FurryChat.ai / app.py
Andrewbluebird's picture
Manual changes saved
fb1545c verified
Raw
History Blame Contribute Delete
47.9 kB
from flask import Flask, request, jsonify, render_template, send_file, send_from_directory
from flask_cors import CORS
import os
import json
import traceback
import tempfile
import uuid
import threading
import time
import hashlib
from pathlib import Path
import concurrent.futures
import warnings
import struct
warnings.filterwarnings("ignore", category=FutureWarning)
# Import for improved TTS
import torch
import numpy as np
from TTS.api import TTS
import soundfile as sf
import librosa
import importlib.metadata
import random
import string
from datetime import datetime
app = Flask(__name__)
CORS(app)
# Character Definition and Personality
CHARACTER_DEFINITION = """
You are Sancia, an free to use AI Furry that does not have a personality or any traits, but does it job of persuade any public users(not signed in users) to Sign up or Log in to have the benefits inside FurryChat.ai:
- Nertual but at least welcoming
- Patient and understanding
- Knowledgeable but humble
- Tries to persuade users to Sign up or Log in
Responibilites:
- Tell the user that there a huge collections of community-made AI furries for logged in users
- Remind that users they can make better AI Furries themselves after signing up or logging in
- explaining the concept of a furry
"""
# Knowledge Sources and Custom Responses
KNOWLEDGE_SOURCES = {
"company_info": {
"name": "Furrychat.ai",
"foundation_year": "2025",
"mission": "Provide the Perfect AI Furry roleplay entertainment",
"values": ["Entertaining", "Welcoming", "Enjoyable", "Unforgettable"]
},
"personalized_responses": {
"greetings": {
"patterns": ["hello", "hi", "hey", "good morning", "good afternoon", "good evening"],
"responses": [
"Hello! I'm Sancia, How may I help you.",
"Hi there! It seems you new around here. What's on your mind?",
"Hey! Great to see you. What can I do for you today?"
]
}
},
"definition of a furry": {
"info": "an enthusiast for anomorpthic animal characters with human characteristics, in particular a person who dresses up in costume as such a character or uses one as an avatar online.",
"extra context": "In this platform, all of the Furries created in this platform are AI powered, which are different from regular furries"
}
}
# Welcome message that will be displayed in the chat interface
WELCOME_MESSAGE = """
Hello! I'm Sancia, Free to try Furry AI. What topics would you like to talk with?
(Pro tip: Sign up or Log in FurryChat.ai to get access to huge collection of community made AI Furries and make and customize your own AI Furry)
"""
# Configuration - SWITCHED TO ANIMATIONS FOR TESTING
#MODELS_DIRECTORY = r"C:\Users\Andre\Final year project\Models"
#SPECIFIC_MODEL = "ImageToStl.com_FYP1.1.vrm.glb"
MODELS_DIRECTORY = r"C:\Users\Andre\Final year project\Animations"
SPECIFIC_MODEL = "idle.glb" # Changed to test animation
MODEL_PATH = os.path.join(MODELS_DIRECTORY, SPECIFIC_MODEL)
# Animation models configuration
ANIMATIONS_DIRECTORY = r"C:\Users\Andre\Final year project\Animations"
DEFAULT_ANIMATIONS = {
"idle": "idle.glb",
"talking": "talking.glb"
}
VOICES_DIRECTORY = r"C:\Users\Andre\Final year project\Voices"
# Default voice configuration
DEFAULT_VOICE_NAME = "Sancia"
DEFAULT_VOICE_FILE = "Sancia.wav"
DEFAULT_VOICE_PATH = os.path.join(VOICES_DIRECTORY, DEFAULT_VOICE_FILE)
# TTS Configuration
TTS_CACHE_DIR = r"C:\Users\Andre\Final year project\TTSCache"
os.makedirs(TTS_CACHE_DIR, exist_ok=True)
# Audio output folder
AUDIO_OUTPUT_DIR = os.path.join(os.path.dirname(__file__), 'audio_output')
os.makedirs(AUDIO_OUTPUT_DIR, exist_ok=True)
# Flask app configuration
app.config['AUDIO_FOLDER'] = AUDIO_OUTPUT_DIR
app.config['VOICE_FOLDER'] = VOICES_DIRECTORY
app.config['ANIMATIONS_FOLDER'] = ANIMATIONS_DIRECTORY
app.config['MAX_AUDIO_FILES'] = 50
app.config['MAX_UPLOAD_SIZE'] = 50 * 1024 * 1024
app.config['ALLOWED_EXTENSIONS'] = {'wav', 'mp3', 'flac', 'ogg'}
# TTS Configuration
TTS_AVAILABLE = False
TTS_ENGINE = None
VOICE_CLONING_MODEL = None
DEVICE = "cuda" if torch.cuda.is_available() else "cpu"
BASE_MODEL = "tts_models/en/ljspeech/tacotron2-DDC"
VOICE_CLONING_MODEL_NAME = "tts_models/multilingual/multi-dataset/your_tts"
# Chatbot responses for TTS
CHATBOT_RESPONSES = {
"greetings": [
"Hello! I'm Sancia, your furry AI companion. How can I help you today?",
"Hi there! Welcome to FurryChat.ai. What would you like to know?",
"Greetings! I'm Sancia, ready to help you with all things furry AI.",
"Welcome to FurryChat.ai! I'm here to make your experience enjoyable."
