fix(openai): gate Codex image bridge injection

Ultraworked with [Sisyphus](https://github.com/code-yeongyu/oh-my-openagent)

Co-authored-by: Sisyphus <clio-agent@sisyphuslabs.ai>
This commit is contained in:
Jlypx
2026-05-07 00:10:20 +08:00
parent a1106e8167
commit 26043a8f29
4 changed files with 254 additions and 7 deletions

View File

@@ -611,6 +611,9 @@ type GatewayConfig struct {
// ForceCodexCLI: 强制将 OpenAI `/v1/responses` 请求按 Codex CLI 处理。
// 用于网关未透传/改写 User-Agent 时的兼容兜底(默认关闭,避免影响其他客户端)。
ForceCodexCLI bool `mapstructure:"force_codex_cli"`
// CodexImageGenerationBridgeEnabled: 是否为 Codex `/v1/responses` 自动注入 image_generation 工具和桥接指令。
// 默认关闭,避免纯文本 Codex 请求被意外改写;显式携带 image_generation 工具的请求仍按分组能力转发。
CodexImageGenerationBridgeEnabled bool `mapstructure:"codex_image_generation_bridge_enabled"`
// ForcedCodexInstructionsTemplateFile: 服务端强制附加到 Codex 顶层 instructions 的模板文件路径。
// 模板渲染后会直接覆盖最终 instructions若需要保留客户端 system 转换结果,请在模板中显式引用 {{ .ExistingInstructions }}。
ForcedCodexInstructionsTemplateFile string `mapstructure:"forced_codex_instructions_template_file"`
@@ -1649,6 +1652,7 @@ func setDefaults() {
viper.SetDefault("gateway.max_account_switches", 10)
viper.SetDefault("gateway.max_account_switches_gemini", 3)
viper.SetDefault("gateway.force_codex_cli", false)
viper.SetDefault("gateway.codex_image_generation_bridge_enabled", false)
viper.SetDefault("gateway.openai_passthrough_allow_timeout_headers", false)
// OpenAI Responses WebSocket默认开启可通过 force_http 紧急回滚)
viper.SetDefault("gateway.openai_ws.enabled", true)

View File

@@ -0,0 +1,64 @@
package service
import "strings"
const featureKeyCodexImageGenerationBridge = "codex_image_generation_bridge"
func boolOverridePtr(v bool) *bool {
return &v
}
func boolOverrideFromMap(values map[string]any, keys ...string) *bool {
if values == nil {
return nil
}
for _, key := range keys {
if v, ok := values[key].(bool); ok {
return boolOverridePtr(v)
}
}
return nil
}
func platformBoolOverride(values map[string]any, key string, platform string) *bool {
if values == nil {
return nil
}
if v, ok := values[key].(bool); ok {
return boolOverridePtr(v)
}
raw, ok := values[key].(map[string]any)
if !ok {
return nil
}
platform = strings.TrimSpace(platform)
if platform == "" {
return nil
}
if v, ok := raw[platform].(bool); ok {
return boolOverridePtr(v)
}
return nil
}
// CodexImageGenerationBridgeOverride returns the channel-level override for Codex
// image_generation bridge injection. Nil means follow the global/account policy.
func (c *Channel) CodexImageGenerationBridgeOverride(platform string) *bool {
if c == nil {
return nil
}
return platformBoolOverride(c.FeaturesConfig, featureKeyCodexImageGenerationBridge, platform)
}
// CodexImageGenerationBridgeOverride returns the account-level override for Codex
// image_generation bridge injection. Nil means follow the channel/global policy.
func (a *Account) CodexImageGenerationBridgeOverride() *bool {
if a == nil || a.Platform != PlatformOpenAI || a.Extra == nil {
return nil
}
if override := boolOverrideFromMap(a.Extra, featureKeyCodexImageGenerationBridge, "codex_image_generation_bridge_enabled"); override != nil {
return override
}
openaiConfig, _ := a.Extra[PlatformOpenAI].(map[string]any)
return boolOverrideFromMap(openaiConfig, featureKeyCodexImageGenerationBridge, "codex_image_generation_bridge_enabled")
}

