/znrFTP9MMRh8trP93mkCiDVeXrui0=
go.opentelemetry.io/proto/otlp v1.3.1/go.mod h1:0X1WI4de4ZsLrrJNLAQbFeLCm3T7yBkR0XqQ7niQU+8=
go.uber.org/atomic v1.10.0 h1:9qC72Qh0+3MqyJbAn8YU5xVq1frD8bn3JtD2oXtafVQ=
@@ -425,20 +454,40 @@ go.uber.org/zap v1.24.0/go.mod h1:2kMP+WWQ8aoFoedH3T2sq6iJ2yDWpHbP0f6MQbS9Gkg=
golang.org/x/arch v0.0.0-20210923205945-b76863e36670/go.mod h1:5om86z9Hs0C8fWVUuoMHwpExlXzs5Tkyp9hOrfG7pp8=
golang.org/x/arch v0.3.0 h1:02VY4/ZcO/gBOH6PUaoiptASxtXU10jazRCP865E97k=
golang.org/x/arch v0.3.0/go.mod h1:5om86z9Hs0C8fWVUuoMHwpExlXzs5Tkyp9hOrfG7pp8=
golang.org/x/crypto v0.0.0-20190308221718-c2843e01d9a2/go.mod h1:djNgcEr1/C05ACkg1iLfiJU5Ep61QUkGW8qpdssI0+w=
golang.org/x/crypto v0.0.0-20200622213623-75b288015ac9/go.mod h1:LzIPMQfyMNhhGPhUkYOs5KpL4U8rLKemX1yGLhDgUto=
golang.org/x/crypto v0.48.0 h1:/VRzVqiRSggnhY7gNRxPauEQ5Drw9haKdM0jqfcCFts=
golang.org/x/crypto v0.48.0/go.mod h1:r0kV5h3qnFPlQnBSrULhlsRfryS2pmewsg+XfMgkVos=
golang.org/x/exp v0.0.0-20190121172915-509febef88a4/go.mod h1:CJ0aWSM057203Lf6IL+f9T1iT9GByDxfZKAQTCR3kQA=
golang.org/x/exp v0.0.0-20251023183803-a4bb9ffd2546 h1:mgKeJMpvi0yx/sU5GsxQ7p6s2wtOnGAHZWCHUM4KGzY=
golang.org/x/exp v0.0.0-20251023183803-a4bb9ffd2546/go.mod h1:j/pmGrbnkbPtQfxEe5D0VQhZC6qKbfKifgD0oM7sR70=
golang.org/x/lint v0.0.0-20181026193005-c67002cb31c3/go.mod h1:UVdnD1Gm6xHRNCYTkRU2/jEulfH38KcIWyp/GAMgvoE=
golang.org/x/lint v0.0.0-20190227174305-5b3e6a55c961/go.mod h1:wehouNa3lNwaWXcvxsM5YxQ5yQlVC4a0KAMCusXpPoU=
golang.org/x/lint v0.0.0-20190313153728-d0100b6bd8b3/go.mod h1:6SW0HCj/g11FgYtHlgUYUwCkIfeOF89ocIRzGO/8vkc=
golang.org/x/mod v0.32.0 h1:9F4d3PHLljb6x//jOyokMv3eX+YDeepZSEo3mFJy93c=
golang.org/x/mod v0.32.0/go.mod h1:SgipZ/3h2Ci89DlEtEXWUk/HteuRin+HHhN+WbNhguU=
golang.org/x/net v0.0.0-20180724234803-3673e40ba225/go.mod h1:mL1N/T3taQHkDXs73rZJwtUhF3w3ftmwwsq0BUmARs4=
golang.org/x/net v0.0.0-20180826012351-8a410e7b638d/go.mod h1:mL1N/T3taQHkDXs73rZJwtUhF3w3ftmwwsq0BUmARs4=
golang.org/x/net v0.0.0-20190213061140-3a22650c66bd/go.mod h1:mL1N/T3taQHkDXs73rZJwtUhF3w3ftmwwsq0BUmARs4=
golang.org/x/net v0.0.0-20190311183353-d8887717615a/go.mod h1:t9HGtf8HONx5eT2rtn7q6eTqICYqUVnKs3thJo3Qplg=
golang.org/x/net v0.0.0-20190404232315-eb5bcb51f2a3/go.mod h1:t9HGtf8HONx5eT2rtn7q6eTqICYqUVnKs3thJo3Qplg=
golang.org/x/net v0.0.0-20201110031124-69a78807bb2b/go.mod h1:sp8m0HH+o8qH0wwXwYZr8TS3Oi6o0r6Gce1SSxlDquU=
golang.org/x/net v0.0.0-20211104170005-ce137452f963/go.mod h1:9nx3DQGgdP8bBQD5qxJ1jj9UTztislL4KSBs9R2vV5Y=
golang.org/x/net v0.51.0 h1:94R/GTO7mt3/4wIKpcR5gkGmRLOuE/2hNGeWq/GBIFo=
golang.org/x/net v0.51.0/go.mod h1:aamm+2QF5ogm02fjy5Bb7CQ0WMt1/WVM7FtyaTLlA9Y=
golang.org/x/oauth2 v0.35.0 h1:Mv2mzuHuZuY2+bkyWXIHMfhNdJAdwW3FuWeCPYN5GVQ=
golang.org/x/oauth2 v0.35.0/go.mod h1:lzm5WQJQwKZ3nwavOZ3IS5Aulzxi68dUSgRHujetwEA=
golang.org/x/net v0.49.0 h1:eeHFmOGUTtaaPSGNmjBKpbng9MulQsJURQUAfUwY++o=
golang.org/x/net v0.49.0/go.mod h1:/ysNB2EvaqvesRkuLAyjI1ycPZlQHM3q01F02UY/MV8=
golang.org/x/oauth2 v0.0.0-20180821212333-d2e6202438be/go.mod h1:N/0e6XlmueqKjAGxoOufVs8QHGRruUQn6yWY3a++T0U=
golang.org/x/oauth2 v0.30.0 h1:dnDm7JmhM45NNpd8FDDeLhK6FwqbOf4MLCM9zb1BOHI=
golang.org/x/oauth2 v0.30.0/go.mod h1:B++QgG3ZKulg6sRPGD/mqlHQs5rB3Ml9erfeDY7xKlU=
golang.org/x/sync v0.0.0-20180314180146-1d60e4601c6f/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
golang.org/x/sync v0.0.0-20181108010431-42b317875d0f/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
golang.org/x/sync v0.0.0-20190423024810-112230192c58/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
golang.org/x/sync v0.19.0 h1:vV+1eWNmZ5geRlYjzm2adRgW2/mcpevXNg50YZtPCE4=
golang.org/x/sync v0.19.0/go.mod h1:9KTHXmSnoGruLpwFjVSX0lNNA75CykiMECbovNTZqGI=
golang.org/x/sys v0.0.0-20180830151530-49385e6e1522/go.mod h1:STP8DvDyc/dI5b8T5hshtkjS+E42TnysNCUPdjciGhY=
golang.org/x/sys v0.0.0-20190215142949-d0b11bdaac8a/go.mod h1:STP8DvDyc/dI5b8T5hshtkjS+E42TnysNCUPdjciGhY=
golang.org/x/sys v0.0.0-20190412213103-97732733099d/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
golang.org/x/sys v0.0.0-20190916202348-b4ddaad3f8a3/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
golang.org/x/sys v0.0.0-20200930185726-fdedc70b468f/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
golang.org/x/sys v0.0.0-20201119102817-f84b799fce68/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
golang.org/x/sys v0.0.0-20201204225414-ed752295db88/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
golang.org/x/sys v0.0.0-20210423082822-04245dca01da/go.mod h1:h1NjWce9XRLGQEsW7wpKNCjG9DtNlClVuFLEZdDNbEs=
@@ -454,29 +503,52 @@ golang.org/x/sys v0.41.0/go.mod h1:OgkHotnGiDImocRcuBABYBEXf8A9a87e/uXjp9XT3ks=
golang.org/x/term v0.0.0-20201126162022-7de9c90e9dd1/go.mod h1:bj7SfCRtBDWHUb9snDiAeCFNEtKQo2Wmx5Cou7ajbmo=
golang.org/x/term v0.40.0 h1:36e4zGLqU4yhjlmxEaagx2KuYbJq3EwY8K943ZsHcvg=
golang.org/x/term v0.40.0/go.mod h1:w2P8uVp06p2iyKKuvXIm7N/y0UCRt3UfJTfZ7oOpglM=
golang.org/x/text v0.3.0/go.mod h1:NqM8EUOU14njkJ3fqMW+pc6Ldnwhi/IjpwHt7yyuwOQ=
golang.org/x/text v0.3.3/go.mod h1:5Zoc/QRtKVWzQhOtBMvqHzDpF6irO9z98xDceosuGiQ=
golang.org/x/text v0.3.6/go.mod h1:5Zoc/QRtKVWzQhOtBMvqHzDpF6irO9z98xDceosuGiQ=
golang.org/x/text v0.34.0 h1:oL/Qq0Kdaqxa1KbNeMKwQq0reLCCaFtqu2eNuSeNHbk=
golang.org/x/text v0.34.0/go.mod h1:homfLqTYRFyVYemLBFl5GgL/DWEiH5wcsQ5gSh1yziA=
golang.org/x/time v0.14.0 h1:MRx4UaLrDotUKUdCIqzPC48t1Y9hANFKIRpNx+Te8PI=
golang.org/x/time v0.14.0/go.mod h1:eL/Oa2bBBK0TkX57Fyni+NgnyQQN4LitPmob2Hjnqw4=
golang.org/x/time v0.12.0 h1:ScB/8o8olJvc+CQPWrK3fPZNfh7qgwCrY0zJmoEQLSE=
golang.org/x/time v0.12.0/go.mod h1:CDIdPxbZBQxdj6cxyCIdrNogrJKMJ7pr37NYpMcMDSg=
golang.org/x/tools v0.0.0-20180917221912-90fa682c2a6e/go.mod h1:n7NCudcB/nEzxVGmLbDWY5pfWTLqBcC2KZ6jyYvM4mQ=
golang.org/x/tools v0.0.0-20190114222345-bf090417da8b/go.mod h1:n7NCudcB/nEzxVGmLbDWY5pfWTLqBcC2KZ6jyYvM4mQ=
golang.org/x/tools v0.0.0-20190226205152-f727befe758c/go.mod h1:9Yl7xja0Znq3iFh3HoIrodX9oNMXvdceNzlUR8zjMvY=
golang.org/x/tools v0.0.0-20190311212946-11955173bddd/go.mod h1:LCzVGOaR6xXOjkQ3onu1FJEFr0SW1gC7cKk1uF8kGRs=
golang.org/x/tools v0.0.0-20190524140312-2c0ae7006135/go.mod h1:RgjU9mgBXZiqYHBnxXauZ1Gv1EHHAz9KjViQ78xBX0Q=
golang.org/x/tools v0.41.0 h1:a9b8iMweWG+S0OBnlU36rzLp20z1Rp10w+IY2czHTQc=
golang.org/x/tools v0.41.0/go.mod h1:XSY6eDqxVNiYgezAVqqCeihT4j1U2CCsqvH3WhQpnlg=
golang.org/x/xerrors v0.0.0-20191204190536-9bdfabe68543/go.mod h1:I/5z698sn9Ka8TeJc9MKroUUfqBBauWjQqLJ2OPfmY0=
gonum.org/v1/gonum v0.16.0 h1:5+ul4Swaf3ESvrOnidPp4GZbzf0mxVQpDCYUQE7OJfk=
gonum.org/v1/gonum v0.16.0/go.mod h1:fef3am4MQ93R2HHpKnLk4/Tbh/s0+wqD5nfa6Pnwy4E=
google.golang.org/api v0.269.0 h1:qDrTOxKUQ/P0MveH6a7vZ+DNHxJQjtGm/uvdbdGXCQg=
google.golang.org/api v0.269.0/go.mod h1:N8Wpcu23Tlccl0zSHEkcAZQKDLdquxK+l9r2LkwAauE=
google.golang.org/genproto v0.0.0-20260128011058-8636f8732409 h1:VQZ/yAbAtjkHgH80teYd2em3xtIkkHd7ZhqfH2N9CsM=
google.golang.org/genproto v0.0.0-20260128011058-8636f8732409/go.mod h1:rxKD3IEILWEu3P44seeNOAwZN4SaoKaQ/2eTg4mM6EM=
google.golang.org/genproto/googleapis/api v0.0.0-20260128011058-8636f8732409 h1:merA0rdPeUV3YIIfHHcH4qBkiQAc1nfCKSI7lB4cV2M=
google.golang.org/genproto/googleapis/api v0.0.0-20260128011058-8636f8732409/go.mod h1:fl8J1IvUjCilwZzQowmw2b7HQB2eAuYBabMXzWurF+I=
google.golang.org/genproto/googleapis/rpc v0.0.0-20260217215200-42d3e9bedb6d h1:t/LOSXPJ9R0B6fnZNyALBRfZBH0Uy0gT+uR+SJ6syqQ=
google.golang.org/genproto/googleapis/rpc v0.0.0-20260217215200-42d3e9bedb6d/go.mod h1:4Hqkh8ycfw05ld/3BWL7rJOSfebL2Q+DVDeRgYgxUU8=
google.golang.org/grpc v1.79.1 h1:zGhSi45ODB9/p3VAawt9a+O/MULLl9dpizzNNpq7flY=
google.golang.org/grpc v1.79.1/go.mod h1:KmT0Kjez+0dde/v2j9vzwoAScgEPx/Bw1CYChhHLrHQ=
google.golang.org/protobuf v1.36.11 h1:fV6ZwhNocDyBLK0dj+fg8ektcVegBBuEolpbTQyBNVE=
google.golang.org/protobuf v1.36.11/go.mod h1:HTf+CrKn2C3g5S8VImy6tdcUvCska2kB7j23XfzDpco=
google.golang.org/api v0.153.0 h1:N1AwGhielyKFaUqH07/ZSIQR3uNPcV7NVw0vj+j4iR4=
google.golang.org/api v0.153.0/go.mod h1:3qNJX5eOmhiWYc67jRA/3GsDw97UFb5ivv7Y2PrriAY=
google.golang.org/appengine v1.1.0/go.mod h1:EbEs0AVv82hx2wNQdGPgUI5lhzA/G0D9YwlJXL52JkM=
google.golang.org/appengine v1.4.0/go.mod h1:xpcJRLb0r/rnEns0DIKYYv+WjYCduHsrkT7/EB5XEv4=
google.golang.org/genproto v0.0.0-20180817151627-c66870c02cf8/go.mod h1:JiN7NxoALGmiZfu7CAH4rXhgtRTLTxftemlI0sWmxmc=
google.golang.org/genproto v0.0.0-20190819201941-24fa4b261c55/go.mod h1:DMBHOl98Agz4BDEuKkezgsaosCRResVns1a3J2ZsMNc=
google.golang.org/genproto v0.0.0-20200526211855-cb27e3aa2013/go.mod h1:NbSheEEYHJ7i3ixzK3sjbqSGDJWnxyFXZblF3eUsNvo=
google.golang.org/genproto/googleapis/api v0.0.0-20250929231259-57b25ae835d4 h1:8XJ4pajGwOlasW+L13MnEGA8W4115jJySQtVfS2/IBU=
google.golang.org/genproto/googleapis/api v0.0.0-20250929231259-57b25ae835d4/go.mod h1:NnuHhy+bxcg30o7FnVAZbXsPHUDQ9qKWAQKCD7VxFtk=
google.golang.org/genproto/googleapis/rpc v0.0.0-20250929231259-57b25ae835d4 h1:i8QOKZfYg6AbGVZzUAY3LrNWCKF8O6zFisU9Wl9RER4=
google.golang.org/genproto/googleapis/rpc v0.0.0-20250929231259-57b25ae835d4/go.mod h1:HSkG/KdJWusxU1F6CNrwNDjBMgisKxGnc5dAZfT0mjQ=
google.golang.org/grpc v1.19.0/go.mod h1:mqu4LbDTu4XGKhr4mRzUsmM4RtVoemTSY81AxZiDr8c=
google.golang.org/grpc v1.23.0/go.mod h1:Y5yQAOtifL1yxbo5wqy6BxZv8vAUGQwXBOALyacEbxg=
google.golang.org/grpc v1.25.1/go.mod h1:c3i+UQWmh7LiEpx4sFZnkU36qjEYZ0imhYfXVyQciAY=
google.golang.org/grpc v1.27.0/go.mod h1:qbnxyOmOxrQa7FizSgH+ReBfzJrCY1pSN7KXBS8abTk=
google.golang.org/grpc v1.33.2/go.mod h1:JMHMWHQWaTccqQQlmk3MJZS+GWXOdAesneDmEnv2fbc=
google.golang.org/grpc v1.75.1 h1:/ODCNEuf9VghjgO3rqLcfg8fiOP0nSluljWFlDxELLI=
google.golang.org/grpc v1.75.1/go.mod h1:JtPAzKiq4v1xcAB2hydNlWI2RnF85XXcV0mhKXr2ecQ=
google.golang.org/protobuf v0.0.0-20200109180630-ec00e32a8dfd/go.mod h1:DFci5gLYBciE7Vtevhsrf46CRTquxDuWsQurQQe4oz8=
google.golang.org/protobuf v0.0.0-20200221191635-4d8936d0db64/go.mod h1:kwYJMbMJ01Woi6D6+Kah6886xMZcty6N08ah7+eCXa0=
google.golang.org/protobuf v0.0.0-20200228230310-ab0ca4ff8a60/go.mod h1:cfTl7dwQJ+fmap5saPgwCLgHXTUD7jkjRqWcaiX5VyM=
google.golang.org/protobuf v1.20.1-0.20200309200217-e05f789c0967/go.mod h1:A+miEFZTKqfCUM6K7xSMQL9OKL/b6hQv+e19PK+JZNE=
google.golang.org/protobuf v1.21.0/go.mod h1:47Nbq4nVaFHyn7ilMalzfO3qCViNmqZ2kzikPIcrTAo=
google.golang.org/protobuf v1.22.0/go.mod h1:EGpADcykh3NcUnDUJcl1+ZksZNG86OlYog2l/sGQquU=
google.golang.org/protobuf v1.23.0/go.mod h1:EGpADcykh3NcUnDUJcl1+ZksZNG86OlYog2l/sGQquU=
google.golang.org/protobuf v1.23.1-0.20200526195155-81db48ad09cc/go.mod h1:EGpADcykh3NcUnDUJcl1+ZksZNG86OlYog2l/sGQquU=
google.golang.org/protobuf v1.25.0/go.mod h1:9JNX74DMeImyA3h4bdi1ymwjUzf21/xIlbajtzgsN7c=
google.golang.org/protobuf v1.36.10 h1:AYd7cD/uASjIL6Q9LiTjz8JLcrh/88q5UObnmY3aOOE=
google.golang.org/protobuf v1.36.10/go.mod h1:HTf+CrKn2C3g5S8VImy6tdcUvCska2kB7j23XfzDpco=
gopkg.in/check.v1 v0.0.0-20161208181325-20d25e280405/go.mod h1:Co6ibVJAznAaIkqp8huTwlJQCZ016jof/cbN4VW5Yz0=
gopkg.in/check.v1 v1.0.0-20201130134442-10cb98267c6c h1:Hei/4ADfdWqJk1ZMxUNpqntNwaWcugrBjAiHlqqRiVk=
gopkg.in/check.v1 v1.0.0-20201130134442-10cb98267c6c/go.mod h1:JHkPIbrfpd72SG/EVd6muEfDQjcINNoR0C8j2r3qZ4Q=
@@ -489,6 +561,8 @@ gopkg.in/yaml.v3 v3.0.1 h1:fxVm/GzAzEWqLHuvctI91KS9hhNmmWOoWu0XTYJS7CA=
gopkg.in/yaml.v3 v3.0.1/go.mod h1:K4uyk7z7BCEPqu6E+C64Yfv1cQ7kz7rIZviUmN+EgEM=
gotest.tools/v3 v3.5.2 h1:7koQfIKdy+I8UTetycgUqXWSDwpgv193Ka+qRsmBY8Q=
gotest.tools/v3 v3.5.2/go.mod h1:LtdLGcnqToBH83WByAAi/wiwSFCArdFIUV/xxN4pcjA=
honnef.co/go/tools v0.0.0-20190102054323-c2f93a96b099/go.mod h1:rf3lG4BRIbNafJWhAfAdb/ePZxsR/4RtNHQocxwk9r4=
honnef.co/go/tools v0.0.0-20190523083050-ea95bdfd59fc/go.mod h1:rf3lG4BRIbNafJWhAfAdb/ePZxsR/4RtNHQocxwk9r4=
modernc.org/cc/v4 v4.27.1 h1:9W30zRlYrefrDV2JE2O8VDtJ1yPGownxciz5rrbQZis=
modernc.org/cc/v4 v4.27.1/go.mod h1:uVtb5OGqUKpoLWhqwNQo/8LwvoiEBLvZXIQ/SmO6mL0=
modernc.org/ccgo/v4 v4.30.1 h1:4r4U1J6Fhj98NKfSjnPUN7Ze2c6MnAdL0hWw6+LrJpc=
backend/internal/handler/admin/account_handler.go
+22 -38
View File
@@ -314,48 +314,32 @@ func (h *AccountHandler) List(c *gin.Context) {
}
}
// 窗口费用获取：lite 模式从快照缓存读取，非 lite 模式执行 PostgreSQL 查询后写入缓存
// 始终获取窗口费用（PostgreSQL 聚合查询）
if len(windowCostAccountIDs) > 0 {
if lite {
// lite 模式：尝试从快照缓存读取
cacheKey := buildWindowCostCacheKey(windowCostAccountIDs)
if cached, ok := accountWindowCostCache.Get(cacheKey); ok {
if costs, ok := cached.Payload.(map[int64]float64); ok {
windowCosts = costs
}
}
// 缓存未命中则 windowCosts 保持 nil（仅发生在服务刚启动时）
} else {
// 非 lite 模式：执行 PostgreSQL 聚合查询（高开销）
windowCosts = make(map[int64]float64)
var mu sync.Mutex
g, gctx := errgroup.WithContext(c.Request.Context())
g.SetLimit(10) // 限制并发数
windowCosts = make(map[int64]float64)
var mu sync.Mutex
g, gctx := errgroup.WithContext(c.Request.Context())
g.SetLimit(10) // 限制并发数
for i := range accounts {
acc := &accounts[i]
if !acc.IsAnthropicOAuthOrSetupToken() || acc.GetWindowCostLimit() <= 0 {
continue
}
accCopy := acc // 闭包捕获
g.Go(func() error {
// 使用统一的窗口开始时间计算逻辑（考虑窗口过期情况）
startTime := accCopy.GetCurrentWindowStartTime()
stats, err := h.accountUsageService.GetAccountWindowStats(gctx, accCopy.ID, startTime)
if err == nil && stats != nil {
mu.Lock()
windowCosts[accCopy.ID] = stats.StandardCost // 使用标准费用
mu.Unlock()
}
return nil // 不返回错误，允许部分失败
})
for i := range accounts {
acc := &accounts[i]
if !acc.IsAnthropicOAuthOrSetupToken() || acc.GetWindowCostLimit() <= 0 {
continue
}
_ = g.Wait()
// 查询完毕后写入快照缓存，供 lite 模式使用
cacheKey := buildWindowCostCacheKey(windowCostAccountIDs)
accountWindowCostCache.Set(cacheKey, windowCosts)
accCopy := acc // 闭包捕获
g.Go(func() error {
// 使用统一的窗口开始时间计算逻辑（考虑窗口过期情况）
startTime := accCopy.GetCurrentWindowStartTime()
stats, err := h.accountUsageService.GetAccountWindowStats(gctx, accCopy.ID, startTime)
if err == nil && stats != nil {
mu.Lock()
windowCosts[accCopy.ID] = stats.StandardCost // 使用标准费用
mu.Unlock()
}
return nil // 不返回错误，允许部分失败
})
}
_ = g.Wait()
}
// 获取亲和客户端数据（Redis Pipeline，低开销）
backend/internal/handler/admin/account_window_cost_cache.go
-25
View File
@@ -1,25 +0,0 @@
package admin
import (
"strconv"
"strings"
"time"
)
var accountWindowCostCache = newSnapshotCache(30 * time.Second)
func buildWindowCostCacheKey(accountIDs []int64) string {
if len(accountIDs) == 0 {
return "accounts_window_cost_empty"
}
var b strings.Builder
b.Grow(len(accountIDs) * 6)
_, _ = b.WriteString("accounts_window_cost:")
for i, id := range accountIDs {
if i > 0 {
_ = b.WriteByte(',')
}
_, _ = b.WriteString(strconv.FormatInt(id, 10))
}
return b.String()
}
backend/internal/handler/dto/mappers.go
+3 -3
View File
@@ -89,9 +89,9 @@ func APIKeyFromService(k *service.APIKey) *APIKey {
RateLimit5h: k.RateLimit5h,
RateLimit1d: k.RateLimit1d,
RateLimit7d: k.RateLimit7d,
Usage5h: k.Usage5h,
Usage1d: k.Usage1d,
Usage7d: k.Usage7d,
Usage5h: k.EffectiveUsage5h(),
Usage1d: k.EffectiveUsage1d(),
Usage7d: k.EffectiveUsage7d(),
Window5hStart: k.Window5hStart,
Window1dStart: k.Window1dStart,
Window7dStart: k.Window7dStart,
backend/internal/handler/gateway_handler.go
+3 -3
View File
@@ -974,7 +974,7 @@ func (h *GatewayHandler) usageQuotaLimited(c *gin.Context, ctx context.Context,
if err == nil && rateLimitData != nil {
var rateLimits []gin.H
if apiKey.RateLimit5h > 0 {
used := rateLimitData.Usage5h
used := rateLimitData.EffectiveUsage5h()
rateLimits = append(rateLimits, gin.H{
"window": "5h",
"limit": apiKey.RateLimit5h,
@@ -984,7 +984,7 @@ func (h *GatewayHandler) usageQuotaLimited(c *gin.Context, ctx context.Context,
})
}
if apiKey.RateLimit1d > 0 {
used := rateLimitData.Usage1d
used := rateLimitData.EffectiveUsage1d()
rateLimits = append(rateLimits, gin.H{
"window": "1d",
"limit": apiKey.RateLimit1d,
@@ -994,7 +994,7 @@ func (h *GatewayHandler) usageQuotaLimited(c *gin.Context, ctx context.Context,
})
}
if apiKey.RateLimit7d > 0 {
used := rateLimitData.Usage7d
used := rateLimitData.EffectiveUsage7d()
rateLimits = append(rateLimits, gin.H{
"window": "7d",
"limit": apiKey.RateLimit7d,
backend/internal/handler/openai_gateway_handler.go
+308 -1
View File
@@ -118,6 +118,20 @@ func (h *OpenAIGatewayHandler) Responses(c *gin.Context) {
}
setOpsRequestContext(c, "", false, body)
sessionHashBody := body
if service.IsOpenAIResponsesCompactPathForTest(c) {
if compactSeed := strings.TrimSpace(gjson.GetBytes(body, "prompt_cache_key").String()); compactSeed != "" {
c.Set(service.OpenAICompactSessionSeedKeyForTest(), compactSeed)
}
normalizedCompactBody, normalizedCompact, compactErr := service.NormalizeOpenAICompactRequestBodyForTest(body)
if compactErr != nil {
h.errorResponse(c, http.StatusBadRequest, "invalid_request_error", "Failed to normalize compact request body")
return
}
if normalizedCompact {
body = normalizedCompactBody
}
}
// 校验请求体 JSON 合法性
if !gjson.ValidBytes(body) {
@@ -193,7 +207,7 @@ func (h *OpenAIGatewayHandler) Responses(c *gin.Context) {
}
// Generate session hash (header first; fallback to prompt_cache_key)
sessionHash := h.gatewayService.GenerateSessionHash(c, body)
sessionHash := h.gatewayService.GenerateSessionHash(c, sessionHashBody)
maxAccountSwitches := h.maxAccountSwitches
switchCount := 0
@@ -305,6 +319,9 @@ func (h *OpenAIGatewayHandler) Responses(c *gin.Context) {
return
}
if result != nil {
if account.Type == service.AccountTypeOAuth {
h.gatewayService.UpdateCodexUsageSnapshotFromHeaders(c.Request.Context(), account.ID, result.ResponseHeaders)
}
h.gatewayService.ReportOpenAIAccountScheduleResult(account.ID, true, result.FirstTokenMs)
} else {
h.gatewayService.ReportOpenAIAccountScheduleResult(account.ID, true, nil)
@@ -424,6 +441,273 @@ func (h *OpenAIGatewayHandler) logOpenAIRemoteCompactOutcome(c *gin.Context, sta
log.Warn("codex.remote_compact.failed")
}
// Messages handles Anthropic Messages API requests routed to OpenAI platform.
