feat: fix frontend and activate imports

This commit is contained in:
2026-03-17 13:53:12 -03:00
parent be699ad6ab
commit 31939e31ad
24 changed files with 34 additions and 2117 deletions
-206
View File
@@ -1,206 +0,0 @@
# Instalación Manual de Bark TTS en t-800
## Opción A: Instalación Automática (Recomendada)
```bash
# 1. Copiar script a t-800
scp scripts/install_bark_tts.sh juan@t-800:/tmp/install_bark_tts.sh
# 2. Conectarse a t-800
ssh juan@t-800
# 3. Ejecutar instalación
chmod +x /tmp/install_bark_tts.sh
sudo /tmp/install_bark_tts.sh
# 4. Verificar instalación
curl http://localhost:8000/health
```
## Opción B: Instalación Paso a Paso
```bash
# Conectarse a t-800
ssh juan@t-800
# Contraseña: apoca11
# Actualizar sistema
sudo apt-get update
sudo apt-get install -y python3 python3-pip python3-venv git ffmpeg curl
# Crear directorio
sudo mkdir -p /opt/bark
sudo chown juan:juan /opt/bark
cd /opt/bark
# Clonar Bark
git clone https://github.com/suno-ai/bark.git
cd bark
# Crear entorno virtual
python3 -m venv venv
source venv/bin/activate
# Instalar dependencias
pip install --upgrade pip
pip install -e .
pip install fastapi uvicorn[standard] python-multipart numpy scipy
# Crear archivo de API
cat > bark_api.py << 'PYEOF'
from fastapi import FastAPI, HTTPException, Query
from fastapi.responses import StreamingResponse
from bark import SAMPLE_RATE, generate_audio, preload_models
from scipy.io.wavfile import write as write_wav
import numpy as np
import io
app = FastAPI(title="Bark TTS API", version="1.0.0")
print("Preloading Bark models...")
preload_models()
print("Models loaded!")
@app.get("/health")
async def health_check():
return {"status": "healthy", "service": "bark-tts"}
@app.get("/api/voices")
async def list_voices():
return {
"voices": [
{"id": "v2/en_speaker_0", "name": "English Speaker 0", "language": "en"},
{"id": "v2/en_speaker_1", "name": "English Speaker 1", "language": "en"},
{"id": "v2/en_speaker_6", "name": "English Speaker 6", "language": "en"},
{"id": "v2/es_speaker_0", "name": "Spanish Speaker 0", "language": "es"},
{"id": "v2/es_speaker_1", "name": "Spanish Speaker 1", "language": "es"},
{"id": "v2/es_speaker_3", "name": "Spanish Speaker 3", "language": "es"},
]
}
@app.post("/api/generate")
async def generate_speech(
text: str = Query(..., min_length=1, max_length=500),
voice: str = Query(default="v2/en_speaker_1"),
speed: float = Query(default=1.0, ge=0.5, le=2.0),
output_format: str = Query(default="wav", regex="^(mp3|wav|ogg)$")
):
try:
audio_array = generate_audio(text, history_prompt=voice)
if speed != 1.0:
new_length = int(len(audio_array) / speed)
audio_array = audio_array[:new_length]
audio_buffer = io.BytesIO()
write_wav(audio_buffer, SAMPLE_RATE, audio_array)
audio_buffer.seek(0)
return StreamingResponse(
audio_buffer,
media_type="audio/wav",
headers={"Content-Disposition": f"attachment; filename=speech.wav"}
)
except Exception as e:
raise HTTPException(status_code=500, detail=str(e))
if __name__ == "__main__":
import uvicorn
uvicorn.run(app, host="0.0.0.0", port=8000)
PYEOF
# Crear servicio systemd
cat > /tmp/bark-tts.service << EOF
[Unit]
Description=Bark TTS API Server
After=network.target
[Service]
Type=simple
User=juan
Group=juan
WorkingDirectory=/opt/bark/bark
Environment="PATH=/opt/bark/bark/venv/bin"
ExecStart=/opt/bark/bark/venv/bin/uvicorn bark_api:app --host 0.0.0.0 --port 8000 --workers 1
Restart=always
RestartSec=10
MemoryMax=4G
MemoryHigh=3G
CPUQuota=80%
[Install]
WantedBy=multi-user.target
EOF
sudo mv /tmp/bark-tts.service /etc/systemd/system/
sudo systemctl daemon-reload
sudo systemctl enable bark-tts
sudo systemctl start bark-tts
# Verificar
sudo systemctl status bark-tts
curl http://localhost:8000/health
```
## Prueba de Funcionamiento
```bash
# Test básico
curl "http://localhost:8000/api/generate?text=Hello%20World&voice=v2/en_speaker_1" -o test.wav
# Test desde OpenCCB
curl http://t-800:8000/health
# Ver logs
sudo journalctl -u bark-tts -f
```
## Configuración en OpenCCB
Agregar a `.env`:
```bash
