feat: get image blob in backend

plugin/crawler/website.go → plugin/http_getter/html_meta.go

@@ -1,9 +1,10 @@
-package crawler
+package getter
 
 import (
+	"fmt"
 	"io"
 	"net/http"
-	urlUtil "net/url"
+	"net/url"
 
 	"golang.org/x/net/html"
 	"golang.org/x/net/html/atom"
@@ -15,19 +16,26 @@ type HTMLMeta struct {
 	Image       string `json:"image"`
 }
 
-func GetWebsiteMeta(url string) (*HTMLMeta, error) {
-	if _, err := urlUtil.Parse(url); err != nil {
+func GetHTMLMeta(urlStr string) (*HTMLMeta, error) {
+	if _, err := url.Parse(urlStr); err != nil {
 		return nil, err
 	}
 
-	response, err := http.Get(url)
+	response, err := http.Get(urlStr)
 	if err != nil {
 		return nil, err
 	}
 	defer response.Body.Close()
 
-	htmlMeta := extractHTMLMeta(response.Body)
+	mediatype, err := getMediatype(response)
+	if err != nil {
+		return nil, err
+	}
+	if mediatype != "text/html" {
+		return nil, fmt.Errorf("Wrong website mediatype")
+	}
 
+	htmlMeta := extractHTMLMeta(response.Body)
 	return htmlMeta, nil
 }
 

plugin/crawler/website_test.go → plugin/http_getter/html_meta_test.go

@@ -1,24 +1,24 @@
-package crawler
+package getter
 
 import (
 	"testing"
 
 	"github.com/stretchr/testify/require"
 )
 
-func TestGetWebsiteMeta(t *testing.T) {
+func TestGetHTMLMeta(t *testing.T) {
 	tests := []struct {
-		url      string
+		urlStr   string
 		htmlMeta HTMLMeta
 	}{
 		{
-			url: "https://baidu.com",
+			urlStr: "https://baidu.com",
 			htmlMeta: HTMLMeta{
 				Title: "百度一下，你就知道",
 			},
 		},
 		{
-			url: "https://www.bytebase.com/blog/sql-review-tool-for-devs",
+			urlStr: "https://www.bytebase.com/blog/sql-review-tool-for-devs",
 			htmlMeta: HTMLMeta{
 				Title:       "The SQL Review Tool for Developers",
 				Description: "Reviewing SQL can be somewhat tedious, yet is essential to keep your database fleet reliable. At Bytebase, we are building a developer-first SQL review tool to empower the DevOps system.",
@@ -27,7 +27,7 @@ func TestGetWebsiteMeta(t *testing.T) {
 		},
 	}
 	for _, test := range tests {
-		metadata, err := GetWebsiteMeta(test.url)
+		metadata, err := GetHTMLMeta(test.urlStr)
 		require.NoError(t, err)
 		require.Equal(t, test.htmlMeta, *metadata)
 	}

plugin/crawler/crawler.go → plugin/http_getter/http_getter.go

@@ -1,4 +1,4 @@
-// crawler is using to get resources from url.
+// getter is using to get resources from url.
 // * Get metadata for website;
 // * Get image blob to avoid CORS;
-package crawler
+package getter

plugin/http_getter/image.go

@@ -0,0 +1,45 @@
+package getter
+
+import (
+	"fmt"
+	"io"
+	"net/http"
+	"net/url"
+	"strings"
+)
+
+type Image struct {
+	Blob      []byte
+	Mediatype string
+}
+
+func GetImage(urlStr string) (*Image, error) {
+	if _, err := url.Parse(urlStr); err != nil {
+		return nil, err
+	}
+
+	response, err := http.Get(urlStr)
+	if err != nil {
+		return nil, err
+	}
+	defer response.Body.Close()
+
+	mediatype, err := getMediatype(response)
+	if err != nil {
+		return nil, err
+	}
+	if !strings.HasPrefix(mediatype, "image/") {
+		return nil, fmt.Errorf("Wrong image mediatype")
+	}
+
+	bodyBytes, err := io.ReadAll(response.Body)
+	if err != nil {
+		return nil, err
+	}
+
+	image := &Image{
+		Blob:      bodyBytes,
+		Mediatype: mediatype,
+	}
+	return image, nil
+}