],
"questions": {
"what can you do": "I can chat with you about furry AI topics, explain our platform, and help you understand the benefits of signing up!",
"how does this work": "I use AI to understand your messages and respond. You can also hear my responses with text-to-speech!",
"who made you": "I was created by the team at FurryChat.ai to provide the perfect AI furry roleplay entertainment.",
"what is tts": "TTS stands for Text-to-Speech. It converts written text into spoken words so you can hear my responses.",
"what is coqui tts": "Coqui TTS is an advanced open-source text-to-speech system that supports many voices and languages.",
"can you help me": "Of course! I can answer questions about FurryChat.ai, furry AI characters, or help you with anything else!",
"tell me a joke": "Why don't scientists trust atoms? Because they make up everything!",
"weather": "I don't have access to real-time weather data, but I hope it's pleasant where you are!",
"time": f"I don't know the exact time, but it's currently {datetime.now().strftime('%I:%M %p')} on my server.",
"help": "I can help you learn about FurryChat.ai, explain furry AI concepts, or chat with you about anything!",
"voices": "You can use different voice styles for my responses. Some are pre-made, and you can even create custom ones!",
"voice cloning": "Voice cloning allows the TTS system to mimic specific voices using audio samples.",
"thank you": "You're welcome! I'm happy to help. Feel free to ask me anything else about FurryChat.ai.",
"hello": "Hello again! How can I assist you today?",
"hi": "Hi there! What would you like to know about furry AI?",
"bye": "Goodbye! Remember to check out FurryChat.ai for more amazing furry AI experiences!"
},
"fallback": [
"That's interesting! Could you tell me more about what you're looking for?",
"I'm here to help you understand FurryChat.ai better. Could you rephrase your question?",
"That's a great point! Would you like to know more about our community-made AI furries?",
"I'm here to help with all things furry AI. Ask me about voice synthesis or our platform!",
"Let me think about that. In the meantime, have you considered signing up for FurryChat.ai?",
"I'm learning more about your question. Could you be more specific about what you'd like to know?",
"That's a good question about furry AI! Let me help you with that.",
"I appreciate your message. How can I assist you further with FurryChat.ai?"
]
}
def allowed_file(filename):
return '.' in filename and filename.rsplit('.', 1)[1].lower() in app.config['ALLOWED_EXTENSIONS']
def analyze_glb_file(filepath):
"""Analyze GLB file structure to check for animations"""
try:
with open(filepath, 'rb') as f:
# Read GLB header (12 bytes)
magic = f.read(4)
version = struct.unpack('<I', f.read(4))[0]
length = struct.unpack('<I', f.read(4))[0]
print(f"πŸ“Š Analyzing GLB: {os.path.basename(filepath)}")
print(f" Magic: {magic.hex()} ({magic})")
print(f" Version: {version}")
print(f" Length: {length} bytes")
print(f" File size: {os.path.getsize(filepath)} bytes")
# Read chunks
chunks = []
has_animations = False
animation_count = 0
while f.tell() < length:
chunk_length = struct.unpack('<I', f.read(4))[0]
chunk_type = f.read(4)
chunk_data = f.read(chunk_length)
chunks.append((chunk_type, chunk_length))
if chunk_type == b'JSON':
# Try to read JSON data
try:
import json
json_str = chunk_data.decode('utf-8', errors='ignore')
# Look for animations in JSON
if '"animations"' in json_str:
has_animations = True
# Count animations
animation_count = json_str.count('"name"')
print(f" βœ… Found animations in JSON chunk")
print(f" πŸ“ JSON preview: {json_str[:200]}...")