View File

@@ -440,6 +440,21 @@ func (s *OpenAIGatewayService) ResolveChannelMappingAndRestrict(ctx context.Cont
return s.channelService.ResolveChannelMappingAndRestrict(ctx, groupID, model)
}
func (s *OpenAIGatewayService) isCodexImageGenerationBridgeEnabled(ctx context.Context, account *Account, apiKey *APIKey) bool {
if override := account.CodexImageGenerationBridgeOverride(); override != nil {
return *override
}
if s != nil && s.channelService != nil && apiKey != nil && apiKey.GroupID != nil {
ch, err := s.channelService.GetChannelForGroup(ctx, *apiKey.GroupID)
if err != nil {
slog.Warn("failed to resolve codex image generation bridge channel override", "group_id", *apiKey.GroupID, "error", err)
} else if override := ch.CodexImageGenerationBridgeOverride(PlatformOpenAI); override != nil {
return *override
}
}
return s != nil && s.cfg != nil && s.cfg.Gateway.CodexImageGenerationBridgeEnabled
}
func (s *OpenAIGatewayService) checkChannelPricingRestriction(ctx context.Context, groupID *int64, requestedModel string) bool {
if groupID == nil || s.channelService == nil || requestedModel == "" {
return false
@@ -2059,6 +2074,7 @@ func (s *OpenAIGatewayService) Forward(ctx context.Context, c *gin.Context, acco
if apiKey != nil {
imageGenerationAllowed = GroupAllowsImageGeneration(apiKey.Group)
}
codexImageGenerationBridgeEnabled := isCodexCLI && imageGenerationAllowed && s.isCodexImageGenerationBridgeEnabled(ctx, account, apiKey)
if IsImageGenerationIntentMap(openAIResponsesEndpoint, reqModel, reqBody) && !imageGenerationAllowed {
setOpsUpstreamError(c, http.StatusForbidden, ImageGenerationPermissionMessage(), "")
c.JSON(http.StatusForbidden, gin.H{
@@ -2128,7 +2144,7 @@ func (s *OpenAIGatewayService) Forward(ctx context.Context, c *gin.Context, acco
markPatchSet("instructions", "You are a helpful coding assistant.")
}
if isCodexCLI && imageGenerationAllowed && ensureOpenAIResponsesImageGenerationTool(reqBody) {
if codexImageGenerationBridgeEnabled && ensureOpenAIResponsesImageGenerationTool(reqBody) {
bodyModified = true
disablePatch()
logger.LegacyPrintf("service.openai_gateway", "[OpenAI] Injected /responses image_generation tool for Codex client")
@@ -2139,7 +2155,7 @@ func (s *OpenAIGatewayService) Forward(ctx context.Context, c *gin.Context, acco
disablePatch()
logger.LegacyPrintf("service.openai_gateway", "[OpenAI] Normalized /responses image_generation tool payload")
}
if isCodexCLI && imageGenerationAllowed && applyCodexImageGenerationBridgeInstructions(reqBody) {
if codexImageGenerationBridgeEnabled && applyCodexImageGenerationBridgeInstructions(reqBody) {
bodyModified = true
disablePatch()
logger.LegacyPrintf("service.openai_gateway", "[OpenAI] Added Codex image_generation bridge instructions")