BARK_API_URL=http://t-800:8000
# O para producción:
# BARK_API_URL=http://t-800.norteamericano.cl:8000
```
## Solución de Problemas
### Error: "Out of Memory"
```bash
# Reducir límite de memoria en systemd
sudo systemctl edit bark-tts
# Agregar:
# [Service]
# MemoryMax=2G
```
### Error: "Model not found"
```bash
# Reinstalar modelos
cd /opt/bark/bark
source venv/bin/activate
python -c "from bark import preload_models; preload_models()"
```
### Servicio no inicia
```bash
# Ver logs
sudo journalctl -u bark-tts -n 50
# Reiniciar
sudo systemctl restart bark-tts
```
## URLs de Acceso
- **Health**: http://t-800:8000/health
- **Voices**: http://t-800:8000/api/voices
- **Generate**: http://t-800:8000/api/generate?text=Hello&voice=v2/en_speaker_1
## Producción (t-800.norteamericano.cl)
Para producción, asegurar que:
1. El puerto 8000 esté abierto en el firewall
2. El dominio t-800.norteamericano.cl apunte a la IP correcta
3. Usar BARK_API_URL=http://t-800.norteamericano.cl:8000
-221
View File
@@ -1,221 +0,0 @@
# Bark TTS Integration Guide
## Overview
OpenCCB now integrates with **Suno AI's Bark** text-to-speech system for generating audio versions of questions. This allows students to listen to questions instead of just reading them, improving accessibility and supporting different learning styles.
## Architecture
```
┌─────────────────┐ HTTP ┌─────────────────┐
│ OpenCCB CMS │ ────────────> │ Bark TTS API │
│ (PostgreSQL) │ <──────────── │ (Server t-800)│
│ │ Audio │ │
└─────────────────┘ └─────────────────┘
```
## Deployment to t-800 Server
### Prerequisites
- SSH access to t-800 server
- At least 8GB RAM recommended (Bark loads large models)
- 10GB free disk space
- Python 3.8+
- GPU optional (CUDA support for faster generation)
### Quick Deploy
```bash
# From your local machine
cd /home/juan/dev/openccb
./scripts/deploy_to_t800.sh
```
This will:
1. SSH into t-800
2. Install Python dependencies
3. Clone Bark repository
4. Set up systemd service
5. Start the API server
### Manual Deploy
```bash
# SSH into t-800
ssh juan@t-800
# Run installation script
wget https://raw.githubusercontent.com/suno-ai/bark/main/scripts/install.sh
sudo bash install.sh
```
## API Endpoints
Once deployed, Bark API is available at `http://t-800:8000`
### Health Check
```bash
curl http://t-800:8000/health
```
### List Available Voices
```bash
curl http://t-800:8000/api/voices
```
### Generate Speech
```bash
# Basic usage
curl "http://t-800:8000/api/generate?text=What%20color%20is%20the%20sky%3F" \
-o question.wav
# With specific voice and speed
curl "http://t-800:8000/api/generate?text=Hello%20World&voice=v2/en_speaker_6&speed=1.2" \
-o greeting.wav
# Spanish voice
curl "http://t-800:8000/api/generate?text=Hola%20mundo&voice=v2/es_speaker_0" \
-o saludo.wav
```
## Available Voices
### English Voices
- `v2/en_speaker_0` through `v2/en_speaker_9`
### Spanish Voices
- `v2/es_speaker_0` through `v2/es_speaker_9`
## Integration with OpenCCB
### Generate Audio for a Question
```bash
# Via API
curl -X POST "http://localhost:3001/question-bank/{question_id}/generate-audio" \
-H "Authorization: Bearer YOUR_TOKEN" \
-H "Content-Type: application/json" \
-d '{
"text": "What color is the sky?",
"voice": "v2/en_speaker_1",
"speed": 1.0
}'
```
### Automatic Audio Generation
When creating a question:
```json
POST /question-bank
{
"question_text": "What is the capital of France?",
"question_type": "multiple-choice",
"options": ["Paris", "London", "Berlin", "Madrid"],
"correct_answer": 0,
"explanation": "Paris is the capital of France.",
"generate_audio": true // Triggers async audio generation
}
```
## Configuration
### Environment Variables
Add to your `.env` file:
```bash
# Bark TTS API URL
BARK_API_URL=http://t-800:8000
# Optional: Default voice for audio generation
BARK_DEFAULT_VOICE=v2/en_speaker_1
# Optional: Default speed
BARK_DEFAULT_SPEED=1.0
```
## Performance Optimization
### Model Preloading
Bark preloads models on startup (takes ~30 seconds). The systemd service handles this automatically.
