From de292a4146b86057edc0d94b51097a854c4fa25c Mon Sep 17 00:00:00 2001
From: Umputun <umputun@gmail.com>
Date: Sat, 23 Mar 2019 03:47:07 -0500
Subject: [PATCH] add extraction of image ids

---
 backend/app/store/image/image.go      | 24 ++++++++++++++++++++++++
 backend/app/store/image/image_test.go | 10 ++++++++++
 2 files changed, 34 insertions(+)

diff --git a/backend/app/store/image/image.go b/backend/app/store/image/image.go
index d9e58a98..16fa144d 100644
--- a/backend/app/store/image/image.go
+++ b/backend/app/store/image/image.go
@@ -16,6 +16,7 @@ import (
 	"sync"
 	"time"
 
+	"github.com/PuerkitoBio/goquery"
 	log "github.com/go-pkgz/lgr"
 	"github.com/google/uuid"
 	"github.com/pkg/errors"
@@ -186,3 +187,26 @@ func (f *FileSystem) location(base string, id string) string {
 
 	return path.Join(base, user, partition(id), file)
 }
+
+// ExtractPictures gets list of images from the doc html and convert from urls to ids, i.e. user/pic.png
+func ExtractPictures(commentHTML string, match string) (ids []string, err error) {
+
+	doc, err := goquery.NewDocumentFromReader(strings.NewReader(commentHTML))
+	if err != nil {
+		return nil, errors.Wrap(err, "can't create document")
+	}
+	result := []string{}
+	doc.Find("img").Each(func(i int, s *goquery.Selection) {
+		if im, ok := s.Attr("src"); ok {
+			if strings.Contains(im, match) {
+				elems := strings.Split(im, "/")
+				if len(elems) >= 2 {
+					id := elems[len(elems)-2] + "/" + elems[len(elems)-1]
+					result = append(result, id)
+				}
+			}
+		}
+	})
+
+	return result, nil
+}
diff --git a/backend/app/store/image/image_test.go b/backend/app/store/image/image_test.go
index 0179df1d..ef3ec135 100644
--- a/backend/app/store/image/image_test.go
+++ b/backend/app/store/image/image_test.go
@@ -183,6 +183,16 @@ func TestImage_Cleanup(t *testing.T) {
 	assert.NotNil(t, err, "no file on staging anymore")
 }
 
+func TestExtractPictures(t *testing.T) {
+	html := `blah <img src="/blah/user1/pic1.png"/> foo 
+<img src="/blah/user2/pic3.png"/> xyz <p>123</p> <img src="/pic3.png"/>`
+	ids, err := ExtractPictures(html, "/blah/")
+	require.NoError(t, err)
+	assert.Equal(t, 2, len(ids), "two images")
+	assert.Equal(t, "user1/pic1.png", ids[0])
+	assert.Equal(t, "user2/pic3.png", ids[1])
+}
+
 func prepareImageTest(t *testing.T) (svc FileSystem, teardown func()) {
 	loc, err := ioutil.TempDir("", "test_image_r42")
 	require.NoError(t, err, "failed to make temp dir")