add parser for blog articles

Signed-off-by: Tobias Erbshäußer <tobias@tesoft.dev>
2026-05-24 09:22:16 +02:00
parent 5147b61c9d
commit 333211d4d0
5 changed files with 427 additions and 0 deletions
@@ -0,0 +1,151 @@
+package main
+
+import (
+	"archive/tar"
+	"backend/md"
+	"encoding/json"
+	"errors"
+	"fmt"
+	"io"
+	"log"
+	"net/http"
+	"strconv"
+	"strings"
+	"time"
+)
+
+type ArticleStatus = int
+
+const (
+	ArticleStatusDraft ArticleStatus = iota
+	ArticleStatusPublished
+	ArticleStatusOffline
+)
+
+type ArticleProperties struct {
+	Id               int64         `json:"id"`
+	Title            string        `json:"title"`
+	Status           ArticleStatus `json:"status"`
+	Tags             []string      `json:"tags"`
+	ReleaseDate      time.Time     `json:"date"`
+	ModificationDate *time.Time    `json:"mod-date"`
+}
+
+type Article struct {
+	ArticleProperties
+	Content string
+	Files   []ArticleFile
+}
+
+type ArticleFile struct {
+	Id   int
+	Data []byte
+}
+
+func ParseArticle(reader io.Reader, filePrefix string) (*Article, error) {
+	tarFiles := make(map[string][]byte)
+	tarReader := tar.NewReader(reader)
+	for {
+		header, err := tarReader.Next()
+		if err == io.EOF {
+			break
+		}
+		if err != nil {
+			return nil, err
+		}
+
+		content, err := io.ReadAll(tarReader)
+		if err != nil {
+			return nil, err
+		}
+		tarFiles[header.Name] = content
+	}
+
+	readmeBytes, found := tarFiles["README.md"]
+	if !found {
+		return nil, errors.New("README.md not found")
+	}
+
+	usedFiles := make(map[string]ArticleFile)
+	nextFileId := 0
+
+	pc, html, err := md.Parse(func(path string) (string, error) {
+		content, ok := tarFiles[strings.TrimPrefix(path, "./")]
+		if !ok {
+			return "", errors.New("file '" + path + "' not found")
+		}
+
+		usedFiles[path] = ArticleFile{nextFileId, content}
+		nextFileId++
+
+		return filePrefix + strconv.Itoa(nextFileId), nil
+	}, readmeBytes)
+	if err != nil {
+		return nil, err
+	}
+
+	title, err := md.GetProperty(pc, "title")
+	if err != nil {
+		if errors.Is(err, md.PropertyNotFoundError) {
+			return nil, errors.New("title property not found")
+		}
+
+		return nil, err
+	}
+
+	tagsStr, err := md.GetProperty(pc, "tags")
+	if err != nil && !errors.Is(err, md.PropertyNotFoundError) {
+		return nil, err
+	}
+
+	tags := make([]string, 0)
+	if strings.TrimSpace(tagsStr) != "" {
+		for _, tagStr := range strings.Split(tagsStr, ",") {
+			tags = append(tags, strings.TrimSpace(tagStr))
+		}
+	}
+
+	dateStr, err := md.GetProperty(pc, "date")
+	if err != nil {
+		if errors.Is(err, md.PropertyNotFoundError) {
+			return nil, errors.New("date property not found")
+		}
+
+		return nil, err
+	}
+	releaseDate, err := time.Parse(time.DateOnly, dateStr)
+	if err != nil {
+		return nil, fmt.Errorf("invalid date property '%s': %v", dateStr, err)
+	}
+
+	var modificationDate *time.Time
+	dateStr, err = md.GetProperty(pc, "mod-date")
+	if err == nil {
+		tmp, err := time.Parse(time.DateOnly, dateStr)
+		if err != nil {
+			return nil, fmt.Errorf("invalid mod-date property '%s': %v", dateStr, err)
+		}
+
+		modificationDate = &tmp
+	} else if !errors.Is(err, md.PropertyNotFoundError) {
+		return nil, err
+	}
+
+	files := make([]ArticleFile, 0, len(usedFiles))
+	for _, file := range usedFiles {
+		files = append(files, file)
+	}
+
+	return &Article{
+		ArticleProperties{
+			-1,
+			title,
+			ArticleStatusDraft,
+			tags,
+			releaseDate,
+			modificationDate,
+		},
+		string(html),
+		files,
+	}, nil
+}
@@ -0,0 +1,7 @@
+go 1.25
+
+use (
+	.
