#!/bin/bash
# PDF索引脚本 - 将核心技术PDF转文本并建立可搜索索引

INDEX_DIR="/root/personal/learn/pdf_index"
PDF_LISTS=(
  "/root/personal/learn/lectures/BLDC_FOC"
  "/root/personal/learn/bldc_motor_control/pdfs/app_notes"
  "/root/personal/learn/humanoid_robot"
  "/root/personal/learn/oopc"
)

mkdir -p "$INDEX_DIR/txt"
INDEX_FILE="$INDEX_DIR/search_index.txt"
echo "# PDF Search Index - $(date)" > "$INDEX_FILE"
echo "# Format: filename ||| first_line ||| filepath" >> "$INDEX_FILE"
echo "" >> "$INDEX_FILE"

COUNT=0
for base in "${PDF_LISTS[@]}"; do
  find "$base" -name "*.pdf" -type f 2>/dev/null | while read pdf; do
    fname=$(basename "$pdf" .pdf)
    txtpath="$INDEX_DIR/txt/${fname}.txt"
    
    # Convert to text if not already done
    if [ ! -f "$txtpath" ]; then
      pdftotext -layout "$pdf" "$txtpath" 2>/dev/null
    fi
    
    # Get first non-empty line for indexing
    first_line=$(head -5 "$txtpath" 2>/dev/null | tr -d '\000' | sed '/^$/d' | head -1 | tr -d '\n' | cut -c1-80)
    
    echo "${fname} ||| ${first_line} ||| ${pdf}" >> "$INDEX_FILE"
    COUNT=$((COUNT + 1))
  done
done

echo "Indexed $COUNT PDFs" 
echo "Done: $INDEX_FILE"