plugin/http_getter/image_test.go

@@ -0,0 +1,21 @@
+package getter
+
+import (
+	"testing"
+
+	"github.com/stretchr/testify/require"
+)
+
+func TestGetImage(t *testing.T) {
+	tests := []struct {
+		urlStr string
+	}{
+		{
+			urlStr: "https://star-history.com/bytebase.webp",
+		},
+	}
+	for _, test := range tests {
+		_, err := GetImage(test.urlStr)
+		require.NoError(t, err)
+	}
+}

plugin/http_getter/util.go

@@ -0,0 +1,15 @@
+package getter
+
+import (
+	"mime"
+	"net/http"
+)
+
+func getMediatype(response *http.Response) (string, error) {
+	contentType := response.Header.Get("content-type")
+	mediatype, _, err := mime.ParseMediaType(contentType)
+	if err != nil {
+		return "", err
+	}
+	return mediatype, nil
+}

server/http_getter.go

@@ -0,0 +1,70 @@
+package server
+
+import (
+	"encoding/json"
+	"fmt"
+	"net/http"
+	"net/url"
+
+	"github.com/labstack/echo/v4"
+	getter "github.com/usememos/memos/plugin/http_getter"
+	metric "github.com/usememos/memos/plugin/metrics"
+)
+
+func (s *Server) registerCrawlerPublicRoutes(g *echo.Group) {
+	g.GET("/get/httpmeta", func(c echo.Context) error {
+		ctx := c.Request().Context()
+		urlStr := c.QueryParam("url")
+		if urlStr == "" {
+			return echo.NewHTTPError(http.StatusBadRequest, "Missing website url")
+		}
+		if _, err := url.Parse(urlStr); err != nil {
+			return echo.NewHTTPError(http.StatusBadRequest, "Wrong url").SetInternal(err)
+		}
+
+		htmlMeta, err := getter.GetHTMLMeta(urlStr)
+		if err != nil {
+			return echo.NewHTTPError(http.StatusNotAcceptable, fmt.Sprintf("Failed to get website meta with url: %s", urlStr)).SetInternal(err)
+		}
+		s.Collector.Collect(ctx, &metric.Metric{
+			Name: "getter used",
+			Labels: map[string]string{
+				"type": "httpmeta",
+			},
+		})
+
+		c.Response().Header().Set(echo.HeaderContentType, echo.MIMEApplicationJSONCharsetUTF8)
+		if err := json.NewEncoder(c.Response().Writer).Encode(composeResponse(htmlMeta)); err != nil {
+			return echo.NewHTTPError(http.StatusInternalServerError, "Failed to encode website HTML meta").SetInternal(err)
+		}
+		return nil
+	})
+	g.GET("/get/image", func(c echo.Context) error {
+		ctx := c.Request().Context()
+		urlStr := c.QueryParam("url")
+		if urlStr == "" {
+			return echo.NewHTTPError(http.StatusBadRequest, "Missing image url")
+		}
+		if _, err := url.Parse(urlStr); err != nil {
+			return echo.NewHTTPError(http.StatusBadRequest, "Wrong url").SetInternal(err)
+		}
+
+		image, err := getter.GetImage(urlStr)
+		if err != nil {
+			return echo.NewHTTPError(http.StatusNotAcceptable, fmt.Sprintf("Failed to get image url: %s", urlStr)).SetInternal(err)
+		}
+		s.Collector.Collect(ctx, &metric.Metric{
+			Name: "getter used",
+			Labels: map[string]string{
+				"type": "image",
+			},
+		})
+
+		c.Response().Writer.WriteHeader(http.StatusOK)
+		c.Response().Writer.Header().Set("Content-Type", image.Mediatype)
+		if _, err := c.Response().Writer.Write(image.Blob); err != nil {
+			return echo.NewHTTPError(http.StatusInternalServerError, "Failed to write image blob").SetInternal(err)
+		}
+		return nil
+	})
+}

server/resource.go

@@ -161,7 +161,6 @@ func (s *Server) registerResourceRoutes(g *echo.Group) {
 		if _, err := c.Response().Writer.Write(resource.Blob); err != nil {
 			return echo.NewHTTPError(http.StatusInternalServerError, "Failed to write resource blob").SetInternal(err)
 		}
-
 		return nil
 	})
 

web/src/labs/marked/parser/Image.ts

@@ -8,7 +8,8 @@ const renderer = (rawStr: string): string => {
     return rawStr;
   }
 
-  return `<img class='img' src='${escape(matchResult[1])}' />`;
+  // NOTE: Get image blob from backend to avoid CORS.
+  return `<img class='img' src='/o/get/image?url=${escape(matchResult[1])}' />`;
 };
 
 export default {