// POST /v1/messages (when group platform is OpenAI)
func (h *OpenAIGatewayHandler) Messages(c *gin.Context) {
streamStarted := false
defer h.recoverAnthropicMessagesPanic(c, &streamStarted)
requestStart := time.Now()
apiKey, ok := middleware2.GetAPIKeyFromContext(c)
if !ok {
h.anthropicErrorResponse(c, http.StatusUnauthorized, "authentication_error", "Invalid API key")
return
}
subject, ok := middleware2.GetAuthSubjectFromContext(c)
if !ok {
h.anthropicErrorResponse(c, http.StatusInternalServerError, "api_error", "User context not found")
return
}
reqLog := requestLogger(
c,
"handler.openai_gateway.messages",
zap.Int64("user_id", subject.UserID),
zap.Int64("api_key_id", apiKey.ID),
zap.Any("group_id", apiKey.GroupID),
)
if !h.ensureResponsesDependencies(c, reqLog) {
return
}
body, err := pkghttputil.ReadRequestBodyWithPrealloc(c.Request)
if err != nil {
if maxErr, ok := extractMaxBytesError(err); ok {
h.anthropicErrorResponse(c, http.StatusRequestEntityTooLarge, "invalid_request_error", buildBodyTooLargeMessage(maxErr.Limit))
return
}
h.anthropicErrorResponse(c, http.StatusBadRequest, "invalid_request_error", "Failed to read request body")
return
}
if len(body) == 0 {
h.anthropicErrorResponse(c, http.StatusBadRequest, "invalid_request_error", "Request body is empty")
return
}
if !gjson.ValidBytes(body) {
h.anthropicErrorResponse(c, http.StatusBadRequest, "invalid_request_error", "Failed to parse request body")
return
}
modelResult := gjson.GetBytes(body, "model")
if !modelResult.Exists() || modelResult.Type != gjson.String || modelResult.String() == "" {
h.anthropicErrorResponse(c, http.StatusBadRequest, "invalid_request_error", "model is required")
return
}
reqModel := modelResult.String()
reqStream := gjson.GetBytes(body, "stream").Bool()
reqLog = reqLog.With(zap.String("model", reqModel), zap.Bool("stream", reqStream))
setOpsRequestContext(c, reqModel, reqStream, body)
// 绑定错误透传服务，允许 service 层在非 failover 错误场景复用规则。
if h.errorPassthroughService != nil {
service.BindErrorPassthroughService(c, h.errorPassthroughService)
}
subscription, _ := middleware2.GetSubscriptionFromContext(c)
service.SetOpsLatencyMs(c, service.OpsAuthLatencyMsKey, time.Since(requestStart).Milliseconds())
routingStart := time.Now()
userReleaseFunc, acquired := h.acquireResponsesUserSlot(c, subject.UserID, subject.Concurrency, reqStream, &streamStarted, reqLog)
if !acquired {
return
}
if userReleaseFunc != nil {
defer userReleaseFunc()
}
if err := h.billingCacheService.CheckBillingEligibility(c.Request.Context(), apiKey.User, apiKey, apiKey.Group, subscription); err != nil {
reqLog.Info("openai_messages.billing_eligibility_check_failed", zap.Error(err))
status, code, message := billingErrorDetails(err)
h.anthropicStreamingAwareError(c, status, code, message, streamStarted)
return
}
sessionHash := h.gatewayService.GenerateSessionHash(c, body)
promptCacheKey := h.gatewayService.ExtractSessionID(c, body)
maxAccountSwitches := h.maxAccountSwitches
switchCount := 0
failedAccountIDs := make(map[int64]struct{})
var lastFailoverErr *service.UpstreamFailoverError
for {
reqLog.Debug("openai_messages.account_selecting", zap.Int("excluded_account_count", len(failedAccountIDs)))
selection, scheduleDecision, err := h.gatewayService.SelectAccountWithScheduler(
c.Request.Context(),
apiKey.GroupID,
"", // no previous_response_id
sessionHash,
reqModel,
failedAccountIDs,
service.OpenAIUpstreamTransportAny,
)
if err != nil {
reqLog.Warn("openai_messages.account_select_failed",
zap.Error(err),
zap.Int("excluded_account_count", len(failedAccountIDs)),
)
if len(failedAccountIDs) == 0 {
h.anthropicStreamingAwareError(c, http.StatusServiceUnavailable, "api_error", "Service temporarily unavailable", streamStarted)
return
}
if lastFailoverErr != nil {
h.handleAnthropicFailoverExhausted(c, lastFailoverErr, streamStarted)
} else {
h.anthropicStreamingAwareError(c, http.StatusBadGateway, "api_error", "Upstream request failed", streamStarted)
}
return
}
if selection == nil || selection.Account == nil {
h.anthropicStreamingAwareError(c, http.StatusServiceUnavailable, "api_error", "No available accounts", streamStarted)
return
}
account := selection.Account
reqLog.Debug("openai_messages.account_selected", zap.Int64("account_id", account.ID), zap.String("account_name", account.Name))
_ = scheduleDecision
setOpsSelectedAccount(c, account.ID, account.Platform)
accountReleaseFunc, acquired := h.acquireResponsesAccountSlot(c, apiKey.GroupID, sessionHash, selection, reqStream, &streamStarted, reqLog)
if !acquired {
return
}
service.SetOpsLatencyMs(c, service.OpsRoutingLatencyMsKey, time.Since(routingStart).Milliseconds())
forwardStart := time.Now()
result, err := h.gatewayService.ForwardAsAnthropic(c.Request.Context(), c, account, body, promptCacheKey)
forwardDurationMs := time.Since(forwardStart).Milliseconds()
if accountReleaseFunc != nil {
accountReleaseFunc()
}
upstreamLatencyMs, _ := getContextInt64(c, service.OpsUpstreamLatencyMsKey)
responseLatencyMs := forwardDurationMs
if upstreamLatencyMs > 0 && forwardDurationMs > upstreamLatencyMs {
responseLatencyMs = forwardDurationMs - upstreamLatencyMs
}
service.SetOpsLatencyMs(c, service.OpsResponseLatencyMsKey, responseLatencyMs)
if err == nil && result != nil && result.FirstTokenMs != nil {
service.SetOpsLatencyMs(c, service.OpsTimeToFirstTokenMsKey, int64(*result.FirstTokenMs))
}
if err != nil {
var failoverErr *service.UpstreamFailoverError
if errors.As(err, &failoverErr) {
h.gatewayService.ReportOpenAIAccountScheduleResult(account.ID, false, nil)
h.gatewayService.RecordOpenAIAccountSwitch()
failedAccountIDs[account.ID] = struct{}{}
lastFailoverErr = failoverErr
if switchCount >= maxAccountSwitches {
h.handleAnthropicFailoverExhausted(c, failoverErr, streamStarted)
return
}
switchCount++
reqLog.Warn("openai_messages.upstream_failover_switching",
zap.Int64("account_id", account.ID),
zap.Int("upstream_status", failoverErr.StatusCode),
zap.Int("switch_count", switchCount),
zap.Int("max_switches", maxAccountSwitches),
)
continue
}
h.gatewayService.ReportOpenAIAccountScheduleResult(account.ID, false, nil)
wroteFallback := h.ensureAnthropicErrorResponse(c, streamStarted)
reqLog.Warn("openai_messages.forward_failed",
zap.Int64("account_id", account.ID),
zap.Bool("fallback_error_response_written", wroteFallback),
zap.Error(err),
)
return
}
if result != nil {
h.gatewayService.ReportOpenAIAccountScheduleResult(account.ID, true, result.FirstTokenMs)
} else {
h.gatewayService.ReportOpenAIAccountScheduleResult(account.ID, true, nil)
}
userAgent := c.GetHeader("User-Agent")
clientIP := ip.GetClientIP(c)
h.submitUsageRecordTask(func(ctx context.Context) {
if err := h.gatewayService.RecordUsage(ctx, &service.OpenAIRecordUsageInput{
Result: result,
APIKey: apiKey,
User: apiKey.User,
Account: account,
Subscription: subscription,
UserAgent: userAgent,
IPAddress: clientIP,
APIKeyService: h.apiKeyService,
}); err != nil {
logger.L().With(
zap.String("component", "handler.openai_gateway.messages"),
zap.Int64("user_id", subject.UserID),
zap.Int64("api_key_id", apiKey.ID),
zap.Any("group_id", apiKey.GroupID),
zap.String("model", reqModel),
zap.Int64("account_id", account.ID),
).Error("openai_messages.record_usage_failed", zap.Error(err))
}
})
reqLog.Debug("openai_messages.request_completed",
zap.Int64("account_id", account.ID),
zap.Int("switch_count", switchCount),
)
return
}
}
// anthropicErrorResponse writes an error in Anthropic Messages API format.
func (h *OpenAIGatewayHandler) anthropicErrorResponse(c *gin.Context, status int, errType, message string) {
c.JSON(status, gin.H{
"type": "error",
"error": gin.H{
"type": errType,
"message": message,
},
})
}
// anthropicStreamingAwareError handles errors that may occur during streaming,
// using Anthropic SSE error format.
func (h *OpenAIGatewayHandler) anthropicStreamingAwareError(c *gin.Context, status int, errType, message string, streamStarted bool) {
if streamStarted {
flusher, ok := c.Writer.(http.Flusher)
if ok {
errPayload, _ := json.Marshal(gin.H{
"type": "error",
"error": gin.H{
"type": errType,
"message": message,
},
})
fmt.Fprintf(c.Writer, "event: error\ndata: %s\n\n", errPayload) //nolint:errcheck
flusher.Flush()
}
return
}
h.anthropicErrorResponse(c, status, errType, message)
}
// handleAnthropicFailoverExhausted maps upstream failover errors to Anthropic format.
func (h *OpenAIGatewayHandler) handleAnthropicFailoverExhausted(c *gin.Context, failoverErr *service.UpstreamFailoverError, streamStarted bool) {
status, errType, errMsg := h.mapUpstreamError(failoverErr.StatusCode)
h.anthropicStreamingAwareError(c, status, errType, errMsg, streamStarted)
}
// ensureAnthropicErrorResponse writes a fallback Anthropic error if no response was written.
func (h *OpenAIGatewayHandler) ensureAnthropicErrorResponse(c *gin.Context, streamStarted bool) bool {
if c == nil || c.Writer == nil || c.Writer.Written() {
return false
}
h.anthropicStreamingAwareError(c, http.StatusBadGateway, "api_error", "Upstream request failed", streamStarted)
return true
}
func (h *OpenAIGatewayHandler) validateFunctionCallOutputRequest(c *gin.Context, body []byte, reqLog *zap.Logger) bool {
if !gjson.GetBytes(body, `input.#(type=="function_call_output")`).Exists() {
return true
@@ -840,6 +1124,9 @@ func (h *OpenAIGatewayHandler) ResponsesWebSocket(c *gin.Context) {
if turnErr != nil || result == nil {
return
}
if account.Type == service.AccountTypeOAuth {
h.gatewayService.UpdateCodexUsageSnapshotFromHeaders(ctx, account.ID, result.ResponseHeaders)
}
h.gatewayService.ReportOpenAIAccountScheduleResult(account.ID, true, result.FirstTokenMs)
h.submitUsageRecordTask(func(taskCtx context.Context) {
if err := h.gatewayService.RecordUsage(taskCtx, &service.OpenAIRecordUsageInput{
@@ -901,6 +1188,26 @@ func (h *OpenAIGatewayHandler) recoverResponsesPanic(c *gin.Context, streamStart
)
}
// recoverAnthropicMessagesPanic recovers from panics in the Anthropic Messages
// handler and returns an Anthropic-formatted error response.
func (h *OpenAIGatewayHandler) recoverAnthropicMessagesPanic(c *gin.Context, streamStarted *bool) {
recovered := recover()
if recovered == nil {
return
}
started := streamStarted != nil && *streamStarted
requestLogger(c, "handler.openai_gateway.messages").Error(
"openai.messages_panic_recovered",
zap.Bool("stream_started", started),
zap.Any("panic", recovered),
zap.ByteString("stack", debug.Stack()),
)
if !started {
h.anthropicErrorResponse(c, http.StatusInternalServerError, "api_error", "Internal server error")
}
}
func (h *OpenAIGatewayHandler) ensureResponsesDependencies(c *gin.Context, reqLog *zap.Logger) bool {
missing := h.missingResponsesDependencies()
if len(missing) == 0 {
backend/internal/pkg/antigravity/stream_transformer.go
+17 -7
View File
@@ -142,23 +142,33 @@ func (p *StreamingProcessor) ProcessLine(line string) []byte {
return result.Bytes()
}
// Finish 结束处理，返回最终事件和用量
// Finish 结束处理，返回最终事件和用量。
// 若整个流未收到任何可解析的上游数据（messageStartSent == false），
// 则不补发任何结束事件，防止客户端收到没有 message_start 的残缺流。
func (p *StreamingProcessor) Finish() ([]byte, *ClaudeUsage) {
var result bytes.Buffer
if !p.messageStopSent {
_, _ = result.Write(p.emitFinish(""))
}
usage := &ClaudeUsage{
InputTokens: p.inputTokens,
OutputTokens: p.outputTokens,
CacheReadInputTokens: p.cacheReadTokens,
}
if !p.messageStartSent {
return nil, usage
}
var result bytes.Buffer
if !p.messageStopSent {
_, _ = result.Write(p.emitFinish(""))
}
return result.Bytes(), usage
}
// MessageStartSent 报告流中是否已发出过 message_start 事件（即是否收到过有效的上游数据）
func (p *StreamingProcessor) MessageStartSent() bool {
return p.messageStartSent
}
// emitMessageStart 发送 message_start 事件
func (p *StreamingProcessor) emitMessageStart(v1Resp *V1InternalResponse) []byte {
if p.messageStartSent {
backend/internal/pkg/apicompat/anthropic_responses_test.go
+735
View File
@@ -0,0 +1,735 @@
package apicompat
import (
"encoding/json"
"testing"
"github.com/stretchr/testify/assert"
"github.com/stretchr/testify/require"
)
// ---------------------------------------------------------------------------
// AnthropicToResponses tests
// ---------------------------------------------------------------------------
func TestAnthropicToResponses_BasicText(t *testing.T) {
req := &AnthropicRequest{
Model: "gpt-5.2",
MaxTokens: 1024,
Stream: true,
Messages: []AnthropicMessage{
{Role: "user", Content: json.RawMessage(`"Hello"`)},
},
}
resp, err := AnthropicToResponses(req)
require.NoError(t, err)
assert.Equal(t, "gpt-5.2", resp.Model)
assert.True(t, resp.Stream)
assert.Equal(t, 1024, *resp.MaxOutputTokens)
assert.False(t, *resp.Store)
var items []ResponsesInputItem
require.NoError(t, json.Unmarshal(resp.Input, &items))
require.Len(t, items, 1)
assert.Equal(t, "user", items[0].Role)
}
func TestAnthropicToResponses_SystemPrompt(t *testing.T) {
t.Run("string", func(t *testing.T) {
req := &AnthropicRequest{
Model: "gpt-5.2",
MaxTokens: 100,
System: json.RawMessage(`"You are helpful."`),
Messages: []AnthropicMessage{{Role: "user", Content: json.RawMessage(`"Hi"`)}},
}
resp, err := AnthropicToResponses(req)
require.NoError(t, err)
var items []ResponsesInputItem
require.NoError(t, json.Unmarshal(resp.Input, &items))
require.Len(t, items, 2)
assert.Equal(t, "system", items[0].Role)
})
t.Run("array", func(t *testing.T) {
req := &AnthropicRequest{
Model: "gpt-5.2",
MaxTokens: 100,
System: json.RawMessage(`[{"type":"text","text":"Part 1"},{"type":"text","text":"Part 2"}]`),
Messages: []AnthropicMessage{{Role: "user", Content: json.RawMessage(`"Hi"`)}},
}
resp, err := AnthropicToResponses(req)
require.NoError(t, err)
var items []ResponsesInputItem
require.NoError(t, json.Unmarshal(resp.Input, &items))
require.Len(t, items, 2)
assert.Equal(t, "system", items[0].Role)
// System text should be joined with double newline.
var text string
require.NoError(t, json.Unmarshal(items[0].Content, &text))
assert.Equal(t, "Part 1\n\nPart 2", text)
})
}
func TestAnthropicToResponses_ToolUse(t *testing.T) {
req := &AnthropicRequest{
Model: "gpt-5.2",
MaxTokens: 1024,
Messages: []AnthropicMessage{
{Role: "user", Content: json.RawMessage(`"What is the weather?"`)},
{Role: "assistant", Content: json.RawMessage(`[{"type":"text","text":"Let me check."},{"type":"tool_use","id":"call_1","name":"get_weather","input":{"city":"NYC"}}]`)},
{Role: "user", Content: json.RawMessage(`[{"type":"tool_result","tool_use_id":"call_1","content":"Sunny, 72°F"}]`)},
},
Tools: []AnthropicTool{
{Name: "get_weather", Description: "Get weather", InputSchema: json.RawMessage(`{"type":"object","properties":{"city":{"type":"string"}}}`)},
},
}
resp, err := AnthropicToResponses(req)
require.NoError(t, err)
// Check tools
require.Len(t, resp.Tools, 1)
assert.Equal(t, "function", resp.Tools[0].Type)
assert.Equal(t, "get_weather", resp.Tools[0].Name)
// Check input items
var items []ResponsesInputItem
require.NoError(t, json.Unmarshal(resp.Input, &items))
// user + assistant + function_call + function_call_output = 4
require.Len(t, items, 4)
assert.Equal(t, "user", items[0].Role)
assert.Equal(t, "assistant", items[1].Role)
assert.Equal(t, "function_call", items[2].Type)
assert.Equal(t, "fc_call_1", items[2].CallID)
assert.Equal(t, "function_call_output", items[3].Type)
assert.Equal(t, "fc_call_1", items[3].CallID)
assert.Equal(t, "Sunny, 72°F", items[3].Output)
}
func TestAnthropicToResponses_ThinkingIgnored(t *testing.T) {
req := &AnthropicRequest{
Model: "gpt-5.2",
MaxTokens: 1024,
Messages: []AnthropicMessage{
{Role: "user", Content: json.RawMessage(`"Hello"`)},
{Role: "assistant", Content: json.RawMessage(`[{"type":"thinking","thinking":"deep thought"},{"type":"text","text":"Hi!"}]`)},
{Role: "user", Content: json.RawMessage(`"More"`)},
},
}
resp, err := AnthropicToResponses(req)
require.NoError(t, err)
var items []ResponsesInputItem
require.NoError(t, json.Unmarshal(resp.Input, &items))
// user + assistant(text only, thinking ignored) + user = 3
require.Len(t, items, 3)
assert.Equal(t, "assistant", items[1].Role)
// Assistant content should only have text, not thinking.
var parts []ResponsesContentPart
require.NoError(t, json.Unmarshal(items[1].Content, &parts))
require.Len(t, parts, 1)
assert.Equal(t, "output_text", parts[0].Type)
assert.Equal(t, "Hi!", parts[0].Text)
}
func TestAnthropicToResponses_MaxTokensFloor(t *testing.T) {
req := &AnthropicRequest{
Model: "gpt-5.2",
MaxTokens: 10, // below minMaxOutputTokens (128)
Messages: []AnthropicMessage{{Role: "user", Content: json.RawMessage(`"Hi"`)}},
}
resp, err := AnthropicToResponses(req)
require.NoError(t, err)
assert.Equal(t, 128, *resp.MaxOutputTokens)
}
// ---------------------------------------------------------------------------
// ResponsesToAnthropic (non-streaming) tests
// ---------------------------------------------------------------------------
func TestResponsesToAnthropic_TextOnly(t *testing.T) {
resp := &ResponsesResponse{
ID: "resp_123",
Model: "gpt-5.2",
Status: "completed",
Output: []ResponsesOutput{
{
Type: "message",
Content: []ResponsesContentPart{
{Type: "output_text", Text: "Hello there!"},
},
},
},
Usage: &ResponsesUsage{InputTokens: 10, OutputTokens: 5, TotalTokens: 15},
}
anth := ResponsesToAnthropic(resp, "claude-opus-4-6")
assert.Equal(t, "resp_123", anth.ID)
assert.Equal(t, "claude-opus-4-6", anth.Model)
assert.Equal(t, "end_turn", anth.StopReason)
require.Len(t, anth.Content, 1)
assert.Equal(t, "text", anth.Content[0].Type)
assert.Equal(t, "Hello there!", anth.Content[0].Text)
assert.Equal(t, 10, anth.Usage.InputTokens)
assert.Equal(t, 5, anth.Usage.OutputTokens)
}
func TestResponsesToAnthropic_ToolUse(t *testing.T) {
resp := &ResponsesResponse{
ID: "resp_456",
Model: "gpt-5.2",
Status: "completed",
Output: []ResponsesOutput{
{
Type: "message",
Content: []ResponsesContentPart{
{Type: "output_text", Text: "Let me check."},
},
},
{
Type: "function_call",
CallID: "call_1",
Name: "get_weather",
Arguments: `{"city":"NYC"}`,
},
},
}
anth := ResponsesToAnthropic(resp, "claude-opus-4-6")
assert.Equal(t, "tool_use", anth.StopReason)
require.Len(t, anth.Content, 2)
assert.Equal(t, "text", anth.Content[0].Type)
assert.Equal(t, "tool_use", anth.Content[1].Type)
assert.Equal(t, "call_1", anth.Content[1].ID)
assert.Equal(t, "get_weather", anth.Content[1].Name)
}
func TestResponsesToAnthropic_Reasoning(t *testing.T) {
resp := &ResponsesResponse{
ID: "resp_789",
Model: "gpt-5.2",
Status: "completed",
Output: []ResponsesOutput{
{
Type: "reasoning",
Summary: []ResponsesSummary{
{Type: "summary_text", Text: "Thinking about the answer..."},
},
},
{
Type: "message",
Content: []ResponsesContentPart{
{Type: "output_text", Text: "42"},
},
},
},
}
anth := ResponsesToAnthropic(resp, "claude-opus-4-6")
require.Len(t, anth.Content, 2)
assert.Equal(t, "thinking", anth.Content[0].Type)
assert.Equal(t, "Thinking about the answer...", anth.Content[0].Thinking)
assert.Equal(t, "text", anth.Content[1].Type)
assert.Equal(t, "42", anth.Content[1].Text)
}
func TestResponsesToAnthropic_Incomplete(t *testing.T) {
resp := &ResponsesResponse{
ID: "resp_inc",
Model: "gpt-5.2",
Status: "incomplete",
IncompleteDetails: &ResponsesIncompleteDetails{
Reason: "max_output_tokens",
},
Output: []ResponsesOutput{
{
Type: "message",
Content: []ResponsesContentPart{{Type: "output_text", Text: "Partial..."}},
},
},
}
anth := ResponsesToAnthropic(resp, "claude-opus-4-6")
assert.Equal(t, "max_tokens", anth.StopReason)
}
func TestResponsesToAnthropic_EmptyOutput(t *testing.T) {
resp := &ResponsesResponse{
ID: "resp_empty",
Model: "gpt-5.2",
Status: "completed",
Output: []ResponsesOutput{},
}
anth := ResponsesToAnthropic(resp, "claude-opus-4-6")
require.Len(t, anth.Content, 1)
assert.Equal(t, "text", anth.Content[0].Type)
assert.Equal(t, "", anth.Content[0].Text)
}
// ---------------------------------------------------------------------------
// Streaming: ResponsesEventToAnthropicEvents tests
// ---------------------------------------------------------------------------
func TestStreamingTextOnly(t *testing.T) {
state := NewResponsesEventToAnthropicState()
// 1. response.created
events := ResponsesEventToAnthropicEvents(&ResponsesStreamEvent{
Type: "response.created",
Response: &ResponsesResponse{
ID: "resp_1",
Model: "gpt-5.2",
},
}, state)
require.Len(t, events, 1)
assert.Equal(t, "message_start", events[0].Type)
// 2. output_item.added (message)
events = ResponsesEventToAnthropicEvents(&ResponsesStreamEvent{
Type: "response.output_item.added",
OutputIndex: 0,
Item: &ResponsesOutput{Type: "message"},
}, state)
assert.Len(t, events, 0) // message item doesn't emit events
// 3. text delta
events = ResponsesEventToAnthropicEvents(&ResponsesStreamEvent{
Type: "response.output_text.delta",
Delta: "Hello",
}, state)
require.Len(t, events, 2) // content_block_start + content_block_delta
assert.Equal(t, "content_block_start", events[0].Type)
assert.Equal(t, "text", events[0].ContentBlock.Type)
assert.Equal(t, "content_block_delta", events[1].Type)
assert.Equal(t, "text_delta", events[1].Delta.Type)
assert.Equal(t, "Hello", events[1].Delta.Text)
// 4. more text
events = ResponsesEventToAnthropicEvents(&ResponsesStreamEvent{
Type: "response.output_text.delta",
Delta: " world",
}, state)
require.Len(t, events, 1) // only delta, no new block start
assert.Equal(t, "content_block_delta", events[0].Type)
// 5. text done
events = ResponsesEventToAnthropicEvents(&ResponsesStreamEvent{
Type: "response.output_text.done",
}, state)
require.Len(t, events, 1)
assert.Equal(t, "content_block_stop", events[0].Type)
// 6. completed
events = ResponsesEventToAnthropicEvents(&ResponsesStreamEvent{
Type: "response.completed",
Response: &ResponsesResponse{
Status: "completed",
Usage: &ResponsesUsage{InputTokens: 10, OutputTokens: 5},
},
}, state)
require.Len(t, events, 2) // message_delta + message_stop
assert.Equal(t, "message_delta", events[0].Type)
assert.Equal(t, "end_turn", events[0].Delta.StopReason)
assert.Equal(t, 10, events[0].Usage.InputTokens)
assert.Equal(t, 5, events[0].Usage.OutputTokens)
assert.Equal(t, "message_stop", events[1].Type)
}
func TestStreamingToolCall(t *testing.T) {
state := NewResponsesEventToAnthropicState()
// 1. response.created
ResponsesEventToAnthropicEvents(&ResponsesStreamEvent{
Type: "response.created",
Response: &ResponsesResponse{ID: "resp_2", Model: "gpt-5.2"},
}, state)
// 2. function_call added
events := ResponsesEventToAnthropicEvents(&ResponsesStreamEvent{
Type: "response.output_item.added",
OutputIndex: 0,
Item: &ResponsesOutput{Type: "function_call", CallID: "call_1", Name: "get_weather"},
}, state)
require.Len(t, events, 1)
assert.Equal(t, "content_block_start", events[0].Type)
assert.Equal(t, "tool_use", events[0].ContentBlock.Type)
assert.Equal(t, "call_1", events[0].ContentBlock.ID)
// 3. arguments delta
events = ResponsesEventToAnthropicEvents(&ResponsesStreamEvent{
Type: "response.function_call_arguments.delta",
OutputIndex: 0,
Delta: `{"city":`,
}, state)
require.Len(t, events, 1)
assert.Equal(t, "content_block_delta", events[0].Type)
assert.Equal(t, "input_json_delta", events[0].Delta.Type)
assert.Equal(t, `{"city":`, events[0].Delta.PartialJSON)
// 4. arguments done
events = ResponsesEventToAnthropicEvents(&ResponsesStreamEvent{
Type: "response.function_call_arguments.done",
}, state)
require.Len(t, events, 1)
assert.Equal(t, "content_block_stop", events[0].Type)
// 5. completed with tool_calls
events = ResponsesEventToAnthropicEvents(&ResponsesStreamEvent{
Type: "response.completed",
Response: &ResponsesResponse{
Status: "completed",
Usage: &ResponsesUsage{InputTokens: 20, OutputTokens: 10},
},
}, state)
require.Len(t, events, 2)
assert.Equal(t, "tool_use", events[0].Delta.StopReason)
}
func TestStreamingReasoning(t *testing.T) {
state := NewResponsesEventToAnthropicState()
ResponsesEventToAnthropicEvents(&ResponsesStreamEvent{
Type: "response.created",
Response: &ResponsesResponse{ID: "resp_3", Model: "gpt-5.2"},
}, state)
// reasoning item added
events := ResponsesEventToAnthropicEvents(&ResponsesStreamEvent{
Type: "response.output_item.added",
OutputIndex: 0,
Item: &ResponsesOutput{Type: "reasoning"},
}, state)
require.Len(t, events, 1)
assert.Equal(t, "content_block_start", events[0].Type)
assert.Equal(t, "thinking", events[0].ContentBlock.Type)
// reasoning text delta
events = ResponsesEventToAnthropicEvents(&ResponsesStreamEvent{
Type: "response.reasoning_summary_text.delta",
OutputIndex: 0,
Delta: "Let me think...",
}, state)
require.Len(t, events, 1)
assert.Equal(t, "content_block_delta", events[0].Type)
assert.Equal(t, "thinking_delta", events[0].Delta.Type)
assert.Equal(t, "Let me think...", events[0].Delta.Thinking)
// reasoning done
events = ResponsesEventToAnthropicEvents(&ResponsesStreamEvent{
Type: "response.reasoning_summary_text.done",
}, state)
require.Len(t, events, 1)
assert.Equal(t, "content_block_stop", events[0].Type)
}
func TestStreamingIncomplete(t *testing.T) {
state := NewResponsesEventToAnthropicState()
ResponsesEventToAnthropicEvents(&ResponsesStreamEvent{
Type: "response.created",
Response: &ResponsesResponse{ID: "resp_4", Model: "gpt-5.2"},
}, state)
ResponsesEventToAnthropicEvents(&ResponsesStreamEvent{
Type: "response.output_text.delta",
Delta: "Partial output...",
}, state)
events := ResponsesEventToAnthropicEvents(&ResponsesStreamEvent{
Type: "response.incomplete",
Response: &ResponsesResponse{
Status: "incomplete",
IncompleteDetails: &ResponsesIncompleteDetails{Reason: "max_output_tokens"},
Usage: &ResponsesUsage{InputTokens: 100, OutputTokens: 4096},
},
}, state)
// Should close the text block + message_delta + message_stop
require.Len(t, events, 3)
assert.Equal(t, "content_block_stop", events[0].Type)
assert.Equal(t, "message_delta", events[1].Type)
assert.Equal(t, "max_tokens", events[1].Delta.StopReason)
assert.Equal(t, "message_stop", events[2].Type)
}
func TestFinalizeStream_NeverStarted(t *testing.T) {
state := NewResponsesEventToAnthropicState()
events := FinalizeResponsesAnthropicStream(state)
assert.Nil(t, events)
}
func TestFinalizeStream_AlreadyCompleted(t *testing.T) {
state := NewResponsesEventToAnthropicState()
state.MessageStartSent = true
state.MessageStopSent = true
events := FinalizeResponsesAnthropicStream(state)
assert.Nil(t, events)
}
func TestFinalizeStream_AbnormalTermination(t *testing.T) {
state := NewResponsesEventToAnthropicState()
// Simulate a stream that started but never completed
ResponsesEventToAnthropicEvents(&ResponsesStreamEvent{
Type: "response.created",
Response: &ResponsesResponse{ID: "resp_5", Model: "gpt-5.2"},
}, state)
ResponsesEventToAnthropicEvents(&ResponsesStreamEvent{
Type: "response.output_text.delta",
Delta: "Interrupted...",
}, state)
// Stream ends without response.completed
events := FinalizeResponsesAnthropicStream(state)
require.Len(t, events, 3) // content_block_stop + message_delta + message_stop
assert.Equal(t, "content_block_stop", events[0].Type)
assert.Equal(t, "message_delta", events[1].Type)
assert.Equal(t, "end_turn", events[1].Delta.StopReason)
assert.Equal(t, "message_stop", events[2].Type)
}
func TestStreamingEmptyResponse(t *testing.T) {
state := NewResponsesEventToAnthropicState()
ResponsesEventToAnthropicEvents(&ResponsesStreamEvent{
Type: "response.created",
Response: &ResponsesResponse{ID: "resp_6", Model: "gpt-5.2"},
}, state)
events := ResponsesEventToAnthropicEvents(&ResponsesStreamEvent{
Type: "response.completed",
Response: &ResponsesResponse{
Status: "completed",
Usage: &ResponsesUsage{InputTokens: 5, OutputTokens: 0},
},
}, state)
require.Len(t, events, 2) // message_delta + message_stop
assert.Equal(t, "message_delta", events[0].Type)
assert.Equal(t, "end_turn", events[0].Delta.StopReason)
}
func TestResponsesAnthropicEventToSSE(t *testing.T) {
evt := AnthropicStreamEvent{
Type: "message_start",
Message: &AnthropicResponse{
ID: "resp_1",
Type: "message",
Role: "assistant",
},
}
sse, err := ResponsesAnthropicEventToSSE(evt)
require.NoError(t, err)
assert.Contains(t, sse, "event: message_start\n")
assert.Contains(t, sse, "data: ")
assert.Contains(t, sse, `"resp_1"`)
}
// ---------------------------------------------------------------------------
// response.failed tests
// ---------------------------------------------------------------------------
func TestStreamingFailed(t *testing.T) {
state := NewResponsesEventToAnthropicState()
// 1. response.created
ResponsesEventToAnthropicEvents(&ResponsesStreamEvent{
Type: "response.created",
Response: &ResponsesResponse{ID: "resp_fail_1", Model: "gpt-5.2"},
}, state)
// 2. Some text output before failure
ResponsesEventToAnthropicEvents(&ResponsesStreamEvent{
Type: "response.output_text.delta",
Delta: "Partial output before failure",
}, state)
// 3. response.failed
events := ResponsesEventToAnthropicEvents(&ResponsesStreamEvent{
Type: "response.failed",
Response: &ResponsesResponse{
Status: "failed",
Error: &ResponsesError{Code: "server_error", Message: "Internal error"},
Usage: &ResponsesUsage{InputTokens: 50, OutputTokens: 10},
},
}, state)
// Should close text block + message_delta + message_stop
require.Len(t, events, 3)
assert.Equal(t, "content_block_stop", events[0].Type)
assert.Equal(t, "message_delta", events[1].Type)
assert.Equal(t, "end_turn", events[1].Delta.StopReason)
assert.Equal(t, 50, events[1].Usage.InputTokens)
assert.Equal(t, 10, events[1].Usage.OutputTokens)
assert.Equal(t, "message_stop", events[2].Type)
}
func TestStreamingFailedNoOutput(t *testing.T) {
state := NewResponsesEventToAnthropicState()
// 1. response.created
ResponsesEventToAnthropicEvents(&ResponsesStreamEvent{
Type: "response.created",
Response: &ResponsesResponse{ID: "resp_fail_2", Model: "gpt-5.2"},
}, state)
// 2. response.failed with no prior output
events := ResponsesEventToAnthropicEvents(&ResponsesStreamEvent{
Type: "response.failed",
Response: &ResponsesResponse{
Status: "failed",
Error: &ResponsesError{Code: "rate_limit_error", Message: "Too many requests"},
Usage: &ResponsesUsage{InputTokens: 20, OutputTokens: 0},
},
}, state)
// Should emit message_delta + message_stop (no block to close)
require.Len(t, events, 2)
assert.Equal(t, "message_delta", events[0].Type)
assert.Equal(t, "end_turn", events[0].Delta.StopReason)
assert.Equal(t, "message_stop", events[1].Type)
}
func TestResponsesToAnthropic_Failed(t *testing.T) {
resp := &ResponsesResponse{
ID: "resp_fail_3",
Model: "gpt-5.2",
Status: "failed",
Error: &ResponsesError{Code: "server_error", Message: "Something went wrong"},
Output: []ResponsesOutput{},
Usage: &ResponsesUsage{InputTokens: 30, OutputTokens: 0},
}
anth := ResponsesToAnthropic(resp, "claude-opus-4-6")
// Failed status defaults to "end_turn" stop reason
assert.Equal(t, "end_turn", anth.StopReason)
// Should have at least an empty text block
require.Len(t, anth.Content, 1)
assert.Equal(t, "text", anth.Content[0].Type)
}
// ---------------------------------------------------------------------------
// thinking → reasoning conversion tests
// ---------------------------------------------------------------------------
func TestAnthropicToResponses_ThinkingEnabled(t *testing.T) {
req := &AnthropicRequest{
Model: "gpt-5.2",
MaxTokens: 1024,
Messages: []AnthropicMessage{{Role: "user", Content: json.RawMessage(`"Hello"`)}},
Thinking: &AnthropicThinking{Type: "enabled", BudgetTokens: 10000},
}
resp, err := AnthropicToResponses(req)
require.NoError(t, err)
require.NotNil(t, resp.Reasoning)
assert.Equal(t, "high", resp.Reasoning.Effort)
assert.Equal(t, "auto", resp.Reasoning.Summary)
assert.Contains(t, resp.Include, "reasoning.encrypted_content")
assert.NotContains(t, resp.Include, "reasoning.summary")
}
func TestAnthropicToResponses_ThinkingAdaptive(t *testing.T) {
req := &AnthropicRequest{
Model: "gpt-5.2",
MaxTokens: 1024,
Messages: []AnthropicMessage{{Role: "user", Content: json.RawMessage(`"Hello"`)}},
Thinking: &AnthropicThinking{Type: "adaptive", BudgetTokens: 5000},
}
resp, err := AnthropicToResponses(req)
require.NoError(t, err)
require.NotNil(t, resp.Reasoning)
assert.Equal(t, "medium", resp.Reasoning.Effort)
assert.Equal(t, "auto", resp.Reasoning.Summary)
assert.NotContains(t, resp.Include, "reasoning.summary")
}
func TestAnthropicToResponses_ThinkingDisabled(t *testing.T) {
req := &AnthropicRequest{
Model: "gpt-5.2",
MaxTokens: 1024,
Messages: []AnthropicMessage{{Role: "user", Content: json.RawMessage(`"Hello"`)}},
Thinking: &AnthropicThinking{Type: "disabled"},
}
resp, err := AnthropicToResponses(req)
require.NoError(t, err)
assert.Nil(t, resp.Reasoning)
assert.NotContains(t, resp.Include, "reasoning.summary")
}
func TestAnthropicToResponses_NoThinking(t *testing.T) {
req := &AnthropicRequest{
Model: "gpt-5.2",
MaxTokens: 1024,
Messages: []AnthropicMessage{{Role: "user", Content: json.RawMessage(`"Hello"`)}},
}
resp, err := AnthropicToResponses(req)
require.NoError(t, err)
assert.Nil(t, resp.Reasoning)
}
// ---------------------------------------------------------------------------
// tool_choice conversion tests
// ---------------------------------------------------------------------------
func TestAnthropicToResponses_ToolChoiceAuto(t *testing.T) {
req := &AnthropicRequest{
Model: "gpt-5.2",
MaxTokens: 1024,
Messages: []AnthropicMessage{{Role: "user", Content: json.RawMessage(`"Hello"`)}},
ToolChoice: json.RawMessage(`{"type":"auto"}`),
}
resp, err := AnthropicToResponses(req)
require.NoError(t, err)
var tc string
require.NoError(t, json.Unmarshal(resp.ToolChoice, &tc))
assert.Equal(t, "auto", tc)
}
func TestAnthropicToResponses_ToolChoiceAny(t *testing.T) {
req := &AnthropicRequest{
Model: "gpt-5.2",
MaxTokens: 1024,
Messages: []AnthropicMessage{{Role: "user", Content: json.RawMessage(`"Hello"`)}},
ToolChoice: json.RawMessage(`{"type":"any"}`),
}
resp, err := AnthropicToResponses(req)
require.NoError(t, err)
var tc string
require.NoError(t, json.Unmarshal(resp.ToolChoice, &tc))
assert.Equal(t, "required", tc)
}
func TestAnthropicToResponses_ToolChoiceSpecific(t *testing.T) {
req := &AnthropicRequest{
Model: "gpt-5.2",
MaxTokens: 1024,
Messages: []AnthropicMessage{{Role: "user", Content: json.RawMessage(`"Hello"`)}},
ToolChoice: json.RawMessage(`{"type":"tool","name":"get_weather"}`),
}
resp, err := AnthropicToResponses(req)
require.NoError(t, err)
var tc map[string]any
require.NoError(t, json.Unmarshal(resp.ToolChoice, &tc))
assert.Equal(t, "function", tc["type"])
fn, ok := tc["function"].(map[string]any)
require.True(t, ok)
assert.Equal(t, "get_weather", fn["name"])
}
backend/internal/pkg/apicompat/anthropic_to_responses.go
+346
View File
@@ -0,0 +1,346 @@
package apicompat
import (
"encoding/json"
"fmt"
"strings"
)
// AnthropicToResponses converts an Anthropic Messages request directly into
// a Responses API request. This preserves fields that would be lost in a
// Chat Completions intermediary round-trip (e.g. thinking, cache_control,
// structured system prompts).
