Merge pull request #161 from router-for-me/plus

v6.7.40

config.example.yaml

@@ -55,6 +55,10 @@ logging-to-file: false
 # files are deleted until within the limit. Set to 0 to disable.
 logs-max-total-size-mb: 0
 
+# Maximum number of error log files retained when request logging is disabled.
+# When exceeded, the oldest error log files are deleted. Default is 10. Set to 0 to disable cleanup.
+error-logs-max-files: 10
+
 # When false, disable in-memory usage statistics aggregation
 usage-statistics-enabled: false
 

examples/custom-provider/main.go

@@ -205,7 +205,7 @@ func main() {
 			// Optional: add a simple middleware + custom request logger
 			api.WithMiddleware(func(c *gin.Context) { c.Header("X-Example", "custom-provider"); c.Next() }),
 			api.WithRequestLoggerFactory(func(cfg *config.Config, cfgPath string) logging.RequestLogger {
-				return logging.NewFileRequestLogger(true, "logs", filepath.Dir(cfgPath))
+				return logging.NewFileRequestLoggerWithOptions(true, "logs", filepath.Dir(cfgPath), cfg.ErrorLogsMaxFiles)
 			}),
 		).
 		WithHooks(hooks).

internal/api/handlers/management/config_basic.go

@@ -222,6 +222,26 @@ func (h *Handler) PutLogsMaxTotalSizeMB(c *gin.Context) {
 	h.persist(c)
 }
 
+// ErrorLogsMaxFiles
+func (h *Handler) GetErrorLogsMaxFiles(c *gin.Context) {
+	c.JSON(200, gin.H{"error-logs-max-files": h.cfg.ErrorLogsMaxFiles})
+}
+func (h *Handler) PutErrorLogsMaxFiles(c *gin.Context) {
+	var body struct {
+		Value *int `json:"value"`
+	}
+	if errBindJSON := c.ShouldBindJSON(&body); errBindJSON != nil || body.Value == nil {
+		c.JSON(http.StatusBadRequest, gin.H{"error": "invalid body"})
+		return
+	}
+	value := *body.Value
+	if value < 0 {
+		value = 10
+	}
+	h.cfg.ErrorLogsMaxFiles = value
+	h.persist(c)
+}
+
 // Request log
 func (h *Handler) GetRequestLog(c *gin.Context) { c.JSON(200, gin.H{"request-log": h.cfg.RequestLog}) }
 func (h *Handler) PutRequestLog(c *gin.Context) {

internal/api/server.go

@@ -61,9 +61,9 @@ type ServerOption func(*serverOptionConfig)
 func defaultRequestLoggerFactory(cfg *config.Config, configPath string) logging.RequestLogger {
 	configDir := filepath.Dir(configPath)
 	if base := util.WritablePath(); base != "" {
-		return logging.NewFileRequestLogger(cfg.RequestLog, filepath.Join(base, "logs"), configDir)
+		return logging.NewFileRequestLogger(cfg.RequestLog, filepath.Join(base, "logs"), configDir, cfg.ErrorLogsMaxFiles)
 	}
-	return logging.NewFileRequestLogger(cfg.RequestLog, "logs", configDir)
+	return logging.NewFileRequestLogger(cfg.RequestLog, "logs", configDir, cfg.ErrorLogsMaxFiles)
 }
 
 // WithMiddleware appends additional Gin middleware during server construction.
@@ -523,6 +523,10 @@ func (s *Server) registerManagementRoutes() {
 		mgmt.PUT("/logs-max-total-size-mb", s.mgmt.PutLogsMaxTotalSizeMB)
 		mgmt.PATCH("/logs-max-total-size-mb", s.mgmt.PutLogsMaxTotalSizeMB)
 
+		mgmt.GET("/error-logs-max-files", s.mgmt.GetErrorLogsMaxFiles)
+		mgmt.PUT("/error-logs-max-files", s.mgmt.PutErrorLogsMaxFiles)
+		mgmt.PATCH("/error-logs-max-files", s.mgmt.PutErrorLogsMaxFiles)
+
 		mgmt.GET("/usage-statistics-enabled", s.mgmt.GetUsageStatisticsEnabled)
 		mgmt.PUT("/usage-statistics-enabled", s.mgmt.PutUsageStatisticsEnabled)
 		mgmt.PATCH("/usage-statistics-enabled", s.mgmt.PutUsageStatisticsEnabled)
@@ -935,6 +939,15 @@ func (s *Server) UpdateClients(cfg *config.Config) {
 		}
 	}
 
