File size: 9,600 Bytes
87337b1
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
#
#
# Agora Real Time Engagement
# Created by XinHui Li in 2024.
# Copyright (c) 2024 Agora IO. All rights reserved.
#
#

from dataclasses import dataclass
from typing import AsyncIterator, Tuple

from ten_ai_base.config import BaseConfig
from ten import (
    AsyncTenEnv,
)

import copy
import websockets
import uuid
import json
import gzip
import asyncio
import threading
from datetime import datetime


MESSAGE_TYPES = {
    11: "audio-only server response",
    12: "frontend server response",
    15: "error message from server",
}
MESSAGE_TYPE_SPECIFIC_FLAGS = {
    0: "no sequence number",
    1: "sequence number > 0",
    2: "last message from server (seq < 0)",
    3: "sequence number < 0",
}
MESSAGE_SERIALIZATION_METHODS = {0: "no serialization", 1: "JSON", 15: "custom type"}
MESSAGE_COMPRESSIONS = {0: "no compression", 1: "gzip", 15: "custom compression method"}

LATENCY_SAMPLE_INTERVAL_MS = 5


@dataclass
class TTSConfig(BaseConfig):
    # Parameters, refer to: https://www.volcengine.com/docs/6561/79823.
    appid: str = ""
    token: str = ""

    # Refer to: https://www.volcengine.com/docs/6561/1257544.
    voice_type: str = "BV001_streaming"
    sample_rate: int = 16000
    api_url: str = "wss://openspeech.bytedance.com/api/v1/tts/ws_binary"
    cluster: str = "volcano_tts"


class TTSClient:
    def __init__(self, config: TTSConfig, ten_env: AsyncTenEnv) -> None:
        self.config = config
        self.websocket = None
        self.ten_env = ten_env

        # Refer to: https://www.volcengine.com/docs/6561/79823.
        self.request_template = {
            "app": {
                "appid": self.config.appid,
                "token": "access_token",
                "cluster": self.config.cluster,
            },
            "user": {"uid": ""},  # Any non-empty string, used for tracing.
            "audio": {
                "rate": self.config.sample_rate,
                "voice_type": self.config.voice_type,
                "encoding": "pcm",
                "speed_ratio": 1.0,
                "volume_ratio": 1.0,
                "pitch_ratio": 1.0,
            },
            "request": {
                "reqid": "",  # Must be unique for each request.
                "text": "",  # Text to be synthesized.
                "text_type": "plain",
                "operation": "submit",
            },
        }

        # version: b0001 (4 bits)
        # header size: b0001 (4 bits)
        # message type: b0001 (Full client request) (4bits)
        # message type specific flags: b0000 (none) (4bits)
        # message serialization method: b0001 (JSON) (4 bits)
        # message compression: b0001 (gzip) (4bits)
        # reserved data: 0x00 (1 byte)
        self.default_header = bytearray(b"\x11\x10\x11\x00")
        self._cancel = threading.Event()

        # Latency.
        self._latest_record_time = None

    def is_cancelled(self) -> bool:
        return self._cancel.is_set()

    async def cancel(self) -> None:
        self._cancel.set()

    async def connect(self) -> None:
        header = {"Authorization": f"Bearer; {self.config.token}"}
        self.websocket = await websockets.connect(
            self.config.api_url,
            extra_headers=header,
            ping_interval=None,
            close_timeout=1,  # Fast close, as the `flush` cmd will close the connection.
        )
        self.ten_env.log_info("Websocket connection established.")

    async def close(self) -> None:
        if self.websocket is not None:
            await self.websocket.close()
            self.websocket = None
            self.ten_env.log_info("Websocket connection closed.")
        else:
            self.ten_env.log_info("Websocket is not connected.")

    async def reconnect(self) -> None:
        await self.close()
        await self.connect()

