From 8706e767affc6bdc8cb7a67ca3a20a62779ff048 Mon Sep 17 00:00:00 2001
From: 雾聪 <wucong.lyb@alibaba-inc.com>
Date: 星期三, 17 五月 2023 15:45:46 +0800
Subject: [PATCH] Merge branch 'main' of https://github.com/alibaba-damo-academy/FunASR into main
---
funasr/runtime/python/websocket/ws_client.py | 136 ++++++++++++++++++++++++++++++++++-----------
1 files changed, 103 insertions(+), 33 deletions(-)
diff --git a/funasr/runtime/python/websocket/ws_client.py b/funasr/runtime/python/websocket/ws_client.py
index d8bbb65..470b3a8 100644
--- a/funasr/runtime/python/websocket/ws_client.py
+++ b/funasr/runtime/python/websocket/ws_client.py
@@ -7,7 +7,13 @@
import argparse
import json
import traceback
-from multiprocessing import Process
+from multiprocessing import Process
+from funasr.fileio.datadir_writer import DatadirWriter
+
+import logging
+
+logging.basicConfig(level=logging.ERROR)
+
parser = argparse.ArgumentParser()
parser.add_argument("--host",
type=str,
@@ -43,6 +49,10 @@
type=int,
default=100,
help="chunk")
+parser.add_argument("--output_dir",
+ type=str,
+ default=None,
+ help="output_dir")
args = parser.parse_args()
args.chunk_size = [int(x) for x in args.chunk_size.split(",")]
@@ -50,6 +60,11 @@
# voices = asyncio.Queue()
from queue import Queue
voices = Queue()
+
+ibest_writer = None
+if args.output_dir is not None:
+ writer = DatadirWriter(args.output_dir)
+ ibest_writer = writer[f"1best_recog"]
async def record_microphone():
is_finished = False
@@ -69,18 +84,19 @@
rate=RATE,
input=True,
frames_per_buffer=CHUNK)
- is_speaking = True
+
+ message = json.dumps({"chunk_size": args.chunk_size, "chunk_interval": args.chunk_interval, "wav_name": "microphone", "is_speaking": True})
+ voices.put(message)
while True:
data = stream.read(CHUNK)
- data = data.decode('ISO-8859-1')
- message = json.dumps({"chunk_size": args.chunk_size, "chunk_interval": args.chunk_interval, "audio": data, "is_speaking": is_speaking, "is_finished": is_finished})
+ message = data
voices.put(message)
await asyncio.sleep(0.005)
-async def record_from_scp():
+async def record_from_scp(chunk_begin,chunk_size):
import wave
global voices
is_finished = False
@@ -89,9 +105,13 @@
wavs = f_scp.readlines()
else:
wavs = [args.audio_in]
+ if chunk_size>0:
+ wavs=wavs[chunk_begin:chunk_begin+chunk_size]
for wav in wavs:
wav_splits = wav.strip().split()
+ wav_name = wav_splits[0] if len(wav_splits) > 1 else "demo"
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:
@@ -105,23 +125,26 @@
stride = int(60*args.chunk_size[1]/args.chunk_interval/1000*16000*2)
chunk_num = (len(audio_bytes)-1)//stride + 1
# print(stride)
+
+ # send first time
+ message = json.dumps({"chunk_size": args.chunk_size, "chunk_interval": args.chunk_interval, "wav_name": wav_name,"is_speaking": True})
+ voices.put(message)
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_size": args.chunk_size, "chunk_interval": args.chunk_interval, "is_speaking": is_speaking, "audio": data, "is_finished": is_finished})
+ message = data
voices.put(message)
+ if i == chunk_num-1:
+ is_speaking = False
+ message = json.dumps({"is_speaking": is_speaking})
+ voices.put(message)
# print("data_chunk: ", len(data_chunk))
# print(voices.qsize())
sleep_duration = 0.001 if args.send_without_sleep else 60*args.chunk_size[1]/args.chunk_interval/1000
await asyncio.sleep(sleep_duration)
- is_finished = True
- message = json.dumps({"is_finished": is_finished})
