Face Recognition Telegram Bot

Bot Telegram untuk deteksi dan pengenalan wajah menggunakan API Face Recognition.

🚀 Fitur Utama

1. Deteksi Wajah

• Upload foto untuk mendeteksi wajah
• Identifikasi wajah yang sudah terdaftar
• Confidence score untuk setiap deteksi
• Analisis AI dengan Gemini (optional)

2. Manajemen Database Wajah

• Tambah wajah baru ke database
• Update informasi wajah
• Lihat daftar semua wajah
• Lihat detail informasi wajah

3. Video Processing

• Upload video untuk deteksi wajah
• Async processing dengan progress tracking
• Cancel processing yang sedang berjalan
• Download hasil processing

4. Stream Processing

• Proses webcam (lokal/remote)
• Support RTSP stream
• Support HTTP/HTTPS stream
• Adjustable duration dan FPS

📦 Instalasi

Prerequisites

# Node.js 14+ 
node --version

# NPM packages
npm install node-telegram-bot-api form-data

Setup

# Clone atau copy script
git clone <repository-url>
cd face-recognition-bot

# Install dependencies
npm install

# Set environment variables
export TELEGRAM_BOT_TOKEN="your-bot-token"

# Run bot
node bot.js

⚙️ Konfigurasi

Environment Variables

# Required
export TELEGRAM_BOT_TOKEN="7587121696:AAHwkAhyiNI5c5Y62zFAMLtRLmBGULZ8BWw"

# API Server (default sudah di-set)
# Tidak perlu diubah jika server di 10.122.25.138:8192

Config dalam Script

const API_BASE_URL = 'http://10.122.25.138:8192'  // Face Recognition API
const BOT_NAME = 'FaceRecBot'                      // Nama bot
const TYPING_DELAY = 1000                          // Delay typing indicator

📖 Panduan Penggunaan

Perintah Dasar

/start

Memulai bot dan menampilkan welcome message

/start

/help atau /menu

Menampilkan panduan lengkap

/help

/status

Cek status API server

/status

Deteksi Wajah

Deteksi Sederhana

1. Kirim foto ke bot
2. Bot akan memproses dan mendeteksi wajah
3. Hasil berisi: nama, confidence, info (jika ada)

Deteksi dengan AI Analysis

1. Kirim foto dengan caption: analyze
2. Bot akan menggunakan Gemini AI untuk analisis tambahan

[Kirim foto]
Caption: analyze

Deteksi dengan Pertanyaan Custom

[Kirim foto]
Caption: analyze - Apa yang dilakukan orang ini?

Manajemen Wajah

Tambah Wajah Baru

# Method 1: Tanpa info
/addface John Doe
[Kirim foto wajah John Doe]

# Method 2: Dengan info
/addface John Doe | Manager IT Department
[Kirim foto wajah John Doe]

Update Info Wajah

/updateinfo John Doe | Senior Manager IT Department

Lihat Semua Wajah

/listfaces

Info Detail Wajah

/faceinfo John Doe

Video Processing

Upload dan Proses Video

1. Kirim file video ke bot
2. Bot akan upload dan memproses secara async
3. Dapatkan session ID untuk tracking

Monitor Progress

/videostatus <session_id>

Cancel Processing

/videocancel <session_id>

Dapatkan Hasil

/videoresults <session_id>

Stream Processing

Webcam Lokal

# Default: 60 detik, 1 FPS
/stream 0

# Custom: 30 detik, 2 FPS
/stream 0 30 2

RTSP Stream

/stream rtsp://192.168.1.100:554/stream

# Dengan custom duration dan FPS
/stream rtsp://192.168.1.100:554/stream 120 2

HTTP Stream

/stream http://camera-ip/stream.mjpeg

Maintenance

Regenerate Embeddings

/regenerate

🔧 API Integration

Endpoints yang Digunakan

Endpoint Method Fungsi /health GET Health check /faces GET List wajah /faces/add POST Tambah wajah /faces/update/<name> POST Update info /process POST Proses gambar /video/upload POST Upload video /video/status/<id> GET Status video /video/cancel/<id> POST Cancel video /video/results/<id> GET Hasil video /stream/process POST Proses stream /regenerate POST Regenerate embeddings Request Format

Tambah Wajah

FormData:
- image: file (jpg/png)
- name: string
- info: string (optional)

Proses Gambar

FormData:
- image: file (jpg/png)
- analyze_with_gemini: boolean
- custom_question: string (optional)

