#!/bin/bash
# =============================================
#  Мониторинг и очистка OCR — 4 часа таймаут
# =============================================

REPORT_DIR="/home/dayhanbiz/public_html/biblioteka/admin/reports"
TIMESTAMP=$(date "+%Y-%m-%d_%H-%M-%S")
REPORT_FILE="$REPORT_DIR/ocr_report_$TIMESTAMP.txt"

echo "=====================================================" > "$REPORT_FILE"
echo "Отчёт OCR-системы — $TIMESTAMP" >> "$REPORT_FILE"
echo "=====================================================" >> "$REPORT_FILE"
echo "" >> "$REPORT_FILE"

# 1. Свободное место
echo "=== 1. Свободное место на диске ===" >> "$REPORT_FILE"
df -h / >> "$REPORT_FILE"
echo "" >> "$REPORT_FILE"

# 2. Размер таблиц
echo "=== 2. Размер таблиц ===" >> "$REPORT_FILE"
mysql -u dayhanbiz_biblioteka_index -p'HN_+@(ngb25642r1' -D dayhanbiz_biblioteka_index -e "
SELECT 
    table_name AS 'Таблица',
    ROUND((data_length + index_length) / 1024 / 1024, 2) AS 'Размер (MB)',
    table_rows AS 'Строк'
FROM information_schema.tables 
WHERE table_schema = 'dayhanbiz_biblioteka_index' 
  AND table_name IN ('pages', 'books', 'word_index')
ORDER BY (data_length + index_length) DESC;" >> "$REPORT_FILE" 2>&1
echo "" >> "$REPORT_FILE"

# 3. Обработано PDF-страниц
echo "=== 3. Обработано PDF-страниц ===" >> "$REPORT_FILE"
mysql -u dayhanbiz_biblioteka_index -p'HN_+@(ngb25642r1' -D dayhanbiz_biblioteka_index -e "
SELECT COUNT(*) as 'Обработано PDF-страниц' 
FROM pages 
WHERE is_pdf_page = 1;" >> "$REPORT_FILE"
echo "" >> "$REPORT_FILE"

# 4. Очистка временных файлов (старше 4 часов)
echo "=== 4. Очистка временных файлов ===" >> "$REPORT_FILE"
echo "Удалено файлов старше 4 часов:" >> "$REPORT_FILE"
find /home/dayhanbiz/ocr_temp -type f -mmin +240 -delete -print | wc -l >> "$REPORT_FILE"
echo "Очистка завершена." >> "$REPORT_FILE"

echo "" >> "$REPORT_FILE"
echo "Отчёт завершён." >> "$REPORT_FILE"
echo "=====================================================" >> "$REPORT_FILE"

echo "Отчёт сохранён: $REPORT_FILE"
tail -n 20 "$REPORT_FILE"
