package main import ( "bufio" "bytes" "elefant/models" "elefant/storage" "encoding/json" "fmt" "io" "log/slog" "net/http" "os" "strings" "time" "github.com/rivo/tview" ) var httpClient = http.Client{ Timeout: time.Second * 20, } var ( logger *slog.Logger userRole = "user" assistantRole = "assistant" toolRole = "tool" assistantIcon = "<🤖>: " userIcon = ": " // TODO: pass as an cli arg or have config APIURL = "http://localhost:8080/v1/chat/completions" logFileName = "log.txt" showSystemMsgs bool chunkLimit = 1000 activeChatName string chunkChan = make(chan string, 10) streamDone = make(chan bool, 1) chatBody *models.ChatBody store storage.FullRepo defaultFirstMsg = "Hello! What can I do for you?" defaultStarter = []models.MessagesStory{ {Role: "system", Content: systemMsg}, {Role: assistantRole, Content: defaultFirstMsg}, } defaultStarterBytes = []byte{} interruptResp = false ) // ==== func formMsg(chatBody *models.ChatBody, newMsg, role string) io.Reader { if newMsg != "" { // otherwise let the bot continue newMsg := models.MessagesStory{Role: role, Content: newMsg} chatBody.Messages = append(chatBody.Messages, newMsg) } data, err := json.Marshal(chatBody) if err != nil { logger.Error("failed to form a msg", "error", err) return nil } return bytes.NewReader(data) } // func sendMsgToLLM(body io.Reader) (*models.LLMRespChunk, error) { func sendMsgToLLM(body io.Reader) (any, error) { resp, err := httpClient.Post(APIURL, "application/json", body) if err != nil { logger.Error("llamacpp api", "error", err) return nil, err } defer resp.Body.Close() llmResp := []models.LLMRespChunk{} // chunkChan <- assistantIcon reader := bufio.NewReader(resp.Body) counter := 0 for { if interruptResp { interruptResp = false logger.Info("interrupted bot response") break } llmchunk := models.LLMRespChunk{} if counter > chunkLimit { logger.Warn("response hit chunk limit", "limit", chunkLimit) streamDone <- true break } line, err := reader.ReadBytes('\n') if err != nil { streamDone <- true logger.Error("error reading response body", "error", err) } // logger.Info("linecheck", "line", string(line), "len", len(line), "counter", counter) if len(line) <= 1 { continue // skip \n } // starts with -> data: line = line[6:] if err := json.Unmarshal(line, &llmchunk); err != nil { logger.Error("failed to decode", "error", err, "line", string(line)) streamDone <- true return nil, err } llmResp = append(llmResp, llmchunk) // logger.Info("streamview", "chunk", llmchunk) // if llmchunk.Choices[len(llmchunk.Choices)-1].FinishReason != "chat.completion.chunk" { if llmchunk.Choices[len(llmchunk.Choices)-1].FinishReason == "stop" { streamDone <- true // last chunk break } counter++ // bot sends way too many \n answerText := strings.ReplaceAll(llmchunk.Choices[0].Delta.Content, "\n\n", "\n") chunkChan <- answerText } return llmResp, nil } func chatRound(userMsg, role string, tv *tview.TextView) { botRespMode = true reader := formMsg(chatBody, userMsg, role) if reader == nil { logger.Error("empty reader from msgs", "role", role) return } go sendMsgToLLM(reader) if userMsg != "" { // no need to write assistant icon since we continue old message fmt.Fprintf(tv, fmt.Sprintf("(%d) ", len(chatBody.Messages))) fmt.Fprintf(tv, assistantIcon) } respText := strings.Builder{} out: for { select { case chunk := <-chunkChan: // fmt.Printf(chunk) fmt.Fprintf(tv, chunk) respText.WriteString(chunk) tv.ScrollToEnd() case <-streamDone: break out } } botRespMode = false chatBody.Messages = append(chatBody.Messages, models.MessagesStory{ Role: assistantRole, Content: respText.String(), }) // bot msg is done; // now check it for func call // logChat(activeChatName, chatBody.Messages) err := updateStorageChat(activeChatName, chatBody.Messages) if err != nil { logger.Warn("failed to update storage", "error", err, "name", activeChatName) } findCall(respText.String(), tv) } func findCall(msg string, tv *tview.TextView) { // prefix := "__tool_call__\n" // suffix := "\n__tool_call__" // if !strings.HasPrefix(msg, prefix) || // !strings.HasSuffix(msg, suffix) { // return // } // jsStr := strings.TrimSuffix(strings.TrimPrefix(msg, prefix), suffix) fc := models.FuncCall{} jsStr := toolCallRE.FindString(msg) if jsStr == "" { // tool call not found return } if err := json.Unmarshal([]byte(jsStr), &fc); err != nil { logger.Error("failed to unmarshal tool call", "error", err) return } // call a func f, ok := fnMap[fc.Name] if !ok { m := fmt.Sprintf("%s is not implemented", fc.Name) chatRound(m, toolRole, tv) return } resp := f(fc.Args) toolMsg := fmt.Sprintf("tool response: %+v", resp) // reader := formMsg(chatBody, toolMsg, toolRole) // sendMsgToLLM() chatRound(toolMsg, toolRole, tv) // return func result to the llm } func chatToTextSlice(showSys bool) []string { resp := make([]string, len(chatBody.Messages)) for i, msg := range chatBody.Messages { if !showSys && (msg.Role != assistantRole && msg.Role != userRole) { continue } resp[i] = msg.ToText(i) } return resp } func chatToText(showSys bool) string { s := chatToTextSlice(showSys) return strings.Join(s, "") } func textToMsg(rawMsg string) models.MessagesStory { msg := models.MessagesStory{} // system and tool? if strings.HasPrefix(rawMsg, assistantIcon) { msg.Role = assistantRole msg.Content = strings.TrimPrefix(rawMsg, assistantIcon) return msg } if strings.HasPrefix(rawMsg, userIcon) { msg.Role = userRole msg.Content = strings.TrimPrefix(rawMsg, userIcon) return msg } return msg } func textSliceToChat(chat []string) []models.MessagesStory { resp := make([]models.MessagesStory, len(chat)) for i, rawMsg := range chat { msg := textToMsg(rawMsg) resp[i] = msg } return resp } func init() { file, err := os.OpenFile(logFileName, os.O_APPEND|os.O_CREATE|os.O_WRONLY, 0644) if err != nil { logger.Error("failed to open log file", "error", err, "filename", logFileName) return } defaultStarterBytes, err = json.Marshal(defaultStarter) if err != nil { logger.Error("failed to marshal defaultStarter", "error", err) return } logger = slog.New(slog.NewTextHandler(file, nil)) store = storage.NewProviderSQL("test.db", logger) // https://github.com/coreydaley/ggerganov-llama.cpp/blob/master/examples/server/README.md // load all chats in memory loadHistoryChats() lastChat := loadOldChatOrGetNew() logger.Info("loaded history", "chat", lastChat) chatBody = &models.ChatBody{ Model: "modl_name", Stream: true, Messages: lastChat, } }