Commit @9faf3da8d3032b8afd0cc3dbd409703b96c72bdb - yjyoon/whisper_streaming

Quentin Fuxa 01-03

no online conflict when multiple users

@9faf3da8d3032b8afd0cc3dbd409703b96c72bdb

4f252ff

9faf3da

whisper_fastapi_online_server.py

--- whisper_fastapi_online_server.py

+++ whisper_fastapi_online_server.py


 from fastapi.responses import HTMLResponse
 from fastapi.middleware.cors import CORSMiddleware
 
-from whisper_online import asr_factory, add_shared_args
+from whisper_online import backend_factory, online_factory, add_shared_args
 
 app = FastAPI()
 app.add_middleware(

 add_shared_args(parser)
 args = parser.parse_args()
 
-asr, online = asr_factory(args)
+asr, tokenizer = backend_factory(args)
 
 # Load demo HTML for the root endpoint
 with open("src/live_transcription.html", "r") as f:

 
     ffmpeg_process = await start_ffmpeg_decoder()
     pcm_buffer = bytearray()
+    print("Loading online.")
+    online = online_factory(args, asr, tokenizer)
+    print("Online loaded.")
 
     # Continuously read decoded PCM from ffmpeg stdout in a background task
     async def ffmpeg_stdout_reader():

4f252ff

9faf3da

whisper_online.py

--- whisper_online.py

+++ whisper_online.py


         default="DEBUG",
     )
 
-
-def asr_factory(args, logfile=sys.stderr):
-    """
-    Creates and configures an ASR and ASR Online instance based on the specified backend and arguments.
-    """
+def backend_factory(args):
     backend = args.backend
     if backend == "openai-api":
         logger.debug("Using OpenAI API.")

         tokenizer = create_tokenizer(tgt_language)
     else:
         tokenizer = None
+    return asr, tokenizer
 
-    # Create the OnlineASRProcessor
+def online_factory(args, asr, tokenizer, logfile=sys.stderr):
     if args.vac:
-
         online = VACOnlineASRProcessor(
             args.min_chunk_size,
             asr,

             logfile=logfile,
             buffer_trimming=(args.buffer_trimming, args.buffer_trimming_sec),
         )
-
+    return online
+  
+def asr_factory(args, logfile=sys.stderr):
+    """
+    Creates and configures an ASR and ASR Online instance based on the specified backend and arguments.
+    """
+    asr, tokenizer = backend_factory(args)
+    online = online_factory(args, asr, tokenizer, logfile=logfile)
     return asr, online
-
 
 def set_logging(args, logger, other="_server"):
     logging.basicConfig(format="%(levelname)s\t%(message)s")  # format='%(name)s

Add a comment

Open 0
Closed 0

List

...	...	@@ -9,7 +9,7 @@
9	9	from fastapi.responses import HTMLResponse
10	10	from fastapi.middleware.cors import CORSMiddleware
11	11
12		-from whisper_online import asr_factory, add_shared_args
	12	+from whisper_online import backend_factory, online_factory, add_shared_args
13	13
14	14	app = FastAPI()
15	15	app.add_middleware(
...	...	@@ -40,7 +40,7 @@
40	40	add_shared_args(parser)
41	41	args = parser.parse_args()
42	42
43		-asr, online = asr_factory(args)
	43	+asr, tokenizer = backend_factory(args)
44	44
45	45	# Load demo HTML for the root endpoint
46	46	with open("src/live_transcription.html", "r") as f:
...	...	@@ -85,6 +85,9 @@
85	85
86	86	ffmpeg_process = await start_ffmpeg_decoder()
87	87	pcm_buffer = bytearray()
	88	+ print("Loading online.")
	89	+ online = online_factory(args, asr, tokenizer)
	90	+ print("Online loaded.")
88	91
89	92	# Continuously read decoded PCM from ffmpeg stdout in a background task
90	93	async def ffmpeg_stdout_reader():

...	...	@@ -920,11 +920,7 @@
920	920	default="DEBUG",
921	921	)
922	922
923		-
924		-def asr_factory(args, logfile=sys.stderr):
925		- """
926		- Creates and configures an ASR and ASR Online instance based on the specified backend and arguments.
927		- """
	923	+def backend_factory(args):
928	924	backend = args.backend
929	925	if backend == "openai-api":
930	926	logger.debug("Using OpenAI API.")
...	...	@@ -967,10 +963,10 @@
967	963	tokenizer = create_tokenizer(tgt_language)
968	964	else:
969	965	tokenizer = None
	966	+ return asr, tokenizer
970	967
971		- # Create the OnlineASRProcessor
	968	+def online_factory(args, asr, tokenizer, logfile=sys.stderr):
972	969	if args.vac:
973		-
974	970	online = VACOnlineASRProcessor(
975	971	args.min_chunk_size,
976	972	asr,
...	...	@@ -985,9 +981,15 @@
985	981	logfile=logfile,
986	982	buffer_trimming=(args.buffer_trimming, args.buffer_trimming_sec),
987	983	)
988		-
	984	+ return online
	985	+
	986	+def asr_factory(args, logfile=sys.stderr):
	987	+ """
	988	+ Creates and configures an ASR and ASR Online instance based on the specified backend and arguments.
	989	+ """
	990	+ asr, tokenizer = backend_factory(args)
	991	+ online = online_factory(args, asr, tokenizer, logfile=logfile)
989	992	return asr, online
990		-
991	993
992	994	def set_logging(args, logger, other="_server"):
993	995	logging.basicConfig(format="%(levelname)s\t%(message)s") # format='%(name)s

Delete comment