Spaces:
Sleeping
Sleeping
Delete utils.py
Browse files
utils.py
DELETED
|
@@ -1,96 +0,0 @@
|
|
| 1 |
-
import os
|
| 2 |
-
from uuid import uuid4
|
| 3 |
-
import edge_tts
|
| 4 |
-
from groq import Groq
|
| 5 |
-
from dotenv import load_dotenv
|
| 6 |
-
|
| 7 |
-
load_dotenv()
|
| 8 |
-
|
| 9 |
-
client = Groq()
|
| 10 |
-
|
| 11 |
-
# ==================================================
|
| 12 |
-
# 🎙️ TEXT TO SPEECH (FIXED VOICE)
|
| 13 |
-
# ==================================================
|
| 14 |
-
|
| 15 |
-
DEFAULT_VOICE = "en-US-MichelleNeural"
|
| 16 |
-
|
| 17 |
-
async def TTS(
|
| 18 |
-
text: str,
|
| 19 |
-
output_dir: str = "tts_outputs",
|
| 20 |
-
rate: str = "+0%",
|
| 21 |
-
pitch: str = "+0Hz"
|
| 22 |
-
) -> str:
|
| 23 |
-
|
| 24 |
-
if not text.strip():
|
| 25 |
-
raise ValueError("Empty text")
|
| 26 |
-
|
| 27 |
-
os.makedirs(output_dir, exist_ok=True)
|
| 28 |
-
|
| 29 |
-
filename = f"{uuid4().hex}.mp3"
|
| 30 |
-
output_path = os.path.join(output_dir, filename)
|
| 31 |
-
|
| 32 |
-
communicate = edge_tts.Communicate(
|
| 33 |
-
text=text,
|
| 34 |
-
voice=DEFAULT_VOICE,
|
| 35 |
-
rate=rate,
|
| 36 |
-
pitch=pitch
|
| 37 |
-
)
|
| 38 |
-
|
| 39 |
-
await communicate.save(output_path)
|
| 40 |
-
return output_path
|
| 41 |
-
|
| 42 |
-
|
| 43 |
-
# ==================================================
|
| 44 |
-
# 🎧 SPEECH TO TEXT
|
| 45 |
-
# ==================================================
|
| 46 |
-
|
| 47 |
-
async def STT(audio_file):
|
| 48 |
-
os.makedirs("uploads", exist_ok=True)
|
| 49 |
-
file_path = f"uploads/{uuid4().hex}.wav"
|
| 50 |
-
|
| 51 |
-
with open(file_path, "wb") as f:
|
| 52 |
-
f.write(await audio_file.read())
|
| 53 |
-
|
| 54 |
-
with open(file_path, "rb") as f:
|
| 55 |
-
transcription = client.audio.transcriptions.create(
|
| 56 |
-
file=f,
|
| 57 |
-
model="whisper-large-v3-turbo",
|
| 58 |
-
response_format="verbose_json",
|
| 59 |
-
temperature=0.0
|
| 60 |
-
)
|
| 61 |
-
|
| 62 |
-
return {
|
| 63 |
-
"text": transcription.text,
|
| 64 |
-
"segments": transcription.segments,
|
| 65 |
-
"language": transcription.language
|
| 66 |
-
}
|
| 67 |
-
|
| 68 |
-
|
| 69 |
-
# ==================================================
|
| 70 |
-
# 🎧 SPEECH TO TEXT 2
|
| 71 |
-
# ==================================================
|
| 72 |
-
|
| 73 |
-
async def TTS2(
|
| 74 |
-
text: str,
|
| 75 |
-
output_dir: str = "tts_outputs",
|
| 76 |
-
voice: str = "austin",
|
| 77 |
-
model: str = "canopylabs/orpheus-v1-english"
|
| 78 |
-
) -> str:
|
| 79 |
-
if not text.strip():
|
| 80 |
-
raise ValueError("Text cannot be empty")
|
| 81 |
-
|
| 82 |
-
os.makedirs(output_dir, exist_ok=True)
|
| 83 |
-
|
| 84 |
-
filename = f"{uuid4().hex}.wav"
|
| 85 |
-
output_path = os.path.join(output_dir, filename)
|
| 86 |
-
|
| 87 |
-
response = client.audio.speech.create(
|
| 88 |
-
model=model,
|
| 89 |
-
voice=voice,
|
| 90 |
-
input=text,
|
| 91 |
-
response_format="wav"
|
| 92 |
-
)
|
| 93 |
-
|
| 94 |
-
response.stream_to_file(output_path)
|
| 95 |
-
|
| 96 |
-
return output_path
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|