Spaces:
Edmond98
/
Running on A100

Afrinetwork7 commited on
Commit
bc747e6
1 Parent(s): fc0645c

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +41 -11
app.py CHANGED
@@ -1,5 +1,5 @@
1
  from fastapi import FastAPI, HTTPException
2
- from fastapi.responses import JSONResponse, StreamingResponse
3
  from pydantic import BaseModel
4
  import numpy as np
5
  import io
@@ -13,6 +13,11 @@ import magic # For MIME type detection
13
  from pydub import AudioSegment
14
  import traceback
15
  from logging.handlers import RotatingFileHandler
 
 
 
 
 
16
 
17
  # Import functions from other modules
18
  from asr import transcribe, ASR_LANGUAGES
@@ -33,6 +38,20 @@ logger.addHandler(file_handler)
33
 
34
  app = FastAPI(title="MMS: Scaling Speech Technology to 1000+ languages")
35
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
36
  # Define request models
37
  class AudioRequest(BaseModel):
38
  audio: str # Base64 encoded audio or video data
@@ -138,17 +157,28 @@ async def synthesize_speech(request: TTSRequest):
138
  buffer.seek(0)
139
  logger.info(f"Buffer size: {buffer.getbuffer().nbytes} bytes")
140
 
141
- logger.info("Preparing StreamingResponse")
142
- response = StreamingResponse(
143
- buffer,
144
- media_type="audio/wav",
145
- headers={
146
- "Content-Disposition": "attachment; filename=synthesized_audio.wav"
147
- }
148
- )
149
- logger.info("StreamingResponse prepared successfully")
 
 
 
 
 
 
 
 
 
150
 
151
- return response
 
 
152
 
153
  except ValueError as ve:
154
  logger.error(f"ValueError in synthesize_speech: {str(ve)}", exc_info=True)
 
1
  from fastapi import FastAPI, HTTPException
2
+ from fastapi.responses import JSONResponse
3
  from pydantic import BaseModel
4
  import numpy as np
5
  import io
 
13
  from pydub import AudioSegment
14
  import traceback
15
  from logging.handlers import RotatingFileHandler
16
+ import os
17
+ import boto3
18
+ from botocore.exceptions import NoCredentialsError
19
+ from urllib.parse import quote
20
+ import time
21
 
22
  # Import functions from other modules
23
  from asr import transcribe, ASR_LANGUAGES
 
38
 
39
  app = FastAPI(title="MMS: Scaling Speech Technology to 1000+ languages")
40
 
41
+ # S3 Configuration
42
+ S3_BUCKET = "afri"
43
+ S3_REGION = "eu-west-3"
44
+ S3_ACCESS_KEY_ID = os.environ.get("AWS_ACCESS_KEY_ID")
45
+ S3_SECRET_ACCESS_KEY = os.environ.get("AWS_SECRET_ACCESS_KEY")
46
+
47
+ # Initialize S3 client
48
+ s3_client = boto3.client(
49
+ 's3',
50
+ aws_access_key_id=S3_ACCESS_KEY_ID,
51
+ aws_secret_access_key=S3_SECRET_ACCESS_KEY,
52
+ region_name=S3_REGION
53
+ )
54
+
55
  # Define request models
56
  class AudioRequest(BaseModel):
57
  audio: str # Base64 encoded audio or video data
 
157
  buffer.seek(0)
158
  logger.info(f"Buffer size: {buffer.getbuffer().nbytes} bytes")
159
 
160
+ # Generate a unique filename
161
+ filename = f"synthesized_audio_{int(time.time())}.wav"
162
+
163
+ # Upload to S3
164
+ try:
165
+ s3_client.upload_fileobj(buffer, S3_BUCKET, filename)
166
+ logger.info(f"File uploaded successfully to S3: {filename}")
167
+
168
+ # Generate a presigned URL
169
+ url = s3_client.generate_presigned_url('get_object',
170
+ Params={'Bucket': S3_BUCKET,
171
+ 'Key': filename},
172
+ ExpiresIn=3600) # URL expires in 1 hour
173
+
174
+ encoded_url = quote(url, safe=':/?&=')
175
+ logger.info(f"Presigned URL generated: {encoded_url}")
176
+
177
+ return JSONResponse(content={"audio_url": encoded_url})
178
 
179
+ except NoCredentialsError:
180
+ logger.error("AWS credentials not available")
181
+ raise HTTPException(status_code=500, detail="Could not upload file to S3")
182
 
183
  except ValueError as ve:
184
  logger.error(f"ValueError in synthesize_speech: {str(ve)}", exc_info=True)