isbn_extract.go

package main

import (
	"archive/zip"
	"fmt"
	"io/ioutil"
	"os"
	"path/filepath"
	"regexp"
	"strings"
)

func main() {
	rc, err := zip.OpenReader(os.Args[1])
	if err != nil {
		panic(err)
	}
	var isbn []string
	seen := map[string]bool{}
	for _, f := range rc.File {
		if e := findISBN13(f.Name); len(e) == 1 {
			isbn = append(isbn, e[0])
			seen[e[0]] = true
		}
		if ext := filepath.Ext(f.Name); ext != ".html" && ext != ".xhtml" {
			continue
		}
		r, err := f.Open()
		if err != nil {
			panic(err)
		}
		buf, err := ioutil.ReadAll(r)
		if err != nil {
			panic(err)
		}
		for _, v := range findISBN13(string(buf)) {
			if !seen[v] {
				isbn = append(isbn, v)
				seen[v] = true
			}
		}
	}
	fmt.Println(isbn)
}

var isbnRe = regexp.MustCompile(`9(?:[ -]?)(?:78|79)(?:[ -]?[0-9]){10}`)

func findISBN13(txt string) []string {
	isbn := isbnRe.FindAllString(txt, -1)
	r := strings.NewReplacer(" ", "", "-", "")
	for i := range isbn {
		isbn[i] = r.Replace(isbn[i])
	}
	return isbn
}