Update api.go

pkg/inference/scheduling/api.go

@@ -7,38 +7,39 @@ import (
 )
 
 const (
-	// maximumOpenAIInferenceRequestSize is the maximum OpenAI API embedding or
-	// completion request size that Scheduler will allow. This should be large
-	// enough to encompass any real-world request but also small enough to avoid
-	// DoS attacks.
-	maximumOpenAIInferenceRequestSize = 10 * 1024 * 1024
+	// maximumOpenAIInferenceRequestSize defines the maximum size (in bytes) 
+	// allowed for an OpenAI API embedding or completion request.
+	// It should be large enough for real-world usage but small enough 
+	// to mitigate DoS risks.
+	maximumOpenAIInferenceRequestSize = 10 * 1024 * 1024 // 10 MB
 )
 
-// trimRequestPathToOpenAIRoot trims a request path to start at the first
-// instance of /v1/ to appear in the path.
+// trimRequestPathToOpenAIRoot returns the substring of path starting from
+// the first occurrence of "/v1/". If not found, it returns the original path.
 func trimRequestPathToOpenAIRoot(path string) string {
-	index := strings.Index(path, "/v1/")
-	if index == -1 {
-		return path
+	if idx := strings.Index(path, "/v1/"); idx != -1 {
+		return path[idx:]
 	}
-	return path[index:]
+	return path
 }
 
-// backendModeForRequest determines the backend operation mode to handle an
-// OpenAI inference request. Its second parameter is true if and only if a valid
-// mode could be determined.
+// backendModeForRequest maps an OpenAI API path to the appropriate
+// inference backend mode. Returns the mode and true if a valid mode is determined,
+// otherwise returns false.
 func backendModeForRequest(path string) (inference.BackendMode, bool) {
-	if strings.HasSuffix(path, "/v1/chat/completions") || strings.HasSuffix(path, "/v1/completions") {
+	switch {
+	case strings.HasSuffix(path, "/v1/chat/completions"), strings.HasSuffix(path, "/v1/completions"):
 		return inference.BackendModeCompletion, true
-	} else if strings.HasSuffix(path, "/v1/embeddings") {
+	case strings.HasSuffix(path, "/v1/embeddings"):
 		return inference.BackendModeEmbedding, true
+	default:
+		return inference.BackendMode(0), false
 	}
-	return inference.BackendMode(0), false
 }
 
-// OpenAIInferenceRequest is used to extract the model specification from either
-// a chat completion or embedding request in the OpenAI API.
+// OpenAIInferenceRequest represents the model information extracted from
+// a chat completion or embedding request payload to the OpenAI API.
 type OpenAIInferenceRequest struct {
-	// Model is the requested model name.
+	// Model specifies the model name requested.
 	Model string `json:"model"`
 }