//Author:TruthHun
//Email:TruthHun@QQ.COM
//Date:2018-01-21
package converter
import (
"fmt"
"io/ioutil"
"os"
"path/filepath"
"strings"
"time"
"os/exec"
"errors"
"github.com/lifei6671/mindoc/utils/filetil"
"github.com/lifei6671/mindoc/utils/ziptil"
"github.com/lifei6671/mindoc/utils/cryptil"
)
type Converter struct {
BasePath string
OutputPath string
Config Config
Debug bool
GeneratedCover string
}
//目录结构
type Toc struct {
Id int `json:"id"`
Link string `json:"link"`
Pid int `json:"pid"`
Title string `json:"title"`
}
//config.json文件解析结构
type Config struct {
Charset string `json:"charset"` //字符编码,默认utf-8编码
Cover string `json:"cover"` //封面图片,或者封面html文件
Timestamp string `json:"date"` //时间日期,如“2018-01-01 12:12:21”,其实是time.Time格式,但是直接用string就好
Description string `json:"description"` //摘要
Footer string `json:"footer"` //pdf的footer
Header string `json:"header"` //pdf的header
Identifier string `json:"identifier"` //即uuid,留空即可
Language string `json:"language"` //语言,如zh、en、zh-CN、en-US等
Creator string `json:"creator"` //作者,即author
Publisher string `json:"publisher"` //出版单位
Contributor string `json:"contributor"` //同Publisher
Title string `json:"title"` //文档标题
Format []string `json:"format"` //导出格式,可选值:pdf、epub、mobi
FontSize string `json:"font_size"` //默认的pdf导出字体大小
PaperSize string `json:"paper_size"` //页面大小
MarginLeft string `json:"margin_left"` //PDF文档左边距,写数字即可,默认72pt
MarginRight string `json:"margin_right"` //PDF文档左边距,写数字即可,默认72pt
MarginTop string `json:"margin_top"` //PDF文档左边距,写数字即可,默认72pt
MarginBottom string `json:"margin_bottom"` //PDF文档左边距,写数字即可,默认72pt
More []string `json:"more"` //更多导出选项[PDF导出选项,具体参考:https://manual.calibre-ebook.com/generated/en/ebook-convert.html#pdf-output-options]
Toc []Toc `json:"toc"` //目录
///////////////////////////////////////////
Order []string `json:"-"` //这个不需要赋值
}
var (
output = "output" //文档导出文件夹
ebookConvert = "ebook-convert"
)
// 接口文档 https://manual.calibre-ebook.com/generated/en/ebook-convert.html#table-of-contents
//根据json配置文件,创建文档转化对象
func NewConverter(configFile string, debug ...bool) (converter *Converter, err error) {
var (
cfg Config
basepath string
db bool
)
if len(debug) > 0 {
db = debug[0]
}
if cfg, err = parseConfig(configFile); err == nil {
if basepath, err = filepath.Abs(filepath.Dir(configFile)); err == nil {
//设置默认值
if len(cfg.Timestamp) == 0 {
cfg.Timestamp = time.Now().Format("2006-01-02 15:04:05")
}
if len(cfg.Charset) == 0 {
cfg.Charset = "utf-8"
}
converter = &Converter{
Config: cfg,
BasePath: basepath,
Debug: db,
}
}
}
return
}
//执行文档转换
func (this *Converter) Convert() (err error) {
if !this.Debug { //调试模式下不删除生成的文件
defer this.converterDefer() //最后移除创建的多余而文件
}
if err = this.generateMimeType(); err != nil {
return
}
if err = this.generateMetaInfo(); err != nil {
return
}
if err = this.generateTocNcx(); err != nil { //生成目录
return
}
if err = this.generateSummary(); err != nil { //生成文档内目录
return
}
if err = this.generateTitlePage(); err != nil { //生成封面
return
}
if err = this.generateContentOpf(); err != nil { //这个必须是generate*系列方法的最后一个调用
return
}
//将当前文件夹下的所有文件压缩成zip包,然后直接改名成content.epub
f := filepath.Join(this.OutputPath, "content.epub")
os.Remove(f) //如果原文件存在了,则删除;
if err = ziptil.Zip(this.BasePath,f); err == nil {
//创建导出文件夹
