package main

import (
	"bufio"
	"encoding/json"
	"errors"
	"io"
	"log"
	"net/url"
	"os"
	"time"

	"github.com/gorilla/websocket"
	"github.com/labstack/gommon/random"
)

type RecogStart struct {
	Type          string `json:"type"`
	Service       string `json:"service"`
	ShowFinalOnly bool   `json:"showFinalOnly"`
	ShowExtraInfo bool   `json:"showExtraInfo"`
	RecogLongMaxWaitTime int    `json:"recogLongMaxWaitTime"`
	AudioFormat   string `json:"audioFormat"`
	RequestId     string `json:"requestId"`
    PassInfo      PassInfoPayload `json:"passinfo"`
}


type AIaaSInfo struct {
	UserID        string `json:"userId,omitempty"`
	Domain        string `json:"domain,omitempty"`
	Phase         string `json:"phase,omitempty"`
	ShowFinalOnly bool   `json:"showFinalOnly,omitempty"`
	ShowExtraInfo bool   `json:"showExtraInfo,omitempty"`
}

type GivenEPD struct {
	Active bool `json:"active,omitempty"`
	Start  int  `json:"start,omitempty"`
	End    int  `json:"end,omitempty"`
}

type PassInfoPayload struct {
	AIaaSInfo                `json:"aiaasInfo,omitempty"`
	GivenEPD                 `json:"givenepd,omitempty"`
	RequestID                string      `json:"requestId,omitempty"`
	KAccountID               int         `json:"kaccountid,omitempty"`
	AIID                     string      `json:"aiid,omitempty"`
	UserAgent                interface{} `json:"userAgent,omitempty"`
	ServiceID                string      `json:"serviceId,omitempty"`
	ServiceName              string      `json:"serviceName,omitempty"`
	Format                   string      `json:"format,omitempty"`
	Activator                interface{} `json:"activator,omitempty"`
	VoiceProfile             interface{} `json:"voiceProfile,omitempty"`
	Domains                  interface{} `json:"domains,omitempty"`
	Entities                 interface{} `json:"entities,omitempty"`
	PreserveVoiceData        bool        `json:"preserveVoiceData,omitempty"`
	EnableSpeakerDiarization bool        `json:"enableSpeakerDiarization,omitempty"`
	DialogRequestId          string      `json:"dialogRequestId,omitempty"`
	DialogRequestOffset      int         `json:"dialogRequestOffset,omitempty"`
	RecogLongMaxWaitTime     int         `json:"recogLongMaxWaitTime,omitempty"`
	EPDTimeOutMS             int         `json:"epdTimeOutMS,omitempty"`
	EPDHangOverMS            int         `json:"epdHangOverMS,omitempty"`
	SpeakerDiarizationSize   int         `json:"speakerDiarizationSize,omitempty"`
	TTSConfig                interface{} `json:"ttsConfig,omitempty"`
	TTSSource                interface{} `json:"ttsSource,omitempty"`
}

const (
	DefaultPreserveVoiceData    = true
	DefaultRecogLongMaxWaitTime = 60000
	DefaultRecogFormat          = "RAWPCM/16/16000/1/_/_"
    //DefaultRecogFormat          = "RAWPCM/16/8000/1/_/_"
)

func NewPassInfoPayload() PassInfoPayload {
	return PassInfoPayload{
		AIaaSInfo:            AIaaSInfo{},
		Format:               DefaultRecogFormat,
		RecogLongMaxWaitTime: DefaultRecogLongMaxWaitTime,
		PreserveVoiceData:    DefaultPreserveVoiceData,
		RequestID:            "",
	}
}

type RecogEnd struct {
	Type string `json:"type"`
}

type ResponseMessage struct {
	Type  string `json:"type"`
	Value string `json:"value"`
}

func main() {
	log.SetFlags(log.Lmicroseconds)

	idxstr := random.String(15)
	requestId := "GNT-" + idxstr

	// must modify endpoint
    endPointURI := "wss://prod-newtone-api.i.kakao.com/v2/ai/newtone-recognize/ws"

	endPoint, err := url.Parse(endPointURI)
	if err != nil {
		log.Fatalf("[%s] url parse err: %s", requestId, err.Error())
	}

