Files
web-to-markdown/main.go
2026-05-17 11:53:25 +00:00

308 lines
8.0 KiB
Go

// Package main provides a CLI utility to download HTML articles and convert them to Markdown with localized images.
package main
import (
"bytes"
"context"
"crypto/sha256"
"encoding/hex"
"flag"
"fmt"
"io"
"net/http"
"net/url"
"os"
"os/signal"
"path/filepath"
"regexp"
"strings"
"syscall"
"time"
"codeberg.org/readeck/go-readability/v2"
md "github.com/JohannesKaufmann/html-to-markdown"
"github.com/PuerkitoBio/goquery"
"golang.org/x/sync/errgroup"
)
const (
dirPerms = 0755
filePerms = 0644
)
var (
// slugRegex compiled once at package level for performance.
slugRegex = regexp.MustCompile("[^a-z0-9]+")
)
// ImageJob represents a single image to be downloaded concurrently.
type ImageJob struct {
URL string
LocalPath string
}
func main() {
var titleOverride string
var outDirBase string
flag.StringVar(&titleOverride, "title", "", "Override the article title and folder name")
flag.StringVar(&outDirBase, "out", ".", "Base directory to download the article folder into")
flag.Parse()
if flag.NArg() < 1 {
fmt.Fprintf(os.Stderr, "Usage: %s [options] <url>\nOptions:\n", os.Args[0])
flag.PrintDefaults()
os.Exit(1)
}
rawURL := flag.Arg(0)
// Handle graceful shutdown via signals.
ctx, stop := signal.NotifyContext(context.Background(), os.Interrupt, syscall.SIGTERM)
defer stop()
client := &http.Client{Timeout: 30 * time.Second}
if err := run(ctx, client, rawURL, titleOverride, outDirBase); err != nil {
fmt.Fprintf(os.Stderr, "error: %v\n", err)
os.Exit(1)
}
}
// run orchestrates the fetching, parsing, and downloading process.
func run(ctx context.Context, client *http.Client, rawURL, titleOverride, outDirBase string) error {
targetURL, err := url.Parse(rawURL)
if err != nil {
return fmt.Errorf("failed to parse target URL %q: %w", rawURL, err)
}
fmt.Printf("Fetching: %s\n", rawURL)
req, err := http.NewRequestWithContext(ctx, http.MethodGet, rawURL, nil)
if err != nil {
return fmt.Errorf("failed to create request: %w", err)
}
resp, err := client.Do(req)
if err != nil {
return fmt.Errorf("failed to fetch URL: %w", err)
}
defer resp.Body.Close()
if resp.StatusCode != http.StatusOK {
return fmt.Errorf("failed to fetch URL: status code %d", resp.StatusCode)
}
article, err := readability.FromReader(resp.Body, targetURL)
if err != nil {
return fmt.Errorf("failed to parse article: %w", err)
}
finalTitle := article.Title()
if titleOverride != "" {
finalTitle = titleOverride
}
slug := generateSlug(finalTitle)
if slug == "" {
slug = "article"
}
outDir := filepath.Join(outDirBase, slug)
if err := os.MkdirAll(outDir, dirPerms); err != nil {
return fmt.Errorf("failed to create directory %s: %w", outDir, err)
}
var buf bytes.Buffer
if err := article.RenderHTML(&buf); err != nil {
return fmt.Errorf("failed to render article HTML: %w", err)
}
modifiedHTML, imageJobs, err := processHTML(buf.String(), targetURL, outDir)
if err != nil {
return fmt.Errorf("failed to process HTML: %w", err)
}
converter := md.NewConverter("", true, nil)
markdownContent, err := converter.ConvertString(modifiedHTML)
if err != nil {
return fmt.Errorf("failed to convert to markdown: %w", err)
}
frontmatter := fmt.Sprintf("---\ntitle: \"%s\"\nsource: \"%s\"\n---\n\n", finalTitle, rawURL)
markdownContent = frontmatter + markdownContent
if len(imageJobs) > 0 {
fmt.Printf("Downloading %d images concurrently...\n", len(imageJobs))
g, gCtx := errgroup.WithContext(ctx)
for _, job := range imageJobs {
g.Go(func() error {
return downloadImage(gCtx, client, job.URL, job.LocalPath)
})
}
if err := g.Wait(); err != nil {
fmt.Fprintf(os.Stderr, "warning: one or more images failed to download: %v\n", err)
}
}
mdPath := filepath.Join(outDir, "index.md")
if err := os.WriteFile(mdPath, []byte(markdownContent), filePerms); err != nil {
return fmt.Errorf("failed to write markdown file: %w", err)
}
fmt.Printf("Successfully saved article to: %s\n", mdPath)
return nil
}
// processHTML extracts images, rewrites their src attributes for a flat structure,
// and returns the modified HTML string and a slice of ImageJobs to be downloaded.
