#!/bin/bash
#
# Training Hybrid ML Detector su Dati Reali
# Legge credenziali da /opt/ids/.env automaticamente
#

set -e  # Exit on error

echo "======================================================================="
echo "  TRAINING HYBRID ML DETECTOR - DATI REALI"
echo "======================================================================="
echo ""

# Percorsi
IDS_ROOT="/opt/ids"
ENV_FILE="$IDS_ROOT/.env"
PYTHON_ML_DIR="$IDS_ROOT/python_ml"
VENV_PYTHON="$PYTHON_ML_DIR/venv/bin/python"

# Verifica file .env esiste
if [ ! -f "$ENV_FILE" ]; then
    echo "❌ ERRORE: File .env non trovato in $ENV_FILE"
    exit 1
fi

# Carica variabili da .env
echo "📂 Caricamento credenziali database da .env..."
source "$ENV_FILE"

# Estrai credenziali database
DB_HOST="${PGHOST:-localhost}"
DB_PORT="${PGPORT:-5432}"
DB_NAME="${PGDATABASE:-ids}"
DB_USER="${PGUSER:-postgres}"
DB_PASSWORD="${PGPASSWORD}"

# Verifica password estratta
if [ -z "$DB_PASSWORD" ]; then
    echo "❌ ERRORE: PGPASSWORD non trovata nel file .env"
    echo "   Aggiungi: PGPASSWORD=tua_password_qui"
    exit 1
fi

echo "✅ Credenziali caricate:"
echo "   Host: $DB_HOST"
echo "   Port: $DB_PORT"
echo "   Database: $DB_NAME"
echo "   User: $DB_USER"
echo "   Password: ****** (nascosta)"
echo ""

# Parametri training
DAYS="${1:-7}"  # Default 7 giorni, puoi passare come argomento
MAX_SAMPLES="${2:-1000000}"  # Default 1M records max

echo "🎯 Parametri training:"
echo "   Periodo: ultimi $DAYS giorni"
echo "   Max records: $MAX_SAMPLES"
echo ""

# Verifica venv Python
if [ ! -f "$VENV_PYTHON" ]; then
    echo "❌ ERRORE: Virtual environment non trovato in $VENV_PYTHON"
    echo "   Esegui prima: cd $IDS_ROOT && python3 -m venv python_ml/venv"
    exit 1
fi

echo "🐍 Python: $VENV_PYTHON"
echo ""

# Verifica dati disponibili nel database
echo "📊 Verifica dati disponibili nel database..."
PGPASSWORD="$DB_PASSWORD" psql -h "$DB_HOST" -p "$DB_PORT" -U "$DB_USER" -d "$DB_NAME" -c "
SELECT 
  TO_CHAR(MIN(timestamp), 'YYYY-MM-DD HH24:MI:SS') as primo_log,
  TO_CHAR(MAX(timestamp), 'YYYY-MM-DD HH24:MI:SS') as ultimo_log,
  EXTRACT(DAY FROM (MAX(timestamp) - MIN(timestamp))) || ' giorni' as periodo_totale,
  TO_CHAR(COUNT(*), 'FM999,999,999') as totale_records
FROM network_logs;
" 2>/dev/null

if [ $? -ne 0 ]; then
    echo "⚠️  WARNING: Impossibile verificare dati database (continuo comunque...)"
fi

echo ""
echo "🚀 Avvio training..."
echo ""
echo "======================================================================="

# Cambia directory
cd "$PYTHON_ML_DIR"

# Esegui training
"$VENV_PYTHON" train_hybrid.py --train --source database \
  --db-host "$DB_HOST" \
  --db-port "$DB_PORT" \
  --db-name "$DB_NAME" \
  --db-user "$DB_USER" \
  --db-password "$DB_PASSWORD" \
  --days "$DAYS"

# Check exit code
if [ $? -eq 0 ]; then
    echo ""
    echo "======================================================================="
    echo "✅ TRAINING COMPLETATO CON SUCCESSO!"
    echo "======================================================================="
    echo ""
    echo "📁 Modelli salvati in: $PYTHON_ML_DIR/models/"
    echo ""
    echo "🔄 PROSSIMI PASSI:"
    echo "   1. Restart ML backend: sudo systemctl restart ids-ml-backend"
    echo "   2. Verifica caricamento: sudo journalctl -u ids-ml-backend -f"
    echo "   3. Test API: curl http://localhost:8000/health"
    echo ""
else
    echo ""
    echo "======================================================================="
    echo "❌ ERRORE DURANTE IL TRAINING"
    echo "======================================================================="
    echo ""
    echo "Controlla i log sopra per dettagli sull'errore."
    exit 1
fi