Documentation
¶
Overview ¶
Package rapidmlx wraps the OpenAI-compatible HTTP surface exposed by Rapid-MLX (https://github.com/raullenchai/Rapid-MLX). It is a concrete provider type distinct from vLLM so the service layer can keep provider identity separate from utilization probing, which Rapid-MLX exposes on a different observability endpoint family.
Index ¶
Constants ¶
const DefaultBaseURL = "http://localhost:8000/v1"
Variables ¶
var ProtocolCapabilities = openai.OpenAIProtocolCapabilities
ProtocolCapabilities keeps Rapid-MLX on the standard OpenAI-compatible surface. The provider remains distinct from vLLM at the type level.
Functions ¶
Types ¶
type UtilizationProbe ¶
type UtilizationProbe struct {
// contains filtered or unexported fields
}
UtilizationProbe queries Rapid-MLX status endpoints and normalizes them into the shared endpoint utilization shape.
func NewUtilizationProbe ¶
func NewUtilizationProbe(baseURL string, client *http.Client) *UtilizationProbe
NewUtilizationProbe creates a probe for an OpenAI-compatible Rapid-MLX base URL.
func (*UtilizationProbe) Probe ¶
func (p *UtilizationProbe) Probe(ctx context.Context) utilization.EndpointUtilization
Probe fetches /v1/status from the OpenAI-compatible base URL and returns a normalized sample. Failures return stale or unknown utilization instead of surfacing endpoint unavailability.
Source Files