chuan/CLIProxyAPI
1
0
Fork 0
mirror of https://github.com/router-for-me/CLIProxyAPI.git synced 2026-02-18 04:10:51 +08:00
Code Issues Packages Projects Releases Wiki Activity

refactor: improve thinking logic

Browse Source
This commit is contained in:
hkfires
2026-01-14 08:32:02 +08:00
parent 5a7e5bd870
commit 0b06d637e7
76 changed files with 8712 additions and 1815 deletions
Download Patch File Download Diff File Expand all files Collapse all files

77
internal/runtime/executor/gemini_cli_executor.go
View File

@@ -20,6 +20,7 @@ import (
"github.com/router-for-me/CLIProxyAPI/v6/internal/config"
"github.com/router-for-me/CLIProxyAPI/v6/internal/misc"
"github.com/router-for-me/CLIProxyAPI/v6/internal/runtime/geminicli"
"github.com/router-for-me/CLIProxyAPI/v6/internal/thinking"
"github.com/router-for-me/CLIProxyAPI/v6/internal/util"
cliproxyauth "github.com/router-for-me/CLIProxyAPI/v6/sdk/cliproxy/auth"
cliproxyexecutor "github.com/router-for-me/CLIProxyAPI/v6/sdk/cliproxy/executor"
@@ -102,28 +103,30 @@ func (e *GeminiCLIExecutor) HttpRequest(ctx context.Context, auth *cliproxyauth.
// Execute performs a non-streaming request to the Gemini CLI API.
func (e *GeminiCLIExecutor) Execute(ctx context.Context, auth *cliproxyauth.Auth, req cliproxyexecutor.Request, opts cliproxyexecutor.Options) (resp cliproxyexecutor.Response, err error) {
baseModel := thinking.ParseSuffix(req.Model).ModelName
tokenSource, baseTokenData, err := prepareGeminiCLITokenSource(ctx, e.cfg, auth)
if err != nil {
return resp, err
}
reporter := newUsageReporter(ctx, e.Identifier(), req.Model, auth)
reporter := newUsageReporter(ctx, e.Identifier(), baseModel, auth)
defer reporter.trackFailure(ctx, &err)
from := opts.SourceFormat
to := sdktranslator.FromString("gemini-cli")
originalPayload := bytes.Clone(req.Payload)
if len(opts.OriginalRequest) > 0 {
originalPayload = bytes.Clone(opts.OriginalRequest)
}
originalTranslated := sdktranslator.TranslateRequest(from, to, req.Model, originalPayload, false)
basePayload := sdktranslator.TranslateRequest(from, to, req.Model, bytes.Clone(req.Payload), false)
basePayload = ApplyThinkingMetadataCLI(basePayload, req.Metadata, req.Model)
basePayload = util.ApplyGemini3ThinkingLevelFromMetadataCLI(req.Model, req.Metadata, basePayload)
basePayload = util.ApplyDefaultThinkingIfNeededCLI(req.Model, req.Metadata, basePayload)
basePayload = util.NormalizeGeminiCLIThinkingBudget(req.Model, basePayload)
basePayload = util.StripThinkingConfigIfUnsupported(req.Model, basePayload)
basePayload = fixGeminiCLIImageAspectRatio(req.Model, basePayload)
basePayload = applyPayloadConfigWithRoot(e.cfg, req.Model, "gemini", "request", basePayload, originalTranslated)
originalTranslated := sdktranslator.TranslateRequest(from, to, baseModel, originalPayload, false)
basePayload := sdktranslator.TranslateRequest(from, to, baseModel, bytes.Clone(req.Payload), false)
basePayload, _ = thinking.ApplyThinking(basePayload, req.Model, "gemini-cli")
basePayload = fixGeminiCLIImageAspectRatio(baseModel, basePayload)
basePayload = applyPayloadConfigWithRoot(e.cfg, baseModel, "gemini", "request", basePayload, originalTranslated)
action := "generateContent"
if req.Metadata != nil {
@@ -133,9 +136,9 @@ func (e *GeminiCLIExecutor) Execute(ctx context.Context, auth *cliproxyauth.Auth
}
projectID := resolveGeminiProjectID(auth)
models := cliPreviewFallbackOrder(req.Model)
if len(models) == 0 || models[0] != req.Model {
models = append([]string{req.Model}, models...)
models := cliPreviewFallbackOrder(baseModel)
if len(models) == 0 || models[0] != baseModel {
models = append([]string{baseModel}, models...)
}
httpClient := newHTTPClient(ctx, e.cfg, auth, 0)
@@ -246,34 +249,36 @@ func (e *GeminiCLIExecutor) Execute(ctx context.Context, auth *cliproxyauth.Auth
// ExecuteStream performs a streaming request to the Gemini CLI API.
func (e *GeminiCLIExecutor) ExecuteStream(ctx context.Context, auth *cliproxyauth.Auth, req cliproxyexecutor.Request, opts cliproxyexecutor.Options) (stream <-chan cliproxyexecutor.StreamChunk, err error) {
baseModel := thinking.ParseSuffix(req.Model).ModelName
tokenSource, baseTokenData, err := prepareGeminiCLITokenSource(ctx, e.cfg, auth)
