update to latest gpt model

cmd/chat/chat.go

@@ -17,6 +17,7 @@ func main() {
 	fmt.Println("---------------------")
 
 	systemPrompt := ""
+	//systemPrompt := "You are Mixtral, a locally hosted AI assistant."
 
 	messages := []openai.ChatCompletionMessage{
 		{
@@ -52,13 +53,12 @@ func main() {
 		}
 
 		fmt.Println("AI: ", resp.Choices[0].Message.Content)
-		//fmt.Println("Finish Reason: ", resp.Choices[0].FinishReason)
 		messages = append(messages, resp.Choices[0].Message)
 
 		currLength := estimateTokenCount(messages)
 
-		if currLength > 3000 {
-			fmt.Println("Token count exceeded 3000, summarizing context")
+		if currLength > 2000 {
+			fmt.Println("Token count exceeded 2000, summarizing context")
 			summarized, err := summarizeChatSoFar(messages)
 			if err != nil {
 				fmt.Printf("error summarizing chat so far | %v\n", err)

cmd/rateTest/rate.go

@@ -0,0 +1,31 @@
+package main
+
+import (
+	"fmt"
+	"sync"
+	"time"
+)
+
+type Rate struct {
+	RPM int
+	mu  sync.Mutex
+}
+
+func main() {
+	r := Rate{RPM: 10}
+	for i := 0; i < 10; i++ {
+		go func(j int) {
+			r.RateLimit()
+			fmt.Println("Rate limiting")
+		}(i)
+	}
+	time.Sleep(time.Minute)
+}
+
+func (r *Rate) RateLimit() {
+	// Lock for the fraction of a minute based on RPM
+	r.mu.Lock()
+	// Unlock after the fraction of a minute
+	time.Sleep(time.Minute / time.Duration(r.RPM))
+	r.mu.Unlock()
+}