Upload Video

FormData:
- video: file (mp4/avi/mov/mkv)
- fps_process: integer (default: 1)
- process_async: boolean (default: true)

Proses Stream

JSON:
{
  "stream_url": "0" | "rtsp://..." | "http://...",
  "duration": 60,
  "fps_process": 1
}

Response Format

Deteksi Wajah

{
  "face_count": 2,
  "faces": [
    {
      "name": "John Doe",
      "confidence": 0.95,
      "bbox": [100, 50, 200, 150],
      "info": "Manager IT"
    }
  ],
  "result_url": "/images/results/result_xxx.jpg",
  "gemini_analysis": "..."
}

Video Upload

{
  "status": "success",
  "message": "Video berhasil diupload",
  "session_id": "abc-123-def-456"
}

Video Status

{
  "status": "processing",
  "progress": 45,
  "message": "Memproses frame 450/1000",
  "frame_number": 450,
  "processed_count": 45
}

📝 Error Handling

Rate Limiting

• User dibatasi 1 pesan per 2 detik
• Mencegah spam dan overload

API Errors

• Connection timeout → Retry otomatis
• Invalid response → Error message ke user
• Server down → Status warning

File Handling

• Max file size: 100MB (Telegram limit)
• Supported formats:
• Images: JPG, PNG, JPEG
• Videos: MP4, AVI, MOV, MKV

🐛 Troubleshooting

Bot Tidak Respon

# Cek status bot
ps aux | grep node

# Cek log
tail -f logs/bot.log

# Restart bot
pkill -f "node bot.js"
node bot.js

API Connection Failed

# Test API
curl http://10.122.25.138:8192/health

# Cek network
ping 10.122.25.138

# Test port
nc -zv 10.122.25.138 8192

Video Processing Stuck

# Gunakan cancel command
/videocancel <session_id>

# Atau restart API server jika perlu

📊 Logging

Log Location

./logs/bot.log

Log Format

[2025-10-05T10:30:45.123Z] 🖼️ Photo from John
[2025-10-05T10:30:46.456Z] ✅ Response sent to John

Log Monitoring

# Real-time monitoring
tail -f logs/bot.log

# Search errors
grep "❌" logs/bot.log

# Count messages
grep "Photo from" logs/bot.log | wc -l

🔐 Security

API Security

• API hanya di internal network (10.122.25.138)
• Tidak expose ke public internet
• Rate limiting untuk mencegah abuse

Bot Security

• Token disimpan di environment variable
• Tidak di-hardcode di script
• Log tidak menyimpan token

Data Privacy

• Gambar hasil disimpan sementara
• Video sessions dibersihkan otomatis
• User data tidak disimpan permanen

🚀 Production Deployment

PM2 (Recommended)

# Install PM2
npm install -g pm2

# Start bot
pm2 start bot.js --name face-recognition-bot

# Auto restart on reboot
pm2 startup
pm2 save

# Monitor
pm2 monit

# Logs
pm2 logs face-recognition-bot

Systemd Service

[Unit]
Description=Face Recognition Telegram Bot
After=network.target

[Service]
Type=simple
User=youruser
WorkingDirectory=/path/to/bot
Environment="TELEGRAM_BOT_TOKEN=your-token"
ExecStart=/usr/bin/node bot.js
Restart=always

[Install]
WantedBy=multi-user.target

Docker

FROM node:16-alpine
WORKDIR /app
COPY package*.json ./
RUN npm install
COPY . .
ENV TELEGRAM_BOT_TOKEN=""
CMD ["node", "bot.js"]

📈 Monitoring

Health Check

# Bot health
curl http://10.122.25.138:8192/health

# Check response time
time curl http://10.122.25.138:8192/health

Performance Metrics

• Response time: < 2 detik (image)
• Video processing: ~1 menit per 100 frames
• Stream processing: Real-time

🔄 Updates & Maintenance

Update Dependencies

npm update
npm audit fix

Backup

# Backup logs
cp -r logs logs_backup_$(date +%Y%m%d)

# Backup script
cp bot.js bot.js.backup

Clear Cache

# Clear old sessions
rm -rf video_sessions/*

# Clear old logs (older than 7 days)
find logs/ -name "*.log" -mtime +7 -delete

📞 Support

Debugging Mode

Enable verbose logging di script:

const DEBUG = true

Version: 1.0.0

Last Updated: October 2025

Status: Production Ready ✅