Spaces:
Running
Running
Create app.py
Browse files
app.py
ADDED
|
@@ -0,0 +1,64 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import gradio as gr
|
| 2 |
+
import requests
|
| 3 |
+
import tempfile
|
| 4 |
+
import pyttsx3
|
| 5 |
+
from faster_whisper import WhisperModel
|
| 6 |
+
|
| 7 |
+
# === Config ===
|
| 8 |
+
GROQ_API_KEY = "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"
|
| 9 |
+
GROQ_MODEL = "llama3-70b-8192"
|
| 10 |
+
GROQ_API_URL = "https://api.groq.com/openai/v1/chat/completions"
|
| 11 |
+
|
| 12 |
+
# === Init Whisper ===
|
| 13 |
+
whisper = WhisperModel("base", device="cpu", compute_type="int8")
|
| 14 |
+
|
| 15 |
+
# === Init TTS (offline) ===
|
| 16 |
+
tts_engine = pyttsx3.init()
|
| 17 |
+
|
| 18 |
+
def speak_to_file(text, output_file="output.wav"):
|
| 19 |
+
"""Save TTS to a file using pyttsx3."""
|
| 20 |
+
tts_engine.save_to_file(text, output_file)
|
| 21 |
+
tts_engine.runAndWait()
|
| 22 |
+
return output_file
|
| 23 |
+
|
| 24 |
+
def process_audio(audio_file):
|
| 25 |
+
# 1. Speech to Text
|
| 26 |
+
segments, _ = whisper.transcribe(audio_file)
|
| 27 |
+
user_text = " ".join([segment.text for segment in segments])
|
| 28 |
+
|
| 29 |
+
# 2. Groq API Call
|
| 30 |
+
headers = {
|
| 31 |
+
"Authorization": f"Bearer {GROQ_API_KEY}",
|
| 32 |
+
"Content-Type": "application/json"
|
| 33 |
+
}
|
| 34 |
+
payload = {
|
| 35 |
+
"model": GROQ_MODEL,
|
| 36 |
+
"messages": [{"role": "user", "content": user_text}],
|
| 37 |
+
"temperature": 0.5
|
| 38 |
+
}
|
| 39 |
+
|
| 40 |
+
response = requests.post(GROQ_API_URL, headers=headers, json=payload)
|
| 41 |
+
if response.status_code != 200:
|
| 42 |
+
return f"Groq API Error: {response.text}", None
|
| 43 |
+
|
| 44 |
+
reply = response.json()["choices"][0]["message"]["content"]
|
| 45 |
+
|
| 46 |
+
# 3. TTS to file
|
| 47 |
+
audio_output = tempfile.NamedTemporaryFile(suffix=".wav", delete=False)
|
| 48 |
+
speak_to_file(reply, audio_output.name)
|
| 49 |
+
|
| 50 |
+
return reply, audio_output.name
|
| 51 |
+
|
| 52 |
+
iface = gr.Interface(
|
| 53 |
+
fn=process_audio,
|
| 54 |
+
inputs=gr.Audio(source="microphone", type="filepath", label="🎤 Speak your question"),
|
| 55 |
+
outputs=[
|
| 56 |
+
gr.Textbox(label="🧠 Groq Response"),
|
| 57 |
+
gr.Audio(label="🔊 AI Voice Reply")
|
| 58 |
+
],
|
| 59 |
+
title="🗣️ Voice AI Assistant (Groq + Whisper + Offline TTS)",
|
| 60 |
+
description="Use your mic. Whisper transcribes, Groq replies, voice response from pyttsx3 (offline).",
|
| 61 |
+
live=True
|
| 62 |
+
)
|
| 63 |
+
|
| 64 |
+
iface.launch()
|