	// must modify file path
	filePath := "./heykakao_hello.wav" //"{File Path}"
	f, err := os.Open(filePath)
	if err != nil {
		log.Fatalf("[%s] file open err: %s", requestId, err.Error())
	}
	defer f.Close()

	recognize(endPoint, f, requestId)
}

func recognize(u *url.URL, r io.Reader, requestId string) {
	log.Printf("[%s] connecting to %s", requestId, u.String())

	headers := http.Header{}
	headers.Set("Authorization", "KakaoAK ${REST_API_KEY}")
	c, _, err := websocket.DefaultDialer.Dial(u.String(), headers)
	if err != nil {
		log.Fatalf("[%s] dial: %s", requestId, err.Error())
	}
	defer c.Close()

	passinfo1 := NewPassInfoPayload()

	// 1. send recogStart message
	startMsg := RecogStart{
		Type:          "recogStart",
		Service:       "DICTATION",
		ShowFinalOnly: false,
		ShowExtraInfo: false,
		RecogLongMaxWaitTime: 3600000,
		AudioFormat:   "RAWPCM/16/16000/1/_/_",
        //AudioFormat:   "RAWPCM/16/8000/1/_/_",
		RequestId:     requestId,
		PassInfo:      passinfo1,
        //PassInfo: string(passinfo),
	}
	if err := c.WriteJSON(startMsg); err != nil {
		log.Printf("[%s] start err: %s", requestId, err.Error())
		return
	}
	log.Printf("[%s] send: %+v", requestId, startMsg)

	// 2. async read server message
	ticker := time.NewTicker(65 * time.Minute)
	defer ticker.Stop()

	exit := make(chan bool, 1)

	go func() {
		defer func() {
			exit <- true
		}()
		for {
			select {
			case <-ticker.C:
				log.Printf("[%s] exit by ticker", requestId)
				return
			default:
				_, message, err := c.ReadMessage()
				if err != nil {
					log.Printf("[%s] ReadMessage: %s", requestId, err.Error())
					return
				}
				log.Printf("[%s] recv: %s", requestId, message)

				var respMsg ResponseMessage
				json.Unmarshal(message, &respMsg)
				if respMsg.Type == "ELR" {
					return
				}
			}
		}
	}()

	// 3. send audio binary message (file)
	br := bufio.NewReader(r)

	data := make([]byte, 640)

	cumulativeReadSize := 0
	readStartTime := time.Now()
	for {
		rbytes, err := br.Read(data)
		if err != nil {
			if errors.Is(err, io.EOF) {
				log.Printf("[%s] EOF", requestId)
				break
			} else {
				log.Printf("[%s] read err: %s", requestId, err.Error())
				return
			}
		}
		cumulativeReadSize += rbytes
		log.Printf("[%s] cumulative read size: %d", requestId, cumulativeReadSize)

		if err := c.WriteMessage(websocket.BinaryMessage, data[:rbytes]); err != nil {
			log.Printf("[%s] write: %s", requestId, err.Error())
			return
		}

		// control sending speed (640Bytes / 20ms)
		for cumulativeReadSize > 32*int(time.Now().Sub(readStartTime).Milliseconds()) {
			time.Sleep(20 * time.Millisecond)
		}
	}

	// 4. send recogEnd message
	endMsg := RecogEnd{Type: "recogEnd"}
	if err := c.WriteJSON(endMsg); err != nil {
		log.Printf("[%s] end err: %s", requestId, err.Error())
		return
	}
	log.Printf("[%s] send: %+v", requestId, endMsg)