+	if s.requestLogger != nil && (oldCfg == nil || oldCfg.ErrorLogsMaxFiles != cfg.ErrorLogsMaxFiles) {
+		if setter, ok := s.requestLogger.(interface{ SetErrorLogsMaxFiles(int) }); ok {
+			setter.SetErrorLogsMaxFiles(cfg.ErrorLogsMaxFiles)
+		}
+		if oldCfg != nil {
+			log.Debugf("error_logs_max_files updated from %d to %d", oldCfg.ErrorLogsMaxFiles, cfg.ErrorLogsMaxFiles)
+		}
+	}
+
 	if oldCfg == nil || oldCfg.DisableCooling != cfg.DisableCooling {
 		auth.SetQuotaCooldownDisabled(cfg.DisableCooling)
 		if oldCfg != nil {

internal/config/config.go

@@ -51,6 +51,10 @@ type Config struct {
 	// When exceeded, the oldest log files are deleted until within the limit. Set to 0 to disable.
 	LogsMaxTotalSizeMB int `yaml:"logs-max-total-size-mb" json:"logs-max-total-size-mb"`
 
+	// ErrorLogsMaxFiles limits the number of error log files retained when request logging is disabled.
+	// When exceeded, the oldest error log files are deleted. Default is 10. Set to 0 to disable cleanup.
+	ErrorLogsMaxFiles int `yaml:"error-logs-max-files" json:"error-logs-max-files"`
+
 	// UsageStatisticsEnabled toggles in-memory usage aggregation; when false, usage data is discarded.
 	UsageStatisticsEnabled bool `yaml:"usage-statistics-enabled" json:"usage-statistics-enabled"`
 
@@ -554,6 +558,7 @@ func LoadConfigOptional(configFile string, optional bool) (*Config, error) {
 	cfg.Host = "" // Default empty: binds to all interfaces (IPv4 + IPv6)
 	cfg.LoggingToFile = false
 	cfg.LogsMaxTotalSizeMB = 0
+	cfg.ErrorLogsMaxFiles = 10
 	cfg.UsageStatisticsEnabled = false
 	cfg.DisableCooling = false
 	cfg.AmpCode.RestrictManagementToLocalhost = false // Default to false: API key auth is sufficient
@@ -603,6 +608,10 @@ func LoadConfigOptional(configFile string, optional bool) (*Config, error) {
 		cfg.LogsMaxTotalSizeMB = 0
 	}
 
+	if cfg.ErrorLogsMaxFiles < 0 {
+		cfg.ErrorLogsMaxFiles = 10
+	}
+
 	// Sync request authentication providers with inline API keys for backwards compatibility.
 	syncInlineAccessProvider(&cfg)
 

internal/logging/request_logger.go

@@ -132,6 +132,9 @@ type FileRequestLogger struct {
 
 	// logsDir is the directory where log files are stored.
 	logsDir string
+
+	// errorLogsMaxFiles limits the number of error log files retained.
+	errorLogsMaxFiles int
 }
 
 // NewFileRequestLogger creates a new file-based request logger.
@@ -141,10 +144,11 @@ type FileRequestLogger struct {
 //   - logsDir: The directory where log files should be stored (can be relative)
 //   - configDir: The directory of the configuration file; when logsDir is
 //     relative, it will be resolved relative to this directory
+//   - errorLogsMaxFiles: Maximum number of error log files to retain (0 = no cleanup)
 //
 // Returns:
 //   - *FileRequestLogger: A new file-based request logger instance
-func NewFileRequestLogger(enabled bool, logsDir string, configDir string) *FileRequestLogger {
+func NewFileRequestLogger(enabled bool, logsDir string, configDir string, errorLogsMaxFiles int) *FileRequestLogger {
 	// Resolve logsDir relative to the configuration file directory when it's not absolute.
 	if !filepath.IsAbs(logsDir) {
 		// If configDir is provided, resolve logsDir relative to it.
@@ -153,8 +157,9 @@ func NewFileRequestLogger(enabled bool, logsDir string, configDir string) *FileR
 		}
 	}
 	return &FileRequestLogger{
-		enabled: enabled,
-		logsDir: logsDir,
+		enabled:           enabled,
+		logsDir:           logsDir,
+		errorLogsMaxFiles: errorLogsMaxFiles,
 	}
 }
 
