Create app.py
Browse files
app.py
ADDED
|
@@ -0,0 +1,83 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import os
|
| 2 |
+
import requests
|
| 3 |
+
import uvicorn
|
| 4 |
+
from fastapi import FastAPI
|
| 5 |
+
from pydub import AudioSegment
|
| 6 |
+
from uuid import uuid4
|
| 7 |
+
|
| 8 |
+
app = FastAPI()
|
| 9 |
+
|
| 10 |
+
# Your first Hugging Face TTS API URL
|
| 11 |
+
TTS_API_URL = "https://hivecorp-s8test.hf.space/tts"
|
| 12 |
+
MAX_LENGTH = 250 # Max characters per chunk
|
| 13 |
+
|
| 14 |
+
|
| 15 |
+
def split_text(text, max_length=MAX_LENGTH):
|
| 16 |
+
"""Splits text into smaller chunks at sentence boundaries."""
|
| 17 |
+
chunks = []
|
| 18 |
+
current_chunk = ""
|
| 19 |
+
sentences = text.split(". ")
|
| 20 |
+
|
| 21 |
+
for sentence in sentences:
|
| 22 |
+
if len(current_chunk + sentence) > max_length:
|
| 23 |
+
chunks.append(current_chunk.strip())
|
| 24 |
+
current_chunk = ""
|
| 25 |
+
current_chunk += sentence + ". "
|
| 26 |
+
|
| 27 |
+
if current_chunk:
|
| 28 |
+
chunks.append(current_chunk.strip())
|
| 29 |
+
|
| 30 |
+
return chunks
|
| 31 |
+
|
| 32 |
+
|
| 33 |
+
def call_tts_api(text, voice):
|
| 34 |
+
"""Calls the TTS API with the text & voice and returns the generated MP3 file."""
|
| 35 |
+
response = requests.post(TTS_API_URL, json={"text": text, "voice": voice})
|
| 36 |
+
if response.status_code == 200:
|
| 37 |
+
filename = f"audio_{uuid4().hex}.mp3"
|
| 38 |
+
with open(filename, "wb") as f:
|
| 39 |
+
f.write(response.content)
|
| 40 |
+
return filename
|
| 41 |
+
return None
|
| 42 |
+
|
| 43 |
+
|
| 44 |
+
def merge_audio(files):
|
| 45 |
+
"""Merges multiple MP3 files into one using pydub."""
|
| 46 |
+
final_audio = AudioSegment.empty()
|
| 47 |
+
|
| 48 |
+
for file in files:
|
| 49 |
+
audio = AudioSegment.from_file(file, format="mp3")
|
| 50 |
+
final_audio += audio
|
| 51 |
+
|
| 52 |
+
final_filename = "final_audio.mp3"
|
| 53 |
+
final_audio.export(final_filename, format="mp3")
|
| 54 |
+
|
| 55 |
+
# Cleanup temp files
|
| 56 |
+
for file in files:
|
| 57 |
+
os.remove(file)
|
| 58 |
+
|
| 59 |
+
return final_filename
|
| 60 |
+
|
| 61 |
+
|
| 62 |
+
@app.post("/generate-audio")
|
| 63 |
+
async def generate_audio(text: str, voice: str):
|
| 64 |
+
"""Processes text, splits it, calls TTS API, and merges MP3s."""
|
| 65 |
+
text_chunks = split_text(text)
|
| 66 |
+
audio_files = [call_tts_api(chunk, voice) for chunk in text_chunks]
|
| 67 |
+
|
| 68 |
+
if None in audio_files:
|
| 69 |
+
return {"error": "TTS API failed for some parts"}
|
| 70 |
+
|
| 71 |
+
merged_file = merge_audio(audio_files)
|
| 72 |
+
return {"audio_url": f"/download/{merged_file}"}
|
| 73 |
+
|
| 74 |
+
|
| 75 |
+
@app.get("/download/{filename}")
|
| 76 |
+
async def download_file(filename: str):
|
| 77 |
+
"""Endpoint to download the merged MP3 file."""
|
| 78 |
+
return {"file": filename}
|
| 79 |
+
|
| 80 |
+
|
| 81 |
+
# Ensure the app starts when running in Hugging Face Spaces
|
| 82 |
+
if __name__ == "__main__":
|
| 83 |
+
uvicorn.run(app, host="0.0.0.0", port=7860)
|