+	go-sqlite3
+	goldmark
+)
@@ -0,0 +1,50 @@
+package md
+
+import (
+	"bytes"
+	"errors"
+
+	"github.com/yuin/goldmark"
+	"github.com/yuin/goldmark/extension"
+	"github.com/yuin/goldmark/parser"
+)
+
+var PropertyNotFoundError = errors.New("property not found")
+
+func Parse(transformer UrlTransformer, source []byte) (parser.Context, []byte, error) {
+	md := goldmark.New(
+		goldmark.WithExtensions(
+			extension.Footnote, extension.Strikethrough,
+			extension.Table,
+			&propertiesExtension{},
+			&urlTransformerExtension{transformer},
+		),
+		goldmark.WithRendererOptions(),
+	)
+
+	pc := parser.NewContext()
+	var readme bytes.Buffer
+	err := md.Convert(source, &readme, parser.WithContext(pc))
+	if err != nil {
+		return nil, nil, err
+	}
+
+	return pc, readme.Bytes(), nil
+}
+
+func GetProperty(pc parser.Context, key string) (string, error) {
+	data := pc.Get(propertiesContextKey)
+	if data != nil {
+		pData := data.(*propertiesData)
+		if pData.Error != nil {
+			return "", pData.Error
+		}
+
+		value, ok := pData.properties[key]
+		if ok {
+			return value, nil
+		}
+	}
+
+	return "", PropertyNotFoundError
+}
@@ -0,0 +1,109 @@
+package md
+
+import (
+	"errors"
+	"strings"
+
+	"github.com/yuin/goldmark"
+	"github.com/yuin/goldmark/ast"
+	"github.com/yuin/goldmark/parser"
+	"github.com/yuin/goldmark/text"
+	"github.com/yuin/goldmark/util"
+)
+
+var propertiesContextKey = parser.NewContextKey()
+
+type propertiesData struct {
+	properties map[string]string
+	Error      error
+}
+
+type propertiesExtension struct {
+}
+
+func (e *propertiesExtension) Extend(m goldmark.Markdown) {
+	m.Parser().AddOptions(
+		parser.WithBlockParsers(
+			util.Prioritized(&propertiesParser{}, 0),
+		),
+	)
+}
+
+type propertiesParser struct {
+}
+
+func (p *propertiesParser) Trigger() []byte {
+	return []byte{'-'}
+}
+
+func (p *propertiesParser) Open(parent ast.Node, reader text.Reader, pc parser.Context) (ast.Node, parser.State) {
+	lineNumber, _ := reader.Position()
+	if lineNumber != 0 {
+		return nil, parser.NoChildren
+	}
+
+	line, _ := reader.PeekLine()
+	if isSeparator(line) {
+		return ast.NewTextBlock(), parser.NoChildren
+	}
+
+	return nil, parser.NoChildren
+}
+
+func (p *propertiesParser) Continue(node ast.Node, reader text.Reader, pc parser.Context) parser.State {
+	line, segment := reader.PeekLine()
+	if isSeparator(line) && !util.IsBlank(line) {
+		reader.Advance(segment.Len())
+		return parser.Close
+	}
+
+	node.Lines().Append(segment)
+	return parser.Continue | parser.NoChildren
+}
+
+func (p *propertiesParser) Close(node ast.Node, reader text.Reader, pc parser.Context) {
+	lines := node.Lines()
+
+	data := &propertiesData{
+		make(map[string]string),
+		nil,
+	}
+	pc.Set(propertiesContextKey, data)
+
+	for i := 0; i < lines.Len(); i++ {
+		segment := lines.At(i)
+		line := string(segment.Value(reader.Source()))
+
+		index := strings.IndexRune(line, ':')
+		if index == -1 {
+			data.Error = errors.New("invalid property line")
+			break
+		}
+
+		key := strings.TrimSpace(line[:index])
+		value := strings.TrimSpace(line[index+1:])
+		data.properties[key] = value
+	}
+
+	node.Parent().RemoveChild(node.Parent(), node)
