refacto get command, fix: images option

This commit is contained in:
lapwat
2022-01-02 02:16:45 +01:00
parent 29008185a8
commit 5e735f9c52
7 changed files with 103 additions and 201 deletions

View File

@@ -5,18 +5,14 @@ import (
"fmt"
"log"
"os"
"os/exec"
"strings"
md "github.com/JohannesKaufmann/html-to-markdown"
"github.com/PuerkitoBio/goquery"
epub "github.com/bmaupin/go-epub"
"github.com/spf13/cobra"
"github.com/lapwat/papeer/book"
)
var recursive, include, images bool
var recursive, include, images, quiet bool
var format, output, selector, name, author string
var limit, offset, delay, threads int
@@ -79,159 +75,46 @@ var getCmd = &cobra.Command{
url := args[0]
b := book.NewBookFromURL(url, selector, name, author, recursive, include, images, limit, offset, delay, threads)
if len(output) == 0 {
// set default output
output = strings.ReplaceAll(b.Name(), " ", "_")
output = strings.ReplaceAll(output, "/", "")
output = fmt.Sprintf("%s.%s", output, format)
}
fakeConfig := book.NewScrapeConfigFake()
fakeChapter := book.NewChapter("", b.Name(), b.Author(), "", b.Chapters(), fakeConfig)
if format == "stdout" {
for _, c := range b.Chapters() {
// convert to markdown
content, err := md.NewConverter("", true, nil).ConvertString(c.Content())
if err != nil {
log.Fatal(err)
}
text := fmt.Sprintf("%s\n%s\n\n%s\n\n\n", c.Name(), strings.Repeat("=", len(c.Name())), content)
// write to stdout
fmt.Println(text)
}
// TODO: ToMarkdownString
markdown := book.ToMarkdown(fakeChapter)
fmt.Println(markdown)
}
if format == "md" {
// TODO: ToMarkdownFile
markdown := book.ToMarkdown(fakeChapter)
// create markdown file
if len(output) == 0 {
filename := book.Filename(fakeChapter.Name())
output = fmt.Sprintf("%s.md", filename)
}
// write to file
f, err := os.Create(output)
if err != nil {
log.Fatal(err)
}
defer f.Close()
for _, c := range b.Chapters() {
// convert to markdown
content, err := md.NewConverter("", true, nil).ConvertString(c.Content())
if err != nil {
log.Fatal(err)
}
text := fmt.Sprintf("%s\n%s\n\n%s\n\n\n", c.Name(), strings.Repeat("=", len(c.Name())), content)
// write to markdown file
_, err = f.WriteString(text)
if err != nil {
log.Fatal(err)
}
_, err2 := f.WriteString(markdown)
if err2 != nil {
log.Fatal(err2)
}
f.Close()
fmt.Printf("Markdown saved to \"%s\"\n", output)
}
if format == "epub" {
e := epub.NewEpub(b.Name())
e.SetAuthor(b.Author())
for _, c := range b.Chapters() {
var content string
if images == false {
content = c.Content()
}
// parse content
doc, err := goquery.NewDocumentFromReader(strings.NewReader(c.Content()))
if err != nil {
log.Fatal(err)
}
// retrieve images and download it
doc.Find("img").Each(func(i int, s *goquery.Selection) {
src, _ := s.Attr("src")
imagePath, _ := e.AddImage(src, "")
if images {
imageTag, _ := goquery.OuterHtml(s)
content += imageTag
}
content = strings.ReplaceAll(content, src, imagePath)
})
html := fmt.Sprintf("<h1>%s</h1>%s", c.Name(), content)
_, err = e.AddSection(html, c.Name(), "", "")
if err != nil {
log.Fatal(err)
}
}
err := e.Write(output)
if err != nil {
log.Fatal(err)
}
output = book.ToEpub(fakeChapter, output)
fmt.Printf("Ebook saved to \"%s\"\n", output)
}
if format == "mobi" {
e := epub.NewEpub(b.Name())
e.SetAuthor(b.Author())
for _, c := range b.Chapters() {
var content string
if images == false {
content = c.Content()
}
// parse content
doc, err := goquery.NewDocumentFromReader(strings.NewReader(c.Content()))
if err != nil {
log.Fatal(err)
}
// retrieve images and download it
doc.Find("img").Each(func(i int, s *goquery.Selection) {
src, _ := s.Attr("src")
imagePath, _ := e.AddImage(src, "")
if images {
imageTag, _ := goquery.OuterHtml(s)
content += imageTag
}
content = strings.ReplaceAll(content, src, imagePath)
})
html := fmt.Sprintf("<h1>%s</h1>%s", c.Name(), content)
_, err = e.AddSection(html, c.Name(), "", "")
if err != nil {
log.Fatal(err)
}
}
outputEPUB := strings.ReplaceAll(output, ".mobi", ".epub")
err := e.Write(outputEPUB)
if err != nil {
log.Fatal(err)
}
exec.Command("kindlegen", outputEPUB).Run()
// exec command always return status 1 even if it succeed
// if err != nil {
// log.Fatal(err)
// }
output = book.ToMobi(fakeChapter, output)
fmt.Printf("Ebook saved to \"%s\"\n", output)
err = os.Remove(outputEPUB)
if err != nil {
log.Fatal(err)
}
}
},
}