os.Mkdir(this.BasePath+"/"+output, os.ModePerm)
if len(this.Config.Format) > 0 {
var errs []string
for _, v := range this.Config.Format {
fmt.Println("convert to " + v)
switch strings.ToLower(v) {
case "epub":
if err = this.convertToEpub(); err != nil {
errs = append(errs, err.Error())
fmt.Println("转换EPUB文档失败:" + err.Error())
}
case "mobi":
if err = this.convertToMobi(); err != nil {
errs = append(errs, err.Error())
fmt.Println("转换MOBI文档失败:" + err.Error())
}
case "pdf":
if err = this.convertToPdf(); err != nil {
fmt.Println("转换PDF文档失败:" + err.Error())
errs = append(errs, err.Error())
}
case "docx":
if err = this.convertToDocx(); err != nil {
fmt.Println("转换WORD文档失败:" + err.Error())
errs = append(errs, err.Error())
}
}
}
if len(errs) > 0 {
err = errors.New(strings.Join(errs, "\n"))
}
} else {
err = this.convertToPdf()
if err != nil {
fmt.Println(err)
}
}
} else {
fmt.Println("压缩目录出错" + err.Error())
}
return
}
//删除生成导出文档而创建的文件
func (this *Converter) converterDefer() {
//删除不必要的文件
os.RemoveAll(filepath.Join(this.BasePath, "META-INF"))
os.RemoveAll(filepath.Join(this.BasePath, "content.epub"))
os.RemoveAll(filepath.Join(this.BasePath, "mimetype"))
os.RemoveAll(filepath.Join(this.BasePath, "toc.ncx"))
os.RemoveAll(filepath.Join(this.BasePath, "content.opf"))
os.RemoveAll(filepath.Join(this.BasePath, "titlepage.xhtml")) //封面图片待优化
os.RemoveAll(filepath.Join(this.BasePath, "summary.html")) //文档目录
}
//生成metainfo
func (this *Converter) generateMetaInfo() (err error) {
xml := `
`
folder := filepath.Join(this.BasePath, "META-INF")
os.MkdirAll(folder, os.ModePerm)
err = ioutil.WriteFile(filepath.Join(folder, "container.xml"), []byte(xml), os.ModePerm)
return
}
//形成mimetyppe
func (this *Converter) generateMimeType() (err error) {
return ioutil.WriteFile(filepath.Join(this.BasePath, "mimetype"), []byte("application/epub+zip"), os.ModePerm)
}
//生成封面
func (this *Converter) generateTitlePage() (err error) {
if ext := strings.ToLower(filepath.Ext(this.Config.Cover)); !(ext == ".html" || ext == ".xhtml") {
xml := `
Cover
`
if err = ioutil.WriteFile(filepath.Join(this.BasePath, "titlepage.xhtml"), []byte(xml), os.ModePerm); err == nil {
this.GeneratedCover = "titlepage.xhtml"
}
}
return
}
//生成文档目录
func (this *Converter) generateTocNcx() (err error) {
ncx := `
%v
%v
`
codes, _ := this.tocToXml(0, 1)
ncx = fmt.Sprintf(ncx, this.Config.Language, this.Config.Title, strings.Join(codes, ""))
return ioutil.WriteFile(filepath.Join(this.BasePath, "toc.ncx"), []byte(ncx), os.ModePerm)
}
//生成文档目录,即summary.html
func (this *Converter) generateSummary() (err error) {
//目录
summary := `
目录
目 录
%v
`
summary = fmt.Sprintf(summary, strings.Join(this.tocToSummary(0), ""))
return ioutil.WriteFile(filepath.Join(this.BasePath, "summary.html"), []byte(summary), os.ModePerm)
}
//将toc转成toc.ncx文件
func (this *Converter) tocToXml(pid, idx int) (codes []string, next_idx int) {
var code string
for _, toc := range this.Config.Toc {
if toc.Pid == pid {
code, idx = this.getNavPoint(toc, idx)
codes = append(codes, code)
for _, item := range this.Config.Toc {
if item.Pid == toc.Id {
code, idx = this.getNavPoint(item, idx)
codes = append(codes, code)
var code_arr []string
code_arr, idx = this.tocToXml(item.Id, idx)
codes = append(codes, code_arr...)