@@ -175,6 +180,11 @@ func (l *FileRequestLogger) SetEnabled(enabled bool) {
 	l.enabled = enabled
 }
 
+// SetErrorLogsMaxFiles updates the maximum number of error log files to retain.
+func (l *FileRequestLogger) SetErrorLogsMaxFiles(maxFiles int) {
+	l.errorLogsMaxFiles = maxFiles
+}
+
 // LogRequest logs a complete non-streaming request/response cycle to a file.
 //
 // Parameters:
@@ -433,8 +443,12 @@ func (l *FileRequestLogger) sanitizeForFilename(path string) string {
 	return sanitized
 }
 
-// cleanupOldErrorLogs keeps only the newest 10 forced error log files.
+// cleanupOldErrorLogs keeps only the newest errorLogsMaxFiles forced error log files.
 func (l *FileRequestLogger) cleanupOldErrorLogs() error {
+	if l.errorLogsMaxFiles <= 0 {
+		return nil
+	}
+
 	entries, errRead := os.ReadDir(l.logsDir)
 	if errRead != nil {
 		return errRead
@@ -462,7 +476,7 @@ func (l *FileRequestLogger) cleanupOldErrorLogs() error {
 		files = append(files, logFile{name: name, modTime: info.ModTime()})
 	}
 
-	if len(files) <= 10 {
+	if len(files) <= l.errorLogsMaxFiles {
 		return nil
 	}
 
@@ -470,7 +484,7 @@ func (l *FileRequestLogger) cleanupOldErrorLogs() error {
 		return files[i].modTime.After(files[j].modTime)
 	})
 
-	for _, file := range files[10:] {
+	for _, file := range files[l.errorLogsMaxFiles:] {
 		if errRemove := os.Remove(filepath.Join(l.logsDir, file.name)); errRemove != nil {
 			log.WithError(errRemove).Warnf("failed to remove old error log: %s", file.name)
 		}

internal/store/objectstore.go

@@ -386,11 +386,12 @@ func (s *ObjectTokenStore) syncConfigFromBucket(ctx context.Context, example str
 }
 
 func (s *ObjectTokenStore) syncAuthFromBucket(ctx context.Context) error {
-	if err := os.RemoveAll(s.authDir); err != nil {
-		return fmt.Errorf("object store: reset auth directory: %w", err)
-	}
+	// NOTE: We intentionally do NOT use os.RemoveAll here.
+	// Wiping the directory triggers file watcher delete events, which then
+	// propagate deletions to the remote object store (race condition).
+	// Instead, we just ensure the directory exists and overwrite files incrementally.
 	if err := os.MkdirAll(s.authDir, 0o700); err != nil {
-		return fmt.Errorf("object store: recreate auth directory: %w", err)
+		return fmt.Errorf("object store: create auth directory: %w", err)
 	}
 
 	prefix := s.prefixedKey(objectStoreAuthPrefix + "/")

internal/translator/antigravity/openai/chat-completions/antigravity_openai_response.go

@@ -22,8 +22,10 @@ import (
 
 // convertCliResponseToOpenAIChatParams holds parameters for response conversion.
 type convertCliResponseToOpenAIChatParams struct {
-	UnixTimestamp int64
-	FunctionIndex int
+	UnixTimestamp        int64
+	FunctionIndex        int
+	SawToolCall          bool   // Tracks if any tool call was seen in the entire stream
+	UpstreamFinishReason string // Caches the upstream finish reason for final chunk
 }
 
 // functionCallIDCounter provides a process-wide unique counter for function call identifiers.
@@ -79,10 +81,9 @@ func ConvertAntigravityResponseToOpenAI(_ context.Context, _ string, originalReq
 		template, _ = sjson.Set(template, "id", responseIDResult.String())
 	}
 