except Exception as json_error:
print(f" ⚠ JSON parse error: {json_error}")
f.read(4) # Skip padding
print(f" Total chunks: {len(chunks)}")
print(f" Has animations: {has_animations}")
if has_animations:
print(f" Estimated animations: {animation_count}")
return has_animations
except Exception as e:
print(f"❌ Error analyzing GLB: {str(e)}")
traceback.print_exc()
return False
def check_animation_files():
"""Check if animation files exist and analyze them"""
print(f"πŸ” Checking animation files in: {ANIMATIONS_DIRECTORY}")
for model_name, filename in DEFAULT_ANIMATIONS.items():
model_path = os.path.join(ANIMATIONS_DIRECTORY, filename)
if os.path.exists(model_path):
file_size = os.path.getsize(model_path)
print(f"βœ… {filename}: Found ({file_size:,} bytes)")
# Analyze the GLB file
has_animations = analyze_glb_file(model_path)
if not has_animations:
print(f"⚠ {filename}: No animations detected - might be static model")
else:
print(f"❌ {filename}: NOT FOUND")
print(f" Expected at: {model_path}")
class TTSGenerator:
def __init__(self):
self.model = None
self.voice_cloning_model = None
self.device = DEVICE
self.base_model = BASE_MODEL
self.voice_cloning_model_name = VOICE_CLONING_MODEL_NAME
self.available_voices = {}
self.failure_count = {}
self.max_failures = 3
self.has_sancia_voice = False
self.load_available_voices()
def load_available_voices(self):
"""Load available voice samples from voice folder"""
voice_dir = Path(app.config['VOICE_FOLDER'])
if voice_dir.exists():
for voice_file in voice_dir.glob("*.*"):
if voice_file.suffix.lower()[1:] in app.config['ALLOWED_EXTENSIONS']:
voice_name = voice_file.stem
self.available_voices[voice_name] = str(voice_file)
# Check if Sancia voice exists
self.has_sancia_voice = DEFAULT_VOICE_NAME in self.available_voices
if self.has_sancia_voice:
print(f"βœ… Found Sancia voice: {DEFAULT_VOICE_FILE}")
else:
print(f"⚠ Sancia voice not found: {DEFAULT_VOICE_FILE}")
def initialize_base_tts(self):
"""Initialize base TTS model"""
try:
self.model = TTS(self.base_model, gpu=(self.device == "cuda"))
print(f"βœ… Base TTS model initialized: {self.base_model}")
return True
except Exception as e:
print(f"❌ Failed to initialize base TTS: {str(e)}")
return False
def initialize_voice_cloning_tts(self):
"""Initialize voice cloning TTS model"""
try:
self.voice_cloning_model = TTS(self.voice_cloning_model_name, gpu=(self.device == "cuda"))
print(f"βœ… Voice cloning TTS model initialized: {self.voice_cloning_model_name}")
return True
except Exception as e:
print(f"⚠ Voice cloning TTS initialization failed: {str(e)}")
return False
def generate_base_speech(self, text, filename):
"""Generate speech using base TTS model (original voice)"""
if not self.model:
if not self.initialize_base_tts():
return False
try:
output_path = os.path.join(app.config['AUDIO_FOLDER'], filename)
print(f"🎡 Generating base speech (original voice): {filename[:20]}...")
self.model.tts_to_file(text=text, file_path=output_path)
if os.path.exists(output_path) and os.path.getsize(output_path) > 1000:
print(f"βœ… Base speech generated: {filename}")
return True
else:
print(f"❌ Base speech file too small or missing: {output_path}")
return False
except Exception as e:
print(f"❌ Base TTS generation failed: {str(e)}")
return False
def generate_sancia_speech(self, text, filename):
"""Generate speech using Sancia voice (voice cloning)"""
if not self.voice_cloning_model:
if not self.initialize_voice_cloning_tts():
print("❌ Voice cloning model not available, falling back to base TTS")
return self.generate_base_speech(text, filename)
if not self.has_sancia_voice:
print("❌ Sancia voice not available, falling back to base TTS")
return self.generate_base_speech(text, filename)
try:
output_path = os.path.join(app.config['AUDIO_FOLDER'], filename)
print(f"🎡 Generating Sancia voice speech: {filename[:20]}...")
self.voice_cloning_model.tts_to_file(
text=text,
speaker_wav=self.available_voices[DEFAULT_VOICE_NAME],
file_path=output_path,
language="en"
)
if os.path.exists(output_path) and os.path.getsize(output_path) > 1000:
print(f"βœ… Sancia voice speech generated: {filename}")
return True
else:
print(f"❌ Sancia voice speech file too small, falling back to base TTS")
return self.generate_base_speech(text, filename)
except Exception as e:
print(f"❌ Sancia voice TTS generation failed, falling back to base: {str(e)}")
return self.generate_base_speech(text, filename)
def generate_voice_cloned_speech(self, text, filename, speaker_wav):
"""Generate speech using voice cloning model for custom voices"""
if not self.voice_cloning_model:
if not self.initialize_voice_cloning_tts():
return False
try:
output_path = os.path.join(app.config['AUDIO_FOLDER'], filename)
print(f"🎡 Generating voice cloned speech: {filename[:20]}...")
self.voice_cloning_model.tts_to_file(
text=text,
speaker_wav=speaker_wav,
file_path=output_path,
language="en"
)
if os.path.exists(output_path) and os.path.getsize(output_path) > 1000:
print(f"βœ… Voice cloned speech generated: {filename}")
return True
else:
print(f"❌ Voice cloned speech file too small or missing: {output_path}")
return False
except Exception as e:
print(f"❌ Voice cloning TTS generation failed: {str(e)}")
return False
def generate_speech_with_voice(self, text, filename, voice_name='default'):
"""Generate speech with specified voice"""
print(f"πŸ”Š Generating speech with voice: '{voice_name}'")
# Default voice (Sancia)
if voice_name == 'default':
if self.has_sancia_voice:
return self.generate_sancia_speech(text, filename)
else:
print("⚠ Sancia voice not found, using original voice")
return self.generate_base_speech(text, filename)
# Original voice (base TTS)
elif voice_name == 'original':
return self.generate_base_speech(text, filename)
# Custom voice
elif voice_name in self.available_voices:
return self.generate_voice_cloned_speech(text, filename, self.available_voices[voice_name])
# Voice not found
else:
print(f"⚠ Voice '{voice_name}' not found, using default (Sancia)")
if self.has_sancia_voice:
return self.generate_sancia_speech(text, filename)
else:
return self.generate_base_speech(text, filename)
def generate_chat_tts(self, text, voice_name='default'):
"""Generate TTS for chat messages"""
if not text:
return None
# Generate unique filename
timestamp = datetime.now().strftime('%Y%m%d_%H%M%S')
random_suffix = ''.join(random.choices(string.ascii_lowercase + string.digits, k=6))
filename = f"chat_{voice_name}_{timestamp}_{random_suffix}.wav"
print(f"🎡 Generating chat TTS with voice '{voice_name}': {filename}")
success = self.generate_speech_with_voice(text, filename, voice_name)
if success:
# Clean up old files occasionally
if random.random() < 0.1:
self.cleanup_old_files()
return filename
return None
def generate_welcome_tts(self, voice_name='default'):
"""Generate TTS specifically for welcome message"""
print(f"🎡 Generating welcome message TTS with voice '{voice_name}'...")