codes = append(codes, ``)
}
}
codes = append(codes, ``)
}
}
next_idx = idx
return
}
//将toc转成toc.ncx文件
func (this *Converter) tocToSummary(pid int) (summarys []string) {
summarys = append(summarys, "")
for _, toc := range this.Config.Toc {
if toc.Pid == pid {
summarys = append(summarys, fmt.Sprintf(`- %v
`, toc.Link, toc.Title))
for _, item := range this.Config.Toc {
if item.Pid == toc.Id {
summarys = append(summarys, fmt.Sprintf(`- %v
`, item.Link, item.Title))
summarys = append(summarys, "- ")
summarys = append(summarys, this.tocToSummary(item.Id)...)
summarys = append(summarys, "
")
}
}
}
}
summarys = append(summarys, "
")
return
}
//生成navPoint
func (this *Converter) getNavPoint(toc Toc, idx int) (navpoint string, nextidx int) {
navpoint = `
%v
`
navpoint = fmt.Sprintf(navpoint, toc.Id, idx, toc.Title, toc.Link)
this.Config.Order = append(this.Config.Order, toc.Link)
nextidx = idx + 1
return
}
//生成content.opf文件
//倒数第二步调用
func (this *Converter) generateContentOpf() (err error) {
var (
guide string
manifest string
manifestArr []string
spine string //注意:如果存在封面,则需要把封面放在第一个位置
spineArr []string
)
meta := `%v
%v
%v
%v
%v
%v
`
meta = fmt.Sprintf(meta, this.Config.Title, this.Config.Contributor, this.Config.Publisher, this.Config.Description, this.Config.Language, this.Config.Creator, this.Config.Timestamp)
if len(this.Config.Cover) > 0 {
meta = meta + ``
guide = ``
manifest = fmt.Sprintf(` `, this.Config.Cover, GetMediaType(filepath.Ext(this.Config.Cover)))
spineArr = append(spineArr, ``)
}
if _, err := os.Stat(this.BasePath + "/summary.html"); err == nil {
spineArr = append(spineArr, ``) //目录
}
//扫描所有文件
if files, err := filetil.ScanFiles(this.BasePath); err == nil {
basePath := strings.Replace(this.BasePath, "\\", "/", -1)
for _, file := range files {
if !file.IsDir {
ext := strings.ToLower(filepath.Ext(file.Path))
sourcefile := strings.TrimPrefix(file.Path, basePath+"/")
id := "ncx"
if ext != ".ncx" {
if file.Name == "titlepage.xhtml" { //封面
id = "titlepage"
} else if file.Name == "summary.html" { //目录
id = "summary"
} else {
id = cryptil.Md5Crypt(sourcefile)
}
}
if mt := GetMediaType(ext); mt != "" { //不是封面图片,且media-type不为空
if sourcefile != strings.TrimLeft(this.Config.Cover, "./") { //不是封面图片,则追加进来。封面图片前面已经追加进来了
manifestArr = append(manifestArr, fmt.Sprintf(` `, sourcefile, id, mt))
}
}
} else {
fmt.Println(file.Path)
}
}
items := make(map[string]string)
for _, link := range this.Config.Order {
id := cryptil.Md5Crypt(link)
if _, ok := items[id]; !ok { //去重
items[id] = id
spineArr = append(spineArr, fmt.Sprintf(``, id))
}
}
manifest = manifest + strings.Join(manifestArr, "\n")
spine = strings.Join(spineArr, "\n")
} else {
return err
}
pkg := `
%v
%v
%v
%v
`
if len(guide) > 0 {
guide = `` + guide + ``
}
pkg = fmt.Sprintf(pkg, meta, manifest, spine, guide)
return ioutil.WriteFile(filepath.Join(this.BasePath, "content.opf"), []byte(pkg), os.ModePerm)
}
//转成epub
func (this *Converter) convertToEpub() (err error) {
args := []string{
filepath.Join(this.OutputPath, "content.epub"),
filepath.Join(this.OutputPath, output, "book.epub"),
}
cmd := exec.Command(ebookConvert, args...)