func AnthropicToResponses(req *AnthropicRequest) (*ResponsesRequest, error) {
input, err := convertAnthropicToResponsesInput(req.System, req.Messages)
if err != nil {
return nil, err
}
inputJSON, err := json.Marshal(input)
if err != nil {
return nil, err
}
out := &ResponsesRequest{
Model: req.Model,
Input: inputJSON,
Temperature: req.Temperature,
TopP: req.TopP,
Stream: req.Stream,
Include: []string{"reasoning.encrypted_content"},
}
storeFalse := false
out.Store = &storeFalse
if req.MaxTokens > 0 {
v := req.MaxTokens
if v < minMaxOutputTokens {
v = minMaxOutputTokens
}
out.MaxOutputTokens = &v
}
if len(req.Tools) > 0 {
out.Tools = convertAnthropicToolsToResponses(req.Tools)
}
// Convert thinking → reasoning.
// generate_summary="auto" causes the upstream to emit reasoning_summary_text
// streaming events; the include array only needs reasoning.encrypted_content
// (already set above) for content continuity.
if req.Thinking != nil {
switch req.Thinking.Type {
case "enabled":
out.Reasoning = &ResponsesReasoning{Effort: "high", Summary: "auto"}
case "adaptive":
out.Reasoning = &ResponsesReasoning{Effort: "medium", Summary: "auto"}
}
// "disabled" or unknown → omit reasoning
}
// Convert tool_choice
if len(req.ToolChoice) > 0 {
tc, err := convertAnthropicToolChoiceToResponses(req.ToolChoice)
if err != nil {
return nil, fmt.Errorf("convert tool_choice: %w", err)
}
out.ToolChoice = tc
}
return out, nil
}
// convertAnthropicToolChoiceToResponses maps Anthropic tool_choice to Responses format.
//
// {"type":"auto"} → "auto"
// {"type":"any"} → "required"
// {"type":"none"} → "none"
// {"type":"tool","name":"X"} → {"type":"function","function":{"name":"X"}}
func convertAnthropicToolChoiceToResponses(raw json.RawMessage) (json.RawMessage, error) {
var tc struct {
Type string `json:"type"`
Name string `json:"name"`
}
if err := json.Unmarshal(raw, &tc); err != nil {
return nil, err
}
switch tc.Type {
case "auto":
return json.Marshal("auto")
case "any":
return json.Marshal("required")
case "none":
return json.Marshal("none")
case "tool":
return json.Marshal(map[string]any{
"type": "function",
"function": map[string]string{"name": tc.Name},
})
default:
// Pass through unknown types as-is
return raw, nil
}
}
// convertAnthropicToResponsesInput builds the Responses API input items array
// from the Anthropic system field and message list.
func convertAnthropicToResponsesInput(system json.RawMessage, msgs []AnthropicMessage) ([]ResponsesInputItem, error) {
var out []ResponsesInputItem
// System prompt → system role input item.
if len(system) > 0 {
sysText, err := parseAnthropicSystemPrompt(system)
if err != nil {
return nil, err
}
if sysText != "" {
content, _ := json.Marshal(sysText)
out = append(out, ResponsesInputItem{
Role: "system",
Content: content,
})
}
}
for _, m := range msgs {
items, err := anthropicMsgToResponsesItems(m)
if err != nil {
return nil, err
}
out = append(out, items...)
}
return out, nil
}
// parseAnthropicSystemPrompt handles the Anthropic system field which can be
// a plain string or an array of text blocks.
func parseAnthropicSystemPrompt(raw json.RawMessage) (string, error) {
var s string
if err := json.Unmarshal(raw, &s); err == nil {
return s, nil
}
var blocks []AnthropicContentBlock
if err := json.Unmarshal(raw, &blocks); err != nil {
return "", err
}
var parts []string
for _, b := range blocks {
if b.Type == "text" && b.Text != "" {
parts = append(parts, b.Text)
}
}
return strings.Join(parts, "\n\n"), nil
}
// anthropicMsgToResponsesItems converts a single Anthropic message into one
// or more Responses API input items.
func anthropicMsgToResponsesItems(m AnthropicMessage) ([]ResponsesInputItem, error) {
switch m.Role {
case "user":
return anthropicUserToResponses(m.Content)
case "assistant":
return anthropicAssistantToResponses(m.Content)
default:
return anthropicUserToResponses(m.Content)
}
}
// anthropicUserToResponses handles an Anthropic user message. Content can be a
// plain string or an array of blocks. tool_result blocks are extracted into
// function_call_output items.
func anthropicUserToResponses(raw json.RawMessage) ([]ResponsesInputItem, error) {
// Try plain string.
var s string
if err := json.Unmarshal(raw, &s); err == nil {
content, _ := json.Marshal(s)
return []ResponsesInputItem{{Role: "user", Content: content}}, nil
}
var blocks []AnthropicContentBlock
if err := json.Unmarshal(raw, &blocks); err != nil {
return nil, err
}
var out []ResponsesInputItem
// Extract tool_result blocks → function_call_output items.
for _, b := range blocks {
if b.Type != "tool_result" {
continue
}
text := extractAnthropicToolResultText(b)
if text == "" {
// OpenAI Responses API requires "output" field; use placeholder for empty results.
text = "(empty)"
}
out = append(out, ResponsesInputItem{
Type: "function_call_output",
CallID: toResponsesCallID(b.ToolUseID),
Output: text,
})
}
// Remaining text blocks → user message.
text := extractAnthropicTextFromBlocks(blocks)
if text != "" {
content, _ := json.Marshal(text)
out = append(out, ResponsesInputItem{Role: "user", Content: content})
}
return out, nil
}
// anthropicAssistantToResponses handles an Anthropic assistant message.
// Text content → assistant message with output_text parts.
// tool_use blocks → function_call items.
// thinking blocks → ignored (OpenAI doesn't accept them as input).
func anthropicAssistantToResponses(raw json.RawMessage) ([]ResponsesInputItem, error) {
// Try plain string.
var s string
if err := json.Unmarshal(raw, &s); err == nil {
parts := []ResponsesContentPart{{Type: "output_text", Text: s}}
partsJSON, err := json.Marshal(parts)
if err != nil {
return nil, err
}
return []ResponsesInputItem{{Role: "assistant", Content: partsJSON}}, nil
}
var blocks []AnthropicContentBlock
if err := json.Unmarshal(raw, &blocks); err != nil {
return nil, err
}
var items []ResponsesInputItem
// Text content → assistant message with output_text content parts.
text := extractAnthropicTextFromBlocks(blocks)
if text != "" {
parts := []ResponsesContentPart{{Type: "output_text", Text: text}}
partsJSON, err := json.Marshal(parts)
if err != nil {
return nil, err
}
items = append(items, ResponsesInputItem{Role: "assistant", Content: partsJSON})
}
// tool_use → function_call items.
for _, b := range blocks {
if b.Type != "tool_use" {
continue
}
args := "{}"
if len(b.Input) > 0 {
args = string(b.Input)
}
fcID := toResponsesCallID(b.ID)
items = append(items, ResponsesInputItem{
Type: "function_call",
CallID: fcID,
Name: b.Name,
Arguments: args,
ID: fcID,
})
}
return items, nil
}
// toResponsesCallID converts an Anthropic tool ID (toolu_xxx / call_xxx) to a
// Responses API function_call ID that starts with "fc_".
func toResponsesCallID(id string) string {
if strings.HasPrefix(id, "fc_") {
return id
}
return "fc_" + id
}
// fromResponsesCallID reverses toResponsesCallID, stripping the "fc_" prefix
// that was added during request conversion.
func fromResponsesCallID(id string) string {
if after, ok := strings.CutPrefix(id, "fc_"); ok {
// Only strip if the remainder doesn't look like it was already "fc_" prefixed.
// E.g. "fc_toolu_xxx" → "toolu_xxx", "fc_call_xxx" → "call_xxx"
if strings.HasPrefix(after, "toolu_") || strings.HasPrefix(after, "call_") {
return after
}
}
return id
}
// extractAnthropicToolResultText gets the text content from a tool_result block.
func extractAnthropicToolResultText(b AnthropicContentBlock) string {
if len(b.Content) == 0 {
return ""
}
var s string
if err := json.Unmarshal(b.Content, &s); err == nil {
return s
}
var inner []AnthropicContentBlock
if err := json.Unmarshal(b.Content, &inner); err == nil {
var parts []string
for _, ib := range inner {
if ib.Type == "text" && ib.Text != "" {
parts = append(parts, ib.Text)
}
}
return strings.Join(parts, "\n\n")
}
return ""
}
// extractAnthropicTextFromBlocks joins all text blocks, ignoring thinking/
// tool_use/tool_result blocks.
func extractAnthropicTextFromBlocks(blocks []AnthropicContentBlock) string {
var parts []string
for _, b := range blocks {
if b.Type == "text" && b.Text != "" {
parts = append(parts, b.Text)
}
}
return strings.Join(parts, "\n\n")
}
// convertAnthropicToolsToResponses maps Anthropic tool definitions to
// Responses API tools. Server-side tools like web_search are mapped to their
// OpenAI equivalents; regular tools become function tools.
func convertAnthropicToolsToResponses(tools []AnthropicTool) []ResponsesTool {
var out []ResponsesTool
for _, t := range tools {
// Anthropic server tools like "web_search_20250305" → OpenAI {"type":"web_search"}
if strings.HasPrefix(t.Type, "web_search") {
out = append(out, ResponsesTool{Type: "web_search"})
continue
}
out = append(out, ResponsesTool{
Type: "function",
Name: t.Name,
Description: t.Description,
Parameters: t.InputSchema,
})
}
return out
}
backend/internal/pkg/apicompat/responses_to_anthropic.go
+516
View File
@@ -0,0 +1,516 @@
package apicompat
import (
"encoding/json"
"fmt"
"time"
)
// ---------------------------------------------------------------------------
// Non-streaming: ResponsesResponse → AnthropicResponse
// ---------------------------------------------------------------------------
// ResponsesToAnthropic converts a Responses API response directly into an
// Anthropic Messages response. Reasoning output items are mapped to thinking
// blocks; function_call items become tool_use blocks.
func ResponsesToAnthropic(resp *ResponsesResponse, model string) *AnthropicResponse {
out := &AnthropicResponse{
ID: resp.ID,
Type: "message",
Role: "assistant",
Model: model,
}
var blocks []AnthropicContentBlock
for _, item := range resp.Output {
switch item.Type {
case "reasoning":
summaryText := ""
for _, s := range item.Summary {
if s.Type == "summary_text" && s.Text != "" {
summaryText += s.Text
}
}
if summaryText != "" {
blocks = append(blocks, AnthropicContentBlock{
Type: "thinking",
Thinking: summaryText,
})
}
case "message":
for _, part := range item.Content {
if part.Type == "output_text" && part.Text != "" {
blocks = append(blocks, AnthropicContentBlock{
Type: "text",
Text: part.Text,
})
}
}
case "function_call":
blocks = append(blocks, AnthropicContentBlock{
Type: "tool_use",
ID: fromResponsesCallID(item.CallID),
Name: item.Name,
Input: json.RawMessage(item.Arguments),
})
case "web_search_call":
toolUseID := "srvtoolu_" + item.ID
query := ""
if item.Action != nil {
query = item.Action.Query
}
inputJSON, _ := json.Marshal(map[string]string{"query": query})
blocks = append(blocks, AnthropicContentBlock{
Type: "server_tool_use",
ID: toolUseID,
Name: "web_search",
Input: inputJSON,
})
emptyResults, _ := json.Marshal([]struct{}{})
blocks = append(blocks, AnthropicContentBlock{
Type: "web_search_tool_result",
ToolUseID: toolUseID,
Content: emptyResults,
})
}
}
if len(blocks) == 0 {
blocks = append(blocks, AnthropicContentBlock{Type: "text", Text: ""})
}
out.Content = blocks
out.StopReason = responsesStatusToAnthropicStopReason(resp.Status, resp.IncompleteDetails, blocks)
if resp.Usage != nil {
out.Usage = AnthropicUsage{
InputTokens: resp.Usage.InputTokens,
OutputTokens: resp.Usage.OutputTokens,
}
if resp.Usage.InputTokensDetails != nil {
out.Usage.CacheReadInputTokens = resp.Usage.InputTokensDetails.CachedTokens
}
}
return out
}
func responsesStatusToAnthropicStopReason(status string, details *ResponsesIncompleteDetails, blocks []AnthropicContentBlock) string {
switch status {
case "incomplete":
if details != nil && details.Reason == "max_output_tokens" {
return "max_tokens"
}
return "end_turn"
case "completed":
if len(blocks) > 0 && blocks[len(blocks)-1].Type == "tool_use" {
return "tool_use"
}
return "end_turn"
default:
return "end_turn"
}
}
// ---------------------------------------------------------------------------
// Streaming: ResponsesStreamEvent → []AnthropicStreamEvent (stateful converter)
// ---------------------------------------------------------------------------
// ResponsesEventToAnthropicState tracks state for converting a sequence of
// Responses SSE events directly into Anthropic SSE events.
type ResponsesEventToAnthropicState struct {
MessageStartSent bool
MessageStopSent bool
ContentBlockIndex int
ContentBlockOpen bool
CurrentBlockType string // "text" | "thinking" | "tool_use"
// OutputIndexToBlockIdx maps Responses output_index → Anthropic content block index.
OutputIndexToBlockIdx map[int]int
InputTokens int
OutputTokens int
CacheReadInputTokens int
ResponseID string
Model string
Created int64
}
// NewResponsesEventToAnthropicState returns an initialised stream state.
func NewResponsesEventToAnthropicState() *ResponsesEventToAnthropicState {
return &ResponsesEventToAnthropicState{
OutputIndexToBlockIdx: make(map[int]int),
Created: time.Now().Unix(),
}
}
// ResponsesEventToAnthropicEvents converts a single Responses SSE event into
// zero or more Anthropic SSE events, updating state as it goes.
func ResponsesEventToAnthropicEvents(
evt *ResponsesStreamEvent,
state *ResponsesEventToAnthropicState,
) []AnthropicStreamEvent {
switch evt.Type {
case "response.created":
return resToAnthHandleCreated(evt, state)
case "response.output_item.added":
return resToAnthHandleOutputItemAdded(evt, state)
case "response.output_text.delta":
return resToAnthHandleTextDelta(evt, state)
case "response.output_text.done":
return resToAnthHandleBlockDone(state)
case "response.function_call_arguments.delta":
return resToAnthHandleFuncArgsDelta(evt, state)
case "response.function_call_arguments.done":
return resToAnthHandleBlockDone(state)
case "response.output_item.done":
return resToAnthHandleOutputItemDone(evt, state)
case "response.reasoning_summary_text.delta":
return resToAnthHandleReasoningDelta(evt, state)
case "response.reasoning_summary_text.done":
return resToAnthHandleBlockDone(state)
case "response.completed", "response.incomplete", "response.failed":
return resToAnthHandleCompleted(evt, state)
default:
return nil
}
}
// FinalizeResponsesAnthropicStream emits synthetic termination events if the
// stream ended without a proper completion event.
func FinalizeResponsesAnthropicStream(state *ResponsesEventToAnthropicState) []AnthropicStreamEvent {
if !state.MessageStartSent || state.MessageStopSent {
return nil
}
var events []AnthropicStreamEvent
events = append(events, closeCurrentBlock(state)...)
events = append(events,
AnthropicStreamEvent{
Type: "message_delta",
Delta: &AnthropicDelta{
StopReason: "end_turn",
},
Usage: &AnthropicUsage{
InputTokens: state.InputTokens,
OutputTokens: state.OutputTokens,
CacheReadInputTokens: state.CacheReadInputTokens,
},
},
AnthropicStreamEvent{Type: "message_stop"},
)
state.MessageStopSent = true
return events
}
// ResponsesAnthropicEventToSSE formats an AnthropicStreamEvent as an SSE line pair.
func ResponsesAnthropicEventToSSE(evt AnthropicStreamEvent) (string, error) {
data, err := json.Marshal(evt)
if err != nil {
return "", err
}
return fmt.Sprintf("event: %s\ndata: %s\n\n", evt.Type, data), nil
}
// --- internal handlers ---
func resToAnthHandleCreated(evt *ResponsesStreamEvent, state *ResponsesEventToAnthropicState) []AnthropicStreamEvent {
if evt.Response != nil {
state.ResponseID = evt.Response.ID
// Only use upstream model if no override was set (e.g. originalModel)
if state.Model == "" {
state.Model = evt.Response.Model
}
}
if state.MessageStartSent {
return nil
}
state.MessageStartSent = true
return []AnthropicStreamEvent{{
Type: "message_start",
Message: &AnthropicResponse{
ID: state.ResponseID,
Type: "message",
Role: "assistant",
Content: []AnthropicContentBlock{},
Model: state.Model,
Usage: AnthropicUsage{
InputTokens: 0,
OutputTokens: 0,
},
},
}}
}
func resToAnthHandleOutputItemAdded(evt *ResponsesStreamEvent, state *ResponsesEventToAnthropicState) []AnthropicStreamEvent {
if evt.Item == nil {
return nil
}
switch evt.Item.Type {
case "function_call":
var events []AnthropicStreamEvent
events = append(events, closeCurrentBlock(state)...)
idx := state.ContentBlockIndex
state.OutputIndexToBlockIdx[evt.OutputIndex] = idx
state.ContentBlockOpen = true
state.CurrentBlockType = "tool_use"
events = append(events, AnthropicStreamEvent{
Type: "content_block_start",
Index: &idx,
ContentBlock: &AnthropicContentBlock{
Type: "tool_use",
ID: fromResponsesCallID(evt.Item.CallID),
Name: evt.Item.Name,
Input: json.RawMessage("{}"),
},
})
return events
case "reasoning":
var events []AnthropicStreamEvent
events = append(events, closeCurrentBlock(state)...)
idx := state.ContentBlockIndex
state.OutputIndexToBlockIdx[evt.OutputIndex] = idx
state.ContentBlockOpen = true
state.CurrentBlockType = "thinking"
events = append(events, AnthropicStreamEvent{
Type: "content_block_start",
Index: &idx,
ContentBlock: &AnthropicContentBlock{
Type: "thinking",
Thinking: "",
},
})
return events
case "message":
return nil
}
return nil
}
func resToAnthHandleTextDelta(evt *ResponsesStreamEvent, state *ResponsesEventToAnthropicState) []AnthropicStreamEvent {
if evt.Delta == "" {
return nil
}
var events []AnthropicStreamEvent
if !state.ContentBlockOpen || state.CurrentBlockType != "text" {
events = append(events, closeCurrentBlock(state)...)
idx := state.ContentBlockIndex
state.ContentBlockOpen = true
state.CurrentBlockType = "text"
events = append(events, AnthropicStreamEvent{
Type: "content_block_start",
Index: &idx,
ContentBlock: &AnthropicContentBlock{
Type: "text",
Text: "",
},
})
}
idx := state.ContentBlockIndex
events = append(events, AnthropicStreamEvent{
Type: "content_block_delta",
Index: &idx,
Delta: &AnthropicDelta{
Type: "text_delta",
Text: evt.Delta,
},
})
return events
}
func resToAnthHandleFuncArgsDelta(evt *ResponsesStreamEvent, state *ResponsesEventToAnthropicState) []AnthropicStreamEvent {
if evt.Delta == "" {
return nil
}
blockIdx, ok := state.OutputIndexToBlockIdx[evt.OutputIndex]
if !ok {
return nil
}
return []AnthropicStreamEvent{{
Type: "content_block_delta",
Index: &blockIdx,
Delta: &AnthropicDelta{
Type: "input_json_delta",
PartialJSON: evt.Delta,
},
}}
}
func resToAnthHandleReasoningDelta(evt *ResponsesStreamEvent, state *ResponsesEventToAnthropicState) []AnthropicStreamEvent {
if evt.Delta == "" {
return nil
}
blockIdx, ok := state.OutputIndexToBlockIdx[evt.OutputIndex]
if !ok {
return nil
}
return []AnthropicStreamEvent{{
Type: "content_block_delta",
Index: &blockIdx,
Delta: &AnthropicDelta{
Type: "thinking_delta",
Thinking: evt.Delta,
},
}}
}
func resToAnthHandleBlockDone(state *ResponsesEventToAnthropicState) []AnthropicStreamEvent {
if !state.ContentBlockOpen {
return nil
}
return closeCurrentBlock(state)
}
func resToAnthHandleOutputItemDone(evt *ResponsesStreamEvent, state *ResponsesEventToAnthropicState) []AnthropicStreamEvent {
if evt.Item == nil {
return nil
}
// Handle web_search_call → synthesize server_tool_use + web_search_tool_result blocks.
if evt.Item.Type == "web_search_call" && evt.Item.Status == "completed" {
return resToAnthHandleWebSearchDone(evt, state)
}
if state.ContentBlockOpen {
return closeCurrentBlock(state)
}
return nil
}
// resToAnthHandleWebSearchDone converts an OpenAI web_search_call output item
// into Anthropic server_tool_use + web_search_tool_result content block pairs.
// This allows Claude Code to count the searches performed.
func resToAnthHandleWebSearchDone(evt *ResponsesStreamEvent, state *ResponsesEventToAnthropicState) []AnthropicStreamEvent {
var events []AnthropicStreamEvent
events = append(events, closeCurrentBlock(state)...)
toolUseID := "srvtoolu_" + evt.Item.ID
query := ""
if evt.Item.Action != nil {
query = evt.Item.Action.Query
}
inputJSON, _ := json.Marshal(map[string]string{"query": query})
// Emit server_tool_use block (start + stop).
idx1 := state.ContentBlockIndex
events = append(events, AnthropicStreamEvent{
Type: "content_block_start",
Index: &idx1,
ContentBlock: &AnthropicContentBlock{
Type: "server_tool_use",
ID: toolUseID,
Name: "web_search",
Input: inputJSON,
},
})
events = append(events, AnthropicStreamEvent{
Type: "content_block_stop",
Index: &idx1,
})
state.ContentBlockIndex++
// Emit web_search_tool_result block (start + stop).
// Content is empty because OpenAI does not expose individual search results;
// the model consumes them internally and produces text output.
emptyResults, _ := json.Marshal([]struct{}{})
idx2 := state.ContentBlockIndex
events = append(events, AnthropicStreamEvent{
Type: "content_block_start",
Index: &idx2,
ContentBlock: &AnthropicContentBlock{
Type: "web_search_tool_result",
ToolUseID: toolUseID,
Content: emptyResults,
},
})
events = append(events, AnthropicStreamEvent{
Type: "content_block_stop",
Index: &idx2,
})
state.ContentBlockIndex++
return events
}
func resToAnthHandleCompleted(evt *ResponsesStreamEvent, state *ResponsesEventToAnthropicState) []AnthropicStreamEvent {
if state.MessageStopSent {
return nil
}
var events []AnthropicStreamEvent
events = append(events, closeCurrentBlock(state)...)
stopReason := "end_turn"
if evt.Response != nil {
if evt.Response.Usage != nil {
state.InputTokens = evt.Response.Usage.InputTokens
state.OutputTokens = evt.Response.Usage.OutputTokens
if evt.Response.Usage.InputTokensDetails != nil {
state.CacheReadInputTokens = evt.Response.Usage.InputTokensDetails.CachedTokens
}
}
switch evt.Response.Status {
case "incomplete":
if evt.Response.IncompleteDetails != nil && evt.Response.IncompleteDetails.Reason == "max_output_tokens" {
stopReason = "max_tokens"
}
case "completed":
if state.ContentBlockIndex > 0 && state.CurrentBlockType == "tool_use" {
stopReason = "tool_use"
}
}
}
events = append(events,
AnthropicStreamEvent{
Type: "message_delta",
Delta: &AnthropicDelta{
StopReason: stopReason,
},
Usage: &AnthropicUsage{
InputTokens: state.InputTokens,
OutputTokens: state.OutputTokens,
CacheReadInputTokens: state.CacheReadInputTokens,
},
},
AnthropicStreamEvent{Type: "message_stop"},
)
state.MessageStopSent = true
return events
}
func closeCurrentBlock(state *ResponsesEventToAnthropicState) []AnthropicStreamEvent {
if !state.ContentBlockOpen {
return nil
}
idx := state.ContentBlockIndex
state.ContentBlockOpen = false
state.ContentBlockIndex++
return []AnthropicStreamEvent{{
Type: "content_block_stop",
Index: &idx,
}}
}
backend/internal/pkg/apicompat/types.go
+320
View File
@@ -0,0 +1,320 @@
// Package apicompat provides type definitions and conversion utilities for
// translating between Anthropic Messages and OpenAI Responses API formats.