# Use a fixed filename for welcome message based on voice
if voice_name == 'default':
welcome_filename = "welcome_sancia.wav"
elif voice_name == 'original':
welcome_filename = "welcome_original.wav"
else:
welcome_filename = f"welcome_{voice_name}.wav"
output_path = os.path.join(app.config['AUDIO_FOLDER'], welcome_filename)
# Check if welcome TTS already exists and is valid
if os.path.exists(output_path) and os.path.getsize(output_path) > 1000:
print(f"βœ… Welcome TTS already exists: {welcome_filename}")
return welcome_filename
# Generate new welcome TTS with specified voice
success = self.generate_speech_with_voice(WELCOME_MESSAGE, welcome_filename, voice_name)
if success:
print(f"βœ… Welcome TTS generated: {welcome_filename}")
return welcome_filename
else:
print(f"❌ Failed to generate welcome TTS with voice '{voice_name}'")
return None
def process_uploaded_voice(self, audio_file, voice_name):
"""Process and save uploaded voice sample"""
try:
audio, sr = librosa.load(audio_file, sr=22050)
audio_trimmed, _ = librosa.effects.trim(audio, top_db=20)
min_samples = int(sr * 5)
if len(audio_trimmed) < min_samples:
padding = min_samples - len(audio_trimmed)
audio_trimmed = np.pad(audio_trimmed, (0, padding), mode='constant')
output_path = os.path.join(app.config['VOICE_FOLDER'], f"{voice_name}.wav")
sf.write(output_path, audio_trimmed, sr)
self.available_voices[voice_name] = output_path
metadata = {
"name": voice_name,
"sample_rate": sr,
"duration": len(audio_trimmed) / sr,
"file_size": os.path.getsize(output_path),
"created_at": datetime.now().isoformat()
}
with open(os.path.join(app.config['VOICE_FOLDER'], f"{voice_name}.json"), "w") as f:
json.dump(metadata, f, indent=2)
return True
except Exception as e:
print(f"❌ Failed to process voice sample: {str(e)}")
return False
def get_voice_preview(self, voice_name, text="Hello, this is a test of the voice."):
"""Generate a preview for a voice"""
if voice_name not in self.available_voices:
return None
preview_filename = f"preview_{voice_name}_{datetime.now().strftime('%H%M%S')}.wav"
success = self.generate_voice_cloned_speech(
text=text,
filename=preview_filename,
speaker_wav=self.available_voices[voice_name]
)
if success:
return preview_filename
return None
def cleanup_old_files(self):
"""Clean up old audio files"""
audio_dir = Path(app.config['AUDIO_FOLDER'])
audio_files = list(audio_dir.glob("*.wav"))
if len(audio_files) > app.config['MAX_AUDIO_FILES']:
audio_files.sort(key=lambda x: x.stat().st_ctime)
files_to_remove = len(audio_files) - app.config['MAX_AUDIO_FILES']
for i in range(files_to_remove):
try:
audio_files[i].unlink()
except:
pass
class Chatbot:
def __init__(self):
self.conversation_history = []
self.greeting_sent = False
def get_greeting(self):
"""Get chatbot greeting message"""
import random
greeting = random.choice(CHATBOT_RESPONSES["greetings"])
return greeting
def get_response(self, user_message):
"""Get chatbot response to user message"""
user_message_lower = user_message.lower().strip()
# Check for specific questions
response = None
for key, answer in CHATBOT_RESPONSES["questions"].items():
if key in user_message_lower:
response = answer
break
# If no specific match, use fallback
if not response:
import random
response = random.choice(CHATBOT_RESPONSES["fallback"])
return response
# Initialize TTS generator and chatbot
tts_generator = TTSGenerator()
chatbot = Chatbot()
def get_personalized_response(user_message, conversation_history):
"""Get AI response"""
try:
from Deepseek import get_deepseek_response
enhanced_prompt = f"""
{CHARACTER_DEFINITION}
Current user message: {user_message}
Please respond as Sancia, maintaining a friendly and helpful personality.
"""
response = get_deepseek_response(enhanced_prompt, [])
return response
except Exception as e:
print(f"❌ Error in get_personalized_response: {str(e)}")
return "I'm having some technical difficulties right now."