### Memory Management
The systemd service includes memory limits:
```ini
MemoryMax=4G
MemoryHigh=3G
```
Adjust based on your server's capacity.
### Batch Generation
For importing many questions:
```bash
# Generate audio for multiple questions
curl "http://t-800:8000/api/generate/batch?texts=Question%201&texts=Question%202&voice=v2/en_speaker_1"
```
## Troubleshooting
### Service Not Starting
```bash
# Check status
sudo systemctl status bark-tts
# View logs
sudo journalctl -u bark-tts -f
# Restart service
sudo systemctl restart bark-tts
```
### Out of Memory
If Bark crashes due to memory:
1. Reduce `MemoryMax` in systemd service
2. Use smaller models: `suno/bark-small`
3. Process questions one at a time
### Slow Generation
- GPU acceleration: Install CUDA-enabled PyTorch
- Reduce audio quality settings
- Use shorter text segments
## Testing
```bash
# Test English voice
curl "http://t-800:8000/api/generate?text=The%20quick%20brown%20fox&voice=v2/en_speaker_1" | play -
# Test Spanish voice
curl "http://t-800:8000/api/generate?text=El%20rápido%20zorro%20marrón&voice=v2/es_speaker_0" | play -
```
## Security Notes
- Bark API runs on internal network only
- No authentication required (assumes trusted network)
- Rate limiting handled by OpenCCB
- Audio files stored in `uploads/audio/` directory
## Future Enhancements
- [ ] Add authentication to Bark API
- [ ] Support for custom voice cloning
- [ ] Audio preprocessing (noise reduction, normalization)
- [ ] Caching layer for repeated requests
- [ ] WebSocket support for streaming audio
## References
- [Bark GitHub](https://github.com/suno-ai/bark)
- [Bark Hugging Face](https://huggingface.co/suno/bark)
- [OpenCCB Question Bank Documentation](../docs/question-bank.md)
-40
View File
@@ -101,40 +101,6 @@ Modal para importar preguntas desde MySQL:
- Progreso de importación
- Resultado (éxito/error)
### 5. AudioGeneratorModal (`AudioGeneratorModal.tsx`)
Modal para generar audio con Bark TTS:
**Configuración:**
- Vista previa del texto de la pregunta
- Texto personalizable (opcional)
- Selector de voz (6 opciones: 3 inglés, 3 español)
- Control de velocidad (0.5x - 2.0x)
**Voces disponibles:**
```
Inglés:
- v2/en_speaker_0 (English Speaker 0)
- v2/en_speaker_1 (English Speaker 1) ← default
- v2/en_speaker_6 (English Speaker 6)
Español:
- v2/es_speaker_0 (Spanish Speaker 0)
- v2/es_speaker_1 (Spanish Speaker 1)
- v2/es_speaker_3 (Spanish Speaker 3)
```
**Estados:**
- ⏳ Generando... (polling cada 1s, max 30s)
- ✅ Audio generado (con preview play/pause)
- ❌ Error (mensaje descriptivo)
**Características:**
- Polling automático para verificar estado
- Reproductor de audio integrado
- Botón Play/Pause
- Indicador visual de estado
## Flujos de Usuario
### Crear Pregunta Manualmente
@@ -288,11 +254,6 @@ GET /question-bank/mysql-courses
- Verificar que `MYSQL_DATABASE_URL` esté configurado en `.env`
- Verificar conectividad al servidor MySQL
**Error: "Error al generar audio"**
- Verificar que Bark TTS esté corriendo en t-800
- Verificar que `BARK_API_URL` esté configurado
- Revisar logs de Bark: `ssh juan@t-800 && journalctl -u bark-tts -f`
**Audio no se reproduce**
- Verificar formato de audio (WAV soportado)
- Verificar permisos del navegador
@@ -301,5 +262,4 @@ GET /question-bank/mysql-courses
## Referencias
- [Question Bank Backend](../../services/cms-service/src/handlers_question_bank.rs)
- [Bark TTS Guide](../../docs/BARK_TTS_GUIDE.md)
- [Test Templates UI](./TestTemplates/)