package main

import (
	"bufio"
	"fmt"
	"os"
	"regexp"
	"sort"
	"strings"
)

// 詞頻結果結構體
type WordCount struct {
	Word  string
	Count int
}

// 統(tǒng)計詞頻函數(shù)
func countWordFrequency(text string) map[string]int {
	// 正則去掉標點符號，只保留字母和空格
	reg := regexp.MustCompile(`[^a-zA-Z\s]+`)
	cleanText := reg.ReplaceAllString(text, "")

	// 全部轉小寫，避免大小寫重復
	cleanText = strings.ToLower(cleanText)

	// 按空格分割單詞
	words := strings.Fields(cleanText)

	// 使用 map 統(tǒng)計詞頻
	wordFreq := make(map[string]int)
	for _, word := range words {
		wordFreq[word]++
	}
	return wordFreq
}

func main() {
	// 打開文件
	file, err := os.Open("sample.txt")
	if err != nil {
		fmt.Println("打開文件失?。?, err)
		return
	}
	defer file.Close()

	// 逐行讀取文件
	var content strings.Builder
	scanner := bufio.NewScanner(file)
	for scanner.Scan() {
		content.WriteString(scanner.Text())
		content.WriteString(" ")
	}

	// 統(tǒng)計詞頻
	wordFreq := countWordFrequency(content.String())

	// 轉換為切片，方便排序
	var wcList []WordCount
	for word, count := range wordFreq {
		wcList = append(wcList, WordCount{word, count})
	}

	// 按詞頻排序（降序）
	sort.Slice(wcList, func(i, j int) bool {
		return wcList[i].Count > wcList[j].Count
	})

	// 輸出前10個高頻詞
	fmt.Println("詞頻最高的前10個單詞：")
	for i, wc := range wcList {
		if i >= 10 {
			break
		}
		fmt.Printf("%s: %d\n", wc.Word, wc.Count)
	}
}

Go is an open source programming language.
Go makes it easy to build simple, reliable, and efficient software.