d2/lib/imgbundler/imgbundler.go

239 lines
5.5 KiB
Go
Raw Normal View History

2022-11-26 23:26:14 +00:00
package imgbundler
import (
"bytes"
2022-11-26 23:26:14 +00:00
"context"
"encoding/base64"
"fmt"
2023-01-14 22:20:08 +00:00
"html"
"io"
"mime"
2022-11-26 23:26:14 +00:00
"net/http"
2022-11-28 21:25:40 +00:00
"net/url"
2022-11-27 02:14:41 +00:00
"os"
"path"
"path/filepath"
2022-11-26 23:26:14 +00:00
"regexp"
"strings"
"sync"
"time"
"golang.org/x/xerrors"
2022-11-29 21:19:03 +00:00
2023-10-20 00:44:39 +00:00
"oss.terrastruct.com/d2/lib/simplelog"
2022-12-08 07:22:20 +00:00
"oss.terrastruct.com/util-go/xdefer"
2022-11-26 23:26:14 +00:00
)
2023-06-06 17:43:52 +00:00
var imgCache sync.Map
2022-11-29 22:20:17 +00:00
const maxImageSize int64 = 1 << 25 // 33_554_432
2022-11-29 21:30:22 +00:00
var imageRegex = regexp.MustCompile(`<image href="([^"]+)"`)
2022-11-26 23:26:14 +00:00
func BundleLocal(ctx context.Context, l simplelog.Logger, inputPath string, in []byte, cacheImages bool) ([]byte, error) {
return bundle(ctx, l, inputPath, in, false, cacheImages)
2022-11-26 23:26:14 +00:00
}
2023-10-20 00:44:39 +00:00
func BundleRemote(ctx context.Context, l simplelog.Logger, in []byte, cacheImages bool) ([]byte, error) {
return bundle(ctx, l, "", in, true, cacheImages)
2022-11-27 02:14:41 +00:00
}
type repl struct {
from []byte
to []byte
2022-11-27 02:14:41 +00:00
}
func bundle(ctx context.Context, l simplelog.Logger, inputPath string, svg []byte, isRemote, cacheImages bool) (_ []byte, err error) {
if isRemote {
defer xdefer.Errorf(&err, "failed to bundle remote images")
} else {
defer xdefer.Errorf(&err, "failed to bundle local images")
}
imgs := imageRegex.FindAllSubmatch(svg, -1)
imgs = filterImageElements(imgs, isRemote)
2022-11-26 23:26:14 +00:00
2024-05-23 19:20:06 +00:00
if len(imgs) == 0 {
return svg, nil
}
ctx, cancel := context.WithTimeout(ctx, time.Minute*5)
defer cancel()
return runWorkers(ctx, l, inputPath, svg, imgs, isRemote, cacheImages)
}
2023-06-06 17:43:52 +00:00
// filterImageElements finds all unique image elements in imgs that are
// eligible for bundling in the current context.
func filterImageElements(imgs [][][]byte, isRemote bool) [][][]byte {
2023-06-06 17:43:52 +00:00
unq := make(map[string]struct{})
imgs2 := imgs[:0]
for _, img := range imgs {
href := string(img[1])
2023-06-06 17:43:52 +00:00
if _, ok := unq[href]; ok {
continue
}
unq[href] = struct{}{}
// Skip already bundled images.
if strings.HasPrefix(href, "data:") {
continue
}
2023-01-14 22:20:08 +00:00
u, err := url.Parse(html.UnescapeString(href))
isRemoteImg := err == nil && strings.HasPrefix(u.Scheme, "http")
if isRemoteImg == isRemote {
imgs2 = append(imgs2, img)
}
}
return imgs2
}
func runWorkers(ctx context.Context, l simplelog.Logger, inputPath string, svg []byte, imgs [][][]byte, isRemote, cacheImages bool) (_ []byte, err error) {
2022-11-26 23:26:14 +00:00
var wg sync.WaitGroup
replc := make(chan repl)
wg.Add(len(imgs))
go func() {
wg.Wait()
close(replc)
}()
// Limits the number of workers to 16.
sema := make(chan struct{}, 16)
2022-11-26 23:26:14 +00:00
var errhrefsMu sync.Mutex
var errhrefs []string
// Start workers as the sema allows.