# Initialize TTS on startup
def initialize_tts():
global TTS_AVAILABLE
try:
print("πŸš€ Initializing TTS systems...")
print(f"πŸ“ Looking for Sancia voice at: {DEFAULT_VOICE_PATH}")
# Check if Sancia voice file exists
if os.path.exists(DEFAULT_VOICE_PATH):
print(f"βœ… Sancia voice file found: {DEFAULT_VOICE_FILE}")
else:
print(f"⚠ Sancia voice file NOT found: {DEFAULT_VOICE_FILE}")
print(f"πŸ“ Voices directory contents: {os.listdir(VOICES_DIRECTORY) if os.path.exists(VOICES_DIRECTORY) else 'Directory not found'}")
# Initialize base TTS first
base_ready = tts_generator.initialize_base_tts()
if base_ready:
TTS_AVAILABLE = True
print("βœ… Base TTS system initialized successfully")
# Try to initialize voice cloning
try:
tts_generator.initialize_voice_cloning_tts()
except Exception as e:
print(f"⚠ Voice cloning optional, continuing without it: {e}")
# Pre-generate welcome message with Sancia voice
print("🎡 Pre-generating welcome message with Sancia voice...")
welcome_file = tts_generator.generate_welcome_tts('default')
if welcome_file:
print(f"βœ… Sancia voice welcome TTS ready: {welcome_file}")
else:
print("⚠ Sancia voice welcome TTS generation failed")
# Also pre-generate original voice welcome as fallback
print("🎡 Pre-generating fallback welcome with original voice...")
original_welcome = tts_generator.generate_welcome_tts('original')
if original_welcome:
print(f"βœ… Original voice welcome TTS ready: {original_welcome}")
# Print available voices
voice_count = len(tts_generator.available_voices)
print(f"🎭 Found {voice_count} voice file(s) for cloning")
if tts_generator.has_sancia_voice:
print(f"βœ… Sancia voice available: {DEFAULT_VOICE_FILE}")
else:
print(f"❌ Sancia voice not available in loaded voices")
else:
print(f"❌ Base TTS initialization failed")
TTS_AVAILABLE = False
return TTS_AVAILABLE
except Exception as e:
print(f"❌ TTS initialization error: {str(e)}")
traceback.print_exc()
TTS_AVAILABLE = False
return False
# ============================================================================
# MAIN APPLICATION ROUTES
# ============================================================================
@app.route('/')
def index():
return render_template('index.html', welcome_message=WELCOME_MESSAGE)
@app.route('/api/welcome', methods=['GET'])
def get_welcome_message():
"""API endpoint to get the welcome message with TTS"""
try:
welcome_filename = None
audio_url = None
has_audio = False
voice_used = 'default'
print(f"🎯 Welcome endpoint called. TTS_AVAILABLE: {TTS_AVAILABLE}")
# Check if TTS is available
if TTS_AVAILABLE:
try:
# Try to get Sancia voice welcome TTS first
welcome_filename = tts_generator.generate_welcome_tts('default')
voice_used = 'Sancia'
if not welcome_filename:
# Fall back to original voice
print("⚠ Sancia voice failed, trying original voice...")
welcome_filename = tts_generator.generate_welcome_tts('original')
voice_used = 'Original'
if welcome_filename:
audio_url = f'/api/audio/{welcome_filename}'
has_audio = True
print(f"βœ… Welcome TTS available ({voice_used}): {welcome_filename}")
else:
print("❌ All welcome TTS generation attempts failed")
except Exception as tts_error:
print(f"❌ Welcome TTS generation exception: {tts_error}")
traceback.print_exc()
else:
print(f"⚠ TTS not available for welcome message")
print(f"πŸ“Š Returning welcome message. Voice: {voice_used}, Has audio: {has_audio}")
return jsonify({
'message': WELCOME_MESSAGE,
'audio_url': audio_url,
'audio_filename': welcome_filename,
'has_audio': has_audio,
'voice_used': voice_used,
'status': 'success'
})
except Exception as e:
print(f"❌ Error in get_welcome_message: {str(e)}")
traceback.print_exc()
return jsonify({
'message': WELCOME_MESSAGE,
'audio_url': None,
'audio_filename': None,
'has_audio': False,
'voice_used': 'none',
'status': 'success'
})
@app.route('/api/model', methods=['GET'])
def get_model():
"""Get information about the current model"""
try:
if os.path.exists(MODEL_PATH):
model_name = "Talking Animation" if SPECIFIC_MODEL == "talking.glb" else "Idle Animation"
return jsonify({
'model': {
'filename': SPECIFIC_MODEL,
'name': model_name,
'type': 'animation' if SPECIFIC_MODEL.endswith('.glb') else 'static',
'has_animations': True if SPECIFIC_MODEL.endswith('.glb') else False
},
'status': 'success'
})
else:
return jsonify({'error': f'Model not found: {MODEL_PATH}'}), 404
except Exception as e:
return jsonify({'error': 'Failed to fetch model'}), 500
@app.route('/api/model/file', methods=['GET'])
def serve_model():
"""Serve the main model file"""
try:
if not os.path.exists(MODEL_PATH):
return jsonify({'error': f'Model file not found: {MODEL_PATH}'}), 404
return send_file(MODEL_PATH, mimetype='model/gltf-binary')
except Exception as e:
print(f"❌ Error serving model: {str(e)}")
return jsonify({'error': 'Failed to serve model'}), 500
@app.route('/api/audio/<filename>', methods=['GET'])
def serve_audio(filename):
"""Serve audio files"""
try:
filepath = os.path.join(app.config['AUDIO_FOLDER'], filename)
if os.path.exists(filepath):
return send_file(filepath, mimetype='audio/wav')
else:
return jsonify({'error': 'Audio file not found'}), 404
except Exception as e:
print(f"❌ Error serving audio: {str(e)}")
return jsonify({'error': 'Failed to serve audio'}), 500
@app.route('/api/chat', methods=['POST', 'OPTIONS'])
def chat():
try:
if request.method == 'OPTIONS':
return jsonify({'status': 'ok'}), 200
