From a917d7557dd2b1e5263eeba7e5e4d5a5fc02f69f Mon Sep 17 00:00:00 2001
From: 游雁 <zhifu.gzf@alibaba-inc.com>
Date: 星期四, 27 四月 2023 11:41:16 +0800
Subject: [PATCH] websocket

---
 funasr/runtime/python/websocket/ASR_client.py |  153 ++++++++++++++++++++++++++++++++++++++------------
 1 files changed, 116 insertions(+), 37 deletions(-)

diff --git a/funasr/runtime/python/websocket/ASR_client.py b/funasr/runtime/python/websocket/ASR_client.py
index 7dce880..9a4a148 100644
--- a/funasr/runtime/python/websocket/ASR_client.py
+++ b/funasr/runtime/python/websocket/ASR_client.py
@@ -1,37 +1,46 @@
-import pyaudio
-import websocket #鍖哄埆鏈嶅姟绔繖閲屾槸 websocket-client搴�
+# -*- encoding: utf-8 -*-
 import time
 import websockets
 import asyncio
+# import threading
+import argparse
+import json
+
+parser = argparse.ArgumentParser()
+parser.add_argument("--host",
+                    type=str,
+                    default="localhost",
+                    required=False,
+                    help="host ip, localhost, 0.0.0.0")
+parser.add_argument("--port",
+                    type=int,
+                    default=10095,
+                    required=False,
+                    help="grpc server port")
+parser.add_argument("--chunk_size",
+                    type=int,
+                    default=300,
+                    help="ms")
+parser.add_argument("--audio_in",
+                    type=str,
+                    default=None,
+                    help="audio_in")
+
+args = parser.parse_args()
+
+# voices = asyncio.Queue()
 from queue import Queue
-import threading
 voices = Queue()
-async def hello():
-    global ws # 瀹氫箟涓�涓叏灞�鍙橀噺ws锛岀敤浜庝繚瀛榳ebsocket杩炴帴瀵硅薄
-    uri = "ws://localhost:8899"
-    ws = await websockets.connect(uri, subprotocols=["binary"]) # 鍒涘缓涓�涓暱杩炴帴
-    ws.max_size = 1024 * 1024 * 20
-    print("connected ws server")
-async def send(data):
-    global ws # 寮曠敤鍏ㄥ眬鍙橀噺ws
-    try:
-        await ws.send(data) # 閫氳繃ws瀵硅薄鍙戦�佹暟鎹�
-    except Exception as e:
-        print('Exception occurred:', e)
     
-
-
-asyncio.get_event_loop().run_until_complete(hello()) # 鍚姩鍗忕▼  
-
-
 # 鍏朵粬鍑芥暟鍙互閫氳繃璋冪敤send(data)鏉ュ彂閫佹暟鎹紝渚嬪锛�
-async def test():
+async def record_microphone():
+    import pyaudio
     #print("2")
-    global voices
+    global voices 
     FORMAT = pyaudio.paInt16
     CHANNELS = 1
     RATE = 16000
-    CHUNK = int(RATE / 1000 * 300)
+    CHUNK = int(RATE / 1000 * args.chunk_size)
 
     p = pyaudio.PyAudio()
 
@@ -40,34 +49,104 @@
                     rate=RATE,
                     input=True,
                     frames_per_buffer=CHUNK)
-
+    is_speaking = True
     while True:
 
         data = stream.read(CHUNK)
+        data = data.decode('ISO-8859-1')
+        message = json.dumps({"chunk": args.chunk_size, "is_speaking": is_speaking, "audio": data})
         
-        voices.put(data)
+        voices.put(message)
         #print(voices.qsize())
-        await asyncio.sleep(0.01)
-    
-      
 
+        await asyncio.sleep(0.005)
 
