main.go

package main

import (
	"fmt"
	"log"
	"net/url"
	"os"
	"sort"
	"strconv"
	"time"
)

// PageCount represents a URL and its link count
type PageCount struct {
	URL   string
	Count int
}

func main() {
	if len(os.Args) < 4 {
		fmt.Println("Usage: ./crawler <baseURL> <maxConcurrency> <maxPages>")
		os.Exit(1)
	}

	baseURL, err := url.Parse(os.Args[1])
	if err != nil {
		log.Fatalf("Invalid base URL %q: %v\n", os.Args[1], err)
	}

	maxConcurrency, err := strconv.Atoi(os.Args[2])
	if err != nil {
		log.Fatalf("Invalid max concurrency: %v\n", err)
	}

	maxPages, err := strconv.Atoi(os.Args[3])
	if err != nil {
		log.Fatalf("Invalid max pages: %v\n", err)
	}

	startTime := time.Now()

	// Initialize crawler
	crawler := &Crawler{
		pageCounts:         make(map[string]int),
		baseURL:            baseURL,
		concurrencyLimiter: make(chan struct{}, maxConcurrency),
		maxPages:           maxPages,
	}

	// Start crawling
	crawler.wg.Add(1)
	go crawler.crawlPage(baseURL.String())
	crawler.wg.Wait()

	// Convert map to slice for sorting
	var results []PageCount
	for url, count := range crawler.pageCounts {
		results = append(results, PageCount{URL: url, Count: count})
	}

	// Sort by descending link count
	sort.Slice(results, func(i, j int) bool {
		return results[i].Count > results[j].Count
	})

	// Compute total links
	var totalLinks int
	for _, r := range results {
		totalLinks += r.Count
	}

	duration := time.Since(startTime)

	printReport(baseURL.String(), results, totalLinks, duration)
}

// printReport outputs crawl statistics in a formatted way
func printReport(baseURL string, results []PageCount, totalLinks int, duration time.Duration) {
	fmt.Println("╔══════════════════════════════════════════════════════════════════════╗")
	fmt.Println("║                          CRAWL REPORT                                ║")
	fmt.Println("╚══════════════════════════════════════════════════════════════════════╝")

	fmt.Printf("%-30s : %s\n", "Base URL", baseURL)
	fmt.Printf("%-30s : %d\n", "Unique Pages", len(results))
	fmt.Printf("%-30s : %d\n", "Total Links", totalLinks)
	fmt.Printf("%-30s : %s\n", "Crawl Duration", duration.Round(time.Millisecond))
	fmt.Printf("%-30s : %.2f\n", "Pages/Second", float64(len(results))/duration.Seconds())

	if len(results) > 0 {
		fmt.Printf("%-30s : %s (%d links)\n", "Most Linked Page", results[0].URL, results[0].Count)
		fmt.Printf("%-30s : %s (%d links)\n", "Least Linked Page", results[len(results)-1].URL, results[len(results)-1].Count)
	}

	fmt.Println("\nTop 5 Most Linked Pages:")
	for i := 0; i < 5 && i < len(results); i++ {
		fmt.Printf("  %-2d. %-50s - %d links\n", i+1, results[i].URL, results[i].Count)
	}
}