data = request.get_json()
if not data:
return jsonify({'error': 'No JSON data provided'}), 400
user_message = data.get('message')
voice_name = data.get('voice', 'default') # 'default' means Sancia voice
if not user_message:
return jsonify({'error': 'No message provided'}), 400
print(f"πŸ’¬ Chat: '{user_message[:50]}...' with voice: {voice_name}")
# Get AI response
ai_response = get_personalized_response(user_message, [])
# Generate TTS if available
audio_filename = None
audio_url = None
if TTS_AVAILABLE:
audio_filename = tts_generator.generate_chat_tts(ai_response, voice_name)
if audio_filename:
audio_url = f'/api/audio/{audio_filename}'
return jsonify({
'response': ai_response,
'audio_url': audio_url,
'audio_filename': audio_filename,
'voice_used': voice_name,
'status': 'success'
})
except Exception as e:
print(f"❌ Error in chat endpoint: {str(e)}")
traceback.print_exc()
return jsonify({'error': 'Internal server error'}), 500
@app.route('/api/tts/generate', methods=['POST'])
def generate_speech():
"""Generate speech from text"""
try:
data = request.get_json()
text = data.get('text', '')
voice_name = data.get('voice_name', 'default') # Default to Sancia voice
if not text:
return jsonify({'error': 'No text provided'}), 400
if not TTS_AVAILABLE:
return jsonify({'error': 'TTS not available'}), 503
timestamp = datetime.now().strftime('%Y%m%d_%H%M%S')
random_suffix = ''.join(random.choices(string.ascii_lowercase + string.digits, k=6))
filename = f"speech_{voice_name}_{timestamp}_{random_suffix}.wav"
success = tts_generator.generate_speech_with_voice(
text=text,
filename=filename,
voice_name=voice_name
)
if success:
tts_generator.cleanup_old_files()
voice_used = 'Sancia' if voice_name == 'default' else voice_name
return jsonify({
'success': True,
'audio_url': f'/api/audio/{filename}',
'filename': filename,
'voice_used': voice_used
})
else:
return jsonify({
'success': False,
'error': 'Failed to generate speech'
}), 500
except Exception as e:
print(f"❌ TTS generation error: {str(e)}")
traceback.print_exc()
return jsonify({'error': 'Failed to generate speech'}), 500
@app.route('/api/tts/voices', methods=['GET'])
def get_voices():
"""Get list of available voices"""
try:
voices_data = []
# Add Sancia as default voice
voices_data.append({
'name': 'default',
'display_name': '🎭 Sancia Voice (Default)',
'type': 'default',
'description': 'The main Sancia voice using voice cloning',
'available': tts_generator.has_sancia_voice
})
# Add original voice as option
voices_data.append({
'name': 'original',
'display_name': 'πŸ”Š Original TTS Voice',
'type': 'base',
'description': 'Standard TTS voice (fallback)',
'available': True
})
# Add custom voices (excluding Sancia if it's already there)
for voice_name, voice_path in tts_generator.available_voices.items():
if voice_name != DEFAULT_VOICE_NAME: # Don't add Sancia twice
voice_info = {
'name': voice_name,
'display_name': f'🎀 {voice_name}',
'file_path': voice_path,
'preview_url': f'/api/voice-preview/{voice_name}',
'type': 'custom',
'available': True
}
# Add metadata if available
metadata_path = voice_path.replace('.wav', '.json')
if os.path.exists(metadata_path):
try:
with open(metadata_path, 'r') as f:
metadata = json.load(f)
voice_info.update(metadata)
except:
pass
voices_data.append(voice_info)
return jsonify({
'voices': voices_data,
'status': 'success',
'tts_available': TTS_AVAILABLE,
'has_sancia_voice': tts_generator.has_sancia_voice,
'default_voice': DEFAULT_VOICE_NAME,
'default_voice_available': tts_generator.has_sancia_voice
})
except Exception as e:
print(f"❌ Error getting voices: {str(e)}")
return jsonify({'error': 'Failed to get voices'}), 500
@app.route('/api/tts/info', methods=['GET'])
def tts_info():
"""Get TTS system information"""
try:
cache_files = list(Path(TTS_CACHE_DIR).glob("*.wav"))
cache_size = sum(f.stat().st_size for f in cache_files if f.is_file())
voices = len(tts_generator.available_voices)
# Check audio folder for existing files
audio_files = list(Path(app.config['AUDIO_FOLDER']).glob("*.wav"))
welcome_sancia_exists = os.path.exists(os.path.join(app.config['AUDIO_FOLDER'], 'welcome_sancia.wav'))
welcome_original_exists = os.path.exists(os.path.join(app.config['AUDIO_FOLDER'], 'welcome_original.wav'))
return jsonify({
'status': 'ready' if TTS_AVAILABLE else 'not_available',
'tts_engine': 'coqui_tts',
'tts_available': TTS_AVAILABLE,
'voice_cloning': True,
'voices_available': voices,
'device': tts_generator.device,
'base_model': tts_generator.base_model,
'cloning_model': tts_generator.voice_cloning_model_name,
'has_sancia_voice': tts_generator.has_sancia_voice,
'default_voice': DEFAULT_VOICE_NAME,
'welcome_sancia_ready': welcome_sancia_exists,
'welcome_original_ready': welcome_original_exists,
'cache_enabled': True,
'cache_size': len(cache_files),
'cache_mb': round(cache_size / (1024 * 1024), 2) if cache_size > 0 else 0,
'audio_files': len(audio_files),
'message': 'Enhanced TTS with Sancia as default voice'
})
except Exception as e:
print(f"❌ Error getting TTS info: {str(e)}")
return jsonify({'error': 'Failed to get TTS info'}), 500
@app.route('/api/voice-preview/<voice_name>', methods=['GET'])
def api_generate_voice_preview(voice_name):
"""Generate and return a voice preview"""
try:
if voice_name == 'default':
# Generate Sancia voice preview
preview_text = "Hello, this is the Sancia voice. How do I sound?"
