Update app.py
Browse files
app.py
CHANGED
|
@@ -21,7 +21,6 @@ class SentenceTokenizer:
|
|
| 21 |
def tokenize(self, text: str) -> List[str]:
|
| 22 |
if not text or not text.strip():
|
| 23 |
return []
|
| 24 |
-
# Simple sentence splitting
|
| 25 |
sentences = self.SENTENCE_END.split(text.strip())
|
| 26 |
return [s.strip() for s in sentences if s.strip()]
|
| 27 |
|
|
@@ -35,21 +34,26 @@ class ElevenlabsTTS:
|
|
| 35 |
def __init__(self):
|
| 36 |
self.session = requests.Session()
|
| 37 |
self.session.headers.update({"User-Agent": "Mozilla/5.0"})
|
| 38 |
-
# Use temporary directory instead of fixed cache_dir
|
| 39 |
self.cache_dir = Path(tempfile.gettempdir())
|
| 40 |
self.all_voices = {
|
| 41 |
-
"Brian": "nPczCjzI2devNBz1zQrb",
|
| 42 |
-
"
|
| 43 |
-
"
|
| 44 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 45 |
}
|
| 46 |
self.params = {'allow_unauthenticated': '1'}
|
|
|
|
| 47 |
|
| 48 |
def tts(self, text: str, voice: str = "Brian") -> str:
|
| 49 |
if voice not in self.all_voices:
|
| 50 |
raise ValueError(f"Voice '{voice}' not available")
|
| 51 |
|
| 52 |
-
filename = self.cache_dir / f"tts_{int(time.time())}.mp3"
|
| 53 |
sentences = split_sentences(text)
|
| 54 |
|
| 55 |
audio_chunks = {}
|
|
@@ -64,7 +68,6 @@ class ElevenlabsTTS:
|
|
| 64 |
response.raise_for_status()
|
| 65 |
audio_chunks[i] = response.content
|
| 66 |
|
| 67 |
-
# No need to create directory as tempfile.gettempdir() already exists
|
| 68 |
combined_audio = BytesIO()
|
| 69 |
for i in sorted(audio_chunks.keys()):
|
| 70 |
combined_audio.write(audio_chunks[i])
|
|
@@ -73,20 +76,37 @@ class ElevenlabsTTS:
|
|
| 73 |
f.write(combined_audio.getvalue())
|
| 74 |
return filename.as_posix()
|
| 75 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 76 |
# Web Interface
|
| 77 |
tts_provider = ElevenlabsTTS()
|
| 78 |
|
| 79 |
@app.route('/', methods=['GET', 'POST'])
|
| 80 |
def index():
|
| 81 |
if request.method == 'POST':
|
| 82 |
-
|
| 83 |
-
|
| 84 |
-
|
| 85 |
-
|
| 86 |
-
|
| 87 |
-
|
| 88 |
-
|
| 89 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 90 |
|
| 91 |
if __name__ == "__main__":
|
| 92 |
app.run(host='0.0.0.0', port=5000)
|
|
|
|
| 21 |
def tokenize(self, text: str) -> List[str]:
|
| 22 |
if not text or not text.strip():
|
| 23 |
return []
|
|
|
|
| 24 |
sentences = self.SENTENCE_END.split(text.strip())
|
| 25 |
return [s.strip() for s in sentences if s.strip()]
|
| 26 |
|
|
|
|
| 34 |
def __init__(self):
|
| 35 |
self.session = requests.Session()
|
| 36 |
self.session.headers.update({"User-Agent": "Mozilla/5.0"})
|
|
|
|
| 37 |
self.cache_dir = Path(tempfile.gettempdir())
|
| 38 |
self.all_voices = {
|
| 39 |
+
"Brian": "nPczCjzI2devNBz1zQrb", "Alice": "Xb7hH8MSUJpSbSDYk0k2",
|
| 40 |
+
"Bill": "pqHfZKP75CvOlQylNhV4", "Callum": "N2lVS1w4EtoT3dr4eOWO",
|
| 41 |
+
"Charlie": "IKne3meq5aSn9XLyUdCD", "Charlotte": "XB0fDUnXU5powFXDhCwa",
|
| 42 |
+
"Chris": "iP95p4xoKVk53GoZ742B", "Daniel": "onwK4e9ZLuTAKqWW03F9",
|
| 43 |
+
"Eric": "cjVigY5qzO86Huf0OWal", "George": "JBFqnCBsd6RMkjVDRZzb",
|
| 44 |
+
"Jessica": "cgSgspJ2msm6clMCkdW9", "Laura": "FGY2WhTYpPnrIDTdsKH5",
|
| 45 |
+
"Liam": "TX3LPaxmHKxFdv7VOQHJ", "Lily": "pFZP5JQG7iQjIQuC4Bku",
|
| 46 |
+
"Matilda": "XrExE9yKIg1WjnnlVkGX", "Sarah": "EXAVITQu4vr4xnSDxMaL",
|
| 47 |
+
"Will": "bIHbv24MWmeRgasZH58o", "Neal": "Zp1aWhL05Pi5BkhizFC3"
|
| 48 |
}
|
| 49 |
self.params = {'allow_unauthenticated': '1'}
|
| 50 |
+
self.preview_text = "Hello, this is a sample of my voice."