-	// Extract and set the finish reason.
+	// Cache the finish reason - do NOT set it in output yet (will be set on final chunk)
 	if finishReasonResult := gjson.GetBytes(rawJSON, "response.candidates.0.finishReason"); finishReasonResult.Exists() {
-		template, _ = sjson.Set(template, "choices.0.finish_reason", strings.ToLower(finishReasonResult.String()))
-		template, _ = sjson.Set(template, "choices.0.native_finish_reason", strings.ToLower(finishReasonResult.String()))
+		(*param).(*convertCliResponseToOpenAIChatParams).UpstreamFinishReason = strings.ToUpper(finishReasonResult.String())
 	}
 
 	// Extract and set usage metadata (token counts).
@@ -112,7 +113,6 @@ func ConvertAntigravityResponseToOpenAI(_ context.Context, _ string, originalReq
 
 	// Process the main content part of the response.
 	partsResult := gjson.GetBytes(rawJSON, "response.candidates.0.content.parts")
-	hasFunctionCall := false
 	if partsResult.IsArray() {
 		partResults := partsResult.Array()
 		for i := 0; i < len(partResults); i++ {
@@ -148,7 +148,7 @@ func ConvertAntigravityResponseToOpenAI(_ context.Context, _ string, originalReq
 				template, _ = sjson.Set(template, "choices.0.delta.role", "assistant")
 			} else if functionCallResult.Exists() {
 				// Handle function call content.
-				hasFunctionCall = true
+				(*param).(*convertCliResponseToOpenAIChatParams).SawToolCall = true // Persist across chunks
 				toolCallsResult := gjson.Get(template, "choices.0.delta.tool_calls")
 				functionCallIndex := (*param).(*convertCliResponseToOpenAIChatParams).FunctionIndex
 				(*param).(*convertCliResponseToOpenAIChatParams).FunctionIndex++
@@ -195,9 +195,25 @@ func ConvertAntigravityResponseToOpenAI(_ context.Context, _ string, originalReq
 		}
 	}
 
-	if hasFunctionCall {
-		template, _ = sjson.Set(template, "choices.0.finish_reason", "tool_calls")
-		template, _ = sjson.Set(template, "choices.0.native_finish_reason", "tool_calls")
+	// Determine finish_reason only on the final chunk (has both finishReason and usage metadata)
+	params := (*param).(*convertCliResponseToOpenAIChatParams)
+	upstreamFinishReason := params.UpstreamFinishReason
+	sawToolCall := params.SawToolCall
+
+	usageExists := gjson.GetBytes(rawJSON, "response.usageMetadata").Exists()
+	isFinalChunk := upstreamFinishReason != "" && usageExists
+
+	if isFinalChunk {
+		var finishReason string
+		if sawToolCall {
+			finishReason = "tool_calls"
+		} else if upstreamFinishReason == "MAX_TOKENS" {
+			finishReason = "max_tokens"
+		} else {
+			finishReason = "stop"
+		}
+		template, _ = sjson.Set(template, "choices.0.finish_reason", finishReason)
+		template, _ = sjson.Set(template, "choices.0.native_finish_reason", strings.ToLower(upstreamFinishReason))
 	}
 
 	return []string{template}

internal/translator/antigravity/openai/chat-completions/antigravity_openai_response_test.go