func processHTML(htmlContent string, targetURL *url.URL, outDir string) (string, []ImageJob, error) {
doc, err := goquery.NewDocumentFromReader(strings.NewReader(htmlContent))
if err != nil {
return "", nil, err
}
var imageJobs []ImageJob
doc.Find("img").Each(func(i int, s *goquery.Selection) {
src, exists := s.Attr("src")
if !exists || src == "" {
return
}
imgURL, err := targetURL.Parse(src)
if err != nil {
fmt.Fprintf(os.Stderr, "warning: failed to parse image URL %s: %v\n", src, err)
return
}
altText, _ := s.Attr("alt")
titleText, _ := s.Attr("title")
figCaption := s.ParentsFiltered("figure").Find("figcaption").Text()
// Traverse up DOM to find nearest preceding section header
var headerText string
curr := s
for curr.Length() > 0 && !curr.Is("body") {
h := curr.PrevAllFiltered("h1, h2, h3, h4, h5, h6").First()
if h.Length() > 0 {
headerText = h.Text()
break
}
curr = curr.Parent()
}
rawName := ""
if altText != "" {
rawName = altText
} else if titleText != "" {
rawName = titleText
} else if figCaption != "" {
rawName = figCaption
}
baseName := ""
if rawName != "" {
s.SetAttr("alt", strings.TrimSpace(rawName)) // set for markdown converter
baseName = generateSlug(rawName)
if len(baseName) > 40 {
baseName = baseName[:40]
baseName = strings.TrimRight(baseName, "-")
}
}
if baseName == "" || baseName == "-" {
baseNameStr := strings.TrimSuffix(filepath.Base(imgURL.Path), filepath.Ext(imgURL.Path))
baseName = generateSlug(baseNameStr)
// Fallback if filename is CDN hash (long) or empty
if len(baseName) >= 20 || baseName == "" || baseName == "-" || baseName == "image" {
if headerText != "" {
baseName = generateSlug(headerText)
s.SetAttr("alt", strings.TrimSpace(headerText))
} else {
baseName = fmt.Sprintf("section-image-%d", i+1)
s.SetAttr("alt", fmt.Sprintf("Section Image %d", i+1))
}
} else {
s.SetAttr("alt", strings.ReplaceAll(baseNameStr, "-", " "))
}
}
hash := sha256.Sum256([]byte(imgURL.String()))
hashStr := hex.EncodeToString(hash[:])[:6]
ext := filepath.Ext(imgURL.Path)
if ext == "" {
ext = ".jpg"
}
ext = strings.Split(ext, "?")[0]
// Format output like `capo-notation-a1b2c3.jpg` (using standard url-safe slugs)
filename := fmt.Sprintf("%s-%s%s", baseName, hashStr, ext)
localAbsPath := filepath.Join(outDir, filename)
s.SetAttr("src", filename)
imageJobs = append(imageJobs, ImageJob{
URL: imgURL.String(),
LocalPath: localAbsPath,
})
})
modifiedHTML, err := doc.Find("body").Html()
if err != nil {
return "", nil, fmt.Errorf("failed to render modified HTML: %w", err)
}
return modifiedHTML, imageJobs, nil
}
// downloadImage fetches the image from the given URL and saves it to localPath.
func downloadImage(ctx context.Context, client *http.Client, urlStr, localPath string) (err error) {
req, reqErr := http.NewRequestWithContext(ctx, http.MethodGet, urlStr, nil)
if reqErr != nil {
return fmt.Errorf("create request for %s: %w", urlStr, reqErr)
}
resp, doErr := client.Do(req)
if doErr != nil {
return fmt.Errorf("get %s: %w", urlStr, doErr)
}
defer resp.Body.Close()
if resp.StatusCode != http.StatusOK {
return fmt.Errorf("bad status %d for %s", resp.StatusCode, urlStr)
}
out, createErr := os.Create(localPath)
if createErr != nil {
return fmt.Errorf("create file %s: %w", localPath, createErr)
}
defer func() {
closeErr := out.Close()
if err == nil {
err = closeErr
}
}()
if _, err = io.Copy(out, resp.Body); err != nil {
return fmt.Errorf("write file %s: %w", localPath, err)
}
return nil
}
// generateSlug truncates the string to 5 words and converts it to a safe filesystem slug.
func generateSlug(s string) string {
words := strings.Fields(s)
if len(words) > 5 {
words = words[:5]
}
s = strings.Join(words, " ")
return strings.Trim(slugRegex.ReplaceAllString(strings.ToLower(s), "-"), "-")
}