    def parse_response(self, response: websockets.Data) -> Tuple[bytes, bool]:
        protocol_version = response[0] >> 4
        header_size = response[0] & 0x0F
        message_type = response[1] >> 4
        message_type_specific_flags = response[1] & 0x0F
        serialization_method = response[2] >> 4
        message_compression = response[2] & 0x0F
        reserved = response[3]
        header_extensions = response[4 : header_size * 4]
        payload = response[header_size * 4 :]
        self.ten_env.log_debug(
            f"Protocol version: {protocol_version:#x} - version {protocol_version}"
        )
        self.ten_env.log_debug(
            f"Header size: {header_size:#x} - {header_size * 4} bytes"
        )
        self.ten_env.log_debug(
            f"Message type: {message_type:#x} - {MESSAGE_TYPES[message_type]}"
        )
        self.ten_env.log_debug(
            f"Message type specific flags: {message_type_specific_flags:#x} - {MESSAGE_TYPE_SPECIFIC_FLAGS[message_type_specific_flags]}"
        )
        self.ten_env.log_debug(
            f"Message serialization method: {serialization_method:#x} - {MESSAGE_SERIALIZATION_METHODS[serialization_method]}"
        )
        self.ten_env.log_debug(
            f"Message compression: {message_compression:#x} - {MESSAGE_COMPRESSIONS[message_compression]}"
        )
        self.ten_env.log_debug(f"Reserved: {reserved:#04x}")

        if header_size != 1:
            self.ten_env.log_debug(f"Header extensions: {header_extensions}")

        if message_type == 0xB:  # audio-only server response
            if message_type_specific_flags == 0:  # no sequence number as ACK
                self.ten_env.log_debug("Payload size: 0")
                return None, False
            else:
                sequence_number = int.from_bytes(payload[:4], "big", signed=True)
                payload_size = int.from_bytes(payload[4:8], "big", signed=False)
                payload = payload[8:]
                self.ten_env.log_debug(f"Sequence number: {sequence_number}")
                self.ten_env.log_debug(f"Payload size: {payload_size} bytes")
            if sequence_number < 0:
                return payload, True
            else:
                return payload, False
        elif message_type == 0xF:
            code = int.from_bytes(payload[:4], "big", signed=False)
            msg_size = int.from_bytes(payload[4:8], "big", signed=False)
            error_msg = payload[8:]
            if message_compression == 1:
                error_msg = gzip.decompress(error_msg)
            error_msg = str(error_msg, "utf-8")
            self.ten_env.log_error(f"Error message code: {code}")
            self.ten_env.log_error(f"Error message size: {msg_size} bytes")
            self.ten_env.log_error(f"Error message: {error_msg}")
            return None, True
        elif message_type == 0xC:
            msg_size = int.from_bytes(payload[:4], "big", signed=False)
            payload = payload[4:]
            if message_compression == 1:
                payload = gzip.decompress(payload)
            self.ten_env.log_debug(f"Frontend message: {payload}")
        else:
            self.ten_env.log_error("undefined message type!")
            return None, True

    def record_latency(self, request_id: str, start: datetime) -> None:
        end_time = datetime.now()

        if self._latest_record_time:
            sample_interval = datetime.now() - self._latest_record_time
            if sample_interval.total_seconds() < LATENCY_SAMPLE_INTERVAL_MS:
                return

        self._latest_record_time = end_time
        latency = int((end_time - start).total_seconds() * 1000)
        self.ten_env.log_info(f"Request ({request_id}), ttfb {latency}ms.")

    async def text_to_speech_stream(self, text: str) -> AsyncIterator[bytes]:
        ws = self.websocket
        if ws is None:
            await self.connect()
            ws = self.websocket

        start_ms = datetime.now()
        request_id = str(uuid.uuid4())

        request = copy.deepcopy(self.request_template)
        request["request"]["reqid"] = request_id
        request["request"]["text"] = text
        request["user"]["uid"] = str(uuid.uuid4())

        request_bytes = str.encode(json.dumps(request))
        request_bytes = gzip.compress(request_bytes)
        full_request = bytearray(self.default_header)

        # payload size(4 bytes)
        full_request.extend((len(request_bytes)).to_bytes(4, "big"))

        # payload
        full_request.extend(request_bytes)

        try:
            await ws.send(full_request)
            self.ten_env.log_info(f"Sent: {request}")

            while True:
                if self.is_cancelled():
                    self.ten_env.log_info(f"Request ({request_id}) has been cancelled.")

                    # Current connection should be closed, as the server will not drop the remain data.
                    await self.close()
                    self._cancel.clear()
                    break

                resp = await ws.recv()
                payload, done = self.parse_response(resp)

                if payload:
                    yield payload
                    self.record_latency(request_id, start_ms)

                if done:
                    self.ten_env.log_info(
                        f"Response is completed for request: {request_id}."
                    )
                    break

        except websockets.exceptions.ConnectionClosedError as e:
            self.ten_env.log_error(
                f"Connection is closed with error: {e}, request: {request_id}."
            )
            await self.connect()
        except asyncio.TimeoutError:
            self.ten_env.log_error("Timeout waiting for response.")