	<-exit
}

{
  "type": "ready",
  "sessionId": "${SESSION_ID}"
}

package main

import (
	"bufio"
	"encoding/json"
	"errors"
	"io"
	"log"
	"net/url"
	"os"
	"time"

	"github.com/gorilla/websocket"
	"github.com/labstack/gommon/random"
)

type RecogStart struct {
	Type          string `json:"type"`
	Service       string `json:"service"`
	ShowFinalOnly bool   `json:"showFinalOnly"`
	ShowExtraInfo bool   `json:"showExtraInfo"`
	RecogLongMaxWaitTime int    `json:"recogLongMaxWaitTime"`
	AudioFormat   string `json:"audioFormat"`
	RequestId     string `json:"requestId"`
    PassInfo      PassInfoPayload `json:"passinfo"`
}


type AIaaSInfo struct {
	UserID        string `json:"userId,omitempty"`
	Domain        string `json:"domain,omitempty"`
	Phase         string `json:"phase,omitempty"`
	ShowFinalOnly bool   `json:"showFinalOnly,omitempty"`
	ShowExtraInfo bool   `json:"showExtraInfo,omitempty"`
}

type GivenEPD struct {
	Active bool `json:"active,omitempty"`
	Start  int  `json:"start,omitempty"`
	End    int  `json:"end,omitempty"`
}

type PassInfoPayload struct {
	AIaaSInfo                `json:"aiaasInfo,omitempty"`
	GivenEPD                 `json:"givenepd,omitempty"`
	RequestID                string      `json:"requestId,omitempty"`
	KAccountID               int         `json:"kaccountid,omitempty"`
	AIID                     string      `json:"aiid,omitempty"`
	UserAgent                interface{} `json:"userAgent,omitempty"`
	ServiceID                string      `json:"serviceId,omitempty"`
	ServiceName              string      `json:"serviceName,omitempty"`
	Format                   string      `json:"format,omitempty"`
	Activator                interface{} `json:"activator,omitempty"`
	VoiceProfile             interface{} `json:"voiceProfile,omitempty"`
	Domains                  interface{} `json:"domains,omitempty"`
	Entities                 interface{} `json:"entities,omitempty"`
	PreserveVoiceData        bool        `json:"preserveVoiceData,omitempty"`
	EnableSpeakerDiarization bool        `json:"enableSpeakerDiarization,omitempty"`
	DialogRequestId          string      `json:"dialogRequestId,omitempty"`
	DialogRequestOffset      int         `json:"dialogRequestOffset,omitempty"`
	RecogLongMaxWaitTime     int         `json:"recogLongMaxWaitTime,omitempty"`
	EPDTimeOutMS             int         `json:"epdTimeOutMS,omitempty"`
	EPDHangOverMS            int         `json:"epdHangOverMS,omitempty"`
	SpeakerDiarizationSize   int         `json:"speakerDiarizationSize,omitempty"`
	TTSConfig                interface{} `json:"ttsConfig,omitempty"`
	TTSSource                interface{} `json:"ttsSource,omitempty"`
}

const (
	DefaultPreserveVoiceData    = true
	DefaultRecogLongMaxWaitTime = 60000
	DefaultRecogFormat          = "RAWPCM/16/16000/1/_/_"
    //DefaultRecogFormat          = "RAWPCM/16/8000/1/_/_"
)

func NewPassInfoPayload() PassInfoPayload {
	return PassInfoPayload{
		AIaaSInfo:            AIaaSInfo{},
		Format:               DefaultRecogFormat,
		RecogLongMaxWaitTime: DefaultRecogLongMaxWaitTime,
		PreserveVoiceData:    DefaultPreserveVoiceData,
		RequestID:            "",
	}
}

type RecogEnd struct {
	Type string `json:"type"`
}

type ResponseMessage struct {
	Type  string `json:"type"`
	Value string `json:"value"`
}

func main() {
	log.SetFlags(log.Lmicroseconds)

	idxstr := random.String(15)
	requestId := "GNT-" + idxstr

	// must modify endpoint
    endPointURI := "wss://prod-newtone-api.i.kakao.com/v2/ai/newtone-recognize/long/ws"

	endPoint, err := url.Parse(endPointURI)
	if err != nil {
		log.Fatalf("[%s] url parse err: %s", requestId, err.Error())
	}