View File

@@ -83,12 +83,14 @@ func TestOpenAIGatewayServiceForward_CodexImageInjectionRespectsGroupCapability(
gin.SetMode(gin.TestMode)
tests := []struct {
name string
allowImages bool
wantInjected bool
name string
allowImages bool
bridgeEnabled bool
wantInjected bool
}{
{name: "disabled group skips injection", allowImages: false, wantInjected: false},
{name: "enabled group injects image tool", allowImages: true, wantInjected: true},
{name: "disabled group skips injection", allowImages: false, bridgeEnabled: true, wantInjected: false},
{name: "enabled group skips injection by default", allowImages: true, bridgeEnabled: false, wantInjected: false},
{name: "enabled group injects image tool when bridge enabled", allowImages: true, bridgeEnabled: true, wantInjected: true},
}
for _, tt := range tests {
@@ -101,6 +103,7 @@ func TestOpenAIGatewayServiceForward_CodexImageInjectionRespectsGroupCapability(
},
}
svc := newOpenAIImageGenerationControlTestService(upstream)
svc.cfg.Gateway.CodexImageGenerationBridgeEnabled = tt.bridgeEnabled
c, _ := newOpenAIImageGenerationControlTestContext(tt.allowImages, "codex_cli_rs/0.98.0")
account := newOpenAIImageGenerationControlTestAccount()
@@ -117,6 +120,154 @@ func TestOpenAIGatewayServiceForward_CodexImageInjectionRespectsGroupCapability(
}
}
func TestOpenAIGatewayServiceForward_ExplicitImageToolWorksWithBridgeDisabled(t *testing.T) {
gin.SetMode(gin.TestMode)
upstream := &httpUpstreamRecorder{
resp: &http.Response{
StatusCode: http.StatusOK,
Header: http.Header{"Content-Type": []string{"application/json"}},
Body: io.NopCloser(strings.NewReader(`{"id":"resp_explicit_image","model":"gpt-5.4","usage":{"input_tokens":2,"output_tokens":1}}`)),
},
}
svc := newOpenAIImageGenerationControlTestService(upstream)
c, _ := newOpenAIImageGenerationControlTestContext(true, "codex_cli_rs/0.98.0")
account := newOpenAIImageGenerationControlTestAccount()
body := []byte(`{"model":"gpt-5.4","input":"draw","stream":false,"tools":[{"type":"image_generation","format":"jpeg"}]}`)
result, err := svc.Forward(context.Background(), c, account, body)
require.NoError(t, err)
require.NotNil(t, result)
require.NotNil(t, upstream.lastReq)
require.True(t, gjson.GetBytes(upstream.lastBody, `tools.#(type=="image_generation")`).Exists())
require.Equal(t, "jpeg", gjson.GetBytes(upstream.lastBody, `tools.#(type=="image_generation").output_format`).String())
require.False(t, gjson.GetBytes(upstream.lastBody, `tools.#(type=="image_generation").format`).Exists())
instructions := gjson.GetBytes(upstream.lastBody, "instructions").String()
require.NotContains(t, instructions, "image_generation")
}
func TestOpenAIGatewayServiceForward_ChannelBridgeOverrideEnablesCodexInjection(t *testing.T) {
gin.SetMode(gin.TestMode)
upstream := &httpUpstreamRecorder{
resp: &http.Response{
StatusCode: http.StatusOK,
Header: http.Header{"Content-Type": []string{"application/json"}},
Body: io.NopCloser(strings.NewReader(`{"id":"resp_channel_bridge","model":"gpt-5.4","usage":{"input_tokens":1,"output_tokens":1}}`)),
},
}
svc := newOpenAIImageGenerationControlTestService(upstream)
groupID := int64(4242)
svc.channelService = newOpenAIImageGenerationControlChannelService(groupID, &Channel{
ID: 9001,
Status: StatusActive,
FeaturesConfig: map[string]any{
featureKeyCodexImageGenerationBridge: map[string]any{PlatformOpenAI: true},
},
})
c, _ := newOpenAIImageGenerationControlTestContext(true, "codex_cli_rs/0.98.0")
account := newOpenAIImageGenerationControlTestAccount()