2022-11-26 23:26:14 +00:00
go func() {
for _, img := range imgs {
img := img
sema <- struct{}{}
go func() {
defer func() {
wg.Done()
<-sema
}()
bundledImage, err := worker(ctx, l, inputPath, img[1], isRemote, cacheImages)
if err != nil {
2023-10-20 00:44:39 +00:00
l.Error(fmt.Sprintf("failed to bundle %s: %v", img[1], err))
errhrefsMu.Lock()
errhrefs = append(errhrefs, string(img[1]))
errhrefsMu.Unlock()
return
}
select {
case <-ctx.Done():
case replc <- repl{
from: img[0],
to: bundledImage,
}:
}
}()
2022-11-26 23:26:14 +00:00
}
}()
2022-11-30 22:44:02 +00:00
t := time.NewTicker(time.Second * 5)
defer t.Stop()
for {
select {
case <-ctx.Done():
return svg, xerrors.Errorf("failed to wait for workers: %w", ctx.Err())
2022-11-30 22:44:02 +00:00
case <-t.C:
2023-10-20 00:44:39 +00:00
l.Info("fetching images...")
case repl, ok := <-replc:
if !ok {
if len(errhrefs) > 0 {
return svg, xerrors.Errorf("%v", errhrefs)
}
return svg, nil
}
2023-06-06 17:43:52 +00:00
svg = bytes.Replace(svg, repl.from, repl.to, -1)
}
}
2022-11-26 23:26:14 +00:00
}
func worker(ctx context.Context, l simplelog.Logger, inputPath string, href []byte, isRemote, cacheImages bool) ([]byte, error) {
if cacheImages {
2023-06-06 18:09:19 +00:00
if hit, ok := imgCache.Load(string(href)); ok {
return hit.([]byte), nil
}
2023-06-06 17:43:52 +00:00
}
var buf []byte
var mimeType string
var err error
if isRemote {
2023-10-20 00:44:39 +00:00
l.Debug(fmt.Sprintf("fetching %s remotely", string(href)))
2023-01-14 22:20:08 +00:00
buf, mimeType, err = httpGet(ctx, html.UnescapeString(string(href)))
} else {
2023-10-20 00:44:39 +00:00
l.Debug(fmt.Sprintf("reading %s from disk", string(href)))
path := html.UnescapeString(string(href))
if inputPath != "-" && !filepath.IsAbs(path) {
path = filepath.Join(filepath.Dir(inputPath), path)
}
buf, err = os.ReadFile(path)
}
if err != nil {
return nil, err
}
if mimeType == "" {
mimeType = sniffMimeType(href, buf, isRemote)
}
mimeType = strings.Replace(mimeType, "text/xml", "image/svg+xml", 1)
b64 := base64.StdEncoding.EncodeToString(buf)
2023-06-06 17:43:52 +00:00
out := []byte(fmt.Sprintf(`<image href="data:%s;base64,%s"`, mimeType, b64))
if cacheImages {
2023-06-06 18:09:19 +00:00
imgCache.Store(string(href), out)
}
2023-06-06 17:43:52 +00:00
return out, nil
}
var httpClient = &http.Client{}
func httpGet(ctx context.Context, href string) ([]byte, string, error) {
2022-11-29 20:54:30 +00:00
ctx, cancel := context.WithTimeout(ctx, time.Minute)
defer cancel()
2022-11-26 23:26:14 +00:00
req, err := http.NewRequestWithContext(ctx, "GET", href, nil)
if err != nil {
return nil, "", err
2022-11-26 23:26:14 +00:00
}
resp, err := httpClient.Do(req)
2022-11-26 23:26:14 +00:00
if err != nil {
return nil, "", err
2022-11-26 23:26:14 +00:00
}
defer resp.Body.Close()
if resp.StatusCode != 200 {
return nil, "", fmt.Errorf("expected status 200 but got %d %s", resp.StatusCode, resp.Status)
2022-11-29 21:19:03 +00:00
}
r := http.MaxBytesReader(nil, resp.Body, maxImageSize)
buf, err := io.ReadAll(r)
if err != nil {
return nil, "", err
}
return buf, resp.Header.Get("Content-Type"), nil
}
// sniffMimeType sniffs the mime type of href based on its file extension and contents.
func sniffMimeType(href, buf []byte, isRemote bool) string {
p := string(href)
if isRemote {
2023-01-14 22:20:08 +00:00
u, err := url.Parse(html.UnescapeString(p))
if err != nil {
p = ""
} else {
p = u.Path
}
}
mimeType := mime.TypeByExtension(path.Ext(p))
if mimeType == "" {
mimeType = http.DetectContentType(buf)
}
return mimeType
2022-11-27 02:14:41 +00:00
}