- voices.put(message)
async def ws_send():
global voices
@@ -145,21 +168,40 @@
async def message(id):
global websocket
text_print = ""
+ text_print_2pass_online = ""
+ text_print_2pass_offline = ""
while True:
try:
meg = await websocket.recv()
meg = json.loads(meg)
- # print(meg, end = '')
- # print("\r")
- # print(meg)
- text = meg["text"][0]
+ wav_name = meg.get("wav_name", "demo")
+ # print(wav_name)
+ text = meg["text"]
+ if ibest_writer is not None:
+ ibest_writer["text"][wav_name] = text
+
if meg["mode"] == "online":
text_print += " {}".format(text)
- else:
+ text_print = text_print[-args.words_max_print:]
+ os.system('clear')
+ print("\rpid"+str(id)+": "+text_print)
+ elif meg["mode"] == "online":
text_print += "{}".format(text)
- text_print = text_print[-args.words_max_print:]
- os.system('clear')
- print("\r"+str(id)+":"+text_print)
+ text_print = text_print[-args.words_max_print:]
+ os.system('clear')
+ print("\rpid"+str(id)+": "+text_print)
+ else:
+ if meg["mode"] == "2pass-online":
+ text_print_2pass_online += " {}".format(text)
+ text_print = text_print_2pass_offline + text_print_2pass_online
+ else:
+ text_print_2pass_online = ""
+ text_print = text_print_2pass_offline + "{}".format(text)
+ text_print_2pass_offline += "{}".format(text)
+ text_print = text_print[-args.words_max_print:]
+ os.system('clear')
+ print("\rpid" + str(id) + ": " + text_print)
+
except Exception as e:
print("Exception:", e)
traceback.print_exc()
@@ -177,33 +219,61 @@
traceback.print_exc()
exit(0)
-async def ws_client(id):
+async def ws_client(id,chunk_begin,chunk_size):
global websocket
uri = "ws://{}:{}".format(args.host, args.port)
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())
+ task = asyncio.create_task(record_from_scp(chunk_begin,chunk_size))
else:
task = asyncio.create_task(record_microphone())
task2 = asyncio.create_task(ws_send())
task3 = asyncio.create_task(message(id))
await asyncio.gather(task, task2, task3)
-def one_thread(id):
- asyncio.get_event_loop().run_until_complete(ws_client(id)) # 鍚姩鍗忕▼
+def one_thread(id,chunk_begin,chunk_size):
+ asyncio.get_event_loop().run_until_complete(ws_client(id,chunk_begin,chunk_size))
asyncio.get_event_loop().run_forever()
if __name__ == '__main__':
- process_list = []
- for i in range(args.test_thread_num):
- p = Process(target=one_thread,args=(i,))
- p.start()
- process_list.append(p)
+ # for microphone
+ if args.audio_in is None:
+ p = Process(target=one_thread,args=(0, 0, 0))
+ p.start()
+ p.join()
+ print('end')
+ else:
+ # calculate the number of wavs for each preocess
+ if args.audio_in.endswith(".scp"):
+ f_scp = open(args.audio_in)
+ wavs = f_scp.readlines()
+ else:
+ wavs = [args.audio_in]
+ total_len=len(wavs)
+ if total_len>=args.test_thread_num:
+ chunk_size=int((total_len)/args.test_thread_num)
+ remain_wavs=total_len-chunk_size*args.test_thread_num
+ else:
+ chunk_size=1
+ remain_wavs=0
- for i in process_list:
- p.join()
+ process_list = []
+ chunk_begin=0
+ for i in range(args.test_thread_num):
+ now_chunk_size= chunk_size
+ if remain_wavs>0:
+ now_chunk_size=chunk_size+1
+ remain_wavs=remain_wavs-1
+ # process i handle wavs at chunk_begin and size of now_chunk_size
+ p = Process(target=one_thread,args=(i,chunk_begin,now_chunk_size))
+ chunk_begin=chunk_begin+now_chunk_size
+ p.start()
+ process_list.append(p)
- print('end')
-
+ for i in process_list:
+ p.join()
+
+ print('end')
+
--
Gitblit v1.9.1