// It enables multi-protocol support so that clients using different API
// formats can be served through a unified gateway.
package apicompat
import "encoding/json"
// ---------------------------------------------------------------------------
// Anthropic Messages API types
// ---------------------------------------------------------------------------
// AnthropicRequest is the request body for POST /v1/messages.
type AnthropicRequest struct {
Model string `json:"model"`
MaxTokens int `json:"max_tokens"`
System json.RawMessage `json:"system,omitempty"` // string or []AnthropicContentBlock
Messages []AnthropicMessage `json:"messages"`
Tools []AnthropicTool `json:"tools,omitempty"`
Stream bool `json:"stream,omitempty"`
Temperature *float64 `json:"temperature,omitempty"`
TopP *float64 `json:"top_p,omitempty"`
StopSeqs []string `json:"stop_sequences,omitempty"`
Thinking *AnthropicThinking `json:"thinking,omitempty"`
ToolChoice json.RawMessage `json:"tool_choice,omitempty"`
}
// AnthropicThinking configures extended thinking in the Anthropic API.
type AnthropicThinking struct {
Type string `json:"type"` // "enabled" | "adaptive" | "disabled"
BudgetTokens int `json:"budget_tokens,omitempty"` // max thinking tokens
}
// AnthropicMessage is a single message in the Anthropic conversation.
type AnthropicMessage struct {
Role string `json:"role"` // "user" | "assistant"
Content json.RawMessage `json:"content"`
}
// AnthropicContentBlock is one block inside a message's content array.
type AnthropicContentBlock struct {
Type string `json:"type"`
// type=text
Text string `json:"text,omitempty"`
// type=thinking
Thinking string `json:"thinking,omitempty"`
// type=tool_use
ID string `json:"id,omitempty"`
Name string `json:"name,omitempty"`
Input json.RawMessage `json:"input,omitempty"`
// type=tool_result
ToolUseID string `json:"tool_use_id,omitempty"`
Content json.RawMessage `json:"content,omitempty"` // string or []AnthropicContentBlock
IsError bool `json:"is_error,omitempty"`
}
// AnthropicTool describes a tool available to the model.
type AnthropicTool struct {
Type string `json:"type,omitempty"` // e.g. "web_search_20250305" for server tools
Name string `json:"name"`
Description string `json:"description,omitempty"`
InputSchema json.RawMessage `json:"input_schema"` // JSON Schema object
}
// AnthropicResponse is the non-streaming response from POST /v1/messages.
type AnthropicResponse struct {
ID string `json:"id"`
Type string `json:"type"` // "message"
Role string `json:"role"` // "assistant"
Content []AnthropicContentBlock `json:"content"`
Model string `json:"model"`
StopReason string `json:"stop_reason"`
StopSequence *string `json:"stop_sequence,omitempty"`
Usage AnthropicUsage `json:"usage"`
}
// AnthropicUsage holds token counts in Anthropic format.
type AnthropicUsage struct {
InputTokens int `json:"input_tokens"`
OutputTokens int `json:"output_tokens"`
CacheCreationInputTokens int `json:"cache_creation_input_tokens"`
CacheReadInputTokens int `json:"cache_read_input_tokens"`
}
// ---------------------------------------------------------------------------
// Anthropic SSE event types
// ---------------------------------------------------------------------------
// AnthropicStreamEvent is a single SSE event in the Anthropic streaming protocol.
type AnthropicStreamEvent struct {
Type string `json:"type"`
// message_start
Message *AnthropicResponse `json:"message,omitempty"`
// content_block_start
Index *int `json:"index,omitempty"`
ContentBlock *AnthropicContentBlock `json:"content_block,omitempty"`
// content_block_delta
Delta *AnthropicDelta `json:"delta,omitempty"`
// message_delta
Usage *AnthropicUsage `json:"usage,omitempty"`
}
// AnthropicDelta carries incremental content in streaming events.
type AnthropicDelta struct {
Type string `json:"type,omitempty"` // "text_delta" | "input_json_delta" | "thinking_delta" | "signature_delta"
// text_delta
Text string `json:"text,omitempty"`
// input_json_delta
PartialJSON string `json:"partial_json,omitempty"`
// thinking_delta
Thinking string `json:"thinking,omitempty"`
// signature_delta
Signature string `json:"signature,omitempty"`
// message_delta fields
StopReason string `json:"stop_reason,omitempty"`
StopSequence *string `json:"stop_sequence,omitempty"`
}
// ---------------------------------------------------------------------------
// OpenAI Responses API types
// ---------------------------------------------------------------------------
// ResponsesRequest is the request body for POST /v1/responses.
type ResponsesRequest struct {
Model string `json:"model"`
Input json.RawMessage `json:"input"` // string or []ResponsesInputItem
MaxOutputTokens *int `json:"max_output_tokens,omitempty"`
Temperature *float64 `json:"temperature,omitempty"`
TopP *float64 `json:"top_p,omitempty"`
Stream bool `json:"stream,omitempty"`
Tools []ResponsesTool `json:"tools,omitempty"`
Include []string `json:"include,omitempty"`
Store *bool `json:"store,omitempty"`
Reasoning *ResponsesReasoning `json:"reasoning,omitempty"`
ToolChoice json.RawMessage `json:"tool_choice,omitempty"`
}
// ResponsesReasoning configures reasoning effort in the Responses API.
type ResponsesReasoning struct {
Effort string `json:"effort"` // "low" | "medium" | "high"
Summary string `json:"summary,omitempty"` // "auto" | "concise" | "detailed"
}
// ResponsesInputItem is one item in the Responses API input array.
// The Type field determines which other fields are populated.
type ResponsesInputItem struct {
// Common
Type string `json:"type,omitempty"` // "" for role-based messages
// Role-based messages (system/user/assistant)
Role string `json:"role,omitempty"`
Content json.RawMessage `json:"content,omitempty"` // string or []ResponsesContentPart
// type=function_call
CallID string `json:"call_id,omitempty"`
Name string `json:"name,omitempty"`
Arguments string `json:"arguments,omitempty"`
ID string `json:"id,omitempty"`
// type=function_call_output
Output string `json:"output,omitempty"`
}
// ResponsesContentPart is a typed content part in a Responses message.
type ResponsesContentPart struct {
Type string `json:"type"` // "input_text" | "output_text" | "input_image"
Text string `json:"text,omitempty"`
}
// ResponsesTool describes a tool in the Responses API.
type ResponsesTool struct {
Type string `json:"type"` // "function" | "web_search" | "local_shell" etc.
Name string `json:"name,omitempty"`
Description string `json:"description,omitempty"`
Parameters json.RawMessage `json:"parameters,omitempty"`
Strict *bool `json:"strict,omitempty"`
}
// ResponsesResponse is the non-streaming response from POST /v1/responses.
type ResponsesResponse struct {
ID string `json:"id"`
Object string `json:"object"` // "response"
Model string `json:"model"`
Status string `json:"status"` // "completed" | "incomplete" | "failed"
Output []ResponsesOutput `json:"output"`
Usage *ResponsesUsage `json:"usage,omitempty"`
// incomplete_details is present when status="incomplete"
IncompleteDetails *ResponsesIncompleteDetails `json:"incomplete_details,omitempty"`
// Error is present when status="failed"
Error *ResponsesError `json:"error,omitempty"`
}
// ResponsesError describes an error in a failed response.
type ResponsesError struct {
Code string `json:"code"`
Message string `json:"message"`
}
// ResponsesIncompleteDetails explains why a response is incomplete.
type ResponsesIncompleteDetails struct {
Reason string `json:"reason"` // "max_output_tokens" | "content_filter"
}
// ResponsesOutput is one output item in a Responses API response.
type ResponsesOutput struct {
Type string `json:"type"` // "message" | "reasoning" | "function_call" | "web_search_call"
// type=message
ID string `json:"id,omitempty"`
Role string `json:"role,omitempty"`
Content []ResponsesContentPart `json:"content,omitempty"`
Status string `json:"status,omitempty"`
// type=reasoning
EncryptedContent string `json:"encrypted_content,omitempty"`
Summary []ResponsesSummary `json:"summary,omitempty"`
// type=function_call
CallID string `json:"call_id,omitempty"`
Name string `json:"name,omitempty"`
Arguments string `json:"arguments,omitempty"`
// type=web_search_call
Action *WebSearchAction `json:"action,omitempty"`
}
// WebSearchAction describes the search action in a web_search_call output item.
type WebSearchAction struct {
Type string `json:"type,omitempty"` // "search"
Query string `json:"query,omitempty"` // primary search query
}
// ResponsesSummary is a summary text block inside a reasoning output.
type ResponsesSummary struct {
Type string `json:"type"` // "summary_text"
Text string `json:"text"`
}
// ResponsesUsage holds token counts in Responses API format.
type ResponsesUsage struct {
InputTokens int `json:"input_tokens"`
OutputTokens int `json:"output_tokens"`
TotalTokens int `json:"total_tokens"`
// Optional detailed breakdown
InputTokensDetails *ResponsesInputTokensDetails `json:"input_tokens_details,omitempty"`
OutputTokensDetails *ResponsesOutputTokensDetails `json:"output_tokens_details,omitempty"`
}
// ResponsesInputTokensDetails breaks down input token usage.
type ResponsesInputTokensDetails struct {
CachedTokens int `json:"cached_tokens,omitempty"`
}
// ResponsesOutputTokensDetails breaks down output token usage.
type ResponsesOutputTokensDetails struct {
ReasoningTokens int `json:"reasoning_tokens,omitempty"`
}
// ---------------------------------------------------------------------------
// Responses SSE event types
// ---------------------------------------------------------------------------
// ResponsesStreamEvent is a single SSE event in the Responses streaming protocol.
// The Type field corresponds to the "type" in the JSON payload.
type ResponsesStreamEvent struct {
Type string `json:"type"`
// response.created / response.completed / response.failed / response.incomplete
Response *ResponsesResponse `json:"response,omitempty"`
// response.output_item.added / response.output_item.done
Item *ResponsesOutput `json:"item,omitempty"`
// response.output_text.delta / response.output_text.done
OutputIndex int `json:"output_index,omitempty"`
ContentIndex int `json:"content_index,omitempty"`
Delta string `json:"delta,omitempty"`
Text string `json:"text,omitempty"`
ItemID string `json:"item_id,omitempty"`
// response.function_call_arguments.delta / done
CallID string `json:"call_id,omitempty"`
Name string `json:"name,omitempty"`
Arguments string `json:"arguments,omitempty"`
// response.reasoning_summary_text.delta / done
// Reuses Text/Delta fields above, SummaryIndex identifies which summary part
SummaryIndex int `json:"summary_index,omitempty"`
// error event fields
Code string `json:"code,omitempty"`
Param string `json:"param,omitempty"`
// Sequence number for ordering events
SequenceNumber int `json:"sequence_number,omitempty"`
}
// ---------------------------------------------------------------------------
// Shared constants
// ---------------------------------------------------------------------------
// minMaxOutputTokens is the floor for max_output_tokens in a Responses request.
// Very small values may cause upstream API errors, so we enforce a minimum.
const minMaxOutputTokens = 128
backend/internal/repository/api_key_repo.go
+6 -6
View File
@@ -471,12 +471,12 @@ func (r *apiKeyRepository) UpdateLastUsed(ctx context.Context, id int64, usedAt
func (r *apiKeyRepository) IncrementRateLimitUsage(ctx context.Context, id int64, cost float64) error {
_, err := r.sql.ExecContext(ctx, `
UPDATE api_keys SET
usage_5h = usage_5h + $1,
usage_1d = usage_1d + $1,
usage_7d = usage_7d + $1,
window_5h_start = COALESCE(window_5h_start, NOW()),
window_1d_start = COALESCE(window_1d_start, NOW()),
window_7d_start = COALESCE(window_7d_start, NOW()),
usage_5h = CASE WHEN window_5h_start IS NOT NULL AND window_5h_start + INTERVAL '5 hours' <= NOW() THEN $1 ELSE usage_5h + $1 END,
usage_1d = CASE WHEN window_1d_start IS NOT NULL AND window_1d_start + INTERVAL '24 hours' <= NOW() THEN $1 ELSE usage_1d + $1 END,
usage_7d = CASE WHEN window_7d_start IS NOT NULL AND window_7d_start + INTERVAL '7 days' <= NOW() THEN $1 ELSE usage_7d + $1 END,
window_5h_start = CASE WHEN window_5h_start IS NULL OR window_5h_start + INTERVAL '5 hours' <= NOW() THEN NOW() ELSE window_5h_start END,
window_1d_start = CASE WHEN window_1d_start IS NULL OR window_1d_start + INTERVAL '24 hours' <= NOW() THEN NOW() ELSE window_1d_start END,
window_7d_start = CASE WHEN window_7d_start IS NULL OR window_7d_start + INTERVAL '7 days' <= NOW() THEN NOW() ELSE window_7d_start END,
updated_at = NOW()
WHERE id = $2 AND deleted_at IS NULL`,
cost, id)
backend/internal/server/routes/gateway.go
+33 -2
View File
@@ -43,12 +43,33 @@ func RegisterGatewayRoutes(
gateway.Use(gin.HandlerFunc(apiKeyAuth))
gateway.Use(requireGroupAnthropic)
{
gateway.POST("/messages", h.Gateway.Messages)
gateway.POST("/messages/count_tokens", h.Gateway.CountTokens)
// /v1/messages: auto-route based on group platform
gateway.POST("/messages", func(c *gin.Context) {
if getGroupPlatform(c) == service.PlatformOpenAI {
h.OpenAIGateway.Messages(c)
return
}
h.Gateway.Messages(c)
})
// /v1/messages/count_tokens: OpenAI groups get 404
gateway.POST("/messages/count_tokens", func(c *gin.Context) {
if getGroupPlatform(c) == service.PlatformOpenAI {
c.JSON(http.StatusNotFound, gin.H{
"type": "error",
"error": gin.H{
"type": "not_found_error",
"message": "Token counting is not supported for this platform",
},
})
return
}
h.Gateway.CountTokens(c)
})
gateway.GET("/models", h.Gateway.Models)
gateway.GET("/usage", h.Gateway.Usage)
// OpenAI Responses API
gateway.POST("/responses", h.OpenAIGateway.Responses)
gateway.POST("/responses/*subpath", h.OpenAIGateway.Responses)
gateway.GET("/responses", h.OpenAIGateway.ResponsesWebSocket)
// 明确阻止旧协议入口：OpenAI 仅支持 Responses API，避免客户端误解为会自动路由到其它平台。
gateway.POST("/chat/completions", func(c *gin.Context) {
@@ -77,6 +98,7 @@ func RegisterGatewayRoutes(
// OpenAI Responses API（不带v1前缀的别名）
r.POST("/responses", bodyLimit, clientRequestID, opsErrorLogger, gin.HandlerFunc(apiKeyAuth), requireGroupAnthropic, h.OpenAIGateway.Responses)
r.POST("/responses/*subpath", bodyLimit, clientRequestID, opsErrorLogger, gin.HandlerFunc(apiKeyAuth), requireGroupAnthropic, h.OpenAIGateway.Responses)
r.GET("/responses", bodyLimit, clientRequestID, opsErrorLogger, gin.HandlerFunc(apiKeyAuth), requireGroupAnthropic, h.OpenAIGateway.ResponsesWebSocket)
// Antigravity 模型列表
@@ -132,3 +154,12 @@ func RegisterGatewayRoutes(
// Sora 媒体代理（签名 URL，无需 API Key）
r.GET("/sora/media-signed/*filepath", h.SoraGateway.MediaProxySigned)
}
// getGroupPlatform extracts the group platform from the API Key stored in context.
func getGroupPlatform(c *gin.Context) string {
apiKey, ok := middleware.GetAPIKeyFromContext(c)
if !ok || apiKey.Group == nil {
return ""
}
return apiKey.Group.Platform
}
backend/internal/server/routes/gateway_test.go
+51
View File
@@ -0,0 +1,51 @@
package routes
import (
"net/http"
"net/http/httptest"
"strings"
"testing"
"github.com/Wei-Shaw/sub2api/internal/config"
"github.com/Wei-Shaw/sub2api/internal/handler"
servermiddleware "github.com/Wei-Shaw/sub2api/internal/server/middleware"
"github.com/gin-gonic/gin"
"github.com/stretchr/testify/require"
)
func newGatewayRoutesTestRouter() *gin.Engine {
gin.SetMode(gin.TestMode)
router := gin.New()
RegisterGatewayRoutes(
router,
&handler.Handlers{
Gateway: &handler.GatewayHandler{},
OpenAIGateway: &handler.OpenAIGatewayHandler{},
SoraGateway: &handler.SoraGatewayHandler{},
},
servermiddleware.APIKeyAuthMiddleware(func(c *gin.Context) {
c.Next()
}),
nil,
nil,
nil,
nil,
&config.Config{},
)
return router
}
func TestGatewayRoutesOpenAIResponsesCompactPathIsRegistered(t *testing.T) {
router := newGatewayRoutesTestRouter()
for _, path := range []string{"/v1/responses/compact", "/responses/compact"} {
req := httptest.NewRequest(http.MethodPost, path, strings.NewReader(`{"model":"gpt-5"}`))
req.Header.Set("Content-Type", "application/json")
w := httptest.NewRecorder()
router.ServeHTTP(w, req)
require.NotEqual(t, http.StatusNotFound, w.Code, "path=%s should hit OpenAI responses handler", path)
}
}
backend/internal/service/account_usage_service.go
+324 -24
View File
@@ -1,17 +1,24 @@
package service
import (
"bytes"
"context"
"encoding/json"
"fmt"
"log"
"math/rand/v2"
"net/http"
"strings"
"sync"
"time"
httppool "github.com/Wei-Shaw/sub2api/internal/pkg/httpclient"
openaipkg "github.com/Wei-Shaw/sub2api/internal/pkg/openai"
"github.com/Wei-Shaw/sub2api/internal/pkg/pagination"
"github.com/Wei-Shaw/sub2api/internal/pkg/timezone"
"github.com/Wei-Shaw/sub2api/internal/pkg/usagestats"
"golang.org/x/sync/errgroup"
"golang.org/x/sync/singleflight"
)
type UsageLogRepository interface {
@@ -70,8 +77,10 @@ type accountWindowStatsBatchReader interface {
}
// apiUsageCache 缓存从 Anthropic API 获取的使用率数据（utilization, resets_at）
// 同时支持缓存错误响应（负缓存），防止 429 等错误导致的重试风暴
type apiUsageCache struct {
response *ClaudeUsageResponse
err error // 非 nil 表示缓存的错误（负缓存）
timestamp time.Time
}
@@ -88,15 +97,21 @@ type antigravityUsageCache struct {
}
const (
apiCacheTTL = 3 * time.Minute
windowStatsCacheTTL = 1 * time.Minute
apiCacheTTL = 3 * time.Minute
apiErrorCacheTTL = 1 * time.Minute // 负缓存 TTL：429 等错误缓存 1 分钟
apiQueryMaxJitter = 800 * time.Millisecond // 用量查询最大随机延迟
windowStatsCacheTTL = 1 * time.Minute
openAIProbeCacheTTL = 10 * time.Minute
openAICodexProbeVersion = "0.104.0"
)
// UsageCache 封装账户使用量相关的缓存
type UsageCache struct {
apiCache sync.Map // accountID -> *apiUsageCache
windowStatsCache sync.Map // accountID -> *windowStatsCache
antigravityCache sync.Map // accountID -> *antigravityUsageCache
apiCache sync.Map // accountID -> *apiUsageCache
windowStatsCache sync.Map // accountID -> *windowStatsCache
antigravityCache sync.Map // accountID -> *antigravityUsageCache
apiFlight singleflight.Group // 防止同一账号的并发请求击穿缓存
openAIProbeCache sync.Map // accountID -> time.Time
}
// NewUsageCache 创建 UsageCache 实例
@@ -224,6 +239,14 @@ func (s *AccountUsageService) GetUsage(ctx context.Context, accountID int64) (*U
return nil, fmt.Errorf("get account failed: %w", err)
}
if account.Platform == PlatformOpenAI && account.Type == AccountTypeOAuth {
usage, err := s.getOpenAIUsage(ctx, account)
if err == nil {
s.tryClearRecoverableAccountError(ctx, account)
}
return usage, err
}
if account.Platform == PlatformGemini {
usage, err := s.getGeminiUsage(ctx, account)
if err == nil {
@@ -245,24 +268,65 @@ func (s *AccountUsageService) GetUsage(ctx context.Context, accountID int64) (*U
if account.CanGetUsage() {
var apiResp *ClaudeUsageResponse
// 1. 检查 API 缓存（10 分钟）
// 1. 检查缓存（成功响应 3 分钟 / 错误响应 1 分钟）
if cached, ok := s.cache.apiCache.Load(accountID); ok {
if cache, ok := cached.(*apiUsageCache); ok && time.Since(cache.timestamp) < apiCacheTTL {
apiResp = cache.response
if cache, ok := cached.(*apiUsageCache); ok {
age := time.Since(cache.timestamp)
if cache.err != nil && age < apiErrorCacheTTL {
// 负缓存命中：返回缓存的错误，避免重试风暴
return nil, cache.err
}
if cache.response != nil && age < apiCacheTTL {
apiResp = cache.response
}
}
}
// 2. 如果没有缓存，从 API 获取
// 2. 如果没有有效缓存，通过 singleflight 从 API 获取（防止并发击穿）
if apiResp == nil {
apiResp, err = s.fetchOAuthUsageRaw(ctx, account)
if err != nil {
return nil, err
// 随机延迟：打散多账号并发请求，避免同一时刻大量相同 TLS 指纹请求
// 触发上游反滥用检测。延迟范围 0~800ms，仅在缓存未命中时生效。
jitter := time.Duration(rand.Int64N(int64(apiQueryMaxJitter)))
select {
case <-time.After(jitter):
case <-ctx.Done():
return nil, ctx.Err()
}
// 缓存 API 响应
s.cache.apiCache.Store(accountID, &apiUsageCache{
response: apiResp,
timestamp: time.Now(),
flightKey := fmt.Sprintf("usage:%d", accountID)
result, flightErr, _ := s.cache.apiFlight.Do(flightKey, func() (any, error) {
// 再次检查缓存（可能在等待 singleflight 期间被其他请求填充）
if cached, ok := s.cache.apiCache.Load(accountID); ok {
if cache, ok := cached.(*apiUsageCache); ok {
age := time.Since(cache.timestamp)
if cache.err != nil && age < apiErrorCacheTTL {
return nil, cache.err
}
if cache.response != nil && age < apiCacheTTL {
return cache.response, nil
}
}
}
resp, fetchErr := s.fetchOAuthUsageRaw(ctx, account)
if fetchErr != nil {
// 负缓存：缓存错误响应，防止后续请求重复触发 429
s.cache.apiCache.Store(accountID, &apiUsageCache{
err: fetchErr,
timestamp: time.Now(),
})
return nil, fetchErr
}
// 缓存成功响应
s.cache.apiCache.Store(accountID, &apiUsageCache{
response: resp,
timestamp: time.Now(),
})
return resp, nil
})
if flightErr != nil {
return nil, flightErr
}
apiResp, _ = result.(*ClaudeUsageResponse)
}
// 3. 构建 UsageInfo（每次都重新计算 RemainingSeconds）
@@ -288,6 +352,161 @@ func (s *AccountUsageService) GetUsage(ctx context.Context, accountID int64) (*U
return nil, fmt.Errorf("account type %s does not support usage query", account.Type)
}
func (s *AccountUsageService) getOpenAIUsage(ctx context.Context, account *Account) (*UsageInfo, error) {
now := time.Now()
usage := &UsageInfo{UpdatedAt: &now}
if account == nil {
return usage, nil
}
if progress := buildCodexUsageProgressFromExtra(account.Extra, "5h", now); progress != nil {
usage.FiveHour = progress
}
if progress := buildCodexUsageProgressFromExtra(account.Extra, "7d", now); progress != nil {
usage.SevenDay = progress
}
if (usage.FiveHour == nil || usage.SevenDay == nil) && s.shouldProbeOpenAICodexSnapshot(account.ID, now) {
if updates, err := s.probeOpenAICodexSnapshot(ctx, account); err == nil && len(updates) > 0 {
mergeAccountExtra(account, updates)
if usage.UpdatedAt == nil {
usage.UpdatedAt = &now
}
if progress := buildCodexUsageProgressFromExtra(account.Extra, "5h", now); progress != nil {
usage.FiveHour = progress
}
if progress := buildCodexUsageProgressFromExtra(account.Extra, "7d", now); progress != nil {
usage.SevenDay = progress
}
}
}
if s.usageLogRepo == nil {
return usage, nil
}
if stats, err := s.usageLogRepo.GetAccountWindowStats(ctx, account.ID, now.Add(-5*time.Hour)); err == nil {
windowStats := windowStatsFromAccountStats(stats)
if hasMeaningfulWindowStats(windowStats) {
if usage.FiveHour == nil {
usage.FiveHour = &UsageProgress{Utilization: 0}
}
usage.FiveHour.WindowStats = windowStats
}
}
if stats, err := s.usageLogRepo.GetAccountWindowStats(ctx, account.ID, now.Add(-7*24*time.Hour)); err == nil {
windowStats := windowStatsFromAccountStats(stats)
if hasMeaningfulWindowStats(windowStats) {
if usage.SevenDay == nil {
usage.SevenDay = &UsageProgress{Utilization: 0}
}
usage.SevenDay.WindowStats = windowStats
}
}
return usage, nil
}
func (s *AccountUsageService) shouldProbeOpenAICodexSnapshot(accountID int64, now time.Time) bool {
if s == nil || s.cache == nil || accountID <= 0 {
return true
}
if cached, ok := s.cache.openAIProbeCache.Load(accountID); ok {
if ts, ok := cached.(time.Time); ok && now.Sub(ts) < openAIProbeCacheTTL {
return false
}
}
s.cache.openAIProbeCache.Store(accountID, now)
return true
}
func (s *AccountUsageService) probeOpenAICodexSnapshot(ctx context.Context, account *Account) (map[string]any, error) {
if account == nil || !account.IsOAuth() {
return nil, nil
}
accessToken := account.GetOpenAIAccessToken()
if accessToken == "" {
return nil, fmt.Errorf("no access token available")
}
modelID := openaipkg.DefaultTestModel
payload := createOpenAITestPayload(modelID, true)
payloadBytes, err := json.Marshal(payload)
if err != nil {
return nil, fmt.Errorf("marshal openai probe payload: %w", err)
}
reqCtx, cancel := context.WithTimeout(ctx, 15*time.Second)
defer cancel()
req, err := http.NewRequestWithContext(reqCtx, http.MethodPost, chatgptCodexURL, bytes.NewReader(payloadBytes))
if err != nil {
return nil, fmt.Errorf("create openai probe request: %w", err)
}
req.Host = "chatgpt.com"
req.Header.Set("Content-Type", "application/json")
req.Header.Set("Authorization", "Bearer "+accessToken)
req.Header.Set("Accept", "text/event-stream")
req.Header.Set("OpenAI-Beta", "responses=experimental")
req.Header.Set("Originator", "codex_cli_rs")
req.Header.Set("Version", openAICodexProbeVersion)
req.Header.Set("User-Agent", codexCLIUserAgent)
if s.identityCache != nil {
if fp, fpErr := s.identityCache.GetFingerprint(reqCtx, account.ID); fpErr == nil && fp != nil && strings.TrimSpace(fp.UserAgent) != "" {
req.Header.Set("User-Agent", strings.TrimSpace(fp.UserAgent))
}
}
if chatgptAccountID := account.GetChatGPTAccountID(); chatgptAccountID != "" {
req.Header.Set("chatgpt-account-id", chatgptAccountID)
}
proxyURL := ""
if account.ProxyID != nil && account.Proxy != nil {
proxyURL = account.Proxy.URL()
}
client, err := httppool.GetClient(httppool.Options{
ProxyURL: proxyURL,
Timeout: 15 * time.Second,
ResponseHeaderTimeout: 10 * time.Second,
})
if err != nil {
return nil, fmt.Errorf("build openai probe client: %w", err)
}
resp, err := client.Do(req)
if err != nil {
return nil, fmt.Errorf("openai codex probe request failed: %w", err)
}
defer func() { _ = resp.Body.Close() }()
if resp.StatusCode < 200 || resp.StatusCode >= 300 {
return nil, fmt.Errorf("openai codex probe returned status %d", resp.StatusCode)
}
if snapshot := ParseCodexRateLimitHeaders(resp.Header); snapshot != nil {
updates := buildCodexUsageExtraUpdates(snapshot, time.Now())
if len(updates) > 0 {
go func(accountID int64, updates map[string]any) {
updateCtx, updateCancel := context.WithTimeout(context.Background(), 5*time.Second)
defer updateCancel()
_ = s.accountRepo.UpdateExtra(updateCtx, accountID, updates)
}(account.ID, updates)
return updates, nil
}
}
return nil, nil
}
func mergeAccountExtra(account *Account, updates map[string]any) {
if account == nil || len(updates) == 0 {
return
}
if account.Extra == nil {
account.Extra = make(map[string]any, len(updates))
}
for k, v := range updates {
account.Extra[k] = v
}
}
func (s *AccountUsageService) getGeminiUsage(ctx context.Context, account *Account) (*UsageInfo, error) {
now := time.Now()
usage := &UsageInfo{
@@ -519,6 +738,72 @@ func windowStatsFromAccountStats(stats *usagestats.AccountStats) *WindowStats {
}
}
func hasMeaningfulWindowStats(stats *WindowStats) bool {
if stats == nil {
return false
}
return stats.Requests > 0 || stats.Tokens > 0 || stats.Cost > 0 || stats.StandardCost > 0 || stats.UserCost > 0
}
func buildCodexUsageProgressFromExtra(extra map[string]any, window string, now time.Time) *UsageProgress {
if len(extra) == 0 {
return nil
}
var (
usedPercentKey string
resetAfterKey string
resetAtKey string
)
switch window {
case "5h":
usedPercentKey = "codex_5h_used_percent"
resetAfterKey = "codex_5h_reset_after_seconds"
resetAtKey = "codex_5h_reset_at"
case "7d":
usedPercentKey = "codex_7d_used_percent"
resetAfterKey = "codex_7d_reset_after_seconds"
resetAtKey = "codex_7d_reset_at"
default:
return nil
}
usedRaw, ok := extra[usedPercentKey]
if !ok {
return nil
}
progress := &UsageProgress{Utilization: parseExtraFloat64(usedRaw)}
if resetAtRaw, ok := extra[resetAtKey]; ok {
if resetAt, err := parseTime(fmt.Sprint(resetAtRaw)); err == nil {
progress.ResetsAt = &resetAt
progress.RemainingSeconds = int(time.Until(resetAt).Seconds())
if progress.RemainingSeconds < 0 {
progress.RemainingSeconds = 0
}
}
}
if progress.ResetsAt == nil {
if resetAfterSeconds := parseExtraInt(extra[resetAfterKey]); resetAfterSeconds > 0 {
base := now
if updatedAtRaw, ok := extra["codex_usage_updated_at"]; ok {
if updatedAt, err := parseTime(fmt.Sprint(updatedAtRaw)); err == nil {
base = updatedAt
}
}
resetAt := base.Add(time.Duration(resetAfterSeconds) * time.Second)
progress.ResetsAt = &resetAt
progress.RemainingSeconds = int(time.Until(resetAt).Seconds())
if progress.RemainingSeconds < 0 {
progress.RemainingSeconds = 0
}
}
}
return progress
}
func (s *AccountUsageService) GetAccountUsageStats(ctx context.Context, accountID int64, startTime, endTime time.Time) (*usagestats.AccountUsageStatsResponse, error) {
stats, err := s.usageLogRepo.GetAccountUsageStats(ctx, accountID, startTime, endTime)
if err != nil {
@@ -666,15 +951,30 @@ func (s *AccountUsageService) estimateSetupTokenUsage(account *Account) *UsageIn
remaining = 0
}
// 根据状态估算使用率 (百分比形式，100 = 100%)
// 优先使用响应头中存储的真实 utilization 值（0-1 小数，转为 0-100 百分比）
var utilization float64
switch account.SessionWindowStatus {
case "rejected":
utilization = 100.0
case "allowed_warning":
utilization = 80.0
default:
utilization = 0.0
var found bool
if stored, ok := account.Extra["session_window_utilization"]; ok {
switch v := stored.(type) {
case float64:
utilization = v * 100
found = true
case json.Number:
if f, err := v.Float64(); err == nil {
utilization = f * 100
found = true
}
}
}
// 如果没有存储的 utilization，回退到状态估算
if !found {
switch account.SessionWindowStatus {
case "rejected":
utilization = 100.0
case "allowed_warning":
utilization = 80.0
}
}
info.FiveHour = &UsageProgress{
backend/internal/service/antigravity_gateway_service.go
+9
View File
@@ -3704,6 +3704,15 @@ func (s *AntigravityGatewayService) handleClaudeStreamingResponse(c *gin.Context
finalEvents, agUsage := processor.Finish()
if len(finalEvents) > 0 {
cw.Write(finalEvents)
} else if !processor.MessageStartSent() && !cw.Disconnected() {
// 整个流未收到任何可解析的上游数据（全部 SSE 行均无法被 JSON 解析），
// 触发 failover 在同账号重试，避免向客户端发出缺少 message_start 的残缺流
logger.LegacyPrintf("service.antigravity_gateway", "[antigravity-Claude-Stream] empty stream response (no valid events parsed), triggering failover")
return nil, &UpstreamFailoverError{
StatusCode: http.StatusBadGateway,
ResponseBody: []byte(`{"error":"empty stream response from upstream"}`),
RetryableOnSameAccount: true,
}
}
return &antigravityStreamResult{usage: convertUsage(agUsage), firstTokenMs: firstTokenMs, clientDisconnect: cw.Disconnected()}, nil
}
backend/internal/service/antigravity_gateway_service_test.go
+40
View File
@@ -998,6 +998,46 @@ func TestHandleClaudeStreamingResponse_ClientDisconnect(t *testing.T) {
require.True(t, result.clientDisconnect)
}
// TestHandleClaudeStreamingResponse_EmptyStream
// 验证：上游只返回无法解析的 SSE 行时，触发 UpstreamFailoverError 而不是向客户端发出残缺流
func TestHandleClaudeStreamingResponse_EmptyStream(t *testing.T) {
gin.SetMode(gin.TestMode)
svc := newAntigravityTestService(&config.Config{
Gateway: config.GatewayConfig{MaxLineSize: defaultMaxLineSize},
})
rec := httptest.NewRecorder()
c, _ := gin.CreateTestContext(rec)
c.Request = httptest.NewRequest(http.MethodPost, "/", nil)
pr, pw := io.Pipe()
resp := &http.Response{StatusCode: http.StatusOK, Body: pr, Header: http.Header{}}
go func() {
defer func() { _ = pw.Close() }()
// 所有行均为无法 JSON 解析的内容，ProcessLine 全部返回 nil
fmt.Fprintln(pw, "data: not-valid-json")
fmt.Fprintln(pw, "")
fmt.Fprintln(pw, "data: also-invalid")
fmt.Fprintln(pw, "")
}()
_, err := svc.handleClaudeStreamingResponse(c, resp, time.Now(), "claude-sonnet-4-5")
_ = pr.Close()
// 应当返回 UpstreamFailoverError 而非 nil，以便上层触发 failover
require.Error(t, err)
var failoverErr *UpstreamFailoverError
require.ErrorAs(t, err, &failoverErr)
require.True(t, failoverErr.RetryableOnSameAccount)
// 客户端不应收到任何 SSE 事件（既无 message_start 也无 message_stop）
body := rec.Body.String()
require.NotContains(t, body, "event: message_start")
require.NotContains(t, body, "event: message_stop")
require.NotContains(t, body, "event: message_delta")
}
// TestHandleClaudeStreamingResponse_ContextCanceled
// 验证：context 取消时不注入错误事件
func TestHandleClaudeStreamingResponse_ContextCanceled(t *testing.T) {
backend/internal/service/api_key.go
+36
View File
@@ -14,6 +14,18 @@ const (
StatusAPIKeyExpired = "expired"
)
// Rate limit window durations
const (
RateLimitWindow5h = 5 * time.Hour
RateLimitWindow1d = 24 * time.Hour
RateLimitWindow7d = 7 * 24 * time.Hour
)
// IsWindowExpired returns true if the window starting at windowStart has exceeded the given duration.
