#!/bin/bash

# Wordlist Builder Script
# Downloads, processes, and creates master wordlist for hashcat cracking

set -e

SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
WORKSPACE_DIR="$(dirname "$SCRIPT_DIR")"
WORDLIST_DIR="$WORKSPACE_DIR/wordlists"
TEMP_DIR="$WORKSPACE_DIR/temp"

# Create directories
mkdir -p "$WORDLIST_DIR" "$TEMP_DIR"

echo "=========================================="
echo "Wordlist Builder for Hashcat Cracking"
echo "=========================================="

# Function to download wordlist
download_wordlist() {
    local url="$1"
    local filename="$2"
    
    echo "Downloading: $filename"
    if command -v wget &> /dev/null; then
        wget -q --show-progress -O "$TEMP_DIR/$filename" "$url"
    elif command -v curl &> /dev/null; then
        curl -s -L "$url" -o "$TEMP_DIR/$filename"
    else
        echo "Error: Neither wget nor curl found. Please install one of them."
        exit 1
    fi
}

# Function to process and clean wordlist
clean_wordlist() {
    local input_file="$1"
    local output_file="$2"
    
    echo "Cleaning: $(basename "$input_file")"
    
    # Remove non-printable characters, convert to lowercase, sort, remove duplicates
    iconv -f utf-8 -t utf-8//IGNORE "$input_file" | \
        tr -cd '\11\12\15\40-\176' | \
        tr '[:upper:]' '[:lower:]' | \
        sort -u | \
        grep -v '^$' > "$output_file"
    
    echo "  → Cleaned: $(wc -l < "$output_file") lines"
}

# List of wordlist sources (common breached password lists)
WORDLIST_SOURCES=(
    "https://raw.githubusercontent.com/danielmiessler/SecLists/master/Passwords/Leaked-Databases/rockyou.txt"
    "https://raw.githubusercontent.com/danielmiessler/SecLists/master/Passwords/Common-Credentials/10-million-password-list-top-1000000.txt"
    "https://raw.githubusercontent.com/danielmiessler/SecLists/master/Passwords/Common-Credentials/10k-most-common.txt"
    "https://raw.githubusercontent.com/danielmiessler/SecLists/master/Passwords/darkweb2017-top10000.txt"
    "https://raw.githubusercontent.com/danielmiessler/SecLists/master/Passwords/xato-net-10-million-passwords-1000000.txt"
    "https://raw.githubusercontent.com/berzerk0/Probable-Wordlists/master/Real-Passwords/Top12Thousand-probable-v2.txt"
)

echo "Step 1: Downloading wordlists..."
for url in "${WORDLIST_SOURCES[@]}"; do
    filename=$(basename "$url")
    if [ ! -f "$TEMP_DIR/$filename" ]; then
        download_wordlist "$url" "$filename"
    else
        echo "Already exists: $filename"
    fi
done

echo ""
echo "Step 2: Cleaning and processing wordlists..."
CLEANED_FILES=()
for file in "$TEMP_DIR"/*.txt; do
    if [ -f "$file" ]; then
        clean_filename="clean_$(basename "$file")"
        clean_wordlist "$file" "$WORDLIST_DIR/$clean_filename"
        CLEANED_FILES+=("$WORDLIST_DIR/$clean_filename")
    fi
done

echo ""
echo "Step 3: Creating master wordlist..."
MASTER_WORDLIST="$WORDLIST_DIR/master_wordlist.txt"

# Combine all cleaned wordlists
cat "${CLEANED_FILES[@]}" | sort -u > "$MASTER_WORDLIST.tmp"

# Remove passwords that are too short or too long (WPA/WPA2 typically 8-63 chars)
echo "Filtering by length (8-63 characters)..."
grep -E '^.{8,63}$' "$MASTER_WORDLIST.tmp" > "$MASTER_WORDLIST"

# Clean up
rm -f "$MASTER_WORDLIST.tmp"

echo ""
echo "Step 4: Generating custom wordlists with Crunch..."
if command -v crunch &> /dev/null; then
    # Generate common pattern wordlists
    echo "Generating common patterns..."
    
    # 8-12 character lowercase
    crunch 8 12 -t @@@@@@@@ -o "$WORDLIST_DIR/crunch_lower_8-12.txt" 2>/dev/null || true
    
    # Common substitutions (leet speak)
    echo "password" > "$TEMP_DIR/base.txt"
    echo "admin" >> "$TEMP_DIR/base.txt"
    echo "welcome" >> "$TEMP_DIR/base.txt"
    echo "123456" >> "$TEMP_DIR/base.txt"
    
    # Create rule file for common substitutions
    cat > "$TEMP_DIR/leet.rule" << 'EOF'
:
l
u
c
s$ 
sa@ 
so0 
si1 
se3 
sa4 
sh5 
sg6 
st7 
sb8 
sg9
EOF
    
    # Apply rules to base words
    if command -v hashcat &> /dev/null; then
        hashcat --stdout "$TEMP_DIR/base.txt" -r "$TEMP_DIR/leet.rule" > "$WORDLIST_DIR/leet_variations.txt" 2>/dev/null || true
    fi
else
    echo "Crunch not installed. Skipping pattern generation."
fi

echo ""
echo "Step 5: Final statistics..."
TOTAL_LINES=$(wc -l < "$MASTER_WORDLIST")
TOTAL_SIZE=$(du -h "$MASTER_WORDLIST" | cut -f1)

echo "=========================================="
echo "Wordlist Creation Complete!"
echo "=========================================="
echo "Master wordlist: $MASTER_WORDLIST"
echo "Total entries: $TOTAL_LINES"
echo "File size: $TOTAL_SIZE"
echo ""
echo "Additional wordlists available in: $WORDLIST_DIR/"
ls -la "$WORDLIST_DIR"/*.txt | head -10
echo ""
echo "Place your CAP/PCAP files in: $WORKSPACE_DIR/input/"
echo "Then run the cracking script to begin."