|
| 51 |
|
| 52 |
def tts(self, text: str, voice: str = "Brian") -> str:
|
| 53 |
if voice not in self.all_voices:
|
| 54 |
raise ValueError(f"Voice '{voice}' not available")
|
| 55 |
|
| 56 |
+
filename = self.cache_dir / f"tts_{voice}_{int(time.time())}.mp3"
|
| 57 |
sentences = split_sentences(text)
|
| 58 |
|
| 59 |
audio_chunks = {}
|
|
|
|
| 68 |
response.raise_for_status()
|
| 69 |
audio_chunks[i] = response.content
|
| 70 |
|
|
|
|
| 71 |
combined_audio = BytesIO()
|
| 72 |
for i in sorted(audio_chunks.keys()):
|
| 73 |
combined_audio.write(audio_chunks[i])
|
|
|
|
| 76 |
f.write(combined_audio.getvalue())
|
| 77 |
return filename.as_posix()
|
| 78 |
|
| 79 |
+
def generate_preview(self, voice: str) -> str:
|
| 80 |
+
preview_file = self.cache_dir / f"preview_{voice}.mp3"
|
| 81 |
+
if not preview_file.exists():
|
| 82 |
+
return self.tts(self.preview_text, voice)
|
| 83 |
+
return preview_file.as_posix()
|
| 84 |
+
|
| 85 |
# Web Interface
|
| 86 |
tts_provider = ElevenlabsTTS()
|
| 87 |
|
| 88 |
@app.route('/', methods=['GET', 'POST'])
|
| 89 |
def index():
|
| 90 |
if request.method == 'POST':
|
| 91 |
+
if 'generate' in request.form:
|
| 92 |
+
text = request.form.get('text')
|
| 93 |
+
voice = request.form.get('voice', 'Brian')
|
| 94 |
+
try:
|
| 95 |
+
audio_file = tts_provider.tts(text, voice)
|
| 96 |
+
return send_file(audio_file, mimetype='audio/mpeg', as_attachment=True, download_name=f"{voice}_output.mp3")
|
| 97 |
+
except Exception as e:
|
| 98 |
+
return render_template('index.html', error=str(e), voices=tts_provider.all_voices.keys())
|
| 99 |
+
# Generate previews on startup or first visit
|
| 100 |
+
previews = {voice: tts_provider.generate_preview(voice) for voice in tts_provider.all_voices.keys()}
|
| 101 |
+
return render_template('index.html', voices=tts_provider.all_voices.keys(), previews=previews)
|
| 102 |
+
|
| 103 |
+
@app.route('/preview/<voice>')
|
| 104 |
+
def preview(voice):
|
| 105 |
+
try:
|
| 106 |
+
audio_file = tts_provider.generate_preview(voice)
|
| 107 |
+
return send_file(audio_file, mimetype='audio/mpeg')
|
| 108 |
+
except Exception as e:
|
| 109 |
+
return str(e), 500
|
| 110 |
|
| 111 |
if __name__ == "__main__":
|
| 112 |
app.run(host='0.0.0.0', port=5000)
|