func IsWindowExpired(windowStart *time.Time, duration time.Duration) bool {
return windowStart != nil && time.Since(*windowStart) >= duration
}
type APIKey struct {
ID int64
UserID int64
@@ -98,6 +110,30 @@ func (k *APIKey) GetDaysUntilExpiry() int {
return int(duration.Hours() / 24)
}
// EffectiveUsage5h returns the 5h window usage, or 0 if the window has expired.
func (k *APIKey) EffectiveUsage5h() float64 {
if IsWindowExpired(k.Window5hStart, RateLimitWindow5h) {
return 0
}
return k.Usage5h
}
// EffectiveUsage1d returns the 1d window usage, or 0 if the window has expired.
func (k *APIKey) EffectiveUsage1d() float64 {
if IsWindowExpired(k.Window1dStart, RateLimitWindow1d) {
return 0
}
return k.Usage1d
}
// EffectiveUsage7d returns the 7d window usage, or 0 if the window has expired.
func (k *APIKey) EffectiveUsage7d() float64 {
if IsWindowExpired(k.Window7dStart, RateLimitWindow7d) {
return 0
}
return k.Usage7d
}
// APIKeyListFilters holds optional filtering parameters for listing API keys.
type APIKeyListFilters struct {
Search string
backend/internal/service/api_key_rate_limit_test.go
+245
View File
@@ -0,0 +1,245 @@
package service
import (
"testing"
"time"
)
func TestIsWindowExpired(t *testing.T) {
now := time.Now()
tests := []struct {
name string
start *time.Time
duration time.Duration
want bool
}{
{
name: "nil window start",
start: nil,
duration: RateLimitWindow5h,
want: false,
},
{
name: "active window (started 1h ago, 5h window)",
start: rateLimitTimePtr(now.Add(-1 * time.Hour)),
duration: RateLimitWindow5h,
want: false,
},
{
name: "expired window (started 6h ago, 5h window)",
start: rateLimitTimePtr(now.Add(-6 * time.Hour)),
duration: RateLimitWindow5h,
want: true,
},
{
name: "exactly at boundary (started 5h ago, 5h window)",
start: rateLimitTimePtr(now.Add(-5 * time.Hour)),
duration: RateLimitWindow5h,
want: true,
},
{
name: "active 1d window (started 12h ago)",
start: rateLimitTimePtr(now.Add(-12 * time.Hour)),
duration: RateLimitWindow1d,
want: false,
},
{
name: "expired 1d window (started 25h ago)",
start: rateLimitTimePtr(now.Add(-25 * time.Hour)),
duration: RateLimitWindow1d,
want: true,
},
{
name: "active 7d window (started 3d ago)",
start: rateLimitTimePtr(now.Add(-3 * 24 * time.Hour)),
duration: RateLimitWindow7d,
want: false,
},
{
name: "expired 7d window (started 8d ago)",
start: rateLimitTimePtr(now.Add(-8 * 24 * time.Hour)),
duration: RateLimitWindow7d,
want: true,
},
}
for _, tt := range tests {
t.Run(tt.name, func(t *testing.T) {
got := IsWindowExpired(tt.start, tt.duration)
if got != tt.want {
t.Errorf("IsWindowExpired() = %v, want %v", got, tt.want)
}
})
}
}
func TestAPIKey_EffectiveUsage(t *testing.T) {
now := time.Now()
tests := []struct {
name string
key APIKey
want5h float64
want1d float64
want7d float64
}{
{
name: "all windows active",
key: APIKey{
Usage5h: 5.0,
Usage1d: 10.0,
Usage7d: 50.0,
Window5hStart: rateLimitTimePtr(now.Add(-1 * time.Hour)),
Window1dStart: rateLimitTimePtr(now.Add(-12 * time.Hour)),
Window7dStart: rateLimitTimePtr(now.Add(-3 * 24 * time.Hour)),
},
want5h: 5.0,
want1d: 10.0,
want7d: 50.0,
},
{
name: "all windows expired",
key: APIKey{
Usage5h: 5.0,
Usage1d: 10.0,
Usage7d: 50.0,
Window5hStart: rateLimitTimePtr(now.Add(-6 * time.Hour)),
Window1dStart: rateLimitTimePtr(now.Add(-25 * time.Hour)),
Window7dStart: rateLimitTimePtr(now.Add(-8 * 24 * time.Hour)),
},
want5h: 0,
want1d: 0,
want7d: 0,
},
{
name: "nil window starts return raw usage",
key: APIKey{
Usage5h: 5.0,
Usage1d: 10.0,
Usage7d: 50.0,
Window5hStart: nil,
Window1dStart: nil,
Window7dStart: nil,
},
want5h: 5.0,
want1d: 10.0,
want7d: 50.0,
},
{
name: "mixed: 5h expired, 1d active, 7d nil",
key: APIKey{
Usage5h: 5.0,
Usage1d: 10.0,
Usage7d: 50.0,
Window5hStart: rateLimitTimePtr(now.Add(-6 * time.Hour)),
Window1dStart: rateLimitTimePtr(now.Add(-12 * time.Hour)),
Window7dStart: nil,
},
want5h: 0,
want1d: 10.0,
want7d: 50.0,
},
{
name: "zero usage with active windows",
key: APIKey{
Usage5h: 0,
Usage1d: 0,
Usage7d: 0,
Window5hStart: rateLimitTimePtr(now.Add(-1 * time.Hour)),
Window1dStart: rateLimitTimePtr(now.Add(-1 * time.Hour)),
Window7dStart: rateLimitTimePtr(now.Add(-1 * time.Hour)),
},
want5h: 0,
want1d: 0,
want7d: 0,
},
}
for _, tt := range tests {
t.Run(tt.name, func(t *testing.T) {
if got := tt.key.EffectiveUsage5h(); got != tt.want5h {
t.Errorf("EffectiveUsage5h() = %v, want %v", got, tt.want5h)
}
if got := tt.key.EffectiveUsage1d(); got != tt.want1d {
t.Errorf("EffectiveUsage1d() = %v, want %v", got, tt.want1d)
}
if got := tt.key.EffectiveUsage7d(); got != tt.want7d {
t.Errorf("EffectiveUsage7d() = %v, want %v", got, tt.want7d)
}
})
}
}
func TestAPIKeyRateLimitData_EffectiveUsage(t *testing.T) {
now := time.Now()
tests := []struct {
name string
data APIKeyRateLimitData
want5h float64
want1d float64
want7d float64
}{
{
name: "all windows active",
data: APIKeyRateLimitData{
Usage5h: 3.0,
Usage1d: 8.0,
Usage7d: 40.0,
Window5hStart: rateLimitTimePtr(now.Add(-2 * time.Hour)),
Window1dStart: rateLimitTimePtr(now.Add(-10 * time.Hour)),
Window7dStart: rateLimitTimePtr(now.Add(-2 * 24 * time.Hour)),
},
want5h: 3.0,
want1d: 8.0,
want7d: 40.0,
},
{
name: "all windows expired",
data: APIKeyRateLimitData{
Usage5h: 3.0,
Usage1d: 8.0,
Usage7d: 40.0,
Window5hStart: rateLimitTimePtr(now.Add(-10 * time.Hour)),
Window1dStart: rateLimitTimePtr(now.Add(-48 * time.Hour)),
Window7dStart: rateLimitTimePtr(now.Add(-10 * 24 * time.Hour)),
},
want5h: 0,
want1d: 0,
want7d: 0,
},
{
name: "nil window starts return raw usage",
data: APIKeyRateLimitData{
Usage5h: 3.0,
Usage1d: 8.0,
Usage7d: 40.0,
Window5hStart: nil,
Window1dStart: nil,
Window7dStart: nil,
},
want5h: 3.0,
want1d: 8.0,
want7d: 40.0,
},
}
for _, tt := range tests {
t.Run(tt.name, func(t *testing.T) {
if got := tt.data.EffectiveUsage5h(); got != tt.want5h {
t.Errorf("EffectiveUsage5h() = %v, want %v", got, tt.want5h)
}
if got := tt.data.EffectiveUsage1d(); got != tt.want1d {
t.Errorf("EffectiveUsage1d() = %v, want %v", got, tt.want1d)
}
if got := tt.data.EffectiveUsage7d(); got != tt.want7d {
t.Errorf("EffectiveUsage7d() = %v, want %v", got, tt.want7d)
}
})
}
}
func rateLimitTimePtr(t time.Time) *time.Time {
return &t
}
backend/internal/service/api_key_service.go
+24
View File
@@ -86,6 +86,30 @@ type APIKeyRateLimitData struct {
Window7dStart *time.Time
}
// EffectiveUsage5h returns the 5h window usage, or 0 if the window has expired.
func (d *APIKeyRateLimitData) EffectiveUsage5h() float64 {
if IsWindowExpired(d.Window5hStart, RateLimitWindow5h) {
return 0
}
return d.Usage5h
}
// EffectiveUsage1d returns the 1d window usage, or 0 if the window has expired.
func (d *APIKeyRateLimitData) EffectiveUsage1d() float64 {
if IsWindowExpired(d.Window1dStart, RateLimitWindow1d) {
return 0
}
return d.Usage1d
}
// EffectiveUsage7d returns the 7d window usage, or 0 if the window has expired.
func (d *APIKeyRateLimitData) EffectiveUsage7d() float64 {
if IsWindowExpired(d.Window7dStart, RateLimitWindow7d) {
return 0
}
return d.Usage7d
}
// APIKeyCache defines cache operations for API key service
type APIKeyCache interface {
GetCreateAttemptCount(ctx context.Context, userID int64) (int, error)
backend/internal/service/billing_cache_service.go
+9 -5
View File
@@ -565,15 +565,15 @@ func (s *BillingCacheService) evaluateRateLimits(ctx context.Context, apiKey *AP
needsReset := false
// Reset expired windows in-memory for check purposes
if w5h != nil && time.Since(*w5h) >= 5*time.Hour {
if IsWindowExpired(w5h, RateLimitWindow5h) {
usage5h = 0
needsReset = true
}
if w1d != nil && time.Since(*w1d) >= 24*time.Hour {
if IsWindowExpired(w1d, RateLimitWindow1d) {
usage1d = 0
needsReset = true
}
if w7d != nil && time.Since(*w7d) >= 7*24*time.Hour {
if IsWindowExpired(w7d, RateLimitWindow7d) {
usage7d = 0
needsReset = true
}
@@ -589,12 +589,16 @@ func (s *BillingCacheService) evaluateRateLimits(ctx context.Context, apiKey *AP
if loader, ok := s.apiKeyRateLimitLoader.(interface {
ResetRateLimitWindows(ctx context.Context, id int64) error
}); ok {
_ = loader.ResetRateLimitWindows(resetCtx, keyID)
if err := loader.ResetRateLimitWindows(resetCtx, keyID); err != nil {
logger.LegacyPrintf("service.billing_cache", "Warning: reset rate limit windows failed for api key %d: %v", keyID, err)
}
}
}
// Invalidate cache so next request loads fresh data
if s.cache != nil {
_ = s.cache.InvalidateAPIKeyRateLimit(resetCtx, keyID)
if err := s.cache.InvalidateAPIKeyRateLimit(resetCtx, keyID); err != nil {
logger.LegacyPrintf("service.billing_cache", "Warning: invalidate rate limit cache failed for api key %d: %v", keyID, err)
}
}
}()
}
backend/internal/service/gateway_service.go
+55 -81
View File
@@ -537,33 +537,34 @@ func (s *GatewayService) TempUnscheduleRetryableError(ctx context.Context, accou
// GatewayService handles API gateway operations
type GatewayService struct {
accountRepo AccountRepository
groupRepo GroupRepository
usageLogRepo UsageLogRepository
userRepo UserRepository
userSubRepo UserSubscriptionRepository
userGroupRateRepo UserGroupRateRepository
cache GatewayCache
digestStore *DigestSessionStore
cfg *config.Config
schedulerSnapshot *SchedulerSnapshotService
billingService *BillingService
rateLimitService *RateLimitService
billingCacheService *BillingCacheService
identityService *IdentityService
httpUpstream HTTPUpstream
deferredService *DeferredService
concurrencyService *ConcurrencyService
claudeTokenProvider *ClaudeTokenProvider
sessionLimitCache SessionLimitCache // 会话数量限制缓存（仅 Anthropic OAuth/SetupToken）
rpmCache RPMCache // RPM 计数缓存（仅 Anthropic OAuth/SetupToken）
userGroupRateCache *gocache.Cache
userGroupRateSF singleflight.Group
modelsListCache *gocache.Cache
modelsListCacheTTL time.Duration
responseHeaderFilter *responseheaders.CompiledHeaderFilter
debugModelRouting atomic.Bool
debugClaudeMimic atomic.Bool
accountRepo AccountRepository
groupRepo GroupRepository
usageLogRepo UsageLogRepository
userRepo UserRepository
userSubRepo UserSubscriptionRepository
userGroupRateRepo UserGroupRateRepository
cache GatewayCache
digestStore *DigestSessionStore
cfg *config.Config
schedulerSnapshot *SchedulerSnapshotService
billingService *BillingService
rateLimitService *RateLimitService
billingCacheService *BillingCacheService
identityService *IdentityService
httpUpstream HTTPUpstream
deferredService *DeferredService
concurrencyService *ConcurrencyService
claudeTokenProvider *ClaudeTokenProvider
sessionLimitCache SessionLimitCache // 会话数量限制缓存（仅 Anthropic OAuth/SetupToken）
rpmCache RPMCache // RPM 计数缓存（仅 Anthropic OAuth/SetupToken）
userGroupRateResolver *userGroupRateResolver
userGroupRateCache *gocache.Cache
userGroupRateSF singleflight.Group
modelsListCache *gocache.Cache
modelsListCacheTTL time.Duration
responseHeaderFilter *responseheaders.CompiledHeaderFilter
debugModelRouting atomic.Bool
debugClaudeMimic atomic.Bool
}
// NewGatewayService creates a new GatewayService
@@ -618,6 +619,13 @@ func NewGatewayService(
modelsListCacheTTL: modelsListTTL,
responseHeaderFilter: compileResponseHeaderFilter(cfg),
}
svc.userGroupRateResolver = newUserGroupRateResolver(
userGroupRateRepo,
svc.userGroupRateCache,
userGroupRateTTL,
&svc.userGroupRateSF,
"service.gateway",
)
svc.debugModelRouting.Store(parseDebugEnvBool(os.Getenv("SUB2API_DEBUG_MODEL_ROUTING")))
svc.debugClaudeMimic.Store(parseDebugEnvBool(os.Getenv("SUB2API_DEBUG_CLAUDE_MIMIC")))
return svc
@@ -3490,6 +3498,10 @@ func (s *GatewayService) isModelSupportedByAccount(account *Account, requestedMo
if account.Platform == PlatformSora {
return s.isSoraModelSupportedByAccount(account, requestedModel)
}
// OpenAI 透传模式：仅替换认证，允许所有模型
if account.Platform == PlatformOpenAI && account.IsOpenAIPassthroughEnabled() {
return true
}
// OAuth/SetupToken 账号使用 Anthropic 标准映射（短ID → 长ID）
if account.Platform == PlatformAnthropic && account.Type != AccountTypeAPIKey {
requestedModel = claude.NormalizeModelID(requestedModel)
@@ -5588,6 +5600,11 @@ func extractUpstreamErrorMessage(body []byte) string {
return m
}
// ChatGPT 内部 API 风格：{"detail":"..."}
if d := gjson.GetBytes(body, "detail").String(); strings.TrimSpace(d) != "" {
return d
}
// 兜底：尝试顶层 message
return gjson.GetBytes(body, "message").String()
}
@@ -6517,63 +6534,20 @@ func (s *GatewayService) replaceModelInResponseBody(body []byte, fromModel, toMo
}
func (s *GatewayService) getUserGroupRateMultiplier(ctx context.Context, userID, groupID int64, groupDefaultMultiplier float64) float64 {
if s == nil || userID <= 0 || groupID <= 0 {
if s == nil {
return groupDefaultMultiplier
}
key := fmt.Sprintf("%d:%d", userID, groupID)
if s.userGroupRateCache != nil {
if cached, ok := s.userGroupRateCache.Get(key); ok {
if multiplier, castOK := cached.(float64); castOK {
userGroupRateCacheHitTotal.Add(1)
return multiplier
}
}
resolver := s.userGroupRateResolver
if resolver == nil {
resolver = newUserGroupRateResolver(
s.userGroupRateRepo,
s.userGroupRateCache,
resolveUserGroupRateCacheTTL(s.cfg),
&s.userGroupRateSF,
"service.gateway",
)
}
if s.userGroupRateRepo == nil {
return groupDefaultMultiplier
}
userGroupRateCacheMissTotal.Add(1)
value, err, shared := s.userGroupRateSF.Do(key, func() (any, error) {
if s.userGroupRateCache != nil {
if cached, ok := s.userGroupRateCache.Get(key); ok {
if multiplier, castOK := cached.(float64); castOK {
userGroupRateCacheHitTotal.Add(1)
return multiplier, nil
}
}
}
userGroupRateCacheLoadTotal.Add(1)
userRate, repoErr := s.userGroupRateRepo.GetByUserAndGroup(ctx, userID, groupID)
if repoErr != nil {
return nil, repoErr
}
multiplier := groupDefaultMultiplier
if userRate != nil {
multiplier = *userRate
}
if s.userGroupRateCache != nil {
s.userGroupRateCache.Set(key, multiplier, resolveUserGroupRateCacheTTL(s.cfg))
}
return multiplier, nil
})
if shared {
userGroupRateCacheSFSharedTotal.Add(1)
}
if err != nil {
userGroupRateCacheFallbackTotal.Add(1)
logger.LegacyPrintf("service.gateway", "get user group rate failed, fallback to group default: user=%d group=%d err=%v", userID, groupID, err)
return groupDefaultMultiplier
}
multiplier, ok := value.(float64)
if !ok {
userGroupRateCacheFallbackTotal.Add(1)
return groupDefaultMultiplier
}
return multiplier
return resolver.Resolve(ctx, userID, groupID, groupDefaultMultiplier)
}
// RecordUsageInput 记录使用量的输入参数
backend/internal/service/openai_codex_transform.go
+21 -10
View File
@@ -77,7 +77,7 @@ type codexTransformResult struct {
PromptCacheKey string
}
func applyCodexOAuthTransform(reqBody map[string]any, isCodexCLI bool) codexTransformResult {
func applyCodexOAuthTransform(reqBody map[string]any, isCodexCLI bool, isCompact bool) codexTransformResult {
result := codexTransformResult{}
// 工具续链需求会影响存储策略与 input 过滤逻辑。
needsToolContinuation := NeedsToolContinuation(reqBody)
@@ -95,15 +95,26 @@ func applyCodexOAuthTransform(reqBody map[string]any, isCodexCLI bool) codexTran
result.NormalizedModel = normalizedModel
}
// OAuth 走 ChatGPT internal API 时，store 必须为 false；显式 true 也会强制覆盖。
// 避免上游返回 "Store must be set to false"。
if v, ok := reqBody["store"].(bool); !ok || v {
reqBody["store"] = false
result.Modified = true
}
if v, ok := reqBody["stream"].(bool); !ok || !v {
reqBody["stream"] = true
result.Modified = true
if isCompact {
if _, ok := reqBody["store"]; ok {
delete(reqBody, "store")
result.Modified = true
}
if _, ok := reqBody["stream"]; ok {
delete(reqBody, "stream")
result.Modified = true
}
} else {
// OAuth 走 ChatGPT internal API 时，store 必须为 false；显式 true 也会强制覆盖。
// 避免上游返回 "Store must be set to false"。
if v, ok := reqBody["store"].(bool); !ok || v {
reqBody["store"] = false
result.Modified = true
}
if v, ok := reqBody["stream"].(bool); !ok || !v {
reqBody["stream"] = true
result.Modified = true
}
}
// Strip parameters unsupported by codex models via the Responses API.
backend/internal/service/openai_codex_transform_test.go
+25 -9
View File
@@ -18,7 +18,7 @@ func TestApplyCodexOAuthTransform_ToolContinuationPreservesInput(t *testing.T) {
"tool_choice": "auto",
}
applyCodexOAuthTransform(reqBody, false)
applyCodexOAuthTransform(reqBody, false, false)
// 未显式设置 store=true，默认为 false。
store, ok := reqBody["store"].(bool)
@@ -53,7 +53,7 @@ func TestApplyCodexOAuthTransform_ExplicitStoreFalsePreserved(t *testing.T) {
"tool_choice": "auto",
}
applyCodexOAuthTransform(reqBody, false)
applyCodexOAuthTransform(reqBody, false, false)
store, ok := reqBody["store"].(bool)
require.True(t, ok)
@@ -72,13 +72,29 @@ func TestApplyCodexOAuthTransform_ExplicitStoreTrueForcedFalse(t *testing.T) {
"tool_choice": "auto",
}
applyCodexOAuthTransform(reqBody, false)
applyCodexOAuthTransform(reqBody, false, false)
store, ok := reqBody["store"].(bool)
require.True(t, ok)
require.False(t, store)
}
func TestApplyCodexOAuthTransform_CompactForcesNonStreaming(t *testing.T) {
reqBody := map[string]any{
"model": "gpt-5.1-codex",
"store": true,
"stream": true,
}
result := applyCodexOAuthTransform(reqBody, true, true)
_, hasStore := reqBody["store"]
require.False(t, hasStore)
_, hasStream := reqBody["stream"]
require.False(t, hasStream)
require.True(t, result.Modified)
}
func TestApplyCodexOAuthTransform_NonContinuationDefaultsStoreFalseAndStripsIDs(t *testing.T) {
// 非续链场景：未设置 store 时默认 false，并移除 input 中的 id。
@@ -89,7 +105,7 @@ func TestApplyCodexOAuthTransform_NonContinuationDefaultsStoreFalseAndStripsIDs(
},
}
applyCodexOAuthTransform(reqBody, false)
applyCodexOAuthTransform(reqBody, false, false)
store, ok := reqBody["store"].(bool)
require.True(t, ok)
@@ -138,7 +154,7 @@ func TestApplyCodexOAuthTransform_NormalizeCodexTools_PreservesResponsesFunction
},
}
applyCodexOAuthTransform(reqBody, false)
applyCodexOAuthTransform(reqBody, false, false)
tools, ok := reqBody["tools"].([]any)
require.True(t, ok)
@@ -158,7 +174,7 @@ func TestApplyCodexOAuthTransform_EmptyInput(t *testing.T) {
"input": []any{},
}
applyCodexOAuthTransform(reqBody, false)
applyCodexOAuthTransform(reqBody, false, false)
input, ok := reqBody["input"].([]any)
require.True(t, ok)
@@ -193,7 +209,7 @@ func TestApplyCodexOAuthTransform_CodexCLI_PreservesExistingInstructions(t *test
"instructions": "existing instructions",
}
result := applyCodexOAuthTransform(reqBody, true) // isCodexCLI=true
result := applyCodexOAuthTransform(reqBody, true, false) // isCodexCLI=true
instructions, ok := reqBody["instructions"].(string)
require.True(t, ok)
@@ -210,7 +226,7 @@ func TestApplyCodexOAuthTransform_CodexCLI_SuppliesDefaultWhenEmpty(t *testing.T
// 没有 instructions 字段
}
result := applyCodexOAuthTransform(reqBody, true) // isCodexCLI=true
result := applyCodexOAuthTransform(reqBody, true, false) // isCodexCLI=true
instructions, ok := reqBody["instructions"].(string)
require.True(t, ok)
@@ -226,7 +242,7 @@ func TestApplyCodexOAuthTransform_NonCodexCLI_OverridesInstructions(t *testing.T
"instructions": "old instructions",
}
result := applyCodexOAuthTransform(reqBody, false) // isCodexCLI=false
result := applyCodexOAuthTransform(reqBody, false, false) // isCodexCLI=false
instructions, ok := reqBody["instructions"].(string)
require.True(t, ok)
backend/internal/service/openai_gateway_messages.go
+411
View File
@@ -0,0 +1,411 @@
package service
import (
"bufio"
"context"
"encoding/json"
"errors"
"fmt"
"io"
"net/http"
"strings"
"time"
"github.com/Wei-Shaw/sub2api/internal/pkg/apicompat"
"github.com/Wei-Shaw/sub2api/internal/pkg/logger"
"github.com/Wei-Shaw/sub2api/internal/util/responseheaders"
"github.com/gin-gonic/gin"
"go.uber.org/zap"
)
// ForwardAsAnthropic accepts an Anthropic Messages request body, converts it
// to OpenAI Responses API format, forwards to the OpenAI upstream, and converts
// the response back to Anthropic Messages format. This enables Claude Code
// clients to access OpenAI models through the standard /v1/messages endpoint.
func (s *OpenAIGatewayService) ForwardAsAnthropic(
ctx context.Context,
c *gin.Context,
account *Account,
body []byte,
promptCacheKey string,
) (*OpenAIForwardResult, error) {
startTime := time.Now()
// 1. Parse Anthropic request
var anthropicReq apicompat.AnthropicRequest
if err := json.Unmarshal(body, &anthropicReq); err != nil {
return nil, fmt.Errorf("parse anthropic request: %w", err)
}
originalModel := anthropicReq.Model
isStream := anthropicReq.Stream
// 2. Convert Anthropic → Responses
responsesReq, err := apicompat.AnthropicToResponses(&anthropicReq)
if err != nil {
return nil, fmt.Errorf("convert anthropic to responses: %w", err)
}
// 3. Model mapping
mappedModel := account.GetMappedModel(originalModel)
responsesReq.Model = mappedModel
logger.L().Debug("openai messages: model mapping applied",
zap.Int64("account_id", account.ID),
zap.String("original_model", originalModel),
zap.String("mapped_model", mappedModel),
zap.Bool("stream", isStream),
)
// 4. Marshal Responses request body, then apply OAuth codex transform
responsesBody, err := json.Marshal(responsesReq)
if err != nil {
return nil, fmt.Errorf("marshal responses request: %w", err)
}
if account.Type == AccountTypeOAuth {
var reqBody map[string]any
if err := json.Unmarshal(responsesBody, &reqBody); err != nil {
return nil, fmt.Errorf("unmarshal for codex transform: %w", err)
}
applyCodexOAuthTransform(reqBody, false, false)
// OAuth codex transform forces stream=true upstream, so always use
// the streaming response handler regardless of what the client asked.