	// must modify file path
	filePath := "./long_sample.wav" //"{File Path}"
	f, err := os.Open(filePath)
	if err != nil {
		log.Fatalf("[%s] file open err: %s", requestId, err.Error())
	}
	defer f.Close()

	recognize(endPoint, f, requestId)
}

func recognize(u *url.URL, r io.Reader, requestId string) {
	log.Printf("[%s] connecting to %s", requestId, u.String())

	headers := http.Header{}
	headers.Set("Authorization", "KakaoAK ${REST_API_KEY}")
	c, _, err := websocket.DefaultDialer.Dial(u.String(), headers)
	if err != nil {
		log.Fatalf("[%s] dial: %s", requestId, err.Error())
	}
	defer c.Close()

	passinfo1 := NewPassInfoPayload()

	// 1. send recogStart message
	startMsg := RecogStart{
		Type:          "recogStart",
		Service:       "DICTATION",
		ShowFinalOnly: false,
		ShowExtraInfo: false,
		RecogLongMaxWaitTime: 3600000,
		AudioFormat:   "RAWPCM/16/16000/1/_/_",
        //AudioFormat:   "RAWPCM/16/8000/1/_/_",
		RequestId:     requestId,
		PassInfo:      passinfo1,
        //PassInfo: string(passinfo),
	}
	if err := c.WriteJSON(startMsg); err != nil {
		log.Printf("[%s] start err: %s", requestId, err.Error())
		return
	}
	log.Printf("[%s] send: %+v", requestId, startMsg)

	// 2. async read server message
	ticker := time.NewTicker(65 * time.Minute)
	defer ticker.Stop()

	exit := make(chan bool, 1)

	go func() {
		defer func() {
			exit <- true
		}()
		for {
			select {
			case <-ticker.C:
				log.Printf("[%s] exit by ticker", requestId)
				return
			default:
				_, message, err := c.ReadMessage()
				if err != nil {
					log.Printf("[%s] ReadMessage: %s", requestId, err.Error())
					return
				}
				log.Printf("[%s] recv: %s", requestId, message)

				var respMsg ResponseMessage
				json.Unmarshal(message, &respMsg)
				if respMsg.Type == "ELR" {
					return
				}
			}
		}
	}()

	// 3. send audio binary message (file)
	br := bufio.NewReader(r)

	data := make([]byte, 640)

	cumulativeReadSize := 0
	readStartTime := time.Now()
	for {
		rbytes, err := br.Read(data)
		if err != nil {
			if errors.Is(err, io.EOF) {
				log.Printf("[%s] EOF", requestId)
				break
			} else {
				log.Printf("[%s] read err: %s", requestId, err.Error())
				return
			}
		}
		cumulativeReadSize += rbytes
		log.Printf("[%s] cumulative read size: %d", requestId, cumulativeReadSize)

		if err := c.WriteMessage(websocket.BinaryMessage, data[:rbytes]); err != nil {
			log.Printf("[%s] write: %s", requestId, err.Error())
			return
		}

		// control sending speed (640Bytes / 20ms)
		for cumulativeReadSize > 32*int(time.Now().Sub(readStartTime).Milliseconds()) {
			time.Sleep(20 * time.Millisecond)
		}
	}

	// 4. send recogEnd message
	endMsg := RecogEnd{Type: "recogEnd"}
	if err := c.WriteJSON(endMsg); err != nil {
		log.Printf("[%s] end err: %s", requestId, err.Error())
		return
	}
	log.Printf("[%s] send: %+v", requestId, endMsg)

	<-exit
}

{
  "type": "ready",
  "sessionId": "${SESSION_ID}"
}