gateway

const (
	HeaderErrorInvalidRouting = "x-error-invalid-routing-strategy"

	// General Error Headers
	HeaderErrorUser                  = "x-error-user"
	HeaderErrorRouting               = "x-error-routing"
	HeaderErrorRequestBodyProcessing = "x-error-request-body-processing"
	HeaderErrorResponseUnmarshal     = "x-error-response-unmarshal"
	HeaderErrorResponseUnknown       = "x-error-response-unknown"

	// Model & Deployment Headers
	HeaderErrorNoModelInRequest = "x-error-no-model-in-request"
	HeaderErrorNoModelBackends  = "x-error-no-model-backends"

	// Streaming Headers
	HeaderErrorStream                    = "x-error-stream"
	HeaderErrorStreaming                 = "x-error-streaming"
	HeaderErrorStreamOptionsIncludeUsage = "x-error-no-stream-options-include-usage"

	// Multipart/Audio Headers
	HeaderErrorMultipartParsing = "x-error-multipart-parsing"

	// Request & Target Headers
	HeaderWentIntoReqHeaders = "x-went-into-req-headers"
	HeaderTargetPodIP        = "target-pod-ip"
	HeaderTargetPod          = "target-pod"
	HeaderRoutingStrategy    = "routing-strategy"
	HeaderRequestID          = "request-id"
	HeaderModel              = "model"
	HeaderExternalFilter     = "external-filter"
	HeaderConfigProfile      = "config-profile"
	// HeaderSessionID is the header used for session affinity routing.
	// NOTE: If you change this value, you MUST also update sessionIDHeader in
	// pkg/plugins/gateway/algorithms/simple_session_affinity.go
	HeaderSessionID   = "x-session-id"
	HeaderTraceParent = "traceparent"

	// RPM & TPM Update Errors
	HeaderUpdateTPM        = "x-update-tpm"
	HeaderUpdateRPM        = "x-update-rpm"
	HeaderErrorRPMExceeded = "x-error-rpm-exceeded"
	HeaderErrorTPMExceeded = "x-error-tpm-exceeded"
	HeaderErrorIncrRPM     = "x-error-incr-rpm"
	HeaderErrorIncrTPM     = "x-error-incr-tpm"

	// Model RPS Errors
	HeaderErrorModelRPSExceeded = "x-error-model-rps-exceeded"
	HeaderErrorIncrModelRPS     = "x-error-incr-model-rps"

	// Rate Limiting defaults
	DefaultRPM           = 100
	DefaultTPMMultiplier = 1000

	// Envs
	EnvRoutingAlgorithm = "ROUTING_ALGORITHM"

	// OpenAI Error Types
	ErrorTypeInvalidRequest = "invalid_request_error"
	ErrorTypeAuthentication = "authentication_error"
	ErrorTypeRateLimit      = "rate_limit_error"
	ErrorTypeApi            = "api_error"
	ErrorTypeOverloaded     = "overloaded_error"

	// OpenAI Error Codes
	ErrorCodeInvalidAPIKey      = "invalid_api_key"
	ErrorCodeModelNotFound      = "model_not_found"
	ErrorCodeRateLimitExceeded  = "rate_limit_exceeded"
	ErrorCodeServiceUnavailable = "service_unavailable"

	// Embedding Constraints
	// https://github.com/openai/openai-go/blob/main/embedding.go#L126
	MaxInputTokensPerModel = 8192
	MaxTotalTokens         = 300000
	MaxArrayDimensions     = 2048

	// Request Paths
	PathChatCompletions     = "/v1/chat/completions"
	PathResponses           = "/v1/responses"
	PathMessages            = "/v1/messages"
	PathCompletions         = "/v1/completions"
	PathEmbeddings          = "/v1/embeddings"
	PathImagesGenerations   = "/v1/images/generations"
	PathVideoGenerations    = "/v1/video/generations"
	PathAudioTranscriptions = "/v1/audio/transcriptions"
	PathAudioTranslations   = "/v1/audio/translations"
	PathRerank              = "/v1/rerank"
	PathClassify            = "/v1/classify"

	// Engine-specific paths (xdit)
	PathXditGenerate      = "/generate"
	PathXditGenerateVideo = "/generatevideo"

	// Engine Types
	EngineXdit = "xdit"
)