Golang实现简单搜索引擎功能工具

func tokenize(text string) []string {
    // 转小写，去除标点，按空格分割
    re := regexp.MustCompile(`[a-zA-Z]+`)
    words := re.FindAllString(strings.ToLower(text), -1)
    return words
}

type Index map[string][]int

func (idx *Index) Add(docID int, content string) {
    words := tokenize(content)
    seen := make(map[string]bool)
    for _, word := range words {
        if !seen[word] {
            (*idx)[word] = append((*idx)[word], docID)
            seen[word] = true
        }
    }
}

func (idx Index) Search(query string) []int {
    words := tokenize(query)
    if len(words) == 0 {
        return nil
    }

    // 获取第一个词的文档列表作为初始结果
    result := make([]int, len(idx[words[0]]))
    copy(result, idx[words[0]])

    // 与其他词的文档列表求交集
    for _, word := range words[1:] {
        result = intersect(result, idx[word])
    }
    return result
}

func intersect(a, b []int) []int {
    i, j := 0, 0
    var res []int
    for i < len(a) && j < len(b) {
        if a[i] == b[j] {
            res = append(res, a[i])
            i++
            j++
        } else if a[i] < b[j] {
            i++
        } else {
            j++
        }
    }
    return res
}

func main() {
    var index Index = make(map[string][]int)

    docs := []string{
        "Go is a great programming language",
        "Search engine in Go is fun",
        "Simple tools work well",
    }

    // 建立索引
    for i, doc := range docs {
        index.Add(i, doc)
    }

    // 搜索
    query := "go search"
    results := index.Search(query)

    fmt.Printf("Matched documents: %v\n", results)
    for _, id := range results {
        fmt.Printf("Doc[%d]: %s\n", id, docs[id])
    }
}

Matched documents: [1]
Doc[1]: Search engine in Go is fun