result, err := svc.Forward(context.Background(), c, account, []byte(`{"model":"gpt-5.4","input":"write code","stream":false}`))
require.NoError(t, err)
require.NotNil(t, result)
require.NotNil(t, upstream.lastReq)
require.True(t, gjson.GetBytes(upstream.lastBody, `tools.#(type=="image_generation")`).Exists())
instructions := gjson.GetBytes(upstream.lastBody, "instructions").String()
require.Contains(t, instructions, "image_generation")
}
func TestOpenAIGatewayService_CodexImageGenerationBridgeOverridePrecedence(t *testing.T) {
groupID := int64(4242)
tests := []struct {
name string
global bool
channel *Channel
account *Account
want bool
}{
{
name: "global default enables bridge",
global: true,
account: &Account{
Platform: PlatformOpenAI,
},
want: true,
},
{
name: "channel true overrides disabled global",
global: false,
channel: &Channel{ID: 1, Status: StatusActive, FeaturesConfig: map[string]any{
featureKeyCodexImageGenerationBridge: map[string]any{PlatformOpenAI: true},
}},
account: &Account{Platform: PlatformOpenAI},
want: true,
},
{
name: "channel false overrides enabled global",
global: true,
channel: &Channel{ID: 1, Status: StatusActive, FeaturesConfig: map[string]any{
featureKeyCodexImageGenerationBridge: map[string]any{PlatformOpenAI: false},
}},
account: &Account{Platform: PlatformOpenAI},
want: false,
},
{
name: "account false overrides channel and global true",
global: true,
channel: &Channel{ID: 1, Status: StatusActive, FeaturesConfig: map[string]any{
featureKeyCodexImageGenerationBridge: map[string]any{PlatformOpenAI: true},
}},
account: &Account{
Platform: PlatformOpenAI,
Extra: map[string]any{featureKeyCodexImageGenerationBridge: false},
},
want: false,
},
{
name: "nested account true overrides channel false",
global: false,
channel: &Channel{ID: 1, Status: StatusActive, FeaturesConfig: map[string]any{
featureKeyCodexImageGenerationBridge: map[string]any{PlatformOpenAI: false},
}},
account: &Account{
Platform: PlatformOpenAI,
Extra: map[string]any{
PlatformOpenAI: map[string]any{"codex_image_generation_bridge_enabled": true},
},
},
want: true,
},
{
name: "non openai account extra is ignored",
global: false,
account: &Account{
Platform: PlatformAnthropic,
Extra: map[string]any{featureKeyCodexImageGenerationBridge: true},
},
want: false,
},
}
for _, tt := range tests {
t.Run(tt.name, func(t *testing.T) {
svc := newOpenAIImageGenerationControlTestService(&httpUpstreamRecorder{})
svc.cfg.Gateway.CodexImageGenerationBridgeEnabled = tt.global
if tt.channel != nil {
svc.channelService = newOpenAIImageGenerationControlChannelService(groupID, tt.channel)
}
apiKey := &APIKey{GroupID: &groupID}
got := svc.isCodexImageGenerationBridgeEnabled(context.Background(), tt.account, apiKey)
require.Equal(t, tt.want, got)
})
}
}
func TestOpenAIGatewayServiceHandleResponsesImageOutputs_NonStreaming(t *testing.T) {
gin.SetMode(gin.TestMode)
@@ -180,6 +331,18 @@ func newOpenAIImageGenerationControlTestService(upstream *httpUpstreamRecorder)
}
}
func newOpenAIImageGenerationControlChannelService(groupID int64, ch *Channel) *ChannelService {
svc := &ChannelService{}
cache := newEmptyChannelCache()
if ch != nil {
cache.channelByGroupID[groupID] = ch
cache.byID[ch.ID] = ch
}
cache.loadedAt = time.Now()
svc.cache.Store(cache)
return svc
}
func newOpenAIImageGenerationControlTestContext(allowImages bool, userAgent string) (*gin.Context, *httptest.ResponseRecorder) {
recorder := httptest.NewRecorder()
c, _ := gin.CreateTestContext(recorder)