isStream = true
responsesBody, err = json.Marshal(reqBody)
if err != nil {
return nil, fmt.Errorf("remarshal after codex transform: %w", err)
}
}
// 5. Get access token
token, _, err := s.GetAccessToken(ctx, account)
if err != nil {
return nil, fmt.Errorf("get access token: %w", err)
}
// 6. Build upstream request
upstreamReq, err := s.buildUpstreamRequest(ctx, c, account, responsesBody, token, isStream, promptCacheKey, false)
if err != nil {
return nil, fmt.Errorf("build upstream request: %w", err)
}
// 7. Send request
proxyURL := ""
if account.Proxy != nil {
proxyURL = account.Proxy.URL()
}
resp, err := s.httpUpstream.Do(upstreamReq, proxyURL, account.ID, account.Concurrency)
if err != nil {
safeErr := sanitizeUpstreamErrorMessage(err.Error())
setOpsUpstreamError(c, 0, safeErr, "")
appendOpsUpstreamError(c, OpsUpstreamErrorEvent{
Platform: account.Platform,
AccountID: account.ID,
AccountName: account.Name,
UpstreamStatusCode: 0,
Kind: "request_error",
Message: safeErr,
})
writeAnthropicError(c, http.StatusBadGateway, "api_error", "Upstream request failed")
return nil, fmt.Errorf("upstream request failed: %s", safeErr)
}
defer func() { _ = resp.Body.Close() }()
// 8. Handle error response with failover
if resp.StatusCode >= 400 {
if s.shouldFailoverUpstreamError(resp.StatusCode) {
respBody, _ := io.ReadAll(io.LimitReader(resp.Body, 2<<20))
_ = resp.Body.Close()
upstreamMsg := strings.TrimSpace(extractUpstreamErrorMessage(respBody))
upstreamMsg = sanitizeUpstreamErrorMessage(upstreamMsg)
upstreamDetail := ""
if s.cfg != nil && s.cfg.Gateway.LogUpstreamErrorBody {
maxBytes := s.cfg.Gateway.LogUpstreamErrorBodyMaxBytes
if maxBytes <= 0 {
maxBytes = 2048
}
upstreamDetail = truncateString(string(respBody), maxBytes)
}
appendOpsUpstreamError(c, OpsUpstreamErrorEvent{
Platform: account.Platform,
AccountID: account.ID,
AccountName: account.Name,
UpstreamStatusCode: resp.StatusCode,
UpstreamRequestID: resp.Header.Get("x-request-id"),
Kind: "failover",
Message: upstreamMsg,
Detail: upstreamDetail,
})
if s.rateLimitService != nil {
s.rateLimitService.HandleUpstreamError(ctx, account, resp.StatusCode, resp.Header, respBody)
}
return nil, &UpstreamFailoverError{StatusCode: resp.StatusCode, ResponseBody: respBody}
}
// Non-failover error: return Anthropic-formatted error to client
return s.handleAnthropicErrorResponse(resp, c, account)
}
// 9. Handle normal response
var result *OpenAIForwardResult
var handleErr error
if isStream {
result, handleErr = s.handleAnthropicStreamingResponse(resp, c, originalModel, mappedModel, startTime)
} else {
result, handleErr = s.handleAnthropicNonStreamingResponse(resp, c, originalModel, mappedModel, startTime)
}
// Extract and save Codex usage snapshot from response headers (for OAuth accounts)
if handleErr == nil && account.Type == AccountTypeOAuth {
if snapshot := ParseCodexRateLimitHeaders(resp.Header); snapshot != nil {
s.updateCodexUsageSnapshot(ctx, account.ID, snapshot)
}
}
return result, handleErr
}
// handleAnthropicErrorResponse reads an upstream error and returns it in
// Anthropic error format.
func (s *OpenAIGatewayService) handleAnthropicErrorResponse(
resp *http.Response,
c *gin.Context,
account *Account,
) (*OpenAIForwardResult, error) {
body, _ := io.ReadAll(io.LimitReader(resp.Body, 2<<20))
upstreamMsg := strings.TrimSpace(extractUpstreamErrorMessage(body))
if upstreamMsg == "" {
upstreamMsg = fmt.Sprintf("Upstream error: %d", resp.StatusCode)
}
upstreamMsg = sanitizeUpstreamErrorMessage(upstreamMsg)
// Record upstream error details for ops logging
upstreamDetail := ""
if s.cfg != nil && s.cfg.Gateway.LogUpstreamErrorBody {
maxBytes := s.cfg.Gateway.LogUpstreamErrorBodyMaxBytes
if maxBytes <= 0 {
maxBytes = 2048
}
upstreamDetail = truncateString(string(body), maxBytes)
}
setOpsUpstreamError(c, resp.StatusCode, upstreamMsg, upstreamDetail)
// Apply error passthrough rules (matches handleErrorResponse pattern in openai_gateway_service.go)
if status, errType, errMsg, matched := applyErrorPassthroughRule(
c, account.Platform, resp.StatusCode, body,
http.StatusBadGateway, "api_error", "Upstream request failed",
); matched {
writeAnthropicError(c, status, errType, errMsg)
if upstreamMsg == "" {
upstreamMsg = errMsg
}
if upstreamMsg == "" {
return nil, fmt.Errorf("upstream error: %d (passthrough rule matched)", resp.StatusCode)
}
return nil, fmt.Errorf("upstream error: %d (passthrough rule matched) message=%s", resp.StatusCode, upstreamMsg)
}
errType := "api_error"
switch {
case resp.StatusCode == 400:
errType = "invalid_request_error"
case resp.StatusCode == 404:
errType = "not_found_error"
case resp.StatusCode == 429:
errType = "rate_limit_error"
case resp.StatusCode >= 500:
errType = "api_error"
}
writeAnthropicError(c, resp.StatusCode, errType, upstreamMsg)
return nil, fmt.Errorf("upstream error: %d %s", resp.StatusCode, upstreamMsg)
}
// handleAnthropicNonStreamingResponse reads a Responses API JSON response,
// converts it to Anthropic Messages format, and writes it to the client.
func (s *OpenAIGatewayService) handleAnthropicNonStreamingResponse(
resp *http.Response,
c *gin.Context,
originalModel string,
mappedModel string,
startTime time.Time,
) (*OpenAIForwardResult, error) {
requestID := resp.Header.Get("x-request-id")
respBody, err := io.ReadAll(resp.Body)
if err != nil {
return nil, fmt.Errorf("read upstream response: %w", err)
}
var responsesResp apicompat.ResponsesResponse
if err := json.Unmarshal(respBody, &responsesResp); err != nil {
return nil, fmt.Errorf("parse responses response: %w", err)
}
anthropicResp := apicompat.ResponsesToAnthropic(&responsesResp, originalModel)
var usage OpenAIUsage
if responsesResp.Usage != nil {
usage = OpenAIUsage{
InputTokens: responsesResp.Usage.InputTokens,
OutputTokens: responsesResp.Usage.OutputTokens,
}
if responsesResp.Usage.InputTokensDetails != nil {
usage.CacheReadInputTokens = responsesResp.Usage.InputTokensDetails.CachedTokens
}
}
if s.responseHeaderFilter != nil {
responseheaders.WriteFilteredHeaders(c.Writer.Header(), resp.Header, s.responseHeaderFilter)
}
c.JSON(http.StatusOK, anthropicResp)
return &OpenAIForwardResult{
RequestID: requestID,
Usage: usage,
Model: originalModel,
BillingModel: mappedModel,
Stream: false,
Duration: time.Since(startTime),
}, nil
}
// handleAnthropicStreamingResponse reads Responses SSE events from upstream,
// converts each to Anthropic SSE events, and writes them to the client.
func (s *OpenAIGatewayService) handleAnthropicStreamingResponse(
resp *http.Response,
c *gin.Context,
originalModel string,
mappedModel string,
startTime time.Time,
) (*OpenAIForwardResult, error) {
requestID := resp.Header.Get("x-request-id")
if s.responseHeaderFilter != nil {
responseheaders.WriteFilteredHeaders(c.Writer.Header(), resp.Header, s.responseHeaderFilter)
}
c.Writer.Header().Set("Content-Type", "text/event-stream")
c.Writer.Header().Set("Cache-Control", "no-cache")
c.Writer.Header().Set("Connection", "keep-alive")
c.Writer.WriteHeader(http.StatusOK)
state := apicompat.NewResponsesEventToAnthropicState()
state.Model = originalModel
var usage OpenAIUsage
var firstTokenMs *int
firstChunk := true
scanner := bufio.NewScanner(resp.Body)
scanner.Buffer(make([]byte, 0, 64*1024), 1024*1024)
for scanner.Scan() {
line := scanner.Text()
if !strings.HasPrefix(line, "data: ") || line == "data: [DONE]" {
continue
}
payload := line[6:]
if firstChunk {
firstChunk = false
ms := int(time.Since(startTime).Milliseconds())
firstTokenMs = &ms
}
// Parse the Responses SSE event
var event apicompat.ResponsesStreamEvent
if err := json.Unmarshal([]byte(payload), &event); err != nil {
logger.L().Warn("openai messages stream: failed to parse event",
zap.Error(err),
zap.String("request_id", requestID),
)
continue
}
// Extract usage from completion events
if (event.Type == "response.completed" || event.Type == "response.incomplete" || event.Type == "response.failed") &&
event.Response != nil && event.Response.Usage != nil {
usage = OpenAIUsage{
InputTokens: event.Response.Usage.InputTokens,
OutputTokens: event.Response.Usage.OutputTokens,
}
if event.Response.Usage.InputTokensDetails != nil {
usage.CacheReadInputTokens = event.Response.Usage.InputTokensDetails.CachedTokens
}
}
// Convert to Anthropic events
events := apicompat.ResponsesEventToAnthropicEvents(&event, state)
for _, evt := range events {
sse, err := apicompat.ResponsesAnthropicEventToSSE(evt)
if err != nil {
logger.L().Warn("openai messages stream: failed to marshal event",
zap.Error(err),
zap.String("request_id", requestID),
)
continue
}
if _, err := fmt.Fprint(c.Writer, sse); err != nil {
// Client disconnected — return collected usage
logger.L().Info("openai messages stream: client disconnected",
zap.String("request_id", requestID),
)
return &OpenAIForwardResult{
RequestID: requestID,
Usage: usage,
Model: originalModel,
BillingModel: mappedModel,
Stream: true,
Duration: time.Since(startTime),
FirstTokenMs: firstTokenMs,
}, nil
}
}
if len(events) > 0 {
c.Writer.Flush()
}
}
if err := scanner.Err(); err != nil {
if !errors.Is(err, context.Canceled) && !errors.Is(err, context.DeadlineExceeded) {
logger.L().Warn("openai messages stream: read error",
zap.Error(err),
zap.String("request_id", requestID),
)
}
}
// Ensure the Anthropic stream is properly terminated
if finalEvents := apicompat.FinalizeResponsesAnthropicStream(state); len(finalEvents) > 0 {
for _, evt := range finalEvents {
sse, err := apicompat.ResponsesAnthropicEventToSSE(evt)
if err != nil {
continue
}
fmt.Fprint(c.Writer, sse) //nolint:errcheck
}
c.Writer.Flush()
}
return &OpenAIForwardResult{
RequestID: requestID,
Usage: usage,
Model: originalModel,
BillingModel: mappedModel,
Stream: true,
Duration: time.Since(startTime),
FirstTokenMs: firstTokenMs,
}, nil
}
// writeAnthropicError writes an error response in Anthropic Messages API format.
func writeAnthropicError(c *gin.Context, statusCode int, errType, message string) {
c.JSON(statusCode, gin.H{
"type": "error",
"error": gin.H{
"type": errType,
"message": message,
},
})
}
backend/internal/service/openai_gateway_record_usage_test.go
+336
View File
@@ -0,0 +1,336 @@
package service
import (
"context"
"errors"
"testing"
"time"
"github.com/Wei-Shaw/sub2api/internal/config"
"github.com/stretchr/testify/require"
)
type openAIRecordUsageLogRepoStub struct {
UsageLogRepository
inserted bool
err error
calls int
lastLog *UsageLog
}
func (s *openAIRecordUsageLogRepoStub) Create(ctx context.Context, log *UsageLog) (bool, error) {
s.calls++
s.lastLog = log
return s.inserted, s.err
}
type openAIRecordUsageUserRepoStub struct {
UserRepository
deductCalls int
deductErr error
lastAmount float64
}
func (s *openAIRecordUsageUserRepoStub) DeductBalance(ctx context.Context, id int64, amount float64) error {
s.deductCalls++
s.lastAmount = amount
return s.deductErr
}
type openAIRecordUsageSubRepoStub struct {
UserSubscriptionRepository
incrementCalls int
incrementErr error
}
func (s *openAIRecordUsageSubRepoStub) IncrementUsage(ctx context.Context, id int64, costUSD float64) error {
s.incrementCalls++
return s.incrementErr
}
type openAIRecordUsageAPIKeyQuotaStub struct {
quotaCalls int
rateLimitCalls int
err error
lastAmount float64
}
func (s *openAIRecordUsageAPIKeyQuotaStub) UpdateQuotaUsed(ctx context.Context, apiKeyID int64, cost float64) error {
s.quotaCalls++
s.lastAmount = cost
return s.err
}
func (s *openAIRecordUsageAPIKeyQuotaStub) UpdateRateLimitUsage(ctx context.Context, apiKeyID int64, cost float64) error {
s.rateLimitCalls++
s.lastAmount = cost
return s.err
}
type openAIUserGroupRateRepoStub struct {
UserGroupRateRepository
rate *float64
err error
calls int
}
func (s *openAIUserGroupRateRepoStub) GetByUserAndGroup(ctx context.Context, userID, groupID int64) (*float64, error) {
s.calls++
if s.err != nil {
return nil, s.err
}
return s.rate, nil
}
func i64p(v int64) *int64 {
return &v
}
func newOpenAIRecordUsageServiceForTest(usageRepo UsageLogRepository, userRepo UserRepository, subRepo UserSubscriptionRepository, rateRepo UserGroupRateRepository) *OpenAIGatewayService {
cfg := &config.Config{}
cfg.Default.RateMultiplier = 1.1
return &OpenAIGatewayService{
usageLogRepo: usageRepo,
userRepo: userRepo,
userSubRepo: subRepo,
cfg: cfg,
billingService: NewBillingService(cfg, nil),
billingCacheService: &BillingCacheService{},
deferredService: &DeferredService{},
userGroupRateResolver: newUserGroupRateResolver(
rateRepo,
nil,
resolveUserGroupRateCacheTTL(cfg),
nil,
"service.openai_gateway.test",
),
}
}
func expectedOpenAICost(t *testing.T, svc *OpenAIGatewayService, model string, usage OpenAIUsage, multiplier float64) *CostBreakdown {
t.Helper()
cost, err := svc.billingService.CalculateCost(model, UsageTokens{
InputTokens: max(usage.InputTokens-usage.CacheReadInputTokens, 0),
OutputTokens: usage.OutputTokens,
CacheCreationTokens: usage.CacheCreationInputTokens,
CacheReadTokens: usage.CacheReadInputTokens,
}, multiplier)
require.NoError(t, err)
return cost
}
func max(a, b int) int {
if a > b {
return a
}
return b
}
func TestOpenAIGatewayServiceRecordUsage_UsesUserSpecificGroupRate(t *testing.T) {
groupID := int64(11)
groupRate := 1.4
userRate := 1.8
usage := OpenAIUsage{InputTokens: 15, OutputTokens: 4, CacheReadInputTokens: 3}
usageRepo := &openAIRecordUsageLogRepoStub{inserted: true}
userRepo := &openAIRecordUsageUserRepoStub{}
subRepo := &openAIRecordUsageSubRepoStub{}
rateRepo := &openAIUserGroupRateRepoStub{rate: &userRate}
svc := newOpenAIRecordUsageServiceForTest(usageRepo, userRepo, subRepo, rateRepo)
err := svc.RecordUsage(context.Background(), &OpenAIRecordUsageInput{
Result: &OpenAIForwardResult{
RequestID: "resp_user_group_rate",
Usage: usage,
Model: "gpt-5.1",
Duration: time.Second,
},
APIKey: &APIKey{
ID: 1001,
GroupID: i64p(groupID),
Group: &Group{
ID: groupID,
RateMultiplier: groupRate,
},
},
User: &User{ID: 2001},
Account: &Account{ID: 3001},
})
require.NoError(t, err)
require.Equal(t, 1, rateRepo.calls)
require.NotNil(t, usageRepo.lastLog)
require.Equal(t, userRate, usageRepo.lastLog.RateMultiplier)
require.Equal(t, 12, usageRepo.lastLog.InputTokens)
require.Equal(t, 3, usageRepo.lastLog.CacheReadTokens)
expected := expectedOpenAICost(t, svc, "gpt-5.1", usage, userRate)
require.InDelta(t, expected.ActualCost, usageRepo.lastLog.ActualCost, 1e-12)
require.InDelta(t, expected.ActualCost, userRepo.lastAmount, 1e-12)
require.Equal(t, 1, userRepo.deductCalls)
}
func TestOpenAIGatewayServiceRecordUsage_FallsBackToGroupDefaultRateOnResolverError(t *testing.T) {
groupID := int64(12)
groupRate := 1.6
usage := OpenAIUsage{InputTokens: 10, OutputTokens: 5, CacheReadInputTokens: 2}
usageRepo := &openAIRecordUsageLogRepoStub{inserted: true}
userRepo := &openAIRecordUsageUserRepoStub{}
subRepo := &openAIRecordUsageSubRepoStub{}
rateRepo := &openAIUserGroupRateRepoStub{err: errors.New("db unavailable")}
svc := newOpenAIRecordUsageServiceForTest(usageRepo, userRepo, subRepo, rateRepo)
err := svc.RecordUsage(context.Background(), &OpenAIRecordUsageInput{
Result: &OpenAIForwardResult{
RequestID: "resp_group_default_on_error",
Usage: usage,
Model: "gpt-5.1",
Duration: time.Second,
},
APIKey: &APIKey{
ID: 1002,
GroupID: i64p(groupID),
Group: &Group{
ID: groupID,
RateMultiplier: groupRate,
},
},
User: &User{ID: 2002},
Account: &Account{ID: 3002},
})
require.NoError(t, err)
require.Equal(t, 1, rateRepo.calls)
require.NotNil(t, usageRepo.lastLog)
require.Equal(t, groupRate, usageRepo.lastLog.RateMultiplier)
expected := expectedOpenAICost(t, svc, "gpt-5.1", usage, groupRate)
require.InDelta(t, expected.ActualCost, userRepo.lastAmount, 1e-12)
}
func TestOpenAIGatewayServiceRecordUsage_FallsBackToGroupDefaultRateWhenResolverMissing(t *testing.T) {
groupID := int64(13)
groupRate := 1.25
usage := OpenAIUsage{InputTokens: 9, OutputTokens: 4, CacheReadInputTokens: 1}
usageRepo := &openAIRecordUsageLogRepoStub{inserted: true}
userRepo := &openAIRecordUsageUserRepoStub{}
subRepo := &openAIRecordUsageSubRepoStub{}
svc := newOpenAIRecordUsageServiceForTest(usageRepo, userRepo, subRepo, nil)
svc.userGroupRateResolver = nil
err := svc.RecordUsage(context.Background(), &OpenAIRecordUsageInput{
Result: &OpenAIForwardResult{
RequestID: "resp_group_default_nil_resolver",
Usage: usage,
Model: "gpt-5.1",
Duration: time.Second,
},
APIKey: &APIKey{
ID: 1003,
GroupID: i64p(groupID),
Group: &Group{
ID: groupID,
RateMultiplier: groupRate,
},
},
User: &User{ID: 2003},
Account: &Account{ID: 3003},
})
require.NoError(t, err)
require.NotNil(t, usageRepo.lastLog)
require.Equal(t, groupRate, usageRepo.lastLog.RateMultiplier)
}
func TestOpenAIGatewayServiceRecordUsage_DuplicateUsageLogSkipsBilling(t *testing.T) {
usageRepo := &openAIRecordUsageLogRepoStub{inserted: false}
userRepo := &openAIRecordUsageUserRepoStub{}
subRepo := &openAIRecordUsageSubRepoStub{}
svc := newOpenAIRecordUsageServiceForTest(usageRepo, userRepo, subRepo, nil)
err := svc.RecordUsage(context.Background(), &OpenAIRecordUsageInput{
Result: &OpenAIForwardResult{
RequestID: "resp_duplicate",
Usage: OpenAIUsage{
InputTokens: 8,
OutputTokens: 4,
},
Model: "gpt-5.1",
Duration: time.Second,
},
APIKey: &APIKey{ID: 1004},
User: &User{ID: 2004},
Account: &Account{ID: 3004},
})
require.NoError(t, err)
require.Equal(t, 1, usageRepo.calls)
require.Equal(t, 0, userRepo.deductCalls)
require.Equal(t, 0, subRepo.incrementCalls)
}
func TestOpenAIGatewayServiceRecordUsage_UpdatesAPIKeyQuotaWhenConfigured(t *testing.T) {
usage := OpenAIUsage{InputTokens: 10, OutputTokens: 6, CacheReadInputTokens: 2}
usageRepo := &openAIRecordUsageLogRepoStub{inserted: true}
userRepo := &openAIRecordUsageUserRepoStub{}
subRepo := &openAIRecordUsageSubRepoStub{}
quotaSvc := &openAIRecordUsageAPIKeyQuotaStub{}
svc := newOpenAIRecordUsageServiceForTest(usageRepo, userRepo, subRepo, nil)
err := svc.RecordUsage(context.Background(), &OpenAIRecordUsageInput{
Result: &OpenAIForwardResult{
RequestID: "resp_quota_update",
Usage: usage,
Model: "gpt-5.1",
Duration: time.Second,
},
APIKey: &APIKey{
ID: 1005,
Quota: 100,
},
User: &User{ID: 2005},
Account: &Account{ID: 3005},
APIKeyService: quotaSvc,
})
require.NoError(t, err)
require.Equal(t, 1, quotaSvc.quotaCalls)
require.Equal(t, 0, quotaSvc.rateLimitCalls)
expected := expectedOpenAICost(t, svc, "gpt-5.1", usage, 1.1)
require.InDelta(t, expected.ActualCost, quotaSvc.lastAmount, 1e-12)
}
func TestOpenAIGatewayServiceRecordUsage_ClampsActualInputTokensToZero(t *testing.T) {
usageRepo := &openAIRecordUsageLogRepoStub{inserted: true}
userRepo := &openAIRecordUsageUserRepoStub{}
subRepo := &openAIRecordUsageSubRepoStub{}
svc := newOpenAIRecordUsageServiceForTest(usageRepo, userRepo, subRepo, nil)
err := svc.RecordUsage(context.Background(), &OpenAIRecordUsageInput{
Result: &OpenAIForwardResult{
RequestID: "resp_clamp_actual_input",
Usage: OpenAIUsage{
InputTokens: 2,
OutputTokens: 1,
CacheReadInputTokens: 5,
},
Model: "gpt-5.1",
Duration: time.Second,
},
APIKey: &APIKey{ID: 1006},
User: &User{ID: 2006},
Account: &Account{ID: 3006},
})
require.NoError(t, err)
require.NotNil(t, usageRepo.lastLog)
require.Equal(t, 0, usageRepo.lastLog.InputTokens)
}
backend/internal/service/openai_gateway_service.go
+285 -54
View File
@@ -25,6 +25,7 @@ import (
"github.com/Wei-Shaw/sub2api/internal/util/responseheaders"
"github.com/Wei-Shaw/sub2api/internal/util/urlvalidator"
"github.com/gin-gonic/gin"
"github.com/google/uuid"
"github.com/tidwall/gjson"
"github.com/tidwall/sjson"
"go.uber.org/zap"
@@ -49,6 +50,8 @@ const (
openAIWSRetryBackoffInitialDefault = 120 * time.Millisecond
openAIWSRetryBackoffMaxDefault = 2 * time.Second
openAIWSRetryJitterRatioDefault = 0.2
openAICompactSessionSeedKey = "openai_compact_session_seed"
codexCLIVersion = "0.104.0"
)
// OpenAI allowed headers whitelist (for non-passthrough).
@@ -204,12 +207,18 @@ type OpenAIUsage struct {
type OpenAIForwardResult struct {
RequestID string
Usage OpenAIUsage
Model string
Model string // 原始模型（用于响应和日志显示）
// BillingModel is the model used for cost calculation.
// When non-empty, CalculateCost uses this instead of Model.
// This is set by the Anthropic Messages conversion path where
// the mapped upstream model differs from the client-facing model.
BillingModel string
// ReasoningEffort is extracted from request body (reasoning.effort) or derived from model suffix.
// Stored for usage records display; nil means not provided / not applicable.
ReasoningEffort *string
Stream bool
OpenAIWSMode bool
ResponseHeaders http.Header
Duration time.Duration
FirstTokenMs *int
}
@@ -245,23 +254,24 @@ type openAIWSRetryMetrics struct {
// OpenAIGatewayService handles OpenAI API gateway operations
type OpenAIGatewayService struct {
accountRepo AccountRepository
usageLogRepo UsageLogRepository
userRepo UserRepository
userSubRepo UserSubscriptionRepository
cache GatewayCache
cfg *config.Config
codexDetector CodexClientRestrictionDetector
schedulerSnapshot *SchedulerSnapshotService
concurrencyService *ConcurrencyService
billingService *BillingService
rateLimitService *RateLimitService
billingCacheService *BillingCacheService
httpUpstream HTTPUpstream
deferredService *DeferredService
openAITokenProvider *OpenAITokenProvider
toolCorrector *CodexToolCorrector
openaiWSResolver OpenAIWSProtocolResolver
accountRepo AccountRepository
usageLogRepo UsageLogRepository
userRepo UserRepository
userSubRepo UserSubscriptionRepository
cache GatewayCache
cfg *config.Config
codexDetector CodexClientRestrictionDetector
schedulerSnapshot *SchedulerSnapshotService
concurrencyService *ConcurrencyService
billingService *BillingService
rateLimitService *RateLimitService
billingCacheService *BillingCacheService
userGroupRateResolver *userGroupRateResolver
httpUpstream HTTPUpstream
deferredService *DeferredService
openAITokenProvider *OpenAITokenProvider
toolCorrector *CodexToolCorrector
openaiWSResolver OpenAIWSProtocolResolver
openaiWSPoolOnce sync.Once
openaiWSStateStoreOnce sync.Once
@@ -284,6 +294,7 @@ func NewOpenAIGatewayService(
usageLogRepo UsageLogRepository,
userRepo UserRepository,
userSubRepo UserSubscriptionRepository,
userGroupRateRepo UserGroupRateRepository,
cache GatewayCache,
cfg *config.Config,
schedulerSnapshot *SchedulerSnapshotService,
@@ -296,18 +307,25 @@ func NewOpenAIGatewayService(
openAITokenProvider *OpenAITokenProvider,
) *OpenAIGatewayService {
svc := &OpenAIGatewayService{
accountRepo: accountRepo,
usageLogRepo: usageLogRepo,
userRepo: userRepo,
userSubRepo: userSubRepo,
cache: cache,
cfg: cfg,
codexDetector: NewOpenAICodexClientRestrictionDetector(cfg),
schedulerSnapshot: schedulerSnapshot,
concurrencyService: concurrencyService,
billingService: billingService,
rateLimitService: rateLimitService,
billingCacheService: billingCacheService,
accountRepo: accountRepo,
usageLogRepo: usageLogRepo,
userRepo: userRepo,
userSubRepo: userSubRepo,
cache: cache,
cfg: cfg,
codexDetector: NewOpenAICodexClientRestrictionDetector(cfg),
schedulerSnapshot: schedulerSnapshot,
concurrencyService: concurrencyService,
billingService: billingService,
rateLimitService: rateLimitService,
billingCacheService: billingCacheService,
userGroupRateResolver: newUserGroupRateResolver(
userGroupRateRepo,
nil,
resolveUserGroupRateCacheTTL(cfg),
nil,
"service.openai_gateway",
),
httpUpstream: httpUpstream,
deferredService: deferredService,
openAITokenProvider: openAITokenProvider,
@@ -888,6 +906,22 @@ func isOpenAIInstructionsRequiredError(upstreamStatusCode int, upstreamMsg strin
return false
}
// ExtractSessionID extracts the raw session ID from headers or body without hashing.
// Used by ForwardAsAnthropic to pass as prompt_cache_key for upstream cache.
func (s *OpenAIGatewayService) ExtractSessionID(c *gin.Context, body []byte) string {
if c == nil {
return ""
}
sessionID := strings.TrimSpace(c.GetHeader("session_id"))
if sessionID == "" {
sessionID = strings.TrimSpace(c.GetHeader("conversation_id"))
}
if sessionID == "" && len(body) > 0 {
sessionID = strings.TrimSpace(gjson.GetBytes(body, "prompt_cache_key").String())
}
return sessionID
}
// GenerateSessionHash generates a sticky-session hash for OpenAI requests.
//
// Priority:
@@ -1605,7 +1639,7 @@ func (s *OpenAIGatewayService) Forward(ctx context.Context, c *gin.Context, acco
}
if account.Type == AccountTypeOAuth {
codexResult := applyCodexOAuthTransform(reqBody, isCodexCLI)
codexResult := applyCodexOAuthTransform(reqBody, isCodexCLI, isOpenAIResponsesCompactPath(c))
if codexResult.Modified {
bodyModified = true
disablePatch()
@@ -2037,14 +2071,14 @@ func (s *OpenAIGatewayService) forwardOpenAIPassthrough(
return nil, fmt.Errorf("openai passthrough rejected before upstream: %s", rejectReason)
}
normalizedBody, normalized, err := normalizeOpenAIPassthroughOAuthBody(body)
normalizedBody, normalized, err := normalizeOpenAIPassthroughOAuthBody(body, isOpenAIResponsesCompactPath(c))
if err != nil {
return nil, err
}
if normalized {
body = normalizedBody
reqStream = true
}
reqStream = gjson.GetBytes(body, "stream").Bool()
}
logger.LegacyPrintf("service.openai_gateway",
@@ -2209,6 +2243,7 @@ func (s *OpenAIGatewayService) buildUpstreamRequestOpenAIPassthrough(
targetURL = buildOpenAIResponsesURL(validatedURL)
}
}
targetURL = appendOpenAIResponsesRequestPathSuffix(targetURL, openAIResponsesRequestPathSuffix(c))
req, err := http.NewRequestWithContext(ctx, http.MethodPost, targetURL, bytes.NewReader(body))
if err != nil {
@@ -2242,7 +2277,15 @@ func (s *OpenAIGatewayService) buildUpstreamRequestOpenAIPassthrough(
if chatgptAccountID := account.GetChatGPTAccountID(); chatgptAccountID != "" {
req.Header.Set("chatgpt-account-id", chatgptAccountID)
}
if req.Header.Get("accept") == "" {
if isOpenAIResponsesCompactPath(c) {
req.Header.Set("accept", "application/json")
if req.Header.Get("version") == "" {
req.Header.Set("version", codexCLIVersion)
}
if req.Header.Get("session_id") == "" {
req.Header.Set("session_id", resolveOpenAICompactSessionID(c))
}
} else if req.Header.Get("accept") == "" {
req.Header.Set("accept", "text/event-stream")
}
if req.Header.Get("OpenAI-Beta") == "" {
@@ -2589,6 +2632,7 @@ func (s *OpenAIGatewayService) buildUpstreamRequest(ctx context.Context, c *gin.
default:
targetURL = openaiPlatformAPIURL
}
targetURL = appendOpenAIResponsesRequestPathSuffix(targetURL, openAIResponsesRequestPathSuffix(c))
req, err := http.NewRequestWithContext(ctx, "POST", targetURL, bytes.NewReader(body))
if err != nil {
@@ -2625,7 +2669,17 @@ func (s *OpenAIGatewayService) buildUpstreamRequest(ctx context.Context, c *gin.