+# 鍏朵粬鍑芥暟鍙互閫氳繃璋冪敤send(data)鏉ュ彂閫佹暟鎹紝渚嬪锛�
+async def record_from_scp():
+    import wave
+    global voices
+    if args.audio_in.endswith(".scp"):
+        f_scp = open(args.audio_in)
+        wavs = f_scp.readlines()
+    else:
+        wavs = [args.audio_in]
+    for wav in wavs:
+        wav_splits = wav.strip().split()
+        wav_path = wav_splits[1] if len(wav_splits) > 1 else wav_splits[0]
+        # bytes_f = open(wav_path, "rb")
+        # bytes_data = bytes_f.read()
+        with wave.open(wav_path, "rb") as wav_file:
+            # 鑾峰彇闊抽鍙傛暟
+            params = wav_file.getparams()
+            # 鑾峰彇澶翠俊鎭殑闀垮害
+            # header_length = wav_file.getheaders()[0][1]
+            # 璇诲彇闊抽甯ф暟鎹紝璺宠繃澶翠俊鎭�
+            # wav_file.setpos(header_length)
+            frames = wav_file.readframes(wav_file.getnframes())
+
+        # 灏嗛煶棰戝抚鏁版嵁杞崲涓哄瓧鑺傜被鍨嬬殑鏁版嵁
+        audio_bytes = bytes(frames)
+        stride = int(args.chunk_size/1000*16000*2)
+        chunk_num = (len(audio_bytes)-1)//stride + 1
+        print(stride)
+        is_speaking = True
+        for i in range(chunk_num):
+            if i == chunk_num-1:
+                is_speaking = False
+            beg = i*stride
+            data = audio_bytes[beg:beg+stride]
+            data = data.decode('ISO-8859-1')
+            message = json.dumps({"chunk": args.chunk_size, "is_speaking": is_speaking, "audio": data})
+            voices.put(message)
+            # print("data_chunk: ", len(data_chunk))
+            # print(voices.qsize())
+        
+            await asyncio.sleep(args.chunk_size/1000)
+     
 
 async def ws_send():
     global voices
+    global websocket
     print("started to sending data!")
     while True:
         while not voices.empty():
             data = voices.get()
             voices.task_done()
-            await send(data)
-            await asyncio.sleep(0.01)
-        await asyncio.sleep(0.01)
+            try:
+                await websocket.send(data) # 閫氳繃ws瀵硅薄鍙戦�佹暟鎹�
+            except Exception as e:
+                print('Exception occurred:', e)
+            await asyncio.sleep(0.005)
+        await asyncio.sleep(0.005)
 
-async def main():
-    task = asyncio.create_task(test()) # 鍒涘缓涓�涓悗鍙颁换鍔�
-    task2 = asyncio.create_task(ws_send()) # 鍒涘缓涓�涓悗鍙颁换鍔�
-     
-    await asyncio.gather(task, task2)
 
-asyncio.run(main())
\ No newline at end of file
+
+async def message():
+    global websocket
+    while True:
+        try:
+            meg = await websocket.recv()
+            meg = json.loads(meg)
+            print(meg)
+        except Exception as e:
+            print("Exception:", e)          
+        
+
+
+async def ws_client():
+    global websocket # 瀹氫箟涓�涓叏灞�鍙橀噺ws锛岀敤浜庝繚瀛榳ebsocket杩炴帴瀵硅薄
+    # uri = "ws://11.167.134.197:8899"
+    uri = "ws://{}:{}".format(args.host, args.port)
+    #ws = await websockets.connect(uri, subprotocols=["binary"]) # 鍒涘缓涓�涓暱杩炴帴
+    async for websocket in websockets.connect(uri, subprotocols=["binary"], ping_interval=None):
+        if args.audio_in is not None:
+            task = asyncio.create_task(record_from_scp()) # 鍒涘缓涓�涓悗鍙颁换鍔″綍闊�
+        else:
+            task = asyncio.create_task(record_microphone())  # 鍒涘缓涓�涓悗鍙颁换鍔″綍闊�
+        task2 = asyncio.create_task(ws_send()) # 鍒涘缓涓�涓悗鍙颁换鍔″彂閫�
+        task3 = asyncio.create_task(message()) # 鍒涘缓涓�涓悗鍙版帴鏀舵秷鎭殑浠诲姟
+        await asyncio.gather(task, task2, task3)
+
+
+asyncio.get_event_loop().run_until_complete(ws_client()) # 鍚姩鍗忕▼
+asyncio.get_event_loop().run_forever()

--
Gitblit v1.9.1