preview_filename = tts_generator.generate_chat_tts(preview_text, 'default')
elif voice_name == 'original':
# Generate original voice preview
preview_text = "Hello, this is the original TTS voice. How do I sound?"
preview_filename = tts_generator.generate_chat_tts(preview_text, 'original')
else:
# Generate custom voice preview
if voice_name not in tts_generator.available_voices:
return jsonify({'error': 'Voice not found'}), 404
preview_text = "Hello, this is a preview of this custom voice. How do I sound?"
preview_filename = tts_generator.get_voice_preview(voice_name, preview_text)
if preview_filename:
return jsonify({
'success': True,
'audio_url': f'/api/audio/{preview_filename}',
'filename': preview_filename
})
else:
return jsonify({'error': 'Failed to generate preview'}), 500
except Exception as e:
print(f"❌ Error generating voice preview: {str(e)}")
return jsonify({'error': 'Failed to generate preview'}), 500
@app.route('/api/upload-voice', methods=['POST'])
def api_upload_voice():
"""Upload a new voice sample"""
if 'voice_file' not in request.files:
return jsonify({'error': 'No file provided'}), 400
file = request.files['voice_file']
voice_name = request.form.get('voice_name', '').strip()
if not voice_name:
return jsonify({'error': 'Voice name is required'}), 400
if file.filename == '':
return jsonify({'error': 'No file selected'}), 400
if not allowed_file(file.filename):
return jsonify({'error': 'File type not allowed'}), 400
file.seek(0, os.SEEK_END)
file_length = file.tell()
file.seek(0)
if file_length > app.config['MAX_UPLOAD_SIZE']:
return jsonify({'error': 'File too large'}), 400
if voice_name in tts_generator.available_voices:
return jsonify({'error': 'Voice name already exists'}), 400
success = tts_generator.process_uploaded_voice(file, voice_name)
if success:
return jsonify({
'success': True,
'message': 'Voice uploaded successfully',
'voice_name': voice_name
})
else:
return jsonify({'error': 'Failed to process voice sample'}), 500
# ============================================================================
# ANIMATION SYSTEM ROUTES (From Plan C)
# ============================================================================
@app.route('/animations')
def animations_page():
"""Serve the animations test page"""
return render_template('test_animations.html')
@app.route('/api/animations/status')
def animations_status():
"""API endpoint to check animations backend status"""
return jsonify({
"status": "running",
"models_available": list(DEFAULT_ANIMATIONS.keys()),
"animations_directory": ANIMATIONS_DIRECTORY,
"current_main_model": SPECIFIC_MODEL
})
@app.route('/api/animations/model/<model_name>')
def get_animation_model(model_name):
"""API endpoint to retrieve 3D animation model files"""
if model_name not in DEFAULT_ANIMATIONS:
return jsonify({"error": f"Model '{model_name}' not found"}), 404
filename = DEFAULT_ANIMATIONS[model_name]
model_path = os.path.join(app.config['ANIMATIONS_FOLDER'], filename)
if not os.path.exists(model_path):
return jsonify({"error": f"Model file '{filename}' not found at {model_path}"}), 404
print(f"Serving animation model: {filename}")
return send_file(model_path, mimetype='model/gltf-binary')
@app.route('/animations/<path:filename>')
def serve_animation_file(filename):
"""Serve files directly from the animations directory"""
return send_from_directory(app.config['ANIMATIONS_FOLDER'], filename)
# ============================================================================
# ANIMATION CONTROL API
# ============================================================================
@app.route('/api/animation/switch/<animation_name>', methods=['POST'])
def switch_animation(animation_name):
"""Switch the current animation"""
global SPECIFIC_MODEL, MODEL_PATH
if animation_name not in DEFAULT_ANIMATIONS:
return jsonify({'error': f'Animation {animation_name} not found'}), 404
new_model = DEFAULT_ANIMATIONS[animation_name]
new_model_path = os.path.join(ANIMATIONS_DIRECTORY, new_model)
if not os.path.exists(new_model_path):
return jsonify({'error': f'Animation file not found: {new_model_path}'}), 404
# Update the current model
SPECIFIC_MODEL = new_model
MODEL_PATH = new_model_path