+}
+
+func (p *propertiesParser) CanInterruptParagraph() bool {
+	return false
+}
+
+func (p *propertiesParser) CanAcceptIndentedLine() bool {
+	return false
+}
+
+func isSeparator(line []byte) bool {
+	line = util.TrimRightSpace(util.TrimLeftSpace(line))
+
+	for i := 0; i < len(line); i++ {
+		if line[i] != '-' {
+			return false
+		}
+	}
+
+	return true
+}
@@ -0,0 +1,110 @@
+package md
+
+import (
+	"bytes"
+	"strings"
+
+	"github.com/yuin/goldmark"
+	"github.com/yuin/goldmark/ast"
+	"github.com/yuin/goldmark/renderer"
+	"github.com/yuin/goldmark/renderer/html"
+	"github.com/yuin/goldmark/util"
+)
+
+type UrlTransformer = func(string) (string, error)
+
+type urlTransformerExtension struct {
+	transformer UrlTransformer
+}
+
+func (e *urlTransformerExtension) Extend(m goldmark.Markdown) {
+	m.Renderer().AddOptions(renderer.WithNodeRenderers(
+		util.Prioritized(&urlRenderer{e.transformer}, 500),
+	))
+}
+
+type urlRenderer struct {
+	transformer UrlTransformer
+}
+
+func (r *urlRenderer) RegisterFuncs(reg renderer.NodeRendererFuncRegisterer) {
+	reg.Register(ast.KindImage, r.renderImage)
+	reg.Register(ast.KindLink, r.renderLink)
+}
+
+func (r *urlRenderer) renderImage(w util.BufWriter, source []byte, node ast.Node, entering bool) (ast.WalkStatus, error) {
+	if !entering {
+		return ast.WalkContinue, nil
+	}
+
+	n := node.(*ast.Image)
+
+	url, err := r.resolveUrl(string(n.Destination))
+	if err != nil {
+		return ast.WalkStop, err
+	}
+
+	_, _ = w.WriteString(`<img src="`)
+	_, _ = w.WriteString(url)
+	_, _ = w.WriteString(`" alt="`)
+	_, _ = w.Write(nodeToHTMLText(n, source))
+	_ = w.WriteByte('"')
+	if n.Title != nil {
+		_, _ = w.WriteString(` title="`)
+		_, _ = w.Write(n.Title)
+		_ = w.WriteByte('"')
+	}
+	if n.Attributes() != nil {
+		html.RenderAttributes(w, n, html.ImageAttributeFilter)
+	}
+	_, _ = w.WriteString(">")
+	return ast.WalkSkipChildren, nil
+}
+
+func (r *urlRenderer) renderLink(w util.BufWriter, source []byte, node ast.Node, entering bool) (ast.WalkStatus, error) {
+	if !entering {
+		return ast.WalkContinue, nil
+	}
+
+	n := node.(*ast.Link)
+
+	url, err := r.resolveUrl(string(n.Destination))
+	if err != nil {
+		return ast.WalkStop, err
+	}
+
+	_, _ = w.WriteString(`<a href="`)
+	_, _ = w.WriteString(url)
+	_, _ = w.WriteString(`"`)
+	if n.Attributes() != nil {
+		html.RenderAttributes(w, n, html.ImageAttributeFilter)
+	}
+	_, _ = w.WriteString(`>`)
+	if n.Title != nil {
+		_, _ = w.Write(n.Title)
+	}
+	_, _ = w.WriteString("</a>")
+	return ast.WalkSkipChildren, nil
+}
+
+func (r *urlRenderer) resolveUrl(url string) (string, error) {
+	if !strings.HasPrefix(url, "http://") && !strings.HasPrefix(url, "https://") {
+		return r.transformer(url)
+	}
+
+	return url, nil
+}
+
+func nodeToHTMLText(n ast.Node, source []byte) []byte {
+	var buf bytes.Buffer
+	for c := n.FirstChild(); c != nil; c = c.NextSibling() {
+		if s, ok := c.(*ast.String); ok && s.IsCode() {
+			buf.Write(s.Text(source))
+		} else if !c.HasChildren() {
+			buf.Write(util.EscapeHTML(c.Text(source)))
+		} else {
+			buf.Write(nodeToHTMLText(c, source))
+		}
+	}
+	return buf.Bytes()
+}