Create fake users for tombstones whose users can't be inferred

scraper/api_types_v2.go

@@ -454,10 +454,10 @@ func (e APIV2Entry) ToTweetTrove() TweetTrove {
 	} else if e.Content.EntryType == "TimelineTimelineItem" {
 		ret, err := e.Content.ItemContent.TweetResults.ToTweetTrove()
 
+		// Handle tombstones in parent reply thread
 		if errors.Is(err, ErrorIsTombstone) {
-			// Handle tombstones
+			ret = NewTweetTrove() // clear the result just in case there is a TweetID(0) in it
-			ret = NewTweetTrove()                                                         // clear the result just in case
+			tombstoned_tweet := APITweet{}
-			tombstoned_tweet := e.Content.ItemContent.TweetResults.Result.Legacy.APITweet // Will be empty to start
 
 			// Capture the tombstone text
 			var is_ok bool
@@ -480,6 +480,8 @@ func (e APIV2Entry) ToTweetTrove() TweetTrove {
 				panic(err)
 			}
 			ret.Tweets[parsed_tombstone_tweet.ID] = parsed_tombstone_tweet
+		} else if err != nil {
+			panic(err)
 		}
 		return ret
 	}
@@ -626,11 +628,14 @@ func (api_response APIV2Response) ToTweetTrove() (TweetTrove, error) {
 			panic(fmt.Sprintf("Tombstoned tweet has no ID (should be %d)", tweet.InReplyToID))
 		}
 
+		// Fill out the replied tweet's UserID using this tweet's "in_reply_to_user_id".
+		// If this tweet doesn't have it (i.e., this tweet is also a tombstone), create a fake user instead, and add it to the tweet trove.
 		if replied_tweet.UserID == 0 {
 			replied_tweet.UserID = tweet.in_reply_to_user_id
-			if replied_tweet.UserID == 0 { // Still??
+			if replied_tweet.UserID == 0 {
-				log.Warn(fmt.Sprintf("Still couldn't find user for replied tweet %d", tweet.InReplyToID))
+				fake_user := GetUnknownUser()
-				continue
+				ret.Users[fake_user.ID] = fake_user
+				replied_tweet.UserID = fake_user.ID
 			}
 		} // replied_tweet.UserID should now be a real UserID
 

scraper/api_types_v2_test.go

@@ -601,12 +601,21 @@ func TestAPIV2ConversationThreadWithTombstones(t *testing.T) {
 	assert.True(is_ok)
 	assert.True(t1.IsStub)
 	assert.Equal(TweetID(0), t1.InReplyToID)
-	// TODO: assert associated user is fake
+	assert.NotEqual(UserID(0), t1.UserID)
+	t1_user, is_ok := trove.Users[t1.UserID]
+	assert.True(is_ok)
+	assert.True(t1_user.IsIdFake)
+	assert.Equal(UserHandle("<UNKNOWN USER>"), t1_user.Handle)
 
 	t2, is_ok := trove.Tweets[1454521424144654344]
 	assert.True(is_ok)
 	assert.True(t2.IsStub)
 	assert.Equal(TweetID(1454515503242829830), t2.InReplyToID)
+	assert.NotEqual(UserID(0), t2.UserID)
+	t2_user, is_ok := trove.Users[t2.UserID]
+	assert.True(is_ok)
+	assert.True(t2_user.IsIdFake)
+	assert.Equal(UserHandle("<UNKNOWN USER>"), t2_user.Handle)
 
 	t3, is_ok := trove.Tweets[1454522147750260742]
 	assert.True(is_ok)

scraper/user.go

@@ -84,6 +84,24 @@ Joined %s
 	return ret
 }
 
+func GetUnknownUser() User {
+	return User{
+		ID:              UserID(0x4000000000000000), // 2^62
+		DisplayName:     "<Unknown User>",
+		Handle:          UserHandle("<UNKNOWN USER>"),
+		Bio:             "<blank>",
+		FollowersCount:  0,
+		FollowingCount:  0,
+		Location:        "<blank>",
+		Website:         "<blank>",
+		JoinDate:        TimestampFromUnix(0),
+		IsVerified:      false,
+		IsPrivate:       false,
+		IsNeedingFakeID: false,
+		IsIdFake:        true,
+	}
+}
+
 /**
  * Unknown Users with handles are only created by direct GetUser calls (either `twitter fetch_user`
  * subcommand or as part of tombstone user fetching.)