} else {
req.Header.Set("originator", "opencode")
}
req.Header.Set("accept", "text/event-stream")
if isOpenAIResponsesCompactPath(c) {
req.Header.Set("accept", "application/json")
if req.Header.Get("version") == "" {
req.Header.Set("version", codexCLIVersion)
}
if req.Header.Get("session_id") == "" {
req.Header.Set("session_id", resolveOpenAICompactSessionID(c))
}
} else {
req.Header.Set("accept", "text/event-stream")
}
if promptCacheKey != "" {
req.Header.Set("conversation_id", promptCacheKey)
req.Header.Set("session_id", promptCacheKey)
@@ -3261,6 +3315,14 @@ func (s *OpenAIGatewayService) handleOAuthSSEToJSON(resp *http.Response, c *gin.
// Correct tool calls in final response
body = s.correctToolCallsInResponseBody(body)
} else {
terminalType, terminalPayload, terminalOK := extractOpenAISSETerminalEvent(bodyText)
if terminalOK && terminalType == "response.failed" {
msg := extractOpenAISSEErrorMessage(terminalPayload)
if msg == "" {
msg = "Upstream compact response failed"
}
return nil, s.writeOpenAINonStreamingProtocolError(resp, c, msg)
}
usage = s.parseSSEUsageFromBody(bodyText)
if originalModel != mappedModel {
bodyText = s.replaceModelInSSEBody(bodyText, mappedModel, originalModel)
@@ -3282,6 +3344,51 @@ func (s *OpenAIGatewayService) handleOAuthSSEToJSON(resp *http.Response, c *gin.
return usage, nil
}
func extractOpenAISSETerminalEvent(body string) (string, []byte, bool) {
lines := strings.Split(body, "\n")
for _, line := range lines {
data, ok := extractOpenAISSEDataLine(line)
if !ok || data == "" || data == "[DONE]" {
continue
}
eventType := strings.TrimSpace(gjson.Get(data, "type").String())
switch eventType {
case "response.completed", "response.done", "response.failed":
return eventType, []byte(data), true
}
}
return "", nil, false
}
func extractOpenAISSEErrorMessage(payload []byte) string {
if len(payload) == 0 {
return ""
}
for _, path := range []string{"response.error.message", "error.message", "message"} {
if msg := strings.TrimSpace(gjson.GetBytes(payload, path).String()); msg != "" {
return sanitizeUpstreamErrorMessage(msg)
}
}
return sanitizeUpstreamErrorMessage(strings.TrimSpace(extractUpstreamErrorMessage(payload)))
}
func (s *OpenAIGatewayService) writeOpenAINonStreamingProtocolError(resp *http.Response, c *gin.Context, message string) error {
message = sanitizeUpstreamErrorMessage(strings.TrimSpace(message))
if message == "" {
message = "Upstream returned an invalid non-streaming response"
}
setOpsUpstreamError(c, http.StatusBadGateway, message, "")
responseheaders.WriteFilteredHeaders(c.Writer.Header(), resp.Header, s.responseHeaderFilter)
c.Writer.Header().Set("Content-Type", "application/json; charset=utf-8")
c.JSON(http.StatusBadGateway, gin.H{
"error": gin.H{
"type": "upstream_error",
"message": message,
},
})
return fmt.Errorf("non-streaming openai protocol error: %s", message)
}
func extractCodexFinalResponse(body string) ([]byte, bool) {
lines := strings.Split(body, "\n")
for _, line := range lines {
@@ -3363,6 +3470,95 @@ func buildOpenAIResponsesURL(base string) string {
return normalized + "/v1/responses"
}
func IsOpenAIResponsesCompactPathForTest(c *gin.Context) bool {
return isOpenAIResponsesCompactPath(c)
}
func OpenAICompactSessionSeedKeyForTest() string {
return openAICompactSessionSeedKey
}
func NormalizeOpenAICompactRequestBodyForTest(body []byte) ([]byte, bool, error) {
return normalizeOpenAICompactRequestBody(body)
}
func isOpenAIResponsesCompactPath(c *gin.Context) bool {
suffix := strings.TrimSpace(openAIResponsesRequestPathSuffix(c))
return suffix == "/compact" || strings.HasPrefix(suffix, "/compact/")
}
func normalizeOpenAICompactRequestBody(body []byte) ([]byte, bool, error) {
if len(body) == 0 {
return body, false, nil
}
normalized := []byte(`{}`)
for _, field := range []string{"model", "input", "instructions", "previous_response_id"} {
value := gjson.GetBytes(body, field)
if !value.Exists() {
continue
}
next, err := sjson.SetRawBytes(normalized, field, []byte(value.Raw))
if err != nil {
return body, false, fmt.Errorf("normalize compact body %s: %w", field, err)
}
normalized = next
}
if bytes.Equal(bytes.TrimSpace(body), bytes.TrimSpace(normalized)) {
return body, false, nil
}
return normalized, true, nil
}
func resolveOpenAICompactSessionID(c *gin.Context) string {
if c != nil {
if sessionID := strings.TrimSpace(c.GetHeader("session_id")); sessionID != "" {
return sessionID
}
if conversationID := strings.TrimSpace(c.GetHeader("conversation_id")); conversationID != "" {
return conversationID
}
if seed, ok := c.Get(openAICompactSessionSeedKey); ok {
if seedStr, ok := seed.(string); ok && strings.TrimSpace(seedStr) != "" {
return strings.TrimSpace(seedStr)
}
}
}
return uuid.NewString()
}
func openAIResponsesRequestPathSuffix(c *gin.Context) string {
if c == nil || c.Request == nil || c.Request.URL == nil {
return ""
}
normalizedPath := strings.TrimRight(strings.TrimSpace(c.Request.URL.Path), "/")
if normalizedPath == "" {
return ""
}
idx := strings.LastIndex(normalizedPath, "/responses")
if idx < 0 {
return ""
}
suffix := normalizedPath[idx+len("/responses"):]
if suffix == "" || suffix == "/" {
return ""
}
if !strings.HasPrefix(suffix, "/") {
return ""
}
return suffix
}
func appendOpenAIResponsesRequestPathSuffix(baseURL, suffix string) string {
trimmedBase := strings.TrimRight(strings.TrimSpace(baseURL), "/")
trimmedSuffix := strings.TrimSpace(suffix)
if trimmedBase == "" || trimmedSuffix == "" {
return trimmedBase
}
return trimmedBase + trimmedSuffix
}
func (s *OpenAIGatewayService) replaceModelInResponseBody(body []byte, fromModel, toModel string) []byte {
// 使用 gjson/sjson 精确替换 model 字段，避免全量 JSON 反序列化
if m := gjson.GetBytes(body, "model"); m.Exists() && m.Str == fromModel {
@@ -3413,10 +3609,18 @@ func (s *OpenAIGatewayService) RecordUsage(ctx context.Context, input *OpenAIRec
// Get rate multiplier
multiplier := s.cfg.Default.RateMultiplier
if apiKey.GroupID != nil && apiKey.Group != nil {
multiplier = apiKey.Group.RateMultiplier
resolver := s.userGroupRateResolver
if resolver == nil {
resolver = newUserGroupRateResolver(nil, nil, resolveUserGroupRateCacheTTL(s.cfg), nil, "service.openai_gateway")
}
multiplier = resolver.Resolve(ctx, user.ID, *apiKey.GroupID, apiKey.Group.RateMultiplier)
}
cost, err := s.billingService.CalculateCost(result.Model, tokens, multiplier)
billingModel := result.Model
if result.BillingModel != "" {
billingModel = result.BillingModel
}
cost, err := s.billingService.CalculateCost(billingModel, tokens, multiplier)
if err != nil {
cost = &CostBreakdown{ActualCost: 0}
}
@@ -3436,7 +3640,7 @@ func (s *OpenAIGatewayService) RecordUsage(ctx context.Context, input *OpenAIRec
APIKeyID: apiKey.ID,
AccountID: account.ID,
RequestID: result.RequestID,
Model: result.Model,
Model: billingModel,
ReasoningEffort: result.ReasoningEffort,
InputTokens: actualInputTokens,
OutputTokens: result.Usage.OutputTokens,
@@ -3681,6 +3885,15 @@ func (s *OpenAIGatewayService) updateCodexUsageSnapshot(ctx context.Context, acc
}()
}
func (s *OpenAIGatewayService) UpdateCodexUsageSnapshotFromHeaders(ctx context.Context, accountID int64, headers http.Header) {
if accountID <= 0 || headers == nil {
return
}
if snapshot := ParseCodexRateLimitHeaders(headers); snapshot != nil {
s.updateCodexUsageSnapshot(ctx, accountID, snapshot)
}
}
func getOpenAIReasoningEffortFromReqBody(reqBody map[string]any) (value string, present bool) {
if reqBody == nil {
return "", false
@@ -3739,8 +3952,8 @@ func extractOpenAIRequestMetaFromBody(body []byte) (model string, stream bool, p
}
// normalizeOpenAIPassthroughOAuthBody 将透传 OAuth 请求体收敛为旧链路关键行为：
// 1) store=false 2) stream=true
func normalizeOpenAIPassthroughOAuthBody(body []byte) ([]byte, bool, error) {
// 1) store=false 2) 非 compact 保持 stream=true；compact 强制 stream=false
func normalizeOpenAIPassthroughOAuthBody(body []byte, compact bool) ([]byte, bool, error) {
if len(body) == 0 {
return body, false, nil
}
@@ -3748,22 +3961,40 @@ func normalizeOpenAIPassthroughOAuthBody(body []byte) ([]byte, bool, error) {
normalized := body
changed := false
if store := gjson.GetBytes(normalized, "store"); !store.Exists() || store.Type != gjson.False {
next, err := sjson.SetBytes(normalized, "store", false)
if err != nil {
return body, false, fmt.Errorf("normalize passthrough body store=false: %w", err)
if compact {
if store := gjson.GetBytes(normalized, "store"); store.Exists() {
next, err := sjson.DeleteBytes(normalized, "store")
if err != nil {
return body, false, fmt.Errorf("normalize passthrough body delete store: %w", err)
}
normalized = next
changed = true
}
normalized = next
changed = true
}
if stream := gjson.GetBytes(normalized, "stream"); !stream.Exists() || stream.Type != gjson.True {
next, err := sjson.SetBytes(normalized, "stream", true)
if err != nil {
return body, false, fmt.Errorf("normalize passthrough body stream=true: %w", err)
if stream := gjson.GetBytes(normalized, "stream"); stream.Exists() {
next, err := sjson.DeleteBytes(normalized, "stream")
if err != nil {
return body, false, fmt.Errorf("normalize passthrough body delete stream: %w", err)
}
normalized = next
changed = true
}
} else {
if store := gjson.GetBytes(normalized, "store"); !store.Exists() || store.Type != gjson.False {
next, err := sjson.SetBytes(normalized, "store", false)
if err != nil {
return body, false, fmt.Errorf("normalize passthrough body store=false: %w", err)
}
normalized = next
changed = true
}
if stream := gjson.GetBytes(normalized, "stream"); !stream.Exists() || stream.Type != gjson.True {
next, err := sjson.SetBytes(normalized, "stream", true)
if err != nil {
return body, false, fmt.Errorf("normalize passthrough body stream=true: %w", err)
}
normalized = next
changed = true
}
normalized = next
changed = true
}
return normalized, changed, nil
backend/internal/service/openai_gateway_service_test.go
+149
View File
@@ -28,6 +28,22 @@ type stubOpenAIAccountRepo struct {
accounts []Account
}
type snapshotUpdateAccountRepo struct {
stubOpenAIAccountRepo
updateExtraCalls chan map[string]any
}
func (r *snapshotUpdateAccountRepo) UpdateExtra(ctx context.Context, id int64, updates map[string]any) error {
if r.updateExtraCalls != nil {
copied := make(map[string]any, len(updates))
for k, v := range updates {
copied[k] = v
}
r.updateExtraCalls <- copied
}
return nil
}
func (r stubOpenAIAccountRepo) GetByID(ctx context.Context, id int64) (*Account, error) {
for i := range r.accounts {
if r.accounts[i].ID == id {
@@ -1261,8 +1277,117 @@ func TestOpenAIValidateUpstreamBaseURLEnabledEnforcesAllowlist(t *testing.T) {
}
}
func TestOpenAIUpdateCodexUsageSnapshotFromHeaders(t *testing.T) {
repo := &snapshotUpdateAccountRepo{updateExtraCalls: make(chan map[string]any, 1)}
svc := &OpenAIGatewayService{accountRepo: repo}
headers := http.Header{}
headers.Set("x-codex-primary-used-percent", "12")
headers.Set("x-codex-secondary-used-percent", "34")
headers.Set("x-codex-primary-window-minutes", "300")
headers.Set("x-codex-secondary-window-minutes", "10080")
headers.Set("x-codex-primary-reset-after-seconds", "600")
headers.Set("x-codex-secondary-reset-after-seconds", "86400")
svc.UpdateCodexUsageSnapshotFromHeaders(context.Background(), 123, headers)
select {
case updates := <-repo.updateExtraCalls:
require.Equal(t, 12.0, updates["codex_5h_used_percent"])
require.Equal(t, 34.0, updates["codex_7d_used_percent"])
require.Equal(t, 600, updates["codex_5h_reset_after_seconds"])
require.Equal(t, 86400, updates["codex_7d_reset_after_seconds"])
case <-time.After(2 * time.Second):
t.Fatal("expected UpdateExtra to be called")
}
}
func TestOpenAIResponsesRequestPathSuffix(t *testing.T) {
gin.SetMode(gin.TestMode)
rec := httptest.NewRecorder()
c, _ := gin.CreateTestContext(rec)
tests := []struct {
name string
path string
want string
}{
{name: "exact v1 responses", path: "/v1/responses", want: ""},
{name: "compact v1 responses", path: "/v1/responses/compact", want: "/compact"},
{name: "compact alias responses", path: "/responses/compact/", want: "/compact"},
{name: "nested suffix", path: "/openai/v1/responses/compact/detail", want: "/compact/detail"},
{name: "unrelated path", path: "/v1/chat/completions", want: ""},
}
for _, tt := range tests {
t.Run(tt.name, func(t *testing.T) {
c.Request = httptest.NewRequest(http.MethodPost, tt.path, nil)
require.Equal(t, tt.want, openAIResponsesRequestPathSuffix(c))
})
}
}
func TestOpenAIBuildUpstreamRequestOpenAIPassthroughPreservesCompactPath(t *testing.T) {
gin.SetMode(gin.TestMode)
rec := httptest.NewRecorder()
c, _ := gin.CreateTestContext(rec)
c.Request = httptest.NewRequest(http.MethodPost, "/v1/responses/compact", bytes.NewReader([]byte(`{"model":"gpt-5"}`)))
svc := &OpenAIGatewayService{}
account := &Account{Type: AccountTypeOAuth}
req, err := svc.buildUpstreamRequestOpenAIPassthrough(c.Request.Context(), c, account, []byte(`{"model":"gpt-5"}`), "token")
require.NoError(t, err)
require.Equal(t, chatgptCodexURL+"/compact", req.URL.String())
require.Equal(t, "application/json", req.Header.Get("Accept"))
require.Equal(t, codexCLIVersion, req.Header.Get("Version"))
require.NotEmpty(t, req.Header.Get("Session_Id"))
}
func TestOpenAIBuildUpstreamRequestCompactForcesJSONAcceptForOAuth(t *testing.T) {
gin.SetMode(gin.TestMode)
rec := httptest.NewRecorder()
c, _ := gin.CreateTestContext(rec)
c.Request = httptest.NewRequest(http.MethodPost, "/v1/responses/compact", bytes.NewReader([]byte(`{"model":"gpt-5"}`)))
svc := &OpenAIGatewayService{}
account := &Account{
Type: AccountTypeOAuth,
Credentials: map[string]any{"chatgpt_account_id": "chatgpt-acc"},
}
req, err := svc.buildUpstreamRequest(c.Request.Context(), c, account, []byte(`{"model":"gpt-5"}`), "token", false, "", true)
require.NoError(t, err)
require.Equal(t, chatgptCodexURL+"/compact", req.URL.String())
require.Equal(t, "application/json", req.Header.Get("Accept"))
require.Equal(t, codexCLIVersion, req.Header.Get("Version"))
require.NotEmpty(t, req.Header.Get("Session_Id"))
}
func TestOpenAIBuildUpstreamRequestPreservesCompactPathForAPIKeyBaseURL(t *testing.T) {
gin.SetMode(gin.TestMode)
rec := httptest.NewRecorder()
c, _ := gin.CreateTestContext(rec)
c.Request = httptest.NewRequest(http.MethodPost, "/responses/compact", bytes.NewReader([]byte(`{"model":"gpt-5"}`)))
svc := &OpenAIGatewayService{cfg: &config.Config{
Security: config.SecurityConfig{
URLAllowlist: config.URLAllowlistConfig{Enabled: false},
},
}}
account := &Account{
Type: AccountTypeAPIKey,
Platform: PlatformOpenAI,
Credentials: map[string]any{"base_url": "https://example.com/v1"},
}
req, err := svc.buildUpstreamRequest(c.Request.Context(), c, account, []byte(`{"model":"gpt-5"}`), "token", false, "", false)
require.NoError(t, err)
require.Equal(t, "https://example.com/v1/responses/compact", req.URL.String())
}
// ==================== P1-08 修复：model 替换性能优化测试 ====================
// ==================== P1-08 修复：model 替换性能优化测试 =============
func TestReplaceModelInSSELine(t *testing.T) {
svc := &OpenAIGatewayService{}
@@ -1589,3 +1714,27 @@ func TestHandleOAuthSSEToJSON_NoFinalResponseKeepsSSEBody(t *testing.T) {
require.Contains(t, rec.Header().Get("Content-Type"), "text/event-stream")
require.Contains(t, rec.Body.String(), `data: {"type":"response.in_progress"`)
}
func TestHandleOAuthSSEToJSON_ResponseFailedReturnsProtocolError(t *testing.T) {
gin.SetMode(gin.TestMode)
rec := httptest.NewRecorder()
c, _ := gin.CreateTestContext(rec)
c.Request = httptest.NewRequest(http.MethodPost, "/", nil)
svc := &OpenAIGatewayService{cfg: &config.Config{}}
resp := &http.Response{
StatusCode: http.StatusOK,
Header: http.Header{"Content-Type": []string{"text/event-stream"}},
}
body := []byte(strings.Join([]string{
`data: {"type":"response.failed","error":{"message":"upstream rejected request"}}`,
`data: [DONE]`,
}, "\n"))
usage, err := svc.handleOAuthSSEToJSON(resp, c, body, "gpt-4o", "gpt-4o")
require.Nil(t, usage)
require.Error(t, err)
require.Equal(t, http.StatusBadGateway, rec.Code)
require.Contains(t, rec.Body.String(), "upstream rejected request")
require.Contains(t, rec.Header().Get("Content-Type"), "application/json")
}
backend/internal/service/openai_oauth_passthrough_test.go
+54
View File
@@ -236,6 +236,60 @@ func TestOpenAIGatewayService_OAuthPassthrough_StreamKeepsToolNameAndBodyNormali
require.NotContains(t, body, "\"name\":\"edit\"")
}
func TestOpenAIGatewayService_OAuthPassthrough_CompactUsesJSONAndKeepsNonStreaming(t *testing.T) {
gin.SetMode(gin.TestMode)
rec := httptest.NewRecorder()
c, _ := gin.CreateTestContext(rec)
c.Request = httptest.NewRequest(http.MethodPost, "/v1/responses/compact", bytes.NewReader(nil))
c.Request.Header.Set("User-Agent", "codex_cli_rs/0.1.0")
c.Request.Header.Set("Content-Type", "application/json")
originalBody := []byte(`{"model":"gpt-5.1-codex","stream":true,"store":true,"instructions":"local-test-instructions","input":[{"type":"text","text":"compact me"}]}`)
resp := &http.Response{
StatusCode: http.StatusOK,
Header: http.Header{"Content-Type": []string{"application/json"}, "x-request-id": []string{"rid-compact"}},
Body: io.NopCloser(strings.NewReader(`{"id":"cmp_123","usage":{"input_tokens":11,"output_tokens":22}}`)),
}
upstream := &httpUpstreamRecorder{resp: resp}
svc := &OpenAIGatewayService{
cfg: &config.Config{Gateway: config.GatewayConfig{ForceCodexCLI: false}},
httpUpstream: upstream,
}
account := &Account{
ID: 123,
Name: "acc",
Platform: PlatformOpenAI,
Type: AccountTypeOAuth,
Concurrency: 1,
Credentials: map[string]any{"access_token": "oauth-token", "chatgpt_account_id": "chatgpt-acc"},
Extra: map[string]any{"openai_passthrough": true},
Status: StatusActive,
Schedulable: true,
RateMultiplier: f64p(1),
}
result, err := svc.Forward(context.Background(), c, account, originalBody)
require.NoError(t, err)
require.NotNil(t, result)
require.False(t, result.Stream)
require.False(t, gjson.GetBytes(upstream.lastBody, "store").Exists())
require.False(t, gjson.GetBytes(upstream.lastBody, "stream").Exists())
require.Equal(t, "gpt-5.1-codex", gjson.GetBytes(upstream.lastBody, "model").String())
require.Equal(t, "compact me", gjson.GetBytes(upstream.lastBody, "input.0.text").String())
require.Equal(t, "local-test-instructions", strings.TrimSpace(gjson.GetBytes(upstream.lastBody, "instructions").String()))
require.Equal(t, "application/json", upstream.lastReq.Header.Get("Accept"))
require.Equal(t, codexCLIVersion, upstream.lastReq.Header.Get("Version"))
require.NotEmpty(t, upstream.lastReq.Header.Get("Session_Id"))
require.Equal(t, "chatgpt.com", upstream.lastReq.Host)
require.Equal(t, "chatgpt-acc", upstream.lastReq.Header.Get("chatgpt-account-id"))
require.Contains(t, rec.Body.String(), `"id":"cmp_123"`)
}
func TestOpenAIGatewayService_OAuthPassthrough_CodexMissingInstructionsRejectedBeforeUpstream(t *testing.T) {
gin.SetMode(gin.TestMode)
logSink, restore := captureStructuredLog(t)
backend/internal/service/openai_ws_forwarder.go
+2
View File
@@ -2309,6 +2309,7 @@ func (s *OpenAIGatewayService) forwardOpenAIWSV2(
ReasoningEffort: extractOpenAIReasoningEffort(reqBody, originalModel),
Stream: reqStream,
OpenAIWSMode: true,
ResponseHeaders: lease.HandshakeHeaders(),
Duration: time.Since(startTime),
FirstTokenMs: firstTokenMs,
}, nil
@@ -2919,6 +2920,7 @@ func (s *OpenAIGatewayService) ProxyResponsesWebSocketFromClient(
ReasoningEffort: extractOpenAIReasoningEffortFromBody(payload, originalModel),
Stream: reqStream,
OpenAIWSMode: true,
ResponseHeaders: lease.HandshakeHeaders(),
Duration: time.Since(turnStart),
FirstTokenMs: firstTokenMs,
}, nil
backend/internal/service/openai_ws_pool.go
+7
View File
@@ -126,6 +126,13 @@ func (l *openAIWSConnLease) HandshakeHeader(name string) string {
return l.conn.handshakeHeader(name)
}
func (l *openAIWSConnLease) HandshakeHeaders() http.Header {
if l == nil || l.conn == nil {
return nil
}
return cloneHeader(l.conn.handshakeHeaders)
}
func (l *openAIWSConnLease) IsPrewarmed() bool {
if l == nil || l.conn == nil {
return false
backend/internal/service/openai_ws_protocol_forward_test.go
+1
View File
@@ -391,6 +391,7 @@ func TestNewOpenAIGatewayService_InitializesOpenAIWSResolver(t *testing.T) {
nil,
nil,
nil,
nil,
cfg,
nil,
nil,
backend/internal/service/openai_ws_v2_passthrough_adapter.go
+12 -10
View File
@@ -177,11 +177,12 @@ func (s *OpenAIGatewayService) proxyResponsesWebSocketV2Passthrough(
CacheCreationInputTokens: turn.Usage.CacheCreationInputTokens,
CacheReadInputTokens: turn.Usage.CacheReadInputTokens,
},
Model: turn.RequestModel,
Stream: true,
OpenAIWSMode: true,
Duration: turn.Duration,
FirstTokenMs: turn.FirstTokenMs,
Model: turn.RequestModel,
Stream: true,
OpenAIWSMode: true,
ResponseHeaders: cloneHeader(handshakeHeaders),
Duration: turn.Duration,
FirstTokenMs: turn.FirstTokenMs,
}
logOpenAIWSV2Passthrough(
"relay_turn_completed account_id=%d turn=%d request_id=%s terminal_event=%s duration_ms=%d first_token_ms=%d input_tokens=%d output_tokens=%d cache_read_tokens=%d",
@@ -223,11 +224,12 @@ func (s *OpenAIGatewayService) proxyResponsesWebSocketV2Passthrough(
CacheCreationInputTokens: relayResult.Usage.CacheCreationInputTokens,
CacheReadInputTokens: relayResult.Usage.CacheReadInputTokens,
},
Model: relayResult.RequestModel,
Stream: true,
OpenAIWSMode: true,
Duration: relayResult.Duration,
FirstTokenMs: relayResult.FirstTokenMs,
Model: relayResult.RequestModel,
Stream: true,
OpenAIWSMode: true,
ResponseHeaders: cloneHeader(handshakeHeaders),
Duration: relayResult.Duration,
FirstTokenMs: relayResult.FirstTokenMs,
}
turnCount := int(completedTurns.Load())
backend/internal/service/ratelimit_service.go
+15
View File
@@ -971,12 +971,27 @@ func (s *RateLimitService) UpdateSessionWindow(ctx context.Context, account *Acc
windowStart = &start
windowEnd = &end
slog.Info("account_session_window_initialized", "account_id", account.ID, "window_start", start, "window_end", end, "status", status)
// 窗口重置时清除旧的 utilization，避免残留上个窗口的数据
_ = s.accountRepo.UpdateExtra(ctx, account.ID, map[string]any{
"session_window_utilization": nil,
})
}
if err := s.accountRepo.UpdateSessionWindow(ctx, account.ID, windowStart, windowEnd, status); err != nil {
slog.Warn("session_window_update_failed", "account_id", account.ID, "error", err)
}
// 存储真实的 utilization 值（0-1 小数），供 estimateSetupTokenUsage 使用
if utilStr := headers.Get("anthropic-ratelimit-unified-5h-utilization"); utilStr != "" {
if util, err := strconv.ParseFloat(utilStr, 64); err == nil {
if err := s.accountRepo.UpdateExtra(ctx, account.ID, map[string]any{
"session_window_utilization": util,
}); err != nil {
slog.Warn("session_window_utilization_update_failed", "account_id", account.ID, "error", err)
}
}
}
// 如果状态为allowed且之前有限流，说明窗口已重置，清除限流状态
if status == "allowed" && account.IsRateLimited() {
if err := s.ClearRateLimit(ctx, account.ID); err != nil {
backend/internal/service/user_group_rate_resolver.go
+103
View File
@@ -0,0 +1,103 @@
package service
import (
"context"
"fmt"
"time"
"github.com/Wei-Shaw/sub2api/internal/pkg/logger"
gocache "github.com/patrickmn/go-cache"
"golang.org/x/sync/singleflight"
)
type userGroupRateResolver struct {
repo UserGroupRateRepository
cache *gocache.Cache
cacheTTL time.Duration
sf *singleflight.Group
logComponent string
}
func newUserGroupRateResolver(repo UserGroupRateRepository, cache *gocache.Cache, cacheTTL time.Duration, sf *singleflight.Group, logComponent string) *userGroupRateResolver {
if cacheTTL <= 0 {
cacheTTL = defaultUserGroupRateCacheTTL
}
if cache == nil {
cache = gocache.New(cacheTTL, time.Minute)
}
if logComponent == "" {
logComponent = "service.gateway"
}
if sf == nil {
sf = &singleflight.Group{}
}
return &userGroupRateResolver{
repo: repo,
cache: cache,
cacheTTL: cacheTTL,
sf: sf,
logComponent: logComponent,
}
}
func (r *userGroupRateResolver) Resolve(ctx context.Context, userID, groupID int64, groupDefaultMultiplier float64) float64 {
if r == nil || userID <= 0 || groupID <= 0 {
return groupDefaultMultiplier
}
key := fmt.Sprintf("%d:%d", userID, groupID)
if r.cache != nil {
if cached, ok := r.cache.Get(key); ok {
if multiplier, castOK := cached.(float64); castOK {
userGroupRateCacheHitTotal.Add(1)
return multiplier
}
}
}
if r.repo == nil {
return groupDefaultMultiplier
}
userGroupRateCacheMissTotal.Add(1)
value, err, shared := r.sf.Do(key, func() (any, error) {
if r.cache != nil {
if cached, ok := r.cache.Get(key); ok {
if multiplier, castOK := cached.(float64); castOK {
userGroupRateCacheHitTotal.Add(1)
return multiplier, nil
}
}
}
userGroupRateCacheLoadTotal.Add(1)
userRate, repoErr := r.repo.GetByUserAndGroup(ctx, userID, groupID)
if repoErr != nil {
return nil, repoErr
}
multiplier := groupDefaultMultiplier
if userRate != nil {
multiplier = *userRate
}
if r.cache != nil {
r.cache.Set(key, multiplier, r.cacheTTL)
}
return multiplier, nil
})
if shared {
userGroupRateCacheSFSharedTotal.Add(1)
}
if err != nil {
userGroupRateCacheFallbackTotal.Add(1)
logger.LegacyPrintf(r.logComponent, "get user group rate failed, fallback to group default: user=%d group=%d err=%v", userID, groupID, err)
return groupDefaultMultiplier
}
multiplier, ok := value.(float64)
if !ok {
userGroupRateCacheFallbackTotal.Add(1)
return groupDefaultMultiplier
}
return multiplier
}
backend/internal/service/user_group_rate_resolver_test.go
+83
View File
@@ -0,0 +1,83 @@
package service
import (
"context"
"testing"
"time"
gocache "github.com/patrickmn/go-cache"
"github.com/stretchr/testify/require"
)
type userGroupRateResolverRepoStub struct {
UserGroupRateRepository
rate *float64
err error
calls int
}
func (s *userGroupRateResolverRepoStub) GetByUserAndGroup(ctx context.Context, userID, groupID int64) (*float64, error) {
s.calls++
if s.err != nil {
return nil, s.err
}
return s.rate, nil
}
func TestNewUserGroupRateResolver_Defaults(t *testing.T) {
resolver := newUserGroupRateResolver(nil, nil, 0, nil, "")
