code-snatch

#!/bin/bash

# code-snatch - Semantic code search CLI
# Author: manusiele254@gmail.com

set -e

# Configuration
GITHUB_USER="manusiele"
REPO_NAME="semantic-code-index"
DB_URL="https://raw.githubusercontent.com/${GITHUB_USER}/${REPO_NAME}/main/data/snippets_db.json"
CACHE_DIR="$HOME/.semantic-code-cli"
CACHE_FILE="$CACHE_DIR/snippets_db.json"
CACHE_AGE_HOURS=24

# Colors
RED='\033[0;31m'
GREEN='\033[0;32m'
YELLOW='\033[1;33m'
BLUE='\033[0;34m'
CYAN='\033[0;36m'
BOLD='\033[1m'
NC='\033[0m'

# Helper functions
error() {
    echo -e "${RED}❌ Error: $1${NC}" >&2
    exit 1
}

success() {
    echo -e "${GREEN}✓ $1${NC}"
}

info() {
    echo -e "${CYAN}ℹ $1${NC}"
}

warning() {
    echo -e "${YELLOW}⚠ $1${NC}"
}

print_header() {
    echo -e "${BOLD}${BLUE}━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━${NC}"
    echo -e "${BOLD}${BLUE}$1${NC}"
    echo -e "${BOLD}${BLUE}━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━${NC}"
}

# Check dependencies
check_dependencies() {
    local missing=()
    
    for cmd in curl jq; do
        if ! command -v $cmd &> /dev/null; then
            missing+=($cmd)
        fi
    done
    
    if [ ${#missing[@]} -gt 0 ]; then
        error "Missing dependencies: ${missing[*]}\n  Install: sudo apt install ${missing[*]} (Ubuntu) or brew install ${missing[*]} (macOS)"
    fi
}

# Check if cache is fresh
is_cache_fresh() {
    if [ ! -f "$CACHE_FILE" ]; then
        return 1
    fi
    
    if [[ "$OSTYPE" == "darwin"* ]]; then
        cache_time=$(stat -f %m "$CACHE_FILE")
    else
        cache_time=$(stat -c %Y "$CACHE_FILE")
    fi
    
    current_time=$(date +%s)
    age=$((current_time - cache_time))
    max_age=$((CACHE_AGE_HOURS * 3600))
    
    [ $age -lt $max_age ]
}

# Download database
download_database() {
    info "Downloading database from GitHub..."
    echo "  Source: $DB_URL"
    
    mkdir -p "$CACHE_DIR"
    
    if curl -sS -f -L "$DB_URL" -o "$CACHE_FILE.tmp"; then
        mv "$CACHE_FILE.tmp" "$CACHE_FILE"
        success "Database downloaded and cached"
        
        local count=$(jq '. | length' "$CACHE_FILE" 2>/dev/null || echo "unknown")
        echo "  📦 Total snippets: $count"
        return 0
    else
        rm -f "$CACHE_FILE.tmp"
        error "Failed to download database. Check your internet connection and repo setup."
        return 1
    fi
}

# Update cache
update_cache() {
    local force=$1
    
    if [ "$force" = "true" ] || ! is_cache_fresh; then
        download_database
    else
        success "Using cached database (fresh)"
    fi
}

# Search snippets
search_snippets() {
    local query="$1"
    
    if [ ! -f "$CACHE_FILE" ]; then
        error "No cache found. Run: code-snatch --update"
    fi
    
    info "Searching for: \"$query\""
    echo ""
    
    local results=$(jq -c --arg query "$query" '
        [
            .[] | 
            select(
                (.docstring // "" | ascii_downcase | contains($query | ascii_downcase)) or
                (.function_name // "" | ascii_downcase | contains($query | ascii_downcase))
            )
        ] | .[0:5]
    ' "$CACHE_FILE")
    
    local count=$(echo "$results" | jq '. | length')
    
    if [ "$count" -eq 0 ]; then
        warning "No results found for \"$query\""
        echo ""
        echo "This might be because:"
        echo "  • Query is about a non-Python language (only Python is indexed)"
        echo "  • Query is too specific (try simpler keywords)"
        echo "  • Topic not in indexed repos (Django, Flask, FastAPI, Requests, CPython)"
        echo ""
        echo "Try simpler keywords like: 'async file', 'error handling', 'flask route'"
        exit 0
    fi
    
