linkpreview pkg

2025-06-10 18:10:49 +09:00 · 2019-10-28 14:03:05 +03:00 · 2019-10-28 14:03:05 +03:00 · 319eb5e591
commit 319eb5e591
parent 0ad2ccc431
4 changed files with 191 additions and 0 deletions
--- a/util/linkpreview/linkpreview.go
+++ b/util/linkpreview/linkpreview.go
@ -0,0 +1,93 @@
+package linkpreview
+
+import (
+	"context"
+	"net/http"
+	"path/filepath"
+	"strings"
+
+	"github.com/otiai10/opengraph"
+)
+
+func New() LinkPreview {
+	return &linkPreview{}
+}
+
+type LinkType string
+
+const (
+	LinkTypeHtml       LinkType = "html"
+	LinkTypeImage      LinkType = "image"
+	LinkTypeVideo      LinkType = "video"
+	LinkTypeText       LinkType = "text"
+	LinkTypeUnexpected LinkType = "unexpected"
+
+	// read no more than 400 kb
+	maxBytesToRead = 400000
+)
+
+type LinkPreview interface {
+	Fetch(ctx context.Context, url string) (Info, error)
+}
+
+type Info struct {
+	Title       string
+	Description string
+	ImageUrl    string
+	Type        LinkType
+}
+
+type linkPreview struct{}
+
+func (l *linkPreview) Fetch(ctx context.Context, url string) (Info, error) {
+	rt := &proxyRoundTripper{RoundTripper: http.DefaultTransport}
+	client := &http.Client{Transport: rt}
+	og, err := opengraph.FetchWithContext(ctx, url, client)
+	if err != nil {
+		if resp := rt.lastResponse; resp != nil && resp.StatusCode == http.StatusOK {
+			return l.makeNonHtml(url, resp)
+		}
+		return Info{}, err
+	}
+	return l.convertOGToInfo(og), nil
+}
+
+func (l *linkPreview) convertOGToInfo(og *opengraph.OpenGraph) (i Info) {
+	i = Info{
+		Title:       og.Title,
+		Description: og.Description,
+		Type:        LinkTypeHtml,
+	}
+	if len(og.Image) != 0 {
+		i.ImageUrl = og.Image[0].URL
+	}
+	return
+}
+
+func (l *linkPreview) makeNonHtml(url string, resp *http.Response) (i Info, err error) {
+	ct := resp.Header.Get("Content-Type")
+	i.Title = filepath.Base(url)
+	if strings.HasPrefix(ct, "image/") {
+		i.Type = LinkTypeImage
+		i.ImageUrl = url
+	} else if strings.HasPrefix(ct, "text/") {
+		i.Type = LinkTypeText
+	} else {
+		i.Type = LinkTypeUnexpected
+	}
+	return
+}
+
+type proxyRoundTripper struct {
+	http.RoundTripper
+	lastResponse *http.Response
+}
+
+func (p *proxyRoundTripper) RoundTrip(req *http.Request) (*http.Response, error) {
+	resp, err := p.RoundTripper.RoundTrip(req)
+	if err == nil {
+		p.lastResponse = resp
+		resp.Body = http.MaxBytesReader(nil, resp.Body, maxBytesToRead)
+	}
+	return resp, err
+}
--- a/util/linkpreview/linkpreview_test.go
+++ b/util/linkpreview/linkpreview_test.go
@ -0,0 +1,83 @@
+package linkpreview
+
+import (
+	"context"
+	"io"
+	"net/http"
+	"net/http/httptest"
+	"strings"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+	"github.com/stretchr/testify/require"
+)
+
+var ctx = context.Background()
+
+func TestLinkPreview_Fetch(t *testing.T) {
+	t.Run("html page", func(t *testing.T) {
+		ts := newTestServer("text/html", strings.NewReader(tetsHtml))
+		defer ts.Close()
+		lp := New()
+
+		info, err := lp.Fetch(ctx, ts.URL)
+		require.NoError(t, err)
+		assert.Equal(t, Info{
+			Title:       "Title",
+			Description: "Description",
+			ImageUrl:    "http://site.com/images/example.jpg",
+			Type:        LinkTypeHtml,
+		}, info)
+	})
+
+	t.Run("binary image", func(t *testing.T) {
+		tr := testReader(0)
+		ts := newTestServer("image/jpg", &tr)
+		defer ts.Close()
+		url := ts.URL + "/filename.jpg"
+		lp := New()
+		info, err := lp.Fetch(ctx, url)
+		require.NoError(t, err)
+		assert.Equal(t, Info{
+			Title:    "filename.jpg",
+			ImageUrl: url,
+			Type:     LinkTypeImage,
+		}, info)
+		assert.True(t, int(tr) <= maxBytesToRead)
+	})
+
+	t.Run("binary", func(t *testing.T) {
+		tr := testReader(0)
+		ts := newTestServer("binary/octed-stream", &tr)
+		defer ts.Close()
+		url := ts.URL + "/filename.jpg"
+		lp := New()
+		info, err := lp.Fetch(ctx, url)
+		require.NoError(t, err)
+		assert.Equal(t, Info{
+			Title: "filename.jpg",
+			Type:  LinkTypeUnexpected,
+		}, info)
+		assert.True(t, int(tr) <= maxBytesToRead)
+	})
+}
+
+func newTestServer(contentType string, data io.Reader) *httptest.Server {
+	return httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		w.Header().Set("Content-Type", contentType)
+		io.Copy(w, data)
+	}))
+}
+
+const tetsHtml = `<html><head>
+<title>Title</title>
+<meta name="description" content="Description">
+<meta property="og:image" content="http://site.com/images/example.jpg" />
+</head></html>`
+
+type testReader int
+
+func (t *testReader) Read(p []byte) (n int, err error) {
+	*t += testReader(len(p))
+	return len(p), nil
+}