require.NotNil(t, resolver)
require.NotNil(t, resolver.cache)
require.Equal(t, defaultUserGroupRateCacheTTL, resolver.cacheTTL)
require.NotNil(t, resolver.sf)
require.Equal(t, "service.gateway", resolver.logComponent)
}
func TestUserGroupRateResolverResolve_FallbackForNilResolverAndInvalidIDs(t *testing.T) {
var nilResolver *userGroupRateResolver
require.Equal(t, 1.4, nilResolver.Resolve(context.Background(), 101, 202, 1.4))
resolver := newUserGroupRateResolver(nil, nil, time.Second, nil, "service.test")
require.Equal(t, 1.4, resolver.Resolve(context.Background(), 0, 202, 1.4))
require.Equal(t, 1.4, resolver.Resolve(context.Background(), 101, 0, 1.4))
}
func TestUserGroupRateResolverResolve_InvalidCacheEntryLoadsRepoAndCaches(t *testing.T) {
resetGatewayHotpathStatsForTest()
rate := 1.7
repo := &userGroupRateResolverRepoStub{rate: &rate}
cache := gocache.New(time.Minute, time.Minute)
cache.Set("101:202", "bad-cache", time.Minute)
resolver := newUserGroupRateResolver(repo, cache, time.Minute, nil, "service.test")
got := resolver.Resolve(context.Background(), 101, 202, 1.2)
require.Equal(t, rate, got)
require.Equal(t, 1, repo.calls)
cached, ok := cache.Get("101:202")
require.True(t, ok)
require.Equal(t, rate, cached)
hit, miss, load, _, fallback := GatewayUserGroupRateCacheStats()
require.Equal(t, int64(0), hit)
require.Equal(t, int64(1), miss)
require.Equal(t, int64(1), load)
require.Equal(t, int64(0), fallback)
}
func TestGatewayServiceGetUserGroupRateMultiplier_FallbacksAndUsesExistingResolver(t *testing.T) {
var nilSvc *GatewayService
require.Equal(t, 1.3, nilSvc.getUserGroupRateMultiplier(context.Background(), 101, 202, 1.3))
rate := 1.9
repo := &userGroupRateResolverRepoStub{rate: &rate}
resolver := newUserGroupRateResolver(repo, nil, time.Minute, nil, "service.gateway")
svc := &GatewayService{userGroupRateResolver: resolver}
got := svc.getUserGroupRateMultiplier(context.Background(), 101, 202, 1.2)
require.Equal(t, rate, got)
require.Equal(t, 1, repo.calls)
}
backend/internal/web/embed_on.go
+15 -16
View File
@@ -83,14 +83,7 @@ func (s *FrontendServer) Middleware() gin.HandlerFunc {
path := c.Request.URL.Path
// Skip API routes
if strings.HasPrefix(path, "/api/") ||
strings.HasPrefix(path, "/v1/") ||
strings.HasPrefix(path, "/v1beta/") ||
strings.HasPrefix(path, "/sora/") ||
strings.HasPrefix(path, "/antigravity/") ||
strings.HasPrefix(path, "/setup/") ||
path == "/health" ||
path == "/responses" {
if shouldBypassEmbeddedFrontend(path) {
c.Next()
return
}
@@ -207,14 +200,7 @@ func ServeEmbeddedFrontend() gin.HandlerFunc {
return func(c *gin.Context) {
path := c.Request.URL.Path
if strings.HasPrefix(path, "/api/") ||
strings.HasPrefix(path, "/v1/") ||
strings.HasPrefix(path, "/v1beta/") ||
strings.HasPrefix(path, "/sora/") ||
strings.HasPrefix(path, "/antigravity/") ||
strings.HasPrefix(path, "/setup/") ||
path == "/health" ||
path == "/responses" {
if shouldBypassEmbeddedFrontend(path) {
c.Next()
return
}
@@ -235,6 +221,19 @@ func ServeEmbeddedFrontend() gin.HandlerFunc {
}
}
func shouldBypassEmbeddedFrontend(path string) bool {
trimmed := strings.TrimSpace(path)
return strings.HasPrefix(trimmed, "/api/") ||
strings.HasPrefix(trimmed, "/v1/") ||
strings.HasPrefix(trimmed, "/v1beta/") ||
strings.HasPrefix(trimmed, "/sora/") ||
strings.HasPrefix(trimmed, "/antigravity/") ||
strings.HasPrefix(trimmed, "/setup/") ||
trimmed == "/health" ||
trimmed == "/responses" ||
strings.HasPrefix(trimmed, "/responses/")
}
func serveIndexHTML(c *gin.Context, fsys fs.FS) {
file, err := fsys.Open("index.html")
if err != nil {
backend/internal/web/embed_test.go
+28
View File
@@ -367,6 +367,7 @@ func TestFrontendServer_Middleware(t *testing.T) {
"/setup/init",
"/health",
"/responses",
"/responses/compact",
}
for _, path := range apiPaths {
@@ -388,6 +389,32 @@ func TestFrontendServer_Middleware(t *testing.T) {
}
})
t.Run("skips_responses_compact_post_routes", func(t *testing.T) {
provider := &mockSettingsProvider{
settings: map[string]string{"test": "value"},
}
server, err := NewFrontendServer(provider)
require.NoError(t, err)
router := gin.New()
router.Use(server.Middleware())
nextCalled := false
router.POST("/responses/compact", func(c *gin.Context) {
nextCalled = true
c.String(http.StatusOK, `{"ok":true}`)
})
w := httptest.NewRecorder()
req := httptest.NewRequest(http.MethodPost, "/responses/compact", strings.NewReader(`{"model":"gpt-5"}`))
req.Header.Set("Content-Type", "application/json")
router.ServeHTTP(w, req)
assert.True(t, nextCalled, "next handler should be called for compact API route")
assert.Equal(t, http.StatusOK, w.Code)
assert.JSONEq(t, `{"ok":true}`, w.Body.String())
})
t.Run("serves_index_for_spa_routes", func(t *testing.T) {
provider := &mockSettingsProvider{
settings: map[string]string{"test": "value"},
@@ -543,6 +570,7 @@ func TestServeEmbeddedFrontend(t *testing.T) {
"/setup/init",
"/health",
"/responses",
"/responses/compact",
}
for _, path := range apiPaths {
deploy/Dockerfile
+1 -1
View File
@@ -7,7 +7,7 @@
# =============================================================================
ARG NODE_IMAGE=node:24-alpine
ARG GOLANG_IMAGE=golang:1.25.5-alpine
ARG GOLANG_IMAGE=golang:1.26.1-alpine
ARG ALPINE_IMAGE=alpine:3.20
ARG GOPROXY=https://goproxy.cn,direct
ARG GOSUMDB=sum.golang.google.cn
frontend/src/components/account/AccountUsageCell.vue
+38
View File
@@ -90,6 +90,36 @@
color="emerald"
/>
</div>
<div v-else-if="loading" class="space-y-1.5">
<div class="flex items-center gap-1">
<div class="h-3 w-[32px] animate-pulse rounded bg-gray-200 dark:bg-gray-700"></div>
<div class="h-1.5 w-8 animate-pulse rounded-full bg-gray-200 dark:bg-gray-700"></div>
<div class="h-3 w-[32px] animate-pulse rounded bg-gray-200 dark:bg-gray-700"></div>
</div>
<div class="flex items-center gap-1">
<div class="h-3 w-[32px] animate-pulse rounded bg-gray-200 dark:bg-gray-700"></div>
<div class="h-1.5 w-8 animate-pulse rounded-full bg-gray-200 dark:bg-gray-700"></div>
<div class="h-3 w-[32px] animate-pulse rounded bg-gray-200 dark:bg-gray-700"></div>
</div>
</div>
<div v-else-if="hasOpenAIUsageFallback" class="space-y-1">
<UsageProgressBar
v-if="usageInfo?.five_hour"
label="5h"
:utilization="usageInfo.five_hour.utilization"
:resets-at="usageInfo.five_hour.resets_at"
:window-stats="usageInfo.five_hour.window_stats"
color="indigo"
/>
<UsageProgressBar
v-if="usageInfo?.seven_day"
label="7d"
:utilization="usageInfo.seven_day.utilization"
:resets-at="usageInfo.seven_day.resets_at"
:window-stats="usageInfo.seven_day.window_stats"
color="emerald"
/>
</div>
<div v-else class="text-xs text-gray-400">-</div>
</template>
@@ -317,6 +347,9 @@ const shouldFetchUsage = computed(() => {
if (props.account.platform === 'antigravity') {
return props.account.type === 'oauth'
}
if (props.account.platform === 'openai') {
return props.account.type === 'oauth'
}
return false
})
@@ -339,6 +372,11 @@ const hasCodexUsage = computed(() => {
return codex5hWindow.value.usedPercent !== null || codex7dWindow.value.usedPercent !== null
})
const hasOpenAIUsageFallback = computed(() => {
if (props.account.platform !== 'openai' || props.account.type !== 'oauth') return false
return !!usageInfo.value?.five_hour || !!usageInfo.value?.seven_day
})
const codex5hUsedPercent = computed(() => codex5hWindow.value.usedPercent)
const codex5hResetAt = computed(() => codex5hWindow.value.resetAt)
const codex7dUsedPercent = computed(() => codex7dWindow.value.usedPercent)
frontend/src/components/account/CreateAccountModal.vue
+152
View File
@@ -1230,6 +1230,142 @@
<!-- API Key 账号配额限制 -->
<QuotaLimitCard v-if="form.type === 'apikey'" :totalLimit="editQuotaLimit" :dailyLimit="editQuotaDailyLimit" :weeklyLimit="editQuotaWeeklyLimit" @update:totalLimit="editQuotaLimit = $event" @update:dailyLimit="editQuotaDailyLimit = $event" @update:weeklyLimit="editQuotaWeeklyLimit = $event" />
<!-- OpenAI OAuth Model Mapping (OAuth 类型没有 apikey 容器，需要独立的模型映射区域) -->
<div
v-if="form.platform === 'openai' && accountCategory === 'oauth-based'"
class="border-t border-gray-200 pt-4 dark:border-dark-600"
>
<label class="input-label">{{ t('admin.accounts.modelRestriction') }}</label>
<div
v-if="isOpenAIModelRestrictionDisabled"
class="mb-3 rounded-lg bg-amber-50 p-3 dark:bg-amber-900/20"
>
<p class="text-xs text-amber-700 dark:text-amber-400">
{{ t('admin.accounts.openai.modelRestrictionDisabledByPassthrough') }}
</p>
</div>
<template v-else>
<!-- Mode Toggle -->
<div class="mb-4 flex gap-2">
<button
type="button"
@click="modelRestrictionMode = 'whitelist'"
:class="[
'flex-1 rounded-lg px-4 py-2 text-sm font-medium transition-all',
modelRestrictionMode === 'whitelist'
? 'bg-primary-100 text-primary-700 dark:bg-primary-900/30 dark:text-primary-400'
: 'bg-gray-100 text-gray-600 hover:bg-gray-200 dark:bg-dark-600 dark:text-gray-400 dark:hover:bg-dark-500'
]"
>
{{ t('admin.accounts.modelWhitelist') }}
</button>
<button
type="button"
@click="modelRestrictionMode = 'mapping'"
:class="[
'flex-1 rounded-lg px-4 py-2 text-sm font-medium transition-all',
modelRestrictionMode === 'mapping'
? 'bg-purple-100 text-purple-700 dark:bg-purple-900/30 dark:text-purple-400'
: 'bg-gray-100 text-gray-600 hover:bg-gray-200 dark:bg-dark-600 dark:text-gray-400 dark:hover:bg-dark-500'
]"
>
{{ t('admin.accounts.modelMapping') }}
</button>
</div>
<!-- Whitelist Mode -->
<div v-if="modelRestrictionMode === 'whitelist'">
<ModelWhitelistSelector v-model="allowedModels" :platform="form.platform" />
<p class="text-xs text-gray-500 dark:text-gray-400">
{{ t('admin.accounts.selectedModels', { count: allowedModels.length }) }}
<span v-if="allowedModels.length === 0">{{
t('admin.accounts.supportsAllModels')
}}</span>
</p>
</div>
<!-- Mapping Mode -->
<div v-else>
<div class="mb-3 rounded-lg bg-purple-50 p-3 dark:bg-purple-900/20">
<p class="text-xs text-purple-700 dark:text-purple-400">
{{ t('admin.accounts.mapRequestModels') }}
</p>
</div>
<div v-if="modelMappings.length > 0" class="mb-3 space-y-2">
<div
v-for="(mapping, index) in modelMappings"
:key="'oauth-' + getModelMappingKey(mapping)"
class="flex items-center gap-2"
>
<input
v-model="mapping.from"
type="text"
class="input flex-1"
:placeholder="t('admin.accounts.requestModel')"
/>
<svg
class="h-4 w-4 flex-shrink-0 text-gray-400"
fill="none"
viewBox="0 0 24 24"
stroke="currentColor"
>
<path
stroke-linecap="round"
stroke-linejoin="round"
stroke-width="2"
d="M14 5l7 7m0 0l-7 7m7-7H3"
/>
</svg>
<input
v-model="mapping.to"
type="text"
class="input flex-1"
:placeholder="t('admin.accounts.actualModel')"
/>
<button
type="button"
@click="removeModelMapping(index)"
class="rounded-lg p-2 text-red-500 transition-colors hover:bg-red-50 hover:text-red-600 dark:hover:bg-red-900/20"
>
<svg class="h-4 w-4" fill="none" viewBox="0 0 24 24" stroke="currentColor">
<path
stroke-linecap="round"
stroke-linejoin="round"
stroke-width="2"
d="M19 7l-.867 12.142A2 2 0 0116.138 21H7.862a2 2 0 01-1.995-1.858L5 7m5 4v6m4-6v6m1-10V4a1 1 0 00-1-1h-4a1 1 0 00-1 1v3M4 7h16"
/>
</svg>
</button>
</div>
</div>
<button
type="button"
@click="addModelMapping"
class="mb-3 w-full rounded-lg border-2 border-dashed border-gray-300 px-4 py-2 text-gray-600 transition-colors hover:border-gray-400 hover:text-gray-700 dark:border-dark-500 dark:text-gray-400 dark:hover:border-dark-400 dark:hover:text-gray-300"
>
+ {{ t('admin.accounts.addMapping') }}
</button>
<!-- Quick Add Buttons -->
<div class="flex flex-wrap gap-2">
<button
v-for="preset in presetMappings"
:key="'oauth-' + preset.label"
type="button"
@click="addPresetMapping(preset.from, preset.to)"
:class="['rounded-lg px-3 py-1 text-xs transition-colors', preset.color]"
>
+ {{ preset.label }}
</button>
</div>
</div>
</template>
</div>
<!-- Temp Unschedulable Rules -->
<div class="border-t border-gray-200 pt-4 dark:border-dark-600 space-y-4">
<div class="mb-3 flex items-center justify-between">
@@ -3665,6 +3801,14 @@ const handleOpenAIExchange = async (authCode: string) => {
const shouldCreateOpenAI = form.platform === 'openai'
const shouldCreateSora = form.platform === 'sora'
// Add model mapping for OpenAI OAuth accounts（透传模式下不应用）
if (shouldCreateOpenAI && !isOpenAIModelRestrictionDisabled.value) {
const modelMapping = buildModelMappingObject(modelRestrictionMode.value, allowedModels.value, modelMappings.value)
if (modelMapping) {
credentials.model_mapping = modelMapping
}
}
// 应用临时不可调度配置
if (!applyTempUnschedConfig(credentials)) {
return
@@ -3775,6 +3919,14 @@ const handleOpenAIValidateRT = async (refreshTokenInput: string) => {
const oauthExtra = oauthClient.buildExtraInfo(tokenInfo) as Record<string, unknown> | undefined
const extra = buildOpenAIExtra(oauthExtra)
// Add model mapping for OpenAI OAuth accounts（透传模式下不应用）
if (shouldCreateOpenAI && !isOpenAIModelRestrictionDisabled.value) {
const modelMapping = buildModelMappingObject(modelRestrictionMode.value, allowedModels.value, modelMappings.value)
if (modelMapping) {
credentials.model_mapping = modelMapping
}
}
// Generate account name with index for batch
const accountName = refreshTokens.length > 1 ? `${form.name} #${i + 1}` : form.name
frontend/src/components/account/EditAccountModal.vue
+191 -3
View File
@@ -351,6 +351,142 @@
</div>
<!-- OpenAI OAuth Model Mapping (OAuth 类型没有 apikey 容器，需要独立的模型映射区域) -->
<div
v-if="account.platform === 'openai' && account.type === 'oauth'"
class="border-t border-gray-200 pt-4 dark:border-dark-600"
>
<label class="input-label">{{ t('admin.accounts.modelRestriction') }}</label>
<div
v-if="isOpenAIModelRestrictionDisabled"
class="mb-3 rounded-lg bg-amber-50 p-3 dark:bg-amber-900/20"
>
<p class="text-xs text-amber-700 dark:text-amber-400">
{{ t('admin.accounts.openai.modelRestrictionDisabledByPassthrough') }}
</p>
</div>
<template v-else>
<!-- Mode Toggle -->
<div class="mb-4 flex gap-2">
<button
type="button"
@click="modelRestrictionMode = 'whitelist'"
:class="[
'flex-1 rounded-lg px-4 py-2 text-sm font-medium transition-all',
modelRestrictionMode === 'whitelist'
? 'bg-primary-100 text-primary-700 dark:bg-primary-900/30 dark:text-primary-400'
: 'bg-gray-100 text-gray-600 hover:bg-gray-200 dark:bg-dark-600 dark:text-gray-400 dark:hover:bg-dark-500'
]"
>
{{ t('admin.accounts.modelWhitelist') }}
</button>
<button
type="button"
@click="modelRestrictionMode = 'mapping'"
:class="[
'flex-1 rounded-lg px-4 py-2 text-sm font-medium transition-all',
modelRestrictionMode === 'mapping'
? 'bg-purple-100 text-purple-700 dark:bg-purple-900/30 dark:text-purple-400'
: 'bg-gray-100 text-gray-600 hover:bg-gray-200 dark:bg-dark-600 dark:text-gray-400 dark:hover:bg-dark-500'
]"
>
{{ t('admin.accounts.modelMapping') }}
</button>
</div>
<!-- Whitelist Mode -->
<div v-if="modelRestrictionMode === 'whitelist'">
<ModelWhitelistSelector v-model="allowedModels" :platform="account?.platform || 'anthropic'" />
<p class="text-xs text-gray-500 dark:text-gray-400">
{{ t('admin.accounts.selectedModels', { count: allowedModels.length }) }}
<span v-if="allowedModels.length === 0">{{
t('admin.accounts.supportsAllModels')
}}</span>
</p>
</div>
<!-- Mapping Mode -->
<div v-else>
<div class="mb-3 rounded-lg bg-purple-50 p-3 dark:bg-purple-900/20">
<p class="text-xs text-purple-700 dark:text-purple-400">
{{ t('admin.accounts.mapRequestModels') }}
</p>
</div>
<div v-if="modelMappings.length > 0" class="mb-3 space-y-2">
<div
v-for="(mapping, index) in modelMappings"
:key="'oauth-' + getModelMappingKey(mapping)"
class="flex items-center gap-2"
>
<input
v-model="mapping.from"
type="text"
class="input flex-1"
:placeholder="t('admin.accounts.requestModel')"
/>
<svg
class="h-4 w-4 flex-shrink-0 text-gray-400"
fill="none"
viewBox="0 0 24 24"
stroke="currentColor"
>
<path
stroke-linecap="round"
stroke-linejoin="round"
stroke-width="2"
d="M14 5l7 7m0 0l-7 7m7-7H3"
/>
</svg>
<input
v-model="mapping.to"
type="text"
class="input flex-1"
:placeholder="t('admin.accounts.actualModel')"
/>
<button
type="button"
@click="removeModelMapping(index)"
class="rounded-lg p-2 text-red-500 transition-colors hover:bg-red-50 hover:text-red-600 dark:hover:bg-red-900/20"
>
<svg class="h-4 w-4" fill="none" viewBox="0 0 24 24" stroke="currentColor">
<path
stroke-linecap="round"
stroke-linejoin="round"
stroke-width="2"
d="M19 7l-.867 12.142A2 2 0 0116.138 21H7.862a2 2 0 01-1.995-1.858L5 7m5 4v6m4-6v6m1-10V4a1 1 0 00-1-1h-4a1 1 0 00-1 1v3M4 7h16"
/>
</svg>
</button>
</div>
</div>
<button
type="button"
@click="addModelMapping"
class="mb-3 w-full rounded-lg border-2 border-dashed border-gray-300 px-4 py-2 text-gray-600 transition-colors hover:border-gray-400 hover:text-gray-700 dark:border-dark-500 dark:text-gray-400 dark:hover:border-dark-400 dark:hover:text-gray-300"
>
+ {{ t('admin.accounts.addMapping') }}
</button>
<!-- Quick Add Buttons -->
<div class="flex flex-wrap gap-2">
<button
v-for="preset in presetMappings"
:key="'oauth-' + preset.label"
type="button"
@click="addPresetMapping(preset.from, preset.to)"
:class="['rounded-lg px-3 py-1 text-xs transition-colors', preset.color]"
>
+ {{ preset.label }}
</button>
</div>
</div>
</template>
</div>
<!-- Upstream fields (only for upstream type) -->
<div v-if="account.type === 'upstream'" class="space-y-4">
<div>
@@ -1698,9 +1834,33 @@ watch(
? 'https://generativelanguage.googleapis.com'
: 'https://api.anthropic.com'
editBaseUrl.value = platformDefaultUrl
modelRestrictionMode.value = 'whitelist'
modelMappings.value = []
allowedModels.value = []
// Load model mappings for OpenAI OAuth accounts
if (newAccount.platform === 'openai' && newAccount.credentials) {
const oauthCredentials = newAccount.credentials as Record<string, unknown>
const existingMappings = oauthCredentials.model_mapping as Record<string, string> | undefined
if (existingMappings && typeof existingMappings === 'object') {
const entries = Object.entries(existingMappings)
const isWhitelistMode = entries.length > 0 && entries.every(([from, to]) => from === to)
if (isWhitelistMode) {
modelRestrictionMode.value = 'whitelist'
allowedModels.value = entries.map(([from]) => from)
modelMappings.value = []
} else {
modelRestrictionMode.value = 'mapping'
modelMappings.value = entries.map(([from, to]) => ({ from, to }))
allowedModels.value = []
}
} else {
modelRestrictionMode.value = 'whitelist'
modelMappings.value = []
allowedModels.value = []
}
} else {
modelRestrictionMode.value = 'whitelist'
modelMappings.value = []
allowedModels.value = []
}
customErrorCodesEnabled.value = false
selectedErrorCodes.value = []
}
@@ -2144,6 +2304,7 @@ const handleSubmit = async () => {
// Always update credentials for apikey type to handle model mapping changes
const newCredentials: Record<string, unknown> = {
...currentCredentials,
base_url: newBaseUrl
}
@@ -2164,6 +2325,8 @@ const handleSubmit = async () => {
const modelMapping = buildModelMappingObject(modelRestrictionMode.value, allowedModels.value, modelMappings.value)
if (modelMapping) {
newCredentials.model_mapping = modelMapping
} else {
delete newCredentials.model_mapping
}
} else if (currentCredentials.model_mapping) {
newCredentials.model_mapping = currentCredentials.model_mapping
@@ -2173,6 +2336,9 @@ const handleSubmit = async () => {
if (customErrorCodesEnabled.value) {
newCredentials.custom_error_codes_enabled = true
newCredentials.custom_error_codes = [...selectedErrorCodes.value]
} else {
delete newCredentials.custom_error_codes_enabled
delete newCredentials.custom_error_codes
}
// Add intercept warmup requests setting
@@ -2213,6 +2379,28 @@ const handleSubmit = async () => {
updatePayload.credentials = newCredentials
}
// OpenAI OAuth: persist model mapping to credentials
if (props.account.platform === 'openai' && props.account.type === 'oauth') {
const currentCredentials = (updatePayload.credentials as Record<string, unknown>) ||
((props.account.credentials as Record<string, unknown>) || {})
const newCredentials: Record<string, unknown> = { ...currentCredentials }
const shouldApplyModelMapping = !openaiPassthroughEnabled.value
if (shouldApplyModelMapping) {
const modelMapping = buildModelMappingObject(modelRestrictionMode.value, allowedModels.value, modelMappings.value)
if (modelMapping) {
newCredentials.model_mapping = modelMapping
} else {
delete newCredentials.model_mapping
}
} else if (currentCredentials.model_mapping) {
// 透传模式保留现有映射
newCredentials.model_mapping = currentCredentials.model_mapping
}
updatePayload.credentials = newCredentials
}
// Antigravity: persist model mapping to credentials (applies to all antigravity types)
// Antigravity 只支持映射模式
if (props.account.platform === 'antigravity') {
frontend/src/components/account/__tests__/AccountUsageCell.spec.ts
+55
View File
@@ -71,4 +71,59 @@ describe('AccountUsageCell', () => {
expect(wrapper.text()).toContain('admin.accounts.usageWindow.gemini3Image|70|2026-03-01T09:00:00Z')
})
it('OpenAI OAuth 在无 codex 快照时会回退显示 usage 接口窗口', async () => {
getUsage.mockResolvedValue({
five_hour: {
utilization: 0,
resets_at: null,
remaining_seconds: 0,
window_stats: {
requests: 2,
tokens: 27700,
cost: 0.06,
standard_cost: 0.06,
user_cost: 0.06
}
},
seven_day: {
utilization: 0,
resets_at: null,
remaining_seconds: 0,
window_stats: {
requests: 2,
tokens: 27700,
cost: 0.06,
standard_cost: 0.06,
user_cost: 0.06
}
}
})
const wrapper = mount(AccountUsageCell, {
props: {
account: {
id: 2002,
platform: 'openai',
type: 'oauth',
extra: {}
} as any
},
global: {
stubs: {
UsageProgressBar: {
props: ['label', 'utilization', 'resetsAt', 'windowStats', 'color'],
template: '<div class="usage-bar">{{ label }}|{{ utilization }}|{{ windowStats?.tokens }}</div>'
},
AccountQuotaInfo: true
}
}
})
await flushPromises()
expect(getUsage).toHaveBeenCalledWith(2002)
expect(wrapper.text()).toContain('5h|0|27700')
expect(wrapper.text()).toContain('7d|0|27700')
})
})
frontend/src/components/keys/UseKeyModal.vue
+10
View File
@@ -269,6 +269,7 @@ const clientTabs = computed((): TabConfig[] => {
return [
{ id: 'codex', label: t('keys.useKeyModal.cliTabs.codexCli'), icon: TerminalIcon },
{ id: 'codex-ws', label: t('keys.useKeyModal.cliTabs.codexCliWs'), icon: TerminalIcon },
{ id: 'claude', label: t('keys.useKeyModal.cliTabs.claudeCode'), icon: TerminalIcon },
{ id: 'opencode', label: t('keys.useKeyModal.cliTabs.opencode'), icon: TerminalIcon }
]
case 'gemini':
@@ -316,6 +317,9 @@ const currentTabs = computed(() => {
const platformDescription = computed(() => {
switch (props.platform) {
case 'openai':
if (activeClientTab.value === 'claude') {
return t('keys.useKeyModal.description')
}
return t('keys.useKeyModal.openai.description')
case 'gemini':
return t('keys.useKeyModal.gemini.description')
@@ -329,6 +333,9 @@ const platformDescription = computed(() => {
const platformNote = computed(() => {
switch (props.platform) {
case 'openai':
if (activeClientTab.value === 'claude') {
return t('keys.useKeyModal.note')
}
return activeTab.value === 'windows'
? t('keys.useKeyModal.openai.noteWindows')
: t('keys.useKeyModal.openai.note')
@@ -402,6 +409,9 @@ const currentFiles = computed((): FileConfig[] => {
switch (props.platform) {
case 'openai':
if (activeClientTab.value === 'claude') {
return generateAnthropicFiles(baseUrl, apiKey)
}
if (activeClientTab.value === 'codex-ws') {
return generateOpenAIWsFiles(baseUrl, apiKey)
}
frontend/src/composables/useModelWhitelist.ts
+4 -1
View File
@@ -280,7 +280,10 @@ const openaiPresetMappings = [
{ label: 'GPT-5.1', from: 'gpt-5.1', to: 'gpt-5.1', color: 'bg-orange-100 text-orange-700 hover:bg-orange-200 dark:bg-orange-900/30 dark:text-orange-400' },
{ label: 'GPT-5.2', from: 'gpt-5.2', to: 'gpt-5.2', color: 'bg-red-100 text-red-700 hover:bg-red-200 dark:bg-red-900/30 dark:text-red-400' },
{ label: 'GPT-5.4', from: 'gpt-5.4', to: 'gpt-5.4', color: 'bg-rose-100 text-rose-700 hover:bg-rose-200 dark:bg-rose-900/30 dark:text-rose-400' },
{ label: 'GPT-5.1 Codex', from: 'gpt-5.1-codex', to: 'gpt-5.1-codex', color: 'bg-cyan-100 text-cyan-700 hover:bg-cyan-200 dark:bg-cyan-900/30 dark:text-cyan-400' }
{ label: 'GPT-5.1 Codex', from: 'gpt-5.1-codex', to: 'gpt-5.1-codex', color: 'bg-cyan-100 text-cyan-700 hover:bg-cyan-200 dark:bg-cyan-900/30 dark:text-cyan-400' },
{ label: 'Haiku→5.4', from: 'claude-haiku-4-5-20251001', to: 'gpt-5.4', color: 'bg-emerald-100 text-emerald-700 hover:bg-emerald-200 dark:bg-emerald-900/30 dark:text-emerald-400' },
{ label: 'Opus→5.4', from: 'claude-opus-4-6', to: 'gpt-5.4', color: 'bg-purple-100 text-purple-700 hover:bg-purple-200 dark:bg-purple-900/30 dark:text-purple-400' },
{ label: 'Sonnet→5.4', from: 'claude-sonnet-4-6', to: 'gpt-5.4', color: 'bg-blue-100 text-blue-700 hover:bg-blue-200 dark:bg-blue-900/30 dark:text-blue-400' }
]
const soraPresetMappings: { label: string; from: string; to: string; color: string }[] = []
frontend/src/i18n/locales/zh.ts
+6
View File
@@ -2652,6 +2652,12 @@ export default {
allProtocols: '全部协议',
allStatus: '全部状态',
searchProxies: '搜索代理...',
protocols: {
http: 'HTTP',
https: 'HTTPS',
socks5: 'SOCKS5',
socks5h: 'SOCKS5H (远程 DNS)',
},
name: '名称',
protocol: '协议',
host: '主机',
frontend/src/views/user/KeysView.vue
+8 -4
View File
@@ -1638,17 +1638,21 @@ const executeCcsImport = (row: ApiKey, clientType: 'claude' | 'gemini') => {
headers: { "Authorization": "Bearer {{apiKey}}" }
},
extractor: function(response) {
const remaining = response?.remaining ?? response?.quota?.remaining ?? response?.balance;
const unit = response?.unit ?? response?.quota?.unit ?? "USD";
return {
isValid: response.is_active || true,
remaining: response.balance,
unit: "USD"
isValid: response?.is_active ?? response?.isValid ?? true,
remaining,
unit
};
}
})`
const providerName = (publicSettings.value?.site_name || 'sub2api').trim() || 'sub2api'
const params = new URLSearchParams({
resource: 'provider',
app: app,
name: 'sub2api',
name: providerName,
homepage: baseUrl,
endpoint: endpoint,
apiKey: row.key,