    # Display results
    echo "$results" | jq -r 'to_entries[] | 
        "\n\u001b[1;32m━━━ Match \(.key + 1) ━━━\u001b[0m",
        "\u001b[1mFunction:\u001b[0m \(.value.function_name)",
        "\u001b[1mSource:\u001b[0m \(.value.source_repo) - \(.value.filepath)",
        "",
        "\u001b[1mDescription:\u001b[0m",
        "  \(.value.docstring[:200])...",
        "",
        "\u001b[1mCode:\u001b[0m",
        "\(.value.code)",
        "",
        "\u001b[2m🔗 \(.value.source_url // "N/A")\u001b[0m",
        (if .value.llm_verified then "\u001b[32m✓ LLM Verified\u001b[0m (Quality: \(.value.llm_analysis.quality_score)/10)" else "" end),
        ""
    '
}

# Show cache info
show_info() {
    print_header "Cache Information"
    
    if [ -f "$CACHE_FILE" ]; then
        local count=$(jq '. | length' "$CACHE_FILE" 2>/dev/null || echo "error")
        local size=$(du -h "$CACHE_FILE" 2>/dev/null | cut -f1 || echo "unknown")
        
        if [[ "$OSTYPE" == "darwin"* ]]; then
            local modified=$(stat -f "%Sm" -t "%Y-%m-%d %H:%M:%S" "$CACHE_FILE" 2>/dev/null || echo "unknown")
        else
            local modified=$(stat -c "%y" "$CACHE_FILE" 2>/dev/null | cut -d'.' -f1 || echo "unknown")
        fi
        
        echo "  📁 Location: $CACHE_FILE"
        echo "  📦 Snippets: $count"
        echo "  💾 Size: $size"
        echo "  🕐 Updated: $modified"
        
        if is_cache_fresh; then
            success "Cache is fresh"
        else
            warning "Cache is stale (older than $CACHE_AGE_HOURS hours)"
            echo "  Run: code-snatch --update"
        fi
    else
        warning "No cache found"
        echo "  Run: code-snatch --update"
    fi
    
    echo ""
}

# Clear cache
clear_cache() {
    if [ -f "$CACHE_FILE" ]; then
        rm "$CACHE_FILE"
        success "Cache cleared"
    else
        warning "No cache to clear"
    fi
}

# Show help
show_help() {
    echo -e "${BOLD}code-snatch${NC} - Python Semantic Code Search"
    echo -e ""
    echo -e "${BOLD}SCOPE:${NC}"
    echo -e "    🐍 Python code only (Django, Flask, FastAPI, Requests, stdlib)"
    echo -e "    ❌ Other languages not supported"
    echo -e ""
    echo -e "${BOLD}USAGE:${NC}"
    echo -e "    code-snatch [OPTIONS] \"python code query\""
    echo -e ""
    echo -e "${BOLD}OPTIONS:${NC}"
    echo -e "    -u, --update        Force update the database from GitHub"
    echo -e "    -i, --info          Show cache information"
    echo -e "    -c, --clear         Clear local cache"
    echo -e "    -h, --help          Show this help message"
    echo -e ""
    echo -e "${BOLD}EXAMPLES:${NC}"
    echo -e "    code-snatch \"flask route with error handling\""
    echo -e "    code-snatch \"async database query\""
    echo -e "    code-snatch \"decorator for timing functions\""
    echo -e "    code-snatch --update"
    echo -e ""
    echo -e "${BOLD}INDEXED REPOSITORIES:${NC}"
    echo -e "    • django/django        (Web framework)"
    echo -e "    • psf/requests         (HTTP library)"
    echo -e "    • pallets/flask        (Micro framework)"
    echo -e "    • fastapi/fastapi      (Async framework)"
    echo -e "    • python/cpython       (Python stdlib)"
    echo -e ""
    echo -e "${BOLD}CONFIGURATION:${NC}"
    echo -e "    Database: https://github.com/${GITHUB_USER}/${REPO_NAME}"
    echo -e "    Cache: ${CACHE_DIR}"
    echo -e ""
    echo -e "${BOLD}REQUIREMENTS:${NC}"
    echo -e "    • curl (for downloading database)"
    echo -e "    • jq (for JSON processing)"
    echo -e ""
    echo -e "${BOLD}TIPS:${NC}"
    echo -e "    • Database updates automatically every 24 hours"
    echo -e "    • Use simple, descriptive keywords"
    echo -e "    • Results are ranked by relevance"
}

# Main function
main() {
    check_dependencies
    
    case "${1:-}" in
        -h|--help)
            show_help
            exit 0
            ;;
        -u|--update)
            update_cache true
            exit 0
            ;;
        -i|--info)
            show_info
            exit 0
            ;;
        -c|--clear)
            clear_cache
            exit 0
            ;;
        "")
            error "No search query provided.\n  Usage: code-snatch \"your query\"\n  Help: code-snatch --help"
            ;;
        -*)
            error "Unknown option: $1\n  Use --help for usage information"
            ;;
        *)
            update_cache false
            search_snippets "$*"
            ;;
    esac
}

main "$@"