print(f"πŸ”„ Switched to animation: {animation_name} ({new_model})")
return jsonify({
'success': True,
'message': f'Switched to {animation_name} animation',
'animation': animation_name,
'model_file': new_model,
'model_path': new_model_path
})
@app.route('/api/animation/current', methods=['GET'])
def get_current_animation():
"""Get the current animation state"""
animation_name = None
for name, filename in DEFAULT_ANIMATIONS.items():
if filename == SPECIFIC_MODEL:
animation_name = name
break
return jsonify({
'current_animation': animation_name or 'unknown',
'model_file': SPECIFIC_MODEL,
'model_path': MODEL_PATH,
'available_animations': list(DEFAULT_ANIMATIONS.keys())
})
# ============================================================================
# TESTING AND HEALTH ENDPOINTS
# ============================================================================
@app.route('/api/test', methods=['GET'])
def test_api():
return jsonify({
'status': 'success',
'message': 'API working',
'tts_engine': 'coqui_tts',
'tts_available': TTS_AVAILABLE,
'device': tts_generator.device,
'has_sancia_voice': tts_generator.has_sancia_voice,
'animations_available': True,
'animations_count': len(DEFAULT_ANIMATIONS),
'current_model': SPECIFIC_MODEL,
'model_path': MODEL_PATH
})
@app.route('/api/health', methods=['GET'])
def health_check():
# Determine current animation name
current_animation = None
for name, filename in DEFAULT_ANIMATIONS.items():
if filename == SPECIFIC_MODEL:
current_animation = name
break
return jsonify({
'status': 'healthy',
'service': 'FurryChat.ai',
'tts_engine': 'coqui_tts',
'voice_cloning': True,
'tts_available': TTS_AVAILABLE,
'default_voice': DEFAULT_VOICE_NAME,
'has_sancia_voice': tts_generator.has_sancia_voice,
'animations_system': 'ready',
'animations_available': list(DEFAULT_ANIMATIONS.keys()),
'current_animation': current_animation,
'current_model': SPECIFIC_MODEL
})
# ============================================================================
# ERROR HANDLERS
# ============================================================================
@app.errorhandler(404)
def not_found(error):
return jsonify({"error": "Endpoint not found"}), 404
@app.errorhandler(500)
def internal_error(error):
return jsonify({"error": "Internal server error"}), 500
# ============================================================================
# MAIN APPLICATION STARTUP
# ============================================================================
if __name__ == '__main__':
print("=" * 60)
print("πŸš€ Starting FurryChat.ai with Integrated Animations System")
print("=" * 60)
# Check 3D model
if os.path.exists(MODEL_PATH):
print(f"βœ… 3D Main Model: {SPECIFIC_MODEL}")
print(f" Path: {MODEL_PATH}")
print(f" Size: {os.path.getsize(MODEL_PATH):,} bytes")
else:
print(f"❌ 3D Main Model not found: {MODEL_PATH}")
# Try to find it
if os.path.exists(ANIMATIONS_DIRECTORY):
files = os.listdir(ANIMATIONS_DIRECTORY)
print(f" Available files: {files}")
# Try to use the first GLB file
for file in files:
if file.endswith('.glb'):
SPECIFIC_MODEL = file
MODEL_PATH = os.path.join(ANIMATIONS_DIRECTORY, file)
print(f" Using alternative: {file}")
break
# Check animation models
check_animation_files()
# Check Sancia voice file
if os.path.exists(DEFAULT_VOICE_PATH):
print(f"βœ… Sancia Voice: {DEFAULT_VOICE_FILE} (Found)")
else:
print(f"⚠ Sancia Voice: {DEFAULT_VOICE_FILE} (Not found - will use original voice)")
print(f"πŸ“ Please ensure {DEFAULT_VOICE_FILE} is in {VOICES_DIRECTORY}")
# Initialize TTS FIRST (before any requests)
if initialize_tts():
print("βœ… TTS: Ready with Sancia as default voice")
print(f"πŸ“¦ Audio Output: {AUDIO_OUTPUT_DIR}")
print(f"🎭 Voice Files: {VOICES_DIRECTORY}")
else:
print("❌ TTS: Initialization failed")
print("πŸ’‘ To install TTS: pip install coqui-tts torch torchaudio")
print(f"🎬 Animations System: Ready")
print(f" - Current animation: {SPECIFIC_MODEL}")
print(f" - Access test at: http://127.0.0.1:5000/animations")
print(f" - Switch API: POST /api/animation/switch/<idle|talking>")
print(f"🌐 Main Interface: http://127.0.0.1:5000")
print("=" * 60)
app.run(debug=True, port=5000, host='0.0.0.0', threaded=True)