From b076f4d2f252f6d9dc1006ced22f2dcecbbdfe2c Mon Sep 17 00:00:00 2001
From: Alessio <me@playfulpachyderm.com>
Date: Sun, 3 Mar 2024 14:56:01 -0800
Subject: [PATCH] Scraper: handle videos being geo-blocked, which was panicking
 the downloader

---
 pkg/persistence/media_download.go |  7 ++++++-
 pkg/scraper/api_types.go          |  4 ++++
 pkg/scraper/video.go              |  2 ++
 pkg/scraper/video_test.go         | 23 +++++++++++++++++++----
 4 files changed, 31 insertions(+), 5 deletions(-)

diff --git a/pkg/persistence/media_download.go b/pkg/persistence/media_download.go
index 502a29a..730a4c4 100644
--- a/pkg/persistence/media_download.go
+++ b/pkg/persistence/media_download.go
@@ -27,7 +27,7 @@ var ErrorDMCA error = errors.New("video is DMCAed, unable to download (HTTP 403
 // - url: the remote file to download
 // - outpath: the path on disk to save it to
 func (d DefaultDownloader) Curl(url string, outpath string) error {
-	println(url)
+	fmt.Println(url)
 	resp, err := http.Get(url)
 	if err != nil {
 		return fmt.Errorf("Error executing HTTP GET(%q):\n  %w", url, err)
@@ -150,6 +150,11 @@ func (p Profile) DownloadTweetContentWithInjector(t *scraper.Tweet, downloader M
 	}
 
 	for i := range t.Videos {
+		// Videos can be geoblocked, and the HTTP response isn't in JSON so it's hard to capture
+		if t.Videos[i].IsGeoblocked {
+			continue
+		}
+
 		err := p.download_tweet_video(&t.Videos[i], downloader)
 		if err != nil {
 			return err
diff --git a/pkg/scraper/api_types.go b/pkg/scraper/api_types.go
index 9caaf36..00d177f 100644
--- a/pkg/scraper/api_types.go
+++ b/pkg/scraper/api_types.go
@@ -38,6 +38,10 @@ type APIExtendedMedia struct {
 		Variants SortableVariants `json:"variants"`
 		Duration int              `json:"duration_millis"`
 	} `json:"video_info"`
+	ExtMediaAvailability struct {
+		Status string `json:"status"`
+		Reason string `json:"reason"`
+	} `json:"ext_media_availability"`
 	OriginalInfo struct {
 		Width  int `json:"width"`
 		Height int `json:"height"`
diff --git a/pkg/scraper/video.go b/pkg/scraper/video.go
index a3c3ae5..947128e 100644
--- a/pkg/scraper/video.go
+++ b/pkg/scraper/video.go
@@ -26,6 +26,7 @@ type Video struct {
 
 	IsDownloaded    bool `db:"is_downloaded"`
 	IsBlockedByDMCA bool `db:"is_blocked_by_dmca"`
+	IsGeoblocked    bool `db:"is_geoblocked"`
 	IsGif           bool `db:"is_gif"`
 }
 
@@ -78,6 +79,7 @@ func ParseAPIVideo(apiVideo APIExtendedMedia, tweet_id TweetID) Video {
 
 		IsDownloaded:    false,
 		IsBlockedByDMCA: false,
+		IsGeoblocked:    apiVideo.ExtMediaAvailability.Reason == "Geoblocked",
 		IsGif:           apiVideo.Type == "animated_gif",
 	}
 }
diff --git a/pkg/scraper/video_test.go b/pkg/scraper/video_test.go
index 1932ce5..a394877 100644
--- a/pkg/scraper/video_test.go
+++ b/pkg/scraper/video_test.go
@@ -13,13 +13,13 @@ import (
 
 func TestParseAPIVideo(t *testing.T) {
 	assert := assert.New(t)
+	require := require.New(t)
 	data, err := os.ReadFile("test_responses/tweet_content/video.json")
-	if err != nil {
-		panic(err)
-	}
+	require.NoError(err)
+
 	var apivideo APIExtendedMedia
 	err = json.Unmarshal(data, &apivideo)
-	require.NoError(t, err)
+	require.NoError(err)
 
 	tweet_id := TweetID(28)
 	video := ParseAPIVideo(apivideo, tweet_id)
@@ -35,3 +35,18 @@ func TestParseAPIVideo(t *testing.T) {
 	assert.Equal(88300, video.Duration)
 	assert.False(video.IsDownloaded)
 }
+
+func TestParseGeoblockedVideo(t *testing.T) {
+	assert := assert.New(t)
+	require := require.New(t)
+	data, err := os.ReadFile("test_responses/tweet_content/video_geoblocked.json")
+	require.NoError(err)
+
+	var apivideo APIExtendedMedia
+	err = json.Unmarshal(data, &apivideo)
+	require.NoError(err)
+
+	tweet_id := TweetID(28)
+	video := ParseAPIVideo(apivideo, tweet_id)
+	assert.True(video.IsGeoblocked)
+}