if this.Debug {
fmt.Println(cmd.Args)
}
return cmd.Run()
//return filetil.CopyFile(filepath.Join(this.OutputPath, "content.epub"),filepath.Join(this.OutputPath, output, "book.epub"))
}
//转成mobi
func (this *Converter) convertToMobi() (err error) {
args := []string{
filepath.Join(this.OutputPath, "content.epub"),
filepath.Join(this.OutputPath, output, "book.mobi"),
}
cmd := exec.Command(ebookConvert, args...)
if this.Debug {
fmt.Println(cmd.Args)
}
return cmd.Run()
}
//转成pdf
func (this *Converter) convertToPdf() (err error) {
args := []string{
filepath.Join(this.OutputPath, "content.epub"),
filepath.Join(this.OutputPath, output, "book.pdf"),
}
//页面大小
if len(this.Config.PaperSize) > 0 {
args = append(args, "--paper-size", this.Config.PaperSize)
}
//文字大小
if len(this.Config.FontSize) > 0 {
args = append(args, "--pdf-default-font-size", this.Config.FontSize)
}
//header template
if len(this.Config.Header) > 0 {
args = append(args, "--pdf-header-template", this.Config.Header)
}
//footer template
if len(this.Config.Footer) > 0 {
args = append(args, "--pdf-footer-template",this.Config.Footer)
}
if strings.Count(this.Config.MarginLeft,"") > 0 {
args = append(args, "--pdf-page-margin-left", this.Config.MarginLeft)
}
if strings.Count(this.Config.MarginTop,"") > 0 {
args = append(args, "--pdf-page-margin-top", this.Config.MarginTop)
}
if strings.Count(this.Config.MarginRight,"") > 0 {
args = append(args, "--pdf-page-margin-right", this.Config.MarginRight)
}
if strings.Count(this.Config.MarginBottom,"") > 0 {
args = append(args, "--pdf-page-margin-bottom", this.Config.MarginBottom)
}
//更多选项
if len(this.Config.More) > 0 {
args = append(args, this.Config.More...)
}
cmd := exec.Command(ebookConvert, args...)
if this.Debug {
fmt.Println(cmd.Args)
}
return cmd.Run()
}
// 转成word
func (this *Converter) convertToDocx() (err error) {
args := []string{
filepath.Join(this.OutputPath , "content.epub"),
filepath.Join(this.OutputPath , output , "book.docx"),
}
args = append(args, "--docx-no-toc")
//页面大小
if len(this.Config.PaperSize) > 0 {
args = append(args, "--docx-page-size", this.Config.PaperSize)
}
if len(this.Config.MarginLeft) > 0 {
args = append(args, "--docx-page-margin-left", this.Config.MarginLeft)
}
if len(this.Config.MarginTop) > 0 {
args = append(args, "--docx-page-margin-top", this.Config.MarginTop)
}
if len(this.Config.MarginRight) > 0 {
args = append(args, "--docx-page-margin-right", this.Config.MarginRight)
}
if len(this.Config.MarginBottom) > 0 {
args = append(args, "--docx-page-margin-bottom", this.Config.MarginBottom)
}
cmd := exec.Command(ebookConvert, args...)
if this.Debug {
fmt.Println(cmd.Args)
}
return cmd.Run()
}