@@ -0,0 +1,128 @@
+package chat_completions
+
+import (
+	"context"
+	"testing"
+
+	"github.com/tidwall/gjson"
+)
+
+func TestFinishReasonToolCallsNotOverwritten(t *testing.T) {
+	ctx := context.Background()
+	var param any
+
+	// Chunk 1: Contains functionCall - should set SawToolCall = true
+	chunk1 := []byte(`{"response":{"candidates":[{"content":{"parts":[{"functionCall":{"name":"list_files","args":{"path":"."}}}]}}]}}`)
+	result1 := ConvertAntigravityResponseToOpenAI(ctx, "model", nil, nil, chunk1, &param)
+
+	// Verify chunk1 has no finish_reason (null)
+	if len(result1) != 1 {
+		t.Fatalf("Expected 1 result from chunk1, got %d", len(result1))
+	}
+	fr1 := gjson.Get(result1[0], "choices.0.finish_reason")
+	if fr1.Exists() && fr1.String() != "" && fr1.Type.String() != "Null" {
+		t.Errorf("Expected finish_reason to be null in chunk1, got: %v", fr1.String())
+	}
+
+	// Chunk 2: Contains finishReason STOP + usage (final chunk, no functionCall)
+	// This simulates what the upstream sends AFTER the tool call chunk
+	chunk2 := []byte(`{"response":{"candidates":[{"finishReason":"STOP"}],"usageMetadata":{"promptTokenCount":10,"candidatesTokenCount":20,"totalTokenCount":30}}}`)
+	result2 := ConvertAntigravityResponseToOpenAI(ctx, "model", nil, nil, chunk2, &param)
+
+	// Verify chunk2 has finish_reason: "tool_calls" (not "stop")
+	if len(result2) != 1 {
+		t.Fatalf("Expected 1 result from chunk2, got %d", len(result2))
+	}
+	fr2 := gjson.Get(result2[0], "choices.0.finish_reason").String()
+	if fr2 != "tool_calls" {
+		t.Errorf("Expected finish_reason 'tool_calls', got: %s", fr2)
+	}
+
+	// Verify native_finish_reason is lowercase upstream value
+	nfr2 := gjson.Get(result2[0], "choices.0.native_finish_reason").String()
+	if nfr2 != "stop" {
+		t.Errorf("Expected native_finish_reason 'stop', got: %s", nfr2)
+	}
+}
+
+func TestFinishReasonStopForNormalText(t *testing.T) {
+	ctx := context.Background()
+	var param any
+
+	// Chunk 1: Text content only
+	chunk1 := []byte(`{"response":{"candidates":[{"content":{"parts":[{"text":"Hello world"}]}}]}}`)
+	ConvertAntigravityResponseToOpenAI(ctx, "model", nil, nil, chunk1, &param)
+
+	// Chunk 2: Final chunk with STOP
+	chunk2 := []byte(`{"response":{"candidates":[{"finishReason":"STOP"}],"usageMetadata":{"promptTokenCount":10,"candidatesTokenCount":5,"totalTokenCount":15}}}`)
+	result2 := ConvertAntigravityResponseToOpenAI(ctx, "model", nil, nil, chunk2, &param)
+
+	// Verify finish_reason is "stop" (no tool calls were made)
+	fr := gjson.Get(result2[0], "choices.0.finish_reason").String()
+	if fr != "stop" {
+		t.Errorf("Expected finish_reason 'stop', got: %s", fr)
+	}
+}
+
+func TestFinishReasonMaxTokens(t *testing.T) {
+	ctx := context.Background()
+	var param any
+
+	// Chunk 1: Text content
+	chunk1 := []byte(`{"response":{"candidates":[{"content":{"parts":[{"text":"Hello"}]}}]}}`)
+	ConvertAntigravityResponseToOpenAI(ctx, "model", nil, nil, chunk1, &param)
+
+	// Chunk 2: Final chunk with MAX_TOKENS
+	chunk2 := []byte(`{"response":{"candidates":[{"finishReason":"MAX_TOKENS"}],"usageMetadata":{"promptTokenCount":10,"candidatesTokenCount":100,"totalTokenCount":110}}}`)
+	result2 := ConvertAntigravityResponseToOpenAI(ctx, "model", nil, nil, chunk2, &param)
+
+	// Verify finish_reason is "max_tokens"
+	fr := gjson.Get(result2[0], "choices.0.finish_reason").String()
+	if fr != "max_tokens" {
+		t.Errorf("Expected finish_reason 'max_tokens', got: %s", fr)
+	}
+}
+
+func TestToolCallTakesPriorityOverMaxTokens(t *testing.T) {
+	ctx := context.Background()
+	var param any
+
+	// Chunk 1: Contains functionCall
+	chunk1 := []byte(`{"response":{"candidates":[{"content":{"parts":[{"functionCall":{"name":"test","args":{}}}]}}]}}`)
+	ConvertAntigravityResponseToOpenAI(ctx, "model", nil, nil, chunk1, &param)
+
+	// Chunk 2: Final chunk with MAX_TOKENS (but we had a tool call, so tool_calls should win)
+	chunk2 := []byte(`{"response":{"candidates":[{"finishReason":"MAX_TOKENS"}],"usageMetadata":{"promptTokenCount":10,"candidatesTokenCount":100,"totalTokenCount":110}}}`)
+	result2 := ConvertAntigravityResponseToOpenAI(ctx, "model", nil, nil, chunk2, &param)
+
+	// Verify finish_reason is "tool_calls" (takes priority over max_tokens)
+	fr := gjson.Get(result2[0], "choices.0.finish_reason").String()
+	if fr != "tool_calls" {
+		t.Errorf("Expected finish_reason 'tool_calls', got: %s", fr)
+	}
+}
+
+func TestNoFinishReasonOnIntermediateChunks(t *testing.T) {
+	ctx := context.Background()
+	var param any
+
+	// Chunk 1: Text content (no finish reason, no usage)
+	chunk1 := []byte(`{"response":{"candidates":[{"content":{"parts":[{"text":"Hello"}]}}]}}`)
+	result1 := ConvertAntigravityResponseToOpenAI(ctx, "model", nil, nil, chunk1, &param)
+
+	// Verify no finish_reason on intermediate chunk
+	fr1 := gjson.Get(result1[0], "choices.0.finish_reason")
+	if fr1.Exists() && fr1.String() != "" && fr1.Type.String() != "Null" {
+		t.Errorf("Expected no finish_reason on intermediate chunk, got: %v", fr1)
+	}
+
+	// Chunk 2: More text (no finish reason, no usage)
+	chunk2 := []byte(`{"response":{"candidates":[{"content":{"parts":[{"text":" world"}]}}]}}`)
+	result2 := ConvertAntigravityResponseToOpenAI(ctx, "model", nil, nil, chunk2, &param)
+
+	// Verify no finish_reason on intermediate chunk
+	fr2 := gjson.Get(result2[0], "choices.0.finish_reason")
+	if fr2.Exists() && fr2.String() != "" && fr2.Type.String() != "Null" {
+		t.Errorf("Expected no finish_reason on intermediate chunk, got: %v", fr2)
+	}
+}

