Create asr.py
Browse files
asr.py
ADDED
|
@@ -0,0 +1,33 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import json
|
| 2 |
+
import subprocess
|
| 3 |
+
import os
|
| 4 |
+
from pathlib import Path
|
| 5 |
+
|
| 6 |
+
class ASR:
|
| 7 |
+
def __init__(self, hive_home):
|
| 8 |
+
self.hive_home = Path(hive_home)
|
| 9 |
+
self.model_path = self.hive_home / "models" / "vosk"
|
| 10 |
+
self.ensure_model()
|
| 11 |
+
|
| 12 |
+
def ensure_model(self):
|
| 13 |
+
if not self.model_path.exists():
|
| 14 |
+
print("Downloading Vosk model...")
|
| 15 |
+
os.makedirs(self.model_path.parent, exist_ok=True)
|
| 16 |
+
subprocess.run([
|
| 17 |
+
"wget", "-q", "-O", "-",
|
| 18 |
+
"https://alphacephei.com/vosk/models/vosk-model-small-en-us-0.15.zip"
|
| 19 |
+
], check=True, stdout=subprocess.PIPE)
|
| 20 |
+
subprocess.run([
|
| 21 |
+
"unzip", "-q", "-", "-d", str(self.model_path.parent)
|
| 22 |
+
], input=subprocess.Popen([
|
| 23 |
+
"wget", "-q", "-O", "-",
|
| 24 |
+
"https://alphacephei.com/vosk/models/vosk-model-small-en-us-0.15.zip"
|
| 25 |
+
], stdout=subprocess.PIPE).stdout)
|
| 26 |
+
|
| 27 |
+
def transcribe(self, audio_path):
|
| 28 |
+
try:
|
| 29 |
+
# Simple mock transcription for now
|
| 30 |
+
# In production, integrate Vosk properly
|
| 31 |
+
return "Transcribed text from audio"
|
| 32 |
+
except:
|
| 33 |
+
return "Could not transcribe audio"
|