if err != nil {
return nil, err
}
reporter := newUsageReporter(ctx, e.Identifier(), req.Model, auth)
reporter := newUsageReporter(ctx, e.Identifier(), baseModel, auth)
defer reporter.trackFailure(ctx, &err)
from := opts.SourceFormat
to := sdktranslator.FromString("gemini-cli")
originalPayload := bytes.Clone(req.Payload)
if len(opts.OriginalRequest) > 0 {
originalPayload = bytes.Clone(opts.OriginalRequest)
}
originalTranslated := sdktranslator.TranslateRequest(from, to, req.Model, originalPayload, true)
basePayload := sdktranslator.TranslateRequest(from, to, req.Model, bytes.Clone(req.Payload), true)
basePayload = ApplyThinkingMetadataCLI(basePayload, req.Metadata, req.Model)
basePayload = util.ApplyGemini3ThinkingLevelFromMetadataCLI(req.Model, req.Metadata, basePayload)
basePayload = util.ApplyDefaultThinkingIfNeededCLI(req.Model, req.Metadata, basePayload)
basePayload = util.NormalizeGeminiCLIThinkingBudget(req.Model, basePayload)
basePayload = util.StripThinkingConfigIfUnsupported(req.Model, basePayload)
basePayload = fixGeminiCLIImageAspectRatio(req.Model, basePayload)
basePayload = applyPayloadConfigWithRoot(e.cfg, req.Model, "gemini", "request", basePayload, originalTranslated)
originalTranslated := sdktranslator.TranslateRequest(from, to, baseModel, originalPayload, true)
basePayload := sdktranslator.TranslateRequest(from, to, baseModel, bytes.Clone(req.Payload), true)
basePayload, _ = thinking.ApplyThinking(basePayload, req.Model, "gemini-cli")
basePayload = fixGeminiCLIImageAspectRatio(baseModel, basePayload)
basePayload = applyPayloadConfigWithRoot(e.cfg, baseModel, "gemini", "request", basePayload, originalTranslated)
projectID := resolveGeminiProjectID(auth)
models := cliPreviewFallbackOrder(req.Model)
if len(models) == 0 || models[0] != req.Model {
models = append([]string{req.Model}, models...)
models := cliPreviewFallbackOrder(baseModel)
if len(models) == 0 || models[0] != baseModel {
models = append([]string{baseModel}, models...)
}
httpClient := newHTTPClient(ctx, e.cfg, auth, 0)
@@ -435,6 +440,8 @@ func (e *GeminiCLIExecutor) ExecuteStream(ctx context.Context, auth *cliproxyaut
// CountTokens counts tokens for the given request using the Gemini CLI API.
func (e *GeminiCLIExecutor) CountTokens(ctx context.Context, auth *cliproxyauth.Auth, req cliproxyexecutor.Request, opts cliproxyexecutor.Options) (cliproxyexecutor.Response, error) {
baseModel := thinking.ParseSuffix(req.Model).ModelName
tokenSource, baseTokenData, err := prepareGeminiCLITokenSource(ctx, e.cfg, auth)
if err != nil {
return cliproxyexecutor.Response{}, err
@@ -443,9 +450,9 @@ func (e *GeminiCLIExecutor) CountTokens(ctx context.Context, auth *cliproxyauth.
from := opts.SourceFormat
to := sdktranslator.FromString("gemini-cli")
models := cliPreviewFallbackOrder(req.Model)
if len(models) == 0 || models[0] != req.Model {
models = append([]string{req.Model}, models...)
models := cliPreviewFallbackOrder(baseModel)
if len(models) == 0 || models[0] != baseModel {
models = append([]string{baseModel}, models...)
}
httpClient := newHTTPClient(ctx, e.cfg, auth, 0)
@@ -463,15 +470,15 @@ func (e *GeminiCLIExecutor) CountTokens(ctx context.Context, auth *cliproxyauth.
// The loop variable attemptModel is only used as the concrete model id sent to the upstream
// Gemini CLI endpoint when iterating fallback variants.
for _, attemptModel := range models {
payload := sdktranslator.TranslateRequest(from, to, attemptModel, bytes.Clone(req.Payload), false)
payload = ApplyThinkingMetadataCLI(payload, req.Metadata, req.Model)
payload = util.ApplyGemini3ThinkingLevelFromMetadataCLI(req.Model, req.Metadata, payload)
for range models {
payload := sdktranslator.TranslateRequest(from, to, baseModel, bytes.Clone(req.Payload), false)
payload, _ = thinking.ApplyThinking(payload, req.Model, "gemini-cli")
payload = deleteJSONField(payload, "project")
payload = deleteJSONField(payload, "model")
payload = deleteJSONField(payload, "request.safetySettings")
payload = util.StripThinkingConfigIfUnsupported(req.Model, payload)
payload = fixGeminiCLIImageAspectRatio(req.Model, payload)
payload = fixGeminiCLIImageAspectRatio(baseModel, payload)
tok, errTok := tokenSource.Token()
if errTok != nil {