sdk/api/handlers/handlers.go

@@ -652,7 +652,7 @@ func (h *BaseAPIHandler) getRequestDetails(modelName string) (providers []string
 	}
 
 	if len(providers) == 0 {
-		return nil, "", &interfaces.ErrorMessage{StatusCode: http.StatusBadRequest, Error: fmt.Errorf("unknown provider for model %s", modelName)}
+		return nil, "", &interfaces.ErrorMessage{StatusCode: http.StatusBadGateway, Error: fmt.Errorf("unknown provider for model %s", modelName)}
 	}
 
 	// The thinking suffix is preserved in the model name itself, so no

sdk/logging/request_logger.go

@@ -3,6 +3,8 @@ package logging
 
 import internallogging "github.com/router-for-me/CLIProxyAPI/v6/internal/logging"
 
+const defaultErrorLogsMaxFiles = 10
+
 // RequestLogger defines the interface for logging HTTP requests and responses.
 type RequestLogger = internallogging.RequestLogger
 
@@ -12,7 +14,12 @@ type StreamingLogWriter = internallogging.StreamingLogWriter
 // FileRequestLogger implements RequestLogger using file-based storage.
 type FileRequestLogger = internallogging.FileRequestLogger
 
-// NewFileRequestLogger creates a new file-based request logger.
+// NewFileRequestLogger creates a new file-based request logger with default error log retention (10 files).
 func NewFileRequestLogger(enabled bool, logsDir string, configDir string) *FileRequestLogger {
-	return internallogging.NewFileRequestLogger(enabled, logsDir, configDir)
+	return internallogging.NewFileRequestLogger(enabled, logsDir, configDir, defaultErrorLogsMaxFiles)
+}
+
+// NewFileRequestLoggerWithOptions creates a new file-based request logger with configurable error log retention.
+func NewFileRequestLoggerWithOptions(enabled bool, logsDir string, configDir string, errorLogsMaxFiles int) *FileRequestLogger {
+	return internallogging.NewFileRequestLogger(enabled, logsDir, configDir, errorLogsMaxFiles)
 }