I've been migrating my blog from Blogger as written in Migrating my blog and implementing with VuePress.
There had been no problem for writing new posts, but when it comes to migrate old posts, I had an issue, and eventually I switched from VuePress to nuxt/content.
Why do I keep changing my blog for a long time, though I don't have much passion for studying frontend? I don’t even know the answer.
But I could have migrated it finally, so I'd like to look back.
Actual contents might be still in StackEdit, but some posts are lost due to the updates of StackEdit, or I've been writing old posts on Blogger directly, so I decided to create Markdown posts based on the HTML contents on Blogger.
Blogger posts can be downloaded with XML format as a backup.
I excpected that I could download a big XML file and it could be manipulated by using xmllint, but it didn't work. I couldn't fix garbled characters.
Even if I could make it work fortunately, I should also have to add frontmatter and make directory structure based on creation dates, which must be hard to handle only by CLI. Therefore I wrote a small tool by Golang.
I could create directories and wrote README.md, but as I thought it was difficult to format contents.
If it is output as is, many errors are caused around <pre> tags and the post couldn't be shown.
Eventually, I used html-to-markdown to convert posts.
With this tool I converted HTML to Markdown, but this also had an issue around code blocks, which made all newlines lost and became 1 line.
I couldn't make code blocks converted nicely either way, so finally I converted posts with html-to-markdown and checked all posts one by one, and if I saw a wierd part I copied the original code from Blogger editor screen.
package main
import (
"encoding/xml"
"flag"
"fmt"
"io/ioutil"
"os"
"path/filepath"
"strings"
"time"
md "github.com/JohannesKaufmann/html-to-markdown"
)
var (
converter *md.Converter
)
func init() {
converter = md.NewConverter("", true, nil)
}
type Feed struct {
XMLName xml.Name `xml:"feed"`
EntryList []Entry `xml:"entry"`
}
type Entry struct {
XMLName xml.Name `xml:"entry"`
Title string `xml:"title"`
Published string `xml:"published"`
CategoryList []Category `xml:"category"`
Content string `xml:"content"`
LinkList []Link `xml:"link"`
}
type Category struct {
Scheme string `xml:"scheme,attr"`
Term string `xml:"term,attr"`
}
type Link struct {
Rel string `xml:"rel,attr"`
Href string `xml:"href,attr"`
}
func main() {
var filename string
flag.StringVar(&filename, "f", "", "target file name")
flag.Parse()
file, err := os.Open(filename)
if err != nil {
panic(err)
}
defer file.Close()
data, err := ioutil.ReadAll(file)
if err != nil {
panic(err)
}
feed := Feed{}
err = xml.Unmarshal(data, &feed)
if err != nil {
panic(err)
}
for _, v := range feed.EntryList {
if !v.IsPost() {
continue
}
convert(&v)
}
}
func (e *Entry) IsPost() bool {
for _, c := range e.CategoryList {
if c.Scheme == "http://schemas.google.com/g/2005#kind" &&
c.Term == "http://schemas.google.com/blogger/2008/kind#post" {
return true
}
}
return false
}
func (e *Entry) GetPagePath() string {
for _, l := range e.LinkList {
if l.Rel == "alternate" {
arr := strings.Split(l.Href, "/")
file := arr[len(arr) - 1]
ext := filepath.Ext(file)
return file[0:len(file)-len(ext)]
}
}
panic("valid link not found")
}
func (e *Entry) GetTags() string {
var tags []string
for _, c := range e.CategoryList {
if c.Scheme == "http://www.blogger.com/atom/ns#" {
tags = append(tags, fmt.Sprintf("\"%s\"", c.Term))
}
}
return strings.Join(tags, ",")
}
func (e *Entry) GetContent() string {
markdown, err := converter.ConvertString(e.Content)
if err != nil {
return e.Content
}
return markdown
}
func convert(v *Entry) {
t, _ := time.Parse("2006-01-02T15:04:05.000-07:00", v.Published)
y := t.Local().Year()
m := t.Local().Month()
pathJa := fmt.Sprintf("../../docs/ja/post/%04d/%02d/%s", y, m, v.GetPagePath())
os.MkdirAll(pathJa, os.ModePerm)
content := fmt.Sprintf(`---
title: "%s"
created: %s
tags: [%s]
---
%s
`,
v.Title,
v.Published,
v.GetTags(),
v.GetContent(),
)
readmePath := fmt.Sprintf("%s/README.md", pathJa)
ioutil.WriteFile(readmePath, []byte(content), 0644)
}
Saved this code to a file like tools/converter/main.go, and also located the downloaded XML file as blogger.xml in the same directory.
Then defined a command like below in the package.json
"scripts": {
"blogger:convert": "cd tools/converter; go run main.go -f blogger.xml"
},
and convert posts by running npm run blogger:convert.
With this, I could get Markdown files like docs/ja/post/2010/05/android-project-path/README.md (these files are still for VuePress).
In addition to that, I manually did the following works steadily.
At last I finished this updates and run build (not in dev mode), saw an error in notations like {{ site.github.url }}.
It also occurred in th:text="${{entity.status}}". I thought I managed to avoid this by replacing {{ to {\{...but with this modification the raw text is shown. ${{variable}} notation didn't work neither. Once again I tried to run it on dev mode and there were certainly errors in that page.
vuepressのテンプレート内で二重波括弧をエスケープ
This was the solution.
It seems that inline code should be escaped by using <code v-pre>.
Continues to